![[tokkee]](http://tokkee.org/images/avatar.png){kind=avatar}
1 #!/usr/bin/env perl
2 # Copyright (C) 2006, Eric Wong <normalperson@yhbt.net>
3 # License: GPL v2 or later
4 use warnings;
5 use strict;
6 use vars qw/ $AUTHOR $VERSION
7 $sha1 $sha1_short $_revision $_repository
8 $_q $_authors %users/;
9 $AUTHOR = 'Eric Wong <normalperson@yhbt.net>';
10 $VERSION = '@@GIT_VERSION@@';
12 # From which subdir have we been invoked?
13 my $cmd_dir_prefix = eval {
14 command_oneline([qw/rev-parse --show-prefix/], STDERR => 0)
15 } || '';
17 my $git_dir_user_set = 1 if defined $ENV{GIT_DIR};
18 $ENV{GIT_DIR} ||= '.git';
19 $Git::SVN::default_repo_id = 'svn';
20 $Git::SVN::default_ref_id = $ENV{GIT_SVN_ID} || 'git-svn';
21 $Git::SVN::Ra::_log_window_size = 100;
23 $Git::SVN::Log::TZ = $ENV{TZ};
24 $ENV{TZ} = 'UTC';
25 $| = 1; # unbuffer STDOUT
27 sub fatal (@) { print STDERR "@_\n"; exit 1 }
28 require SVN::Core; # use()-ing this causes segfaults for me... *shrug*
29 require SVN::Ra;
30 require SVN::Delta;
31 if ($SVN::Core::VERSION lt '1.1.0') {
32 fatal "Need SVN::Core 1.1.0 or better (got $SVN::Core::VERSION)";
33 }
34 push @Git::SVN::Ra::ISA, 'SVN::Ra';
35 push @SVN::Git::Editor::ISA, 'SVN::Delta::Editor';
36 push @SVN::Git::Fetcher::ISA, 'SVN::Delta::Editor';
37 use Carp qw/croak/;
38 use Digest::MD5;
39 use IO::File qw//;
40 use File::Basename qw/dirname basename/;
41 use File::Path qw/mkpath/;
42 use Getopt::Long qw/:config gnu_getopt no_ignore_case auto_abbrev/;
43 use IPC::Open3;
44 use Git;
46 BEGIN {
47 # import functions from Git into our packages, en masse
48 no strict 'refs';
49 foreach (qw/command command_oneline command_noisy command_output_pipe
50 command_input_pipe command_close_pipe/) {
51 for my $package ( qw(SVN::Git::Editor SVN::Git::Fetcher
52 Git::SVN::Migration Git::SVN::Log Git::SVN),
53 __PACKAGE__) {
54 *{"${package}::$_"} = \&{"Git::$_"};
55 }
56 }
57 }
59 my ($SVN);
61 $sha1 = qr/[a-f\d]{40}/;
62 $sha1_short = qr/[a-f\d]{4,40}/;
63 my ($_stdin, $_help, $_edit,
64 $_message, $_file,
65 $_template, $_shared,
66 $_version, $_fetch_all, $_no_rebase,
67 $_merge, $_strategy, $_dry_run, $_local,
68 $_prefix, $_no_checkout, $_url, $_verbose,
69 $_git_format);
70 $Git::SVN::_follow_parent = 1;
71 my %remote_opts = ( 'username=s' => \$Git::SVN::Prompt::_username,
72 'config-dir=s' => \$Git::SVN::Ra::config_dir,
73 'no-auth-cache' => \$Git::SVN::Prompt::_no_auth_cache );
74 my %fc_opts = ( 'follow-parent|follow!' => \$Git::SVN::_follow_parent,
75 'authors-file|A=s' => \$_authors,
76 'repack:i' => \$Git::SVN::_repack,
77 'noMetadata' => \$Git::SVN::_no_metadata,
78 'useSvmProps' => \$Git::SVN::_use_svm_props,
79 'useSvnsyncProps' => \$Git::SVN::_use_svnsync_props,
80 'log-window-size=i' => \$Git::SVN::Ra::_log_window_size,
81 'no-checkout' => \$_no_checkout,
82 'quiet|q' => \$_q,
83 'repack-flags|repack-args|repack-opts=s' =>
84 \$Git::SVN::_repack_flags,
85 'use-log-author' => \$Git::SVN::_use_log_author,
86 'add-author-from' => \$Git::SVN::_add_author_from,
87 %remote_opts );
89 my ($_trunk, $_tags, $_branches, $_stdlayout);
90 my %icv;
91 my %init_opts = ( 'template=s' => \$_template, 'shared:s' => \$_shared,
92 'trunk|T=s' => \$_trunk, 'tags|t=s' => \$_tags,
93 'branches|b=s' => \$_branches, 'prefix=s' => \$_prefix,
94 'stdlayout|s' => \$_stdlayout,
95 'minimize-url|m' => \$Git::SVN::_minimize_url,
96 'no-metadata' => sub { $icv{noMetadata} = 1 },
97 'use-svm-props' => sub { $icv{useSvmProps} = 1 },
98 'use-svnsync-props' => sub { $icv{useSvnsyncProps} = 1 },
99 'rewrite-root=s' => sub { $icv{rewriteRoot} = $_[1] },
100 %remote_opts );
101 my %cmt_opts = ( 'edit|e' => \$_edit,
102 'rmdir' => \$SVN::Git::Editor::_rmdir,
103 'find-copies-harder' => \$SVN::Git::Editor::_find_copies_harder,
104 'l=i' => \$SVN::Git::Editor::_rename_limit,
105 'copy-similarity|C=i'=> \$SVN::Git::Editor::_cp_similarity
106 );
108 my %cmd = (
109 fetch => [ \&cmd_fetch, "Download new revisions from SVN",
110 { 'revision|r=s' => \$_revision,
111 'fetch-all|all' => \$_fetch_all,
112 %fc_opts } ],
113 clone => [ \&cmd_clone, "Initialize and fetch revisions",
114 { 'revision|r=s' => \$_revision,
115 %fc_opts, %init_opts } ],
116 init => [ \&cmd_init, "Initialize a repo for tracking" .
117 " (requires URL argument)",
118 \%init_opts ],
119 'multi-init' => [ \&cmd_multi_init,
120 "Deprecated alias for ".
121 "'$0 init -T<trunk> -b<branches> -t<tags>'",
122 \%init_opts ],
123 dcommit => [ \&cmd_dcommit,
124 'Commit several diffs to merge with upstream',
125 { 'merge|m|M' => \$_merge,
126 'strategy|s=s' => \$_strategy,
127 'verbose|v' => \$_verbose,
128 'dry-run|n' => \$_dry_run,
129 'fetch-all|all' => \$_fetch_all,
130 'no-rebase' => \$_no_rebase,
131 %cmt_opts, %fc_opts } ],
132 'set-tree' => [ \&cmd_set_tree,
133 "Set an SVN repository to a git tree-ish",
134 { 'stdin|' => \$_stdin, %cmt_opts, %fc_opts, } ],
135 'create-ignore' => [ \&cmd_create_ignore,
136 'Create a .gitignore per svn:ignore',
137 { 'revision|r=i' => \$_revision
138 } ],
139 'propget' => [ \&cmd_propget,
140 'Print the value of a property on a file or directory',
141 { 'revision|r=i' => \$_revision } ],
142 'proplist' => [ \&cmd_proplist,
143 'List all properties of a file or directory',
144 { 'revision|r=i' => \$_revision } ],
145 'show-ignore' => [ \&cmd_show_ignore, "Show svn:ignore listings",
146 { 'revision|r=i' => \$_revision
147 } ],
148 'show-externals' => [ \&cmd_show_externals, "Show svn:externals listings",
149 { 'revision|r=i' => \$_revision
150 } ],
151 'multi-fetch' => [ \&cmd_multi_fetch,
152 "Deprecated alias for $0 fetch --all",
153 { 'revision|r=s' => \$_revision, %fc_opts } ],
154 'migrate' => [ sub { },
155 # no-op, we automatically run this anyways,
156 'Migrate configuration/metadata/layout from
157 previous versions of git-svn',
158 { 'minimize' => \$Git::SVN::Migration::_minimize,
159 %remote_opts } ],
160 'log' => [ \&Git::SVN::Log::cmd_show_log, 'Show commit logs',
161 { 'limit=i' => \$Git::SVN::Log::limit,
162 'revision|r=s' => \$_revision,
163 'verbose|v' => \$Git::SVN::Log::verbose,
164 'incremental' => \$Git::SVN::Log::incremental,
165 'oneline' => \$Git::SVN::Log::oneline,
166 'show-commit' => \$Git::SVN::Log::show_commit,
167 'non-recursive' => \$Git::SVN::Log::non_recursive,
168 'authors-file|A=s' => \$_authors,
169 'color' => \$Git::SVN::Log::color,
170 'pager=s' => \$Git::SVN::Log::pager
171 } ],
172 'find-rev' => [ \&cmd_find_rev, "Translate between SVN revision numbers and tree-ish",
173 {} ],
174 'rebase' => [ \&cmd_rebase, "Fetch and rebase your working directory",
175 { 'merge|m|M' => \$_merge,
176 'verbose|v' => \$_verbose,
177 'strategy|s=s' => \$_strategy,
178 'local|l' => \$_local,
179 'fetch-all|all' => \$_fetch_all,
180 'dry-run|n' => \$_dry_run,
181 %fc_opts } ],
182 'commit-diff' => [ \&cmd_commit_diff,
183 'Commit a diff between two trees',
184 { 'message|m=s' => \$_message,
185 'file|F=s' => \$_file,
186 'revision|r=s' => \$_revision,
187 %cmt_opts } ],
188 'info' => [ \&cmd_info,
189 "Show info about the latest SVN revision
190 on the current branch",
191 { 'url' => \$_url, } ],
192 'blame' => [ \&Git::SVN::Log::cmd_blame,
193 "Show what revision and author last modified each line of a file",
194 { 'git-format' => \$_git_format } ],
195 );
197 my $cmd;
198 for (my $i = 0; $i < @ARGV; $i++) {
199 if (defined $cmd{$ARGV[$i]}) {
200 $cmd = $ARGV[$i];
201 splice @ARGV, $i, 1;
202 last;
203 }
204 };
206 # make sure we're always running at the top-level working directory
207 unless ($cmd && $cmd =~ /(?:clone|init|multi-init)$/) {
208 unless (-d $ENV{GIT_DIR}) {
209 if ($git_dir_user_set) {
210 die "GIT_DIR=$ENV{GIT_DIR} explicitly set, ",
211 "but it is not a directory\n";
212 }
213 my $git_dir = delete $ENV{GIT_DIR};
214 chomp(my $cdup = command_oneline(qw/rev-parse --show-cdup/));
215 unless (length $cdup) {
216 die "Already at toplevel, but $git_dir ",
217 "not found '$cdup'\n";
218 }
219 chdir $cdup or die "Unable to chdir up to '$cdup'\n";
220 unless (-d $git_dir) {
221 die "$git_dir still not found after going to ",
222 "'$cdup'\n";
223 }
224 $ENV{GIT_DIR} = $git_dir;
225 }
226 $_repository = Git->repository(Repository => $ENV{GIT_DIR});
227 }
229 my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
231 read_repo_config(\%opts);
232 Getopt::Long::Configure('pass_through') if ($cmd && ($cmd eq 'log' || $cmd eq 'blame'));
233 my $rv = GetOptions(%opts, 'help|H|h' => \$_help, 'version|V' => \$_version,
234 'minimize-connections' => \$Git::SVN::Migration::_minimize,
235 'id|i=s' => \$Git::SVN::default_ref_id,
236 'svn-remote|remote|R=s' => sub {
237 $Git::SVN::no_reuse_existing = 1;
238 $Git::SVN::default_repo_id = $_[1] });
239 exit 1 if (!$rv && $cmd && $cmd ne 'log');
241 usage(0) if $_help;
242 version() if $_version;
243 usage(1) unless defined $cmd;
244 load_authors() if $_authors;
246 unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
247 Git::SVN::Migration::migration_check();
248 }
249 Git::SVN::init_vars();
250 eval {
251 Git::SVN::verify_remotes_sanity();
252 $cmd{$cmd}->[0]->(@ARGV);
253 };
254 fatal $@ if $@;
255 post_fetch_checkout();
256 exit 0;
258 ####################### primary functions ######################
259 sub usage {
260 my $exit = shift || 0;
261 my $fd = $exit ? \*STDERR : \*STDOUT;
262 print $fd <<"";
263 git-svn - bidirectional operations between a single Subversion tree and git
264 Usage: $0 <command> [options] [arguments]\n
266 print $fd "Available commands:\n" unless $cmd;
268 foreach (sort keys %cmd) {
269 next if $cmd && $cmd ne $_;
270 next if /^multi-/; # don't show deprecated commands
271 print $fd ' ',pack('A17',$_),$cmd{$_}->[1],"\n";
272 foreach (sort keys %{$cmd{$_}->[2]}) {
273 # mixed-case options are for .git/config only
274 next if /[A-Z]/ && /^[a-z]+$/i;
275 # prints out arguments as they should be passed:
276 my $x = s#[:=]s$## ? '<arg>' : s#[:=]i$## ? '<num>' : '';
277 print $fd ' ' x 21, join(', ', map { length $_ > 1 ?
278 "--$_" : "-$_" }
279 split /\|/,$_)," $x\n";
280 }
281 }
282 print $fd <<"";
283 \nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
284 arbitrary identifier if you're tracking multiple SVN branches/repositories in
285 one git repository and want to keep them separate. See git-svn(1) for more
286 information.
288 exit $exit;
289 }
291 sub version {
292 print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
293 exit 0;
294 }
296 sub do_git_init_db {
297 unless (-d $ENV{GIT_DIR}) {
298 my @init_db = ('init');
299 push @init_db, "--template=$_template" if defined $_template;
300 if (defined $_shared) {
301 if ($_shared =~ /[a-z]/) {
302 push @init_db, "--shared=$_shared";
303 } else {
304 push @init_db, "--shared";
305 }
306 }
307 command_noisy(@init_db);
308 $_repository = Git->repository(Repository => ".git");
309 }
310 my $set;
311 my $pfx = "svn-remote.$Git::SVN::default_repo_id";
312 foreach my $i (keys %icv) {
313 die "'$set' and '$i' cannot both be set\n" if $set;
314 next unless defined $icv{$i};
315 command_noisy('config', "$pfx.$i", $icv{$i});
316 $set = $i;
317 }
318 }
320 sub init_subdir {
321 my $repo_path = shift or return;
322 mkpath([$repo_path]) unless -d $repo_path;
323 chdir $repo_path or die "Couldn't chdir to $repo_path: $!\n";
324 $ENV{GIT_DIR} = '.git';
325 $_repository = Git->repository(Repository => $ENV{GIT_DIR});
326 }
328 sub cmd_clone {
329 my ($url, $path) = @_;
330 if (!defined $path &&
331 (defined $_trunk || defined $_branches || defined $_tags ||
332 defined $_stdlayout) &&
333 $url !~ m#^[a-z\+]+://#) {
334 $path = $url;
335 }
336 $path = basename($url) if !defined $path || !length $path;
337 cmd_init($url, $path);
338 Git::SVN::fetch_all($Git::SVN::default_repo_id);
339 }
341 sub cmd_init {
342 if (defined $_stdlayout) {
343 $_trunk = 'trunk' if (!defined $_trunk);
344 $_tags = 'tags' if (!defined $_tags);
345 $_branches = 'branches' if (!defined $_branches);
346 }
347 if (defined $_trunk || defined $_branches || defined $_tags) {
348 return cmd_multi_init(@_);
349 }
350 my $url = shift or die "SVN repository location required ",
351 "as a command-line argument\n";
352 init_subdir(@_);
353 do_git_init_db();
355 Git::SVN->init($url);
356 }
358 sub cmd_fetch {
359 if (grep /^\d+=./, @_) {
360 die "'<rev>=<commit>' fetch arguments are ",
361 "no longer supported.\n";
362 }
363 my ($remote) = @_;
364 if (@_ > 1) {
365 die "Usage: $0 fetch [--all] [svn-remote]\n";
366 }
367 $remote ||= $Git::SVN::default_repo_id;
368 if ($_fetch_all) {
369 cmd_multi_fetch();
370 } else {
371 Git::SVN::fetch_all($remote, Git::SVN::read_all_remotes());
372 }
373 }
375 sub cmd_set_tree {
376 my (@commits) = @_;
377 if ($_stdin || !@commits) {
378 print "Reading from stdin...\n";
379 @commits = ();
380 while (<STDIN>) {
381 if (/\b($sha1_short)\b/o) {
382 unshift @commits, $1;
383 }
384 }
385 }
386 my @revs;
387 foreach my $c (@commits) {
388 my @tmp = command('rev-parse',$c);
389 if (scalar @tmp == 1) {
390 push @revs, $tmp[0];
391 } elsif (scalar @tmp > 1) {
392 push @revs, reverse(command('rev-list',@tmp));
393 } else {
394 fatal "Failed to rev-parse $c";
395 }
396 }
397 my $gs = Git::SVN->new;
398 my ($r_last, $cmt_last) = $gs->last_rev_commit;
399 $gs->fetch;
400 if (defined $gs->{last_rev} && $r_last != $gs->{last_rev}) {
401 fatal "There are new revisions that were fetched ",
402 "and need to be merged (or acknowledged) ",
403 "before committing.\nlast rev: $r_last\n",
404 " current: $gs->{last_rev}";
405 }
406 $gs->set_tree($_) foreach @revs;
407 print "Done committing ",scalar @revs," revisions to SVN\n";
408 unlink $gs->{index};
409 }
411 sub cmd_dcommit {
412 my $head = shift;
413 git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
414 'Cannot dcommit with a dirty index. Commit your changes first, '
415 . "or stash them with `git stash'.\n";
416 $head ||= 'HEAD';
417 my @refs;
418 my ($url, $rev, $uuid, $gs) = working_head_info($head, \@refs);
419 if ($url) {
420 print "Committing to $url ...\n";
421 }
422 unless ($gs) {
423 die "Unable to determine upstream SVN information from ",
424 "$head history.\nPerhaps the repository is empty.";
425 }
426 my $last_rev;
427 my ($linear_refs, $parents) = linearize_history($gs, \@refs);
428 if ($_no_rebase && scalar(@$linear_refs) > 1) {
429 warn "Attempting to commit more than one change while ",
430 "--no-rebase is enabled.\n",
431 "If these changes depend on each other, re-running ",
432 "without --no-rebase may be required."
433 }
434 while (1) {
435 my $d = shift @$linear_refs or last;
436 unless (defined $last_rev) {
437 (undef, $last_rev, undef) = cmt_metadata("$d~1");
438 unless (defined $last_rev) {
439 fatal "Unable to extract revision information ",
440 "from commit $d~1";
441 }
442 }
443 if ($_dry_run) {
444 print "diff-tree $d~1 $d\n";
445 } else {
446 my $cmt_rev;
447 my %ed_opts = ( r => $last_rev,
448 log => get_commit_entry($d)->{log},
449 ra => Git::SVN::Ra->new($gs->full_url),
450 config => SVN::Core::config_get_config(
451 $Git::SVN::Ra::config_dir
452 ),
453 tree_a => "$d~1",
454 tree_b => $d,
455 editor_cb => sub {
456 print "Committed r$_[0]\n";
457 $cmt_rev = $_[0];
458 },
459 svn_path => '');
460 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
461 print "No changes\n$d~1 == $d\n";
462 } elsif ($parents->{$d} && @{$parents->{$d}}) {
463 $gs->{inject_parents_dcommit}->{$cmt_rev} =
464 $parents->{$d};
465 }
466 $_fetch_all ? $gs->fetch_all : $gs->fetch;
467 $last_rev = $cmt_rev;
468 next if $_no_rebase;
470 # we always want to rebase against the current HEAD,
471 # not any head that was passed to us
472 my @diff = command('diff-tree', $d,
473 $gs->refname, '--');
474 my @finish;
475 if (@diff) {
476 @finish = rebase_cmd();
477 print STDERR "W: $d and ", $gs->refname,
478 " differ, using @finish:\n",
479 join("\n", @diff), "\n";
480 } else {
481 print "No changes between current HEAD and ",
482 $gs->refname,
483 "\nResetting to the latest ",
484 $gs->refname, "\n";
485 @finish = qw/reset --mixed/;
486 }
487 command_noisy(@finish, $gs->refname);
488 if (@diff) {
489 @refs = ();
490 my ($url_, $rev_, $uuid_, $gs_) =
491 working_head_info($head, \@refs);
492 my ($linear_refs_, $parents_) =
493 linearize_history($gs_, \@refs);
494 if (scalar(@$linear_refs) !=
495 scalar(@$linear_refs_)) {
496 fatal "# of revisions changed ",
497 "\nbefore:\n",
498 join("\n", @$linear_refs),
499 "\n\nafter:\n",
500 join("\n", @$linear_refs_), "\n",
501 'If you are attempting to commit ',
502 "merges, try running:\n\t",
503 'git rebase --interactive',
504 '--preserve-merges ',
505 $gs->refname,
506 "\nBefore dcommitting";
507 }
508 if ($url_ ne $url) {
509 fatal "URL mismatch after rebase: ",
510 "$url_ != $url";
511 }
512 if ($uuid_ ne $uuid) {
513 fatal "uuid mismatch after rebase: ",
514 "$uuid_ != $uuid";
515 }
516 # remap parents
517 my (%p, @l, $i);
518 for ($i = 0; $i < scalar @$linear_refs; $i++) {
519 my $new = $linear_refs_->[$i] or next;
520 $p{$new} =
521 $parents->{$linear_refs->[$i]};
522 push @l, $new;
523 }
524 $parents = \%p;
525 $linear_refs = \@l;
526 }
527 }
528 }
529 unlink $gs->{index};
530 }
532 sub cmd_find_rev {
533 my $revision_or_hash = shift or die "SVN or git revision required ",
534 "as a command-line argument\n";
535 my $result;
536 if ($revision_or_hash =~ /^r\d+$/) {
537 my $head = shift;
538 $head ||= 'HEAD';
539 my @refs;
540 my (undef, undef, undef, $gs) = working_head_info($head, \@refs);
541 unless ($gs) {
542 die "Unable to determine upstream SVN information from ",
543 "$head history\n";
544 }
545 my $desired_revision = substr($revision_or_hash, 1);
546 $result = $gs->rev_map_get($desired_revision);
547 } else {
548 my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
549 $result = $rev;
550 }
551 print "$result\n" if $result;
552 }
554 sub cmd_rebase {
555 command_noisy(qw/update-index --refresh/);
556 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
557 unless ($gs) {
558 die "Unable to determine upstream SVN information from ",
559 "working tree history\n";
560 }
561 if ($_dry_run) {
562 print "Remote Branch: " . $gs->refname . "\n";
563 print "SVN URL: " . $url . "\n";
564 return;
565 }
566 if (command(qw/diff-index HEAD --/)) {
567 print STDERR "Cannot rebase with uncommited changes:\n";
568 command_noisy('status');
569 exit 1;
570 }
571 unless ($_local) {
572 # rebase will checkout for us, so no need to do it explicitly
573 $_no_checkout = 'true';
574 $_fetch_all ? $gs->fetch_all : $gs->fetch;
575 }
576 command_noisy(rebase_cmd(), $gs->refname);
577 }
579 sub cmd_show_ignore {
580 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
581 $gs ||= Git::SVN->new;
582 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
583 $gs->prop_walk($gs->{path}, $r, sub {
584 my ($gs, $path, $props) = @_;
585 print STDOUT "\n# $path\n";
586 my $s = $props->{'svn:ignore'} or return;
587 $s =~ s/[\r\n]+/\n/g;
588 chomp $s;
589 $s =~ s#^#$path#gm;
590 print STDOUT "$s\n";
591 });
592 }
594 sub cmd_show_externals {
595 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
596 $gs ||= Git::SVN->new;
597 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
598 $gs->prop_walk($gs->{path}, $r, sub {
599 my ($gs, $path, $props) = @_;
600 print STDOUT "\n# $path\n";
601 my $s = $props->{'svn:externals'} or return;
602 $s =~ s/[\r\n]+/\n/g;
603 chomp $s;
604 $s =~ s#^#$path#gm;
605 print STDOUT "$s\n";
606 });
607 }
609 sub cmd_create_ignore {
610 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
611 $gs ||= Git::SVN->new;
612 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
613 $gs->prop_walk($gs->{path}, $r, sub {
614 my ($gs, $path, $props) = @_;
615 # $path is of the form /path/to/dir/
616 my $ignore = '.' . $path . '.gitignore';
617 my $s = $props->{'svn:ignore'} or return;
618 open(GITIGNORE, '>', $ignore)
619 or fatal("Failed to open `$ignore' for writing: $!");
620 $s =~ s/[\r\n]+/\n/g;
621 chomp $s;
622 # Prefix all patterns so that the ignore doesn't apply
623 # to sub-directories.
624 $s =~ s#^#/#gm;
625 print GITIGNORE "$s\n";
626 close(GITIGNORE)
627 or fatal("Failed to close `$ignore': $!");
628 command_noisy('add', '-f', $ignore);
629 });
630 }
632 sub canonicalize_path {
633 my ($path) = @_;
634 my $dot_slash_added = 0;
635 if (substr($path, 0, 1) ne "/") {
636 $path = "./" . $path;
637 $dot_slash_added = 1;
638 }
639 # File::Spec->canonpath doesn't collapse x/../y into y (for a
640 # good reason), so let's do this manually.
641 $path =~ s#/+#/#g;
642 $path =~ s#/\.(?:/|$)#/#g;
643 $path =~ s#/[^/]+/\.\.##g;
644 $path =~ s#/$##g;
645 $path =~ s#^\./## if $dot_slash_added;
646 return $path;
647 }
649 # get_svnprops(PATH)
650 # ------------------
651 # Helper for cmd_propget and cmd_proplist below.
652 sub get_svnprops {
653 my $path = shift;
654 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
655 $gs ||= Git::SVN->new;
657 # prefix THE PATH by the sub-directory from which the user
658 # invoked us.
659 $path = $cmd_dir_prefix . $path;
660 fatal("No such file or directory: $path") unless -e $path;
661 my $is_dir = -d $path ? 1 : 0;
662 $path = $gs->{path} . '/' . $path;
664 # canonicalize the path (otherwise libsvn will abort or fail to
665 # find the file)
666 $path = canonicalize_path($path);
668 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
669 my $props;
670 if ($is_dir) {
671 (undef, undef, $props) = $gs->ra->get_dir($path, $r);
672 }
673 else {
674 (undef, $props) = $gs->ra->get_file($path, $r, undef);
675 }
676 return $props;
677 }
679 # cmd_propget (PROP, PATH)
680 # ------------------------
681 # Print the SVN property PROP for PATH.
682 sub cmd_propget {
683 my ($prop, $path) = @_;
684 $path = '.' if not defined $path;
685 usage(1) if not defined $prop;
686 my $props = get_svnprops($path);
687 if (not defined $props->{$prop}) {
688 fatal("`$path' does not have a `$prop' SVN property.");
689 }
690 print $props->{$prop} . "\n";
691 }
693 # cmd_proplist (PATH)
694 # -------------------
695 # Print the list of SVN properties for PATH.
696 sub cmd_proplist {
697 my $path = shift;
698 $path = '.' if not defined $path;
699 my $props = get_svnprops($path);
700 print "Properties on '$path':\n";
701 foreach (sort keys %{$props}) {
702 print " $_\n";
703 }
704 }
706 sub cmd_multi_init {
707 my $url = shift;
708 unless (defined $_trunk || defined $_branches || defined $_tags) {
709 usage(1);
710 }
712 # there are currently some bugs that prevent multi-init/multi-fetch
713 # setups from working well without this.
714 $Git::SVN::_minimize_url = 1;
716 $_prefix = '' unless defined $_prefix;
717 if (defined $url) {
718 $url =~ s#/+$##;
719 init_subdir(@_);
720 }
721 do_git_init_db();
722 if (defined $_trunk) {
723 my $trunk_ref = $_prefix . 'trunk';
724 # try both old-style and new-style lookups:
725 my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
726 unless ($gs_trunk) {
727 my ($trunk_url, $trunk_path) =
728 complete_svn_url($url, $_trunk);
729 $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
730 undef, $trunk_ref);
731 }
732 }
733 return unless defined $_branches || defined $_tags;
734 my $ra = $url ? Git::SVN::Ra->new($url) : undef;
735 complete_url_ls_init($ra, $_branches, '--branches/-b', $_prefix);
736 complete_url_ls_init($ra, $_tags, '--tags/-t', $_prefix . 'tags/');
737 }
739 sub cmd_multi_fetch {
740 my $remotes = Git::SVN::read_all_remotes();
741 foreach my $repo_id (sort keys %$remotes) {
742 if ($remotes->{$repo_id}->{url}) {
743 Git::SVN::fetch_all($repo_id, $remotes);
744 }
745 }
746 }
748 # this command is special because it requires no metadata
749 sub cmd_commit_diff {
750 my ($ta, $tb, $url) = @_;
751 my $usage = "Usage: $0 commit-diff -r<revision> ".
752 "<tree-ish> <tree-ish> [<URL>]";
753 fatal($usage) if (!defined $ta || !defined $tb);
754 my $svn_path = '';
755 if (!defined $url) {
756 my $gs = eval { Git::SVN->new };
757 if (!$gs) {
758 fatal("Needed URL or usable git-svn --id in ",
759 "the command-line\n", $usage);
760 }
761 $url = $gs->{url};
762 $svn_path = $gs->{path};
763 }
764 unless (defined $_revision) {
765 fatal("-r|--revision is a required argument\n", $usage);
766 }
767 if (defined $_message && defined $_file) {
768 fatal("Both --message/-m and --file/-F specified ",
769 "for the commit message.\n",
770 "I have no idea what you mean");
771 }
772 if (defined $_file) {
773 $_message = file_to_s($_file);
774 } else {
775 $_message ||= get_commit_entry($tb)->{log};
776 }
777 my $ra ||= Git::SVN::Ra->new($url);
778 my $r = $_revision;
779 if ($r eq 'HEAD') {
780 $r = $ra->get_latest_revnum;
781 } elsif ($r !~ /^\d+$/) {
782 die "revision argument: $r not understood by git-svn\n";
783 }
784 my %ed_opts = ( r => $r,
785 log => $_message,
786 ra => $ra,
787 tree_a => $ta,
788 tree_b => $tb,
789 editor_cb => sub { print "Committed r$_[0]\n" },
790 svn_path => $svn_path );
791 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
792 print "No changes\n$ta == $tb\n";
793 }
794 }
796 sub cmd_info {
797 my $path = canonicalize_path(shift or ".");
798 unless (scalar(@_) == 0) {
799 die "Too many arguments specified\n";
800 }
802 my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
804 if (!$file_type && !$diff_status) {
805 print STDERR "$path: (Not a versioned resource)\n\n";
806 return;
807 }
809 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
810 unless ($gs) {
811 die "Unable to determine upstream SVN information from ",
812 "working tree history\n";
813 }
814 my $full_url = $url . ($path eq "." ? "" : "/$path");
816 if ($_url) {
817 print $full_url, "\n";
818 return;
819 }
821 my $result = "Path: $path\n";
822 $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
823 $result .= "URL: " . $full_url . "\n";
825 eval {
826 my $repos_root = $gs->repos_root;
827 Git::SVN::remove_username($repos_root);
828 $result .= "Repository Root: $repos_root\n";
829 };
830 if ($@) {
831 $result .= "Repository Root: (offline)\n";
832 }
833 $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A";
834 $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
836 $result .= "Node Kind: " .
837 ($file_type eq "dir" ? "directory" : "file") . "\n";
839 my $schedule = $diff_status eq "A"
840 ? "add"
841 : ($diff_status eq "D" ? "delete" : "normal");
842 $result .= "Schedule: $schedule\n";
844 if ($diff_status eq "A") {
845 print $result, "\n";
846 return;
847 }
849 my ($lc_author, $lc_rev, $lc_date_utc);
850 my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $path);
851 my $log = command_output_pipe(@args);
852 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
853 while (<$log>) {
854 if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
855 $lc_author = $1;
856 $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
857 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
858 (undef, $lc_rev, undef) = ::extract_metadata($1);
859 }
860 }
861 close $log;
863 Git::SVN::Log::set_local_timezone();
865 $result .= "Last Changed Author: $lc_author\n";
866 $result .= "Last Changed Rev: $lc_rev\n";
867 $result .= "Last Changed Date: " .
868 Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
870 if ($file_type ne "dir") {
871 my $text_last_updated_date =
872 ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
873 $result .=
874 "Text Last Updated: " .
875 Git::SVN::Log::format_svn_date($text_last_updated_date) .
876 "\n";
877 my $checksum;
878 if ($diff_status eq "D") {
879 my ($fh, $ctx) =
880 command_output_pipe(qw(cat-file blob), "HEAD:$path");
881 if ($file_type eq "link") {
882 my $file_name = <$fh>;
883 $checksum = md5sum("link $file_name");
884 } else {
885 $checksum = md5sum($fh);
886 }
887 command_close_pipe($fh, $ctx);
888 } elsif ($file_type eq "link") {
889 my $file_name =
890 command(qw(cat-file blob), "HEAD:$path");
891 $checksum =
892 md5sum("link " . $file_name);
893 } else {
894 open FILE, "<", $path or die $!;
895 $checksum = md5sum(\*FILE);
896 close FILE or die $!;
897 }
898 $result .= "Checksum: " . $checksum . "\n";
899 }
901 print $result, "\n";
902 }
904 ########################### utility functions #########################
906 sub rebase_cmd {
907 my @cmd = qw/rebase/;
908 push @cmd, '-v' if $_verbose;
909 push @cmd, qw/--merge/ if $_merge;
910 push @cmd, "--strategy=$_strategy" if $_strategy;
911 @cmd;
912 }
914 sub post_fetch_checkout {
915 return if $_no_checkout;
916 my $gs = $Git::SVN::_head or return;
917 return if verify_ref('refs/heads/master^0');
919 my $valid_head = verify_ref('HEAD^0');
920 command_noisy(qw(update-ref refs/heads/master), $gs->refname);
921 return if ($valid_head || !verify_ref('HEAD^0'));
923 return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
924 my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
925 return if -f $index;
927 return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
928 return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
929 command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
930 print STDERR "Checked out HEAD:\n ",
931 $gs->full_url, " r", $gs->last_rev, "\n";
932 }
934 sub complete_svn_url {
935 my ($url, $path) = @_;
936 $path =~ s#/+$##;
937 if ($path !~ m#^[a-z\+]+://#) {
938 if (!defined $url || $url !~ m#^[a-z\+]+://#) {
939 fatal("E: '$path' is not a complete URL ",
940 "and a separate URL is not specified");
941 }
942 return ($url, $path);
943 }
944 return ($path, '');
945 }
947 sub complete_url_ls_init {
948 my ($ra, $repo_path, $switch, $pfx) = @_;
949 unless ($repo_path) {
950 print STDERR "W: $switch not specified\n";
951 return;
952 }
953 $repo_path =~ s#/+$##;
954 if ($repo_path =~ m#^[a-z\+]+://#) {
955 $ra = Git::SVN::Ra->new($repo_path);
956 $repo_path = '';
957 } else {
958 $repo_path =~ s#^/+##;
959 unless ($ra) {
960 fatal("E: '$repo_path' is not a complete URL ",
961 "and a separate URL is not specified");
962 }
963 }
964 my $url = $ra->{url};
965 my $gs = Git::SVN->init($url, undef, undef, undef, 1);
966 my $k = "svn-remote.$gs->{repo_id}.url";
967 my $orig_url = eval { command_oneline(qw/config --get/, $k) };
968 if ($orig_url && ($orig_url ne $gs->{url})) {
969 die "$k already set: $orig_url\n",
970 "wanted to set to: $gs->{url}\n";
971 }
972 command_oneline('config', $k, $gs->{url}) unless $orig_url;
973 my $remote_path = "$ra->{svn_path}/$repo_path";
974 $remote_path =~ s#/+#/#g;
975 $remote_path =~ s#^/##g;
976 $remote_path .= "/*" if $remote_path !~ /\*/;
977 my ($n) = ($switch =~ /^--(\w+)/);
978 if (length $pfx && $pfx !~ m#/$#) {
979 die "--prefix='$pfx' must have a trailing slash '/'\n";
980 }
981 command_noisy('config', "svn-remote.$gs->{repo_id}.$n",
982 "$remote_path:refs/remotes/$pfx*");
983 }
985 sub verify_ref {
986 my ($ref) = @_;
987 eval { command_oneline([ 'rev-parse', '--verify', $ref ],
988 { STDERR => 0 }); };
989 }
991 sub get_tree_from_treeish {
992 my ($treeish) = @_;
993 # $treeish can be a symbolic ref, too:
994 my $type = command_oneline(qw/cat-file -t/, $treeish);
995 my $expected;
996 while ($type eq 'tag') {
997 ($treeish, $type) = command(qw/cat-file tag/, $treeish);
998 }
999 if ($type eq 'commit') {
1000 $expected = (grep /^tree /, command(qw/cat-file commit/,
1001 $treeish))[0];
1002 ($expected) = ($expected =~ /^tree ($sha1)$/o);
1003 die "Unable to get tree from $treeish\n" unless $expected;
1004 } elsif ($type eq 'tree') {
1005 $expected = $treeish;
1006 } else {
1007 die "$treeish is a $type, expected tree, tag or commit\n";
1008 }
1009 return $expected;
1010 }
1012 sub get_commit_entry {
1013 my ($treeish) = shift;
1014 my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
1015 my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
1016 my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
1017 open my $log_fh, '>', $commit_editmsg or croak $!;
1019 my $type = command_oneline(qw/cat-file -t/, $treeish);
1020 if ($type eq 'commit' || $type eq 'tag') {
1021 my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1022 $type, $treeish);
1023 my $in_msg = 0;
1024 my $author;
1025 my $saw_from = 0;
1026 my $msgbuf = "";
1027 while (<$msg_fh>) {
1028 if (!$in_msg) {
1029 $in_msg = 1 if (/^\s*$/);
1030 $author = $1 if (/^author (.*>)/);
1031 } elsif (/^git-svn-id: /) {
1032 # skip this for now, we regenerate the
1033 # correct one on re-fetch anyways
1034 # TODO: set *:merge properties or like...
1035 } else {
1036 if (/^From:/ || /^Signed-off-by:/) {
1037 $saw_from = 1;
1038 }
1039 $msgbuf .= $_;
1040 }
1041 }
1042 $msgbuf =~ s/\s+$//s;
1043 if ($Git::SVN::_add_author_from && defined($author)
1044 && !$saw_from) {
1045 $msgbuf .= "\n\nFrom: $author";
1046 }
1047 print $log_fh $msgbuf or croak $!;
1048 command_close_pipe($msg_fh, $ctx);
1049 }
1050 close $log_fh or croak $!;
1052 if ($_edit || ($type eq 'tree')) {
1053 my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
1054 # TODO: strip out spaces, comments, like git-commit.sh
1055 system($editor, $commit_editmsg);
1056 }
1057 rename $commit_editmsg, $commit_msg or croak $!;
1058 open $log_fh, '<', $commit_msg or croak $!;
1059 { local $/; chomp($log_entry{log} = <$log_fh>); }
1060 close $log_fh or croak $!;
1061 unlink $commit_msg;
1062 \%log_entry;
1063 }
1065 sub s_to_file {
1066 my ($str, $file, $mode) = @_;
1067 open my $fd,'>',$file or croak $!;
1068 print $fd $str,"\n" or croak $!;
1069 close $fd or croak $!;
1070 chmod ($mode &~ umask, $file) if (defined $mode);
1071 }
1073 sub file_to_s {
1074 my $file = shift;
1075 open my $fd,'<',$file or croak "$!: file: $file\n";
1076 local $/;
1077 my $ret = <$fd>;
1078 close $fd or croak $!;
1079 $ret =~ s/\s*$//s;
1080 return $ret;
1081 }
1083 # '<svn username> = real-name <email address>' mapping based on git-svnimport:
1084 sub load_authors {
1085 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1086 my $log = $cmd eq 'log';
1087 while (<$authors>) {
1088 chomp;
1089 next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1090 my ($user, $name, $email) = ($1, $2, $3);
1091 if ($log) {
1092 $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1093 } else {
1094 $users{$user} = [$name, $email];
1095 }
1096 }
1097 close $authors or croak $!;
1098 }
1100 # convert GetOpt::Long specs for use by git-config
1101 sub read_repo_config {
1102 return unless -d $ENV{GIT_DIR};
1103 my $opts = shift;
1104 my @config_only;
1105 foreach my $o (keys %$opts) {
1106 # if we have mixedCase and a long option-only, then
1107 # it's a config-only variable that we don't need for
1108 # the command-line.
1109 push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1110 my $v = $opts->{$o};
1111 my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1112 $key =~ s/-//g;
1113 my $arg = 'git-config';
1114 $arg .= ' --int' if ($o =~ /[:=]i$/);
1115 $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1116 if (ref $v eq 'ARRAY') {
1117 chomp(my @tmp = `$arg --get-all svn.$key`);
1118 @$v = @tmp if @tmp;
1119 } else {
1120 chomp(my $tmp = `$arg --get svn.$key`);
1121 if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1122 $$v = $tmp;
1123 }
1124 }
1125 }
1126 delete @$opts{@config_only} if @config_only;
1127 }
1129 sub extract_metadata {
1130 my $id = shift or return (undef, undef, undef);
1131 my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1132 \s([a-f\d\-]+)$/x);
1133 if (!defined $rev || !$uuid || !$url) {
1134 # some of the original repositories I made had
1135 # identifiers like this:
1136 ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/);
1137 }
1138 return ($url, $rev, $uuid);
1139 }
1141 sub cmt_metadata {
1142 return extract_metadata((grep(/^git-svn-id: /,
1143 command(qw/cat-file commit/, shift)))[-1]);
1144 }
1146 sub working_head_info {
1147 my ($head, $refs) = @_;
1148 my @args = ('log', '--no-color', '--first-parent', '--pretty=medium');
1149 my ($fh, $ctx) = command_output_pipe(@args, $head);
1150 my $hash;
1151 my %max;
1152 while (<$fh>) {
1153 if ( m{^commit ($::sha1)$} ) {
1154 unshift @$refs, $hash if $hash and $refs;
1155 $hash = $1;
1156 next;
1157 }
1158 next unless s{^\s*(git-svn-id:)}{$1};
1159 my ($url, $rev, $uuid) = extract_metadata($_);
1160 if (defined $url && defined $rev) {
1161 next if $max{$url} and $max{$url} < $rev;
1162 if (my $gs = Git::SVN->find_by_url($url)) {
1163 my $c = $gs->rev_map_get($rev);
1164 if ($c && $c eq $hash) {
1165 close $fh; # break the pipe
1166 return ($url, $rev, $uuid, $gs);
1167 } else {
1168 $max{$url} ||= $gs->rev_map_max;
1169 }
1170 }
1171 }
1172 }
1173 command_close_pipe($fh, $ctx);
1174 (undef, undef, undef, undef);
1175 }
1177 sub read_commit_parents {
1178 my ($parents, $c) = @_;
1179 chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1180 $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1181 @{$parents->{$c}} = split(/ /, $p);
1182 }
1184 sub linearize_history {
1185 my ($gs, $refs) = @_;
1186 my %parents;
1187 foreach my $c (@$refs) {
1188 read_commit_parents(\%parents, $c);
1189 }
1191 my @linear_refs;
1192 my %skip = ();
1193 my $last_svn_commit = $gs->last_commit;
1194 foreach my $c (reverse @$refs) {
1195 next if $c eq $last_svn_commit;
1196 last if $skip{$c};
1198 unshift @linear_refs, $c;
1199 $skip{$c} = 1;
1201 # we only want the first parent to diff against for linear
1202 # history, we save the rest to inject when we finalize the
1203 # svn commit
1204 my $fp_a = verify_ref("$c~1");
1205 my $fp_b = shift @{$parents{$c}} if $parents{$c};
1206 if (!$fp_a || !$fp_b) {
1207 die "Commit $c\n",
1208 "has no parent commit, and therefore ",
1209 "nothing to diff against.\n",
1210 "You should be working from a repository ",
1211 "originally created by git-svn\n";
1212 }
1213 if ($fp_a ne $fp_b) {
1214 die "$c~1 = $fp_a, however parsing commit $c ",
1215 "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1216 }
1218 foreach my $p (@{$parents{$c}}) {
1219 $skip{$p} = 1;
1220 }
1221 }
1222 (\@linear_refs, \%parents);
1223 }
1225 sub find_file_type_and_diff_status {
1226 my ($path) = @_;
1227 return ('dir', '') if $path eq '.';
1229 my $diff_output =
1230 command_oneline(qw(diff --cached --name-status --), $path) || "";
1231 my $diff_status = (split(' ', $diff_output))[0] || "";
1233 my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1235 return (undef, undef) if !$diff_status && !$ls_tree;
1237 if ($diff_status eq "A") {
1238 return ("link", $diff_status) if -l $path;
1239 return ("dir", $diff_status) if -d $path;
1240 return ("file", $diff_status);
1241 }
1243 my $mode = (split(' ', $ls_tree))[0] || "";
1245 return ("link", $diff_status) if $mode eq "120000";
1246 return ("dir", $diff_status) if $mode eq "040000";
1247 return ("file", $diff_status);
1248 }
1250 sub md5sum {
1251 my $arg = shift;
1252 my $ref = ref $arg;
1253 my $md5 = Digest::MD5->new();
1254 if ($ref eq 'GLOB' || $ref eq 'IO::File') {
1255 $md5->addfile($arg) or croak $!;
1256 } elsif ($ref eq 'SCALAR') {
1257 $md5->add($$arg) or croak $!;
1258 } elsif (!$ref) {
1259 $md5->add($arg) or croak $!;
1260 } else {
1261 ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
1262 }
1263 return $md5->hexdigest();
1264 }
1266 package Git::SVN;
1267 use strict;
1268 use warnings;
1269 use Fcntl qw/:DEFAULT :seek/;
1270 use constant rev_map_fmt => 'NH40';
1271 use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1272 $_repack $_repack_flags $_use_svm_props $_head
1273 $_use_svnsync_props $no_reuse_existing $_minimize_url
1274 $_use_log_author $_add_author_from/;
1275 use Carp qw/croak/;
1276 use File::Path qw/mkpath/;
1277 use File::Copy qw/copy/;
1278 use IPC::Open3;
1280 my ($_gc_nr, $_gc_period);
1282 # properties that we do not log:
1283 my %SKIP_PROP;
1284 BEGIN {
1285 %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1286 svn:special svn:executable
1287 svn:entry:committed-rev
1288 svn:entry:last-author
1289 svn:entry:uuid
1290 svn:entry:committed-date/;
1292 # some options are read globally, but can be overridden locally
1293 # per [svn-remote "..."] section. Command-line options will *NOT*
1294 # override options set in an [svn-remote "..."] section
1295 no strict 'refs';
1296 for my $option (qw/follow_parent no_metadata use_svm_props
1297 use_svnsync_props/) {
1298 my $key = $option;
1299 $key =~ tr/_//d;
1300 my $prop = "-$option";
1301 *$option = sub {
1302 my ($self) = @_;
1303 return $self->{$prop} if exists $self->{$prop};
1304 my $k = "svn-remote.$self->{repo_id}.$key";
1305 eval { command_oneline(qw/config --get/, $k) };
1306 if ($@) {
1307 $self->{$prop} = ${"Git::SVN::_$option"};
1308 } else {
1309 my $v = command_oneline(qw/config --bool/,$k);
1310 $self->{$prop} = $v eq 'false' ? 0 : 1;
1311 }
1312 return $self->{$prop};
1313 }
1314 }
1315 }
1317 my (%LOCKFILES, %INDEX_FILES);
1318 END {
1319 unlink keys %LOCKFILES if %LOCKFILES;
1320 unlink keys %INDEX_FILES if %INDEX_FILES;
1321 }
1323 sub resolve_local_globs {
1324 my ($url, $fetch, $glob_spec) = @_;
1325 return unless defined $glob_spec;
1326 my $ref = $glob_spec->{ref};
1327 my $path = $glob_spec->{path};
1328 foreach (command(qw#for-each-ref --format=%(refname) refs/remotes#)) {
1329 next unless m#^refs/remotes/$ref->{regex}$#;
1330 my $p = $1;
1331 my $pathname = desanitize_refname($path->full_path($p));
1332 my $refname = desanitize_refname($ref->full_path($p));
1333 if (my $existing = $fetch->{$pathname}) {
1334 if ($existing ne $refname) {
1335 die "Refspec conflict:\n",
1336 "existing: refs/remotes/$existing\n",
1337 " globbed: refs/remotes/$refname\n";
1338 }
1339 my $u = (::cmt_metadata("refs/remotes/$refname"))[0];
1340 $u =~ s!^\Q$url\E(/|$)!! or die
1341 "refs/remotes/$refname: '$url' not found in '$u'\n";
1342 if ($pathname ne $u) {
1343 warn "W: Refspec glob conflict ",
1344 "(ref: refs/remotes/$refname):\n",
1345 "expected path: $pathname\n",
1346 " real path: $u\n",
1347 "Continuing ahead with $u\n";
1348 next;
1349 }
1350 } else {
1351 $fetch->{$pathname} = $refname;
1352 }
1353 }
1354 }
1356 sub parse_revision_argument {
1357 my ($base, $head) = @_;
1358 if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1359 return ($base, $head);
1360 }
1361 return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1362 return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1363 return ($head, $head) if ($::_revision eq 'HEAD');
1364 return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1365 return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1366 die "revision argument: $::_revision not understood by git-svn\n";
1367 }
1369 sub fetch_all {
1370 my ($repo_id, $remotes) = @_;
1371 if (ref $repo_id) {
1372 my $gs = $repo_id;
1373 $repo_id = undef;
1374 $repo_id = $gs->{repo_id};
1375 }
1376 $remotes ||= read_all_remotes();
1377 my $remote = $remotes->{$repo_id} or
1378 die "[svn-remote \"$repo_id\"] unknown\n";
1379 my $fetch = $remote->{fetch};
1380 my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1381 my (@gs, @globs);
1382 my $ra = Git::SVN::Ra->new($url);
1383 my $uuid = $ra->get_uuid;
1384 my $head = $ra->get_latest_revnum;
1385 my $base = defined $fetch ? $head : 0;
1387 # read the max revs for wildcard expansion (branches/*, tags/*)
1388 foreach my $t (qw/branches tags/) {
1389 defined $remote->{$t} or next;
1390 push @globs, $remote->{$t};
1391 my $max_rev = eval { tmp_config(qw/--int --get/,
1392 "svn-remote.$repo_id.${t}-maxRev") };
1393 if (defined $max_rev && ($max_rev < $base)) {
1394 $base = $max_rev;
1395 } elsif (!defined $max_rev) {
1396 $base = 0;
1397 }
1398 }
1400 if ($fetch) {
1401 foreach my $p (sort keys %$fetch) {
1402 my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1403 my $lr = $gs->rev_map_max;
1404 if (defined $lr) {
1405 $base = $lr if ($lr < $base);
1406 }
1407 push @gs, $gs;
1408 }
1409 }
1411 ($base, $head) = parse_revision_argument($base, $head);
1412 $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1413 }
1415 sub read_all_remotes {
1416 my $r = {};
1417 foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1418 if (m!^(.+)\.fetch=\s*(.*)\s*:\s*refs/remotes/(.+)\s*$!) {
1419 my ($remote, $local_ref, $remote_ref) = ($1, $2, $3);
1420 $local_ref =~ s{^/}{};
1421 $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1422 } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1423 $r->{$1}->{url} = $2;
1424 } elsif (m!^(.+)\.(branches|tags)=
1425 (.*):refs/remotes/(.+)\s*$/!x) {
1426 my ($p, $g) = ($3, $4);
1427 my $rs = $r->{$1}->{$2} = {
1428 t => $2,
1429 remote => $1,
1430 path => Git::SVN::GlobSpec->new($p),
1431 ref => Git::SVN::GlobSpec->new($g) };
1432 if (length($rs->{ref}->{right}) != 0) {
1433 die "The '*' glob character must be the last ",
1434 "character of '$g'\n";
1435 }
1436 }
1437 }
1438 $r;
1439 }
1441 sub init_vars {
1442 $_gc_nr = $_gc_period = 1000;
1443 if (defined $_repack || defined $_repack_flags) {
1444 warn "Repack options are obsolete; they have no effect.\n";
1445 }
1446 }
1448 sub verify_remotes_sanity {
1449 return unless -d $ENV{GIT_DIR};
1450 my %seen;
1451 foreach (command(qw/config -l/)) {
1452 if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1453 if ($seen{$1}) {
1454 die "Remote ref refs/remote/$1 is tracked by",
1455 "\n \"$_\"\nand\n \"$seen{$1}\"\n",
1456 "Please resolve this ambiguity in ",
1457 "your git configuration file before ",
1458 "continuing\n";
1459 }
1460 $seen{$1} = $_;
1461 }
1462 }
1463 }
1465 # we allow more chars than remotes2config.sh...
1466 sub sanitize_remote_name {
1467 my ($name) = @_;
1468 $name =~ tr{A-Za-z0-9:,/+-}{.}c;
1469 $name;
1470 }
1472 sub find_existing_remote {
1473 my ($url, $remotes) = @_;
1474 return undef if $no_reuse_existing;
1475 my $existing;
1476 foreach my $repo_id (keys %$remotes) {
1477 my $u = $remotes->{$repo_id}->{url} or next;
1478 next if $u ne $url;
1479 $existing = $repo_id;
1480 last;
1481 }
1482 $existing;
1483 }
1485 sub init_remote_config {
1486 my ($self, $url, $no_write) = @_;
1487 $url =~ s!/+$!!; # strip trailing slash
1488 my $r = read_all_remotes();
1489 my $existing = find_existing_remote($url, $r);
1490 if ($existing) {
1491 unless ($no_write) {
1492 print STDERR "Using existing ",
1493 "[svn-remote \"$existing\"]\n";
1494 }
1495 $self->{repo_id} = $existing;
1496 } elsif ($_minimize_url) {
1497 my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1498 $existing = find_existing_remote($min_url, $r);
1499 if ($existing) {
1500 unless ($no_write) {
1501 print STDERR "Using existing ",
1502 "[svn-remote \"$existing\"]\n";
1503 }
1504 $self->{repo_id} = $existing;
1505 }
1506 if ($min_url ne $url) {
1507 unless ($no_write) {
1508 print STDERR "Using higher level of URL: ",
1509 "$url => $min_url\n";
1510 }
1511 my $old_path = $self->{path};
1512 $self->{path} = $url;
1513 $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1514 if (length $old_path) {
1515 $self->{path} .= "/$old_path";
1516 }
1517 $url = $min_url;
1518 }
1519 }
1520 my $orig_url;
1521 if (!$existing) {
1522 # verify that we aren't overwriting anything:
1523 $orig_url = eval {
1524 command_oneline('config', '--get',
1525 "svn-remote.$self->{repo_id}.url")
1526 };
1527 if ($orig_url && ($orig_url ne $url)) {
1528 die "svn-remote.$self->{repo_id}.url already set: ",
1529 "$orig_url\nwanted to set to: $url\n";
1530 }
1531 }
1532 my ($xrepo_id, $xpath) = find_ref($self->refname);
1533 if (defined $xpath) {
1534 die "svn-remote.$xrepo_id.fetch already set to track ",
1535 "$xpath:refs/remotes/", $self->refname, "\n";
1536 }
1537 unless ($no_write) {
1538 command_noisy('config',
1539 "svn-remote.$self->{repo_id}.url", $url);
1540 $self->{path} =~ s{^/}{};
1541 command_noisy('config', '--add',
1542 "svn-remote.$self->{repo_id}.fetch",
1543 "$self->{path}:".$self->refname);
1544 }
1545 $self->{url} = $url;
1546 }
1548 sub find_by_url { # repos_root and, path are optional
1549 my ($class, $full_url, $repos_root, $path) = @_;
1551 return undef unless defined $full_url;
1552 remove_username($full_url);
1553 remove_username($repos_root) if defined $repos_root;
1554 my $remotes = read_all_remotes();
1555 if (defined $full_url && defined $repos_root && !defined $path) {
1556 $path = $full_url;
1557 $path =~ s#^\Q$repos_root\E(?:/|$)##;
1558 }
1559 foreach my $repo_id (keys %$remotes) {
1560 my $u = $remotes->{$repo_id}->{url} or next;
1561 remove_username($u);
1562 next if defined $repos_root && $repos_root ne $u;
1564 my $fetch = $remotes->{$repo_id}->{fetch} || {};
1565 foreach (qw/branches tags/) {
1566 resolve_local_globs($u, $fetch,
1567 $remotes->{$repo_id}->{$_});
1568 }
1569 my $p = $path;
1570 my $rwr = rewrite_root({repo_id => $repo_id});
1571 unless (defined $p) {
1572 $p = $full_url;
1573 my $z = $u;
1574 if ($rwr) {
1575 $z = $rwr;
1576 }
1577 $p =~ s#^\Q$z\E(?:/|$)## or next;
1578 }
1579 foreach my $f (keys %$fetch) {
1580 next if $f ne $p;
1581 return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1582 }
1583 }
1584 undef;
1585 }
1587 sub init {
1588 my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1589 my $self = _new($class, $repo_id, $ref_id, $path);
1590 if (defined $url) {
1591 $self->init_remote_config($url, $no_write);
1592 }
1593 $self;
1594 }
1596 sub find_ref {
1597 my ($ref_id) = @_;
1598 foreach (command(qw/config -l/)) {
1599 next unless m!^svn-remote\.(.+)\.fetch=
1600 \s*(.*)\s*:\s*refs/remotes/(.+)\s*$!x;
1601 my ($repo_id, $path, $ref) = ($1, $2, $3);
1602 if ($ref eq $ref_id) {
1603 $path = '' if ($path =~ m#^\./?#);
1604 return ($repo_id, $path);
1605 }
1606 }
1607 (undef, undef, undef);
1608 }
1610 sub new {
1611 my ($class, $ref_id, $repo_id, $path) = @_;
1612 if (defined $ref_id && !defined $repo_id && !defined $path) {
1613 ($repo_id, $path) = find_ref($ref_id);
1614 if (!defined $repo_id) {
1615 die "Could not find a \"svn-remote.*.fetch\" key ",
1616 "in the repository configuration matching: ",
1617 "refs/remotes/$ref_id\n";
1618 }
1619 }
1620 my $self = _new($class, $repo_id, $ref_id, $path);
1621 if (!defined $self->{path} || !length $self->{path}) {
1622 my $fetch = command_oneline('config', '--get',
1623 "svn-remote.$repo_id.fetch",
1624 ":refs/remotes/$ref_id\$") or
1625 die "Failed to read \"svn-remote.$repo_id.fetch\" ",
1626 "\":refs/remotes/$ref_id\$\" in config\n";
1627 ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
1628 }
1629 $self->{url} = command_oneline('config', '--get',
1630 "svn-remote.$repo_id.url") or
1631 die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
1632 $self->rebuild;
1633 $self;
1634 }
1636 sub refname {
1637 my ($refname) = "refs/remotes/$_[0]->{ref_id}" ;
1639 # It cannot end with a slash /, we'll throw up on this because
1640 # SVN can't have directories with a slash in their name, either:
1641 if ($refname =~ m{/$}) {
1642 die "ref: '$refname' ends with a trailing slash, this is ",
1643 "not permitted by git nor Subversion\n";
1644 }
1646 # It cannot have ASCII control character space, tilde ~, caret ^,
1647 # colon :, question-mark ?, asterisk *, space, or open bracket [
1648 # anywhere.
1649 #
1650 # Additionally, % must be escaped because it is used for escaping
1651 # and we want our escaped refname to be reversible
1652 $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
1654 # no slash-separated component can begin with a dot .
1655 # /.* becomes /%2E*
1656 $refname =~ s{/\.}{/%2E}g;
1658 # It cannot have two consecutive dots .. anywhere
1659 # .. becomes %2E%2E
1660 $refname =~ s{\.\.}{%2E%2E}g;
1662 return $refname;
1663 }
1665 sub desanitize_refname {
1666 my ($refname) = @_;
1667 $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
1668 return $refname;
1669 }
1671 sub svm_uuid {
1672 my ($self) = @_;
1673 return $self->{svm}->{uuid} if $self->svm;
1674 $self->ra;
1675 unless ($self->{svm}) {
1676 die "SVM UUID not cached, and reading remotely failed\n";
1677 }
1678 $self->{svm}->{uuid};
1679 }
1681 sub svm {
1682 my ($self) = @_;
1683 return $self->{svm} if $self->{svm};
1684 my $svm;
1685 # see if we have it in our config, first:
1686 eval {
1687 my $section = "svn-remote.$self->{repo_id}";
1688 $svm = {
1689 source => tmp_config('--get', "$section.svm-source"),
1690 uuid => tmp_config('--get', "$section.svm-uuid"),
1691 replace => tmp_config('--get', "$section.svm-replace"),
1692 }
1693 };
1694 if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
1695 $self->{svm} = $svm;
1696 }
1697 $self->{svm};
1698 }
1700 sub _set_svm_vars {
1701 my ($self, $ra) = @_;
1702 return $ra if $self->svm;
1704 my @err = ( "useSvmProps set, but failed to read SVM properties\n",
1705 "(svm:source, svm:uuid) ",
1706 "from the following URLs:\n" );
1707 sub read_svm_props {
1708 my ($self, $ra, $path, $r) = @_;
1709 my $props = ($ra->get_dir($path, $r))[2];
1710 my $src = $props->{'svm:source'};
1711 my $uuid = $props->{'svm:uuid'};
1712 return undef if (!$src || !$uuid);
1714 chomp($src, $uuid);
1716 $uuid =~ m{^[0-9a-f\-]{30,}$}
1717 or die "doesn't look right - svm:uuid is '$uuid'\n";
1719 # the '!' is used to mark the repos_root!/relative/path
1720 $src =~ s{/?!/?}{/};
1721 $src =~ s{/+$}{}; # no trailing slashes please
1722 # username is of no interest
1723 $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
1725 my $replace = $ra->{url};
1726 $replace .= "/$path" if length $path;
1728 my $section = "svn-remote.$self->{repo_id}";
1729 tmp_config("$section.svm-source", $src);
1730 tmp_config("$section.svm-replace", $replace);
1731 tmp_config("$section.svm-uuid", $uuid);
1732 $self->{svm} = {
1733 source => $src,
1734 uuid => $uuid,
1735 replace => $replace
1736 };
1737 }
1739 my $r = $ra->get_latest_revnum;
1740 my $path = $self->{path};
1741 my %tried;
1742 while (length $path) {
1743 unless ($tried{"$self->{url}/$path"}) {
1744 return $ra if $self->read_svm_props($ra, $path, $r);
1745 $tried{"$self->{url}/$path"} = 1;
1746 }
1747 $path =~ s#/?[^/]+$##;
1748 }
1749 die "Path: '$path' should be ''\n" if $path ne '';
1750 return $ra if $self->read_svm_props($ra, $path, $r);
1751 $tried{"$self->{url}/$path"} = 1;
1753 if ($ra->{repos_root} eq $self->{url}) {
1754 die @err, (map { " $_\n" } keys %tried), "\n";
1755 }
1757 # nope, make sure we're connected to the repository root:
1758 my $ok;
1759 my @tried_b;
1760 $path = $ra->{svn_path};
1761 $ra = Git::SVN::Ra->new($ra->{repos_root});
1762 while (length $path) {
1763 unless ($tried{"$ra->{url}/$path"}) {
1764 $ok = $self->read_svm_props($ra, $path, $r);
1765 last if $ok;
1766 $tried{"$ra->{url}/$path"} = 1;
1767 }
1768 $path =~ s#/?[^/]+$##;
1769 }
1770 die "Path: '$path' should be ''\n" if $path ne '';
1771 $ok ||= $self->read_svm_props($ra, $path, $r);
1772 $tried{"$ra->{url}/$path"} = 1;
1773 if (!$ok) {
1774 die @err, (map { " $_\n" } keys %tried), "\n";
1775 }
1776 Git::SVN::Ra->new($self->{url});
1777 }
1779 sub svnsync {
1780 my ($self) = @_;
1781 return $self->{svnsync} if $self->{svnsync};
1783 if ($self->no_metadata) {
1784 die "Can't have both 'noMetadata' and ",
1785 "'useSvnsyncProps' options set!\n";
1786 }
1787 if ($self->rewrite_root) {
1788 die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
1789 "options set!\n";
1790 }
1792 my $svnsync;
1793 # see if we have it in our config, first:
1794 eval {
1795 my $section = "svn-remote.$self->{repo_id}";
1797 my $url = tmp_config('--get', "$section.svnsync-url");
1798 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1799 die "doesn't look right - svn:sync-from-url is '$url'\n";
1801 my $uuid = tmp_config('--get', "$section.svnsync-uuid");
1802 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1803 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1805 $svnsync = { url => $url, uuid => $uuid }
1806 };
1807 if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
1808 return $self->{svnsync} = $svnsync;
1809 }
1811 my $err = "useSvnsyncProps set, but failed to read " .
1812 "svnsync property: svn:sync-from-";
1813 my $rp = $self->ra->rev_proplist(0);
1815 my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
1816 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1817 die "doesn't look right - svn:sync-from-url is '$url'\n";
1819 my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
1820 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1821 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1823 my $section = "svn-remote.$self->{repo_id}";
1824 tmp_config('--add', "$section.svnsync-uuid", $uuid);
1825 tmp_config('--add', "$section.svnsync-url", $url);
1826 return $self->{svnsync} = { url => $url, uuid => $uuid };
1827 }
1829 # this allows us to memoize our SVN::Ra UUID locally and avoid a
1830 # remote lookup (useful for 'git svn log').
1831 sub ra_uuid {
1832 my ($self) = @_;
1833 unless ($self->{ra_uuid}) {
1834 my $key = "svn-remote.$self->{repo_id}.uuid";
1835 my $uuid = eval { tmp_config('--get', $key) };
1836 if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/) {
1837 $self->{ra_uuid} = $uuid;
1838 } else {
1839 die "ra_uuid called without URL\n" unless $self->{url};
1840 $self->{ra_uuid} = $self->ra->get_uuid;
1841 tmp_config('--add', $key, $self->{ra_uuid});
1842 }
1843 }
1844 $self->{ra_uuid};
1845 }
1847 sub _set_repos_root {
1848 my ($self, $repos_root) = @_;
1849 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1850 $repos_root ||= $self->ra->{repos_root};
1851 tmp_config($k, $repos_root);
1852 $repos_root;
1853 }
1855 sub repos_root {
1856 my ($self) = @_;
1857 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1858 eval { tmp_config('--get', $k) } || $self->_set_repos_root;
1859 }
1861 sub ra {
1862 my ($self) = shift;
1863 my $ra = Git::SVN::Ra->new($self->{url});
1864 $self->_set_repos_root($ra->{repos_root});
1865 if ($self->use_svm_props && !$self->{svm}) {
1866 if ($self->no_metadata) {
1867 die "Can't have both 'noMetadata' and ",
1868 "'useSvmProps' options set!\n";
1869 } elsif ($self->use_svnsync_props) {
1870 die "Can't have both 'useSvnsyncProps' and ",
1871 "'useSvmProps' options set!\n";
1872 }
1873 $ra = $self->_set_svm_vars($ra);
1874 $self->{-want_revprops} = 1;
1875 }
1876 $ra;
1877 }
1879 sub rel_path {
1880 my ($self) = @_;
1881 my $repos_root = $self->ra->{repos_root};
1882 return $self->{path} if ($self->{url} eq $repos_root);
1883 my $url = $self->{url} .
1884 (length $self->{path} ? "/$self->{path}" : $self->{path});
1885 $url =~ s!^\Q$repos_root\E(?:/+|$)!!g;
1886 $url;
1887 }
1889 # prop_walk(PATH, REV, SUB)
1890 # -------------------------
1891 # Recursively traverse PATH at revision REV and invoke SUB for each
1892 # directory that contains a SVN property. SUB will be invoked as
1893 # follows: &SUB(gs, path, props); where `gs' is this instance of
1894 # Git::SVN, `path' the path to the directory where the properties
1895 # `props' were found. The `path' will be relative to point of checkout,
1896 # that is, if url://repo/trunk is the current Git branch, and that
1897 # directory contains a sub-directory `d', SUB will be invoked with `/d/'
1898 # as `path' (note the trailing `/').
1899 sub prop_walk {
1900 my ($self, $path, $rev, $sub) = @_;
1902 $path =~ s#^/##;
1903 my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
1904 $path =~ s#^/*#/#g;
1905 my $p = $path;
1906 # Strip the irrelevant part of the path.
1907 $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
1908 # Ensure the path is terminated by a `/'.
1909 $p =~ s#/*$#/#;
1911 # The properties contain all the internal SVN stuff nobody
1912 # (usually) cares about.
1913 my $interesting_props = 0;
1914 foreach (keys %{$props}) {
1915 # If it doesn't start with `svn:', it must be a
1916 # user-defined property.
1917 ++$interesting_props and next if $_ !~ /^svn:/;
1918 # FIXME: Fragile, if SVN adds new public properties,
1919 # this needs to be updated.
1920 ++$interesting_props if /^svn:(?:ignore|keywords|executable
1921 |eol-style|mime-type
1922 |externals|needs-lock)$/x;
1923 }
1924 &$sub($self, $p, $props) if $interesting_props;
1926 foreach (sort keys %$dirent) {
1927 next if $dirent->{$_}->{kind} != $SVN::Node::dir;
1928 $self->prop_walk($self->{path} . $p . $_, $rev, $sub);
1929 }
1930 }
1932 sub last_rev { ($_[0]->last_rev_commit)[0] }
1933 sub last_commit { ($_[0]->last_rev_commit)[1] }
1935 # returns the newest SVN revision number and newest commit SHA1
1936 sub last_rev_commit {
1937 my ($self) = @_;
1938 if (defined $self->{last_rev} && defined $self->{last_commit}) {
1939 return ($self->{last_rev}, $self->{last_commit});
1940 }
1941 my $c = ::verify_ref($self->refname.'^0');
1942 if ($c && !$self->use_svm_props && !$self->no_metadata) {
1943 my $rev = (::cmt_metadata($c))[1];
1944 if (defined $rev) {
1945 ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
1946 return ($rev, $c);
1947 }
1948 }
1949 my $map_path = $self->map_path;
1950 unless (-e $map_path) {
1951 ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
1952 return (undef, undef);
1953 }
1954 my ($rev, $commit) = $self->rev_map_max(1);
1955 ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
1956 return ($rev, $commit);
1957 }
1959 sub get_fetch_range {
1960 my ($self, $min, $max) = @_;
1961 $max ||= $self->ra->get_latest_revnum;
1962 $min ||= $self->rev_map_max;
1963 (++$min, $max);
1964 }
1966 sub tmp_config {
1967 my (@args) = @_;
1968 my $old_def_config = "$ENV{GIT_DIR}/svn/config";
1969 my $config = "$ENV{GIT_DIR}/svn/.metadata";
1970 if (! -f $config && -f $old_def_config) {
1971 rename $old_def_config, $config or
1972 die "Failed rename $old_def_config => $config: $!\n";
1973 }
1974 my $old_config = $ENV{GIT_CONFIG};
1975 $ENV{GIT_CONFIG} = $config;
1976 $@ = undef;
1977 my @ret = eval {
1978 unless (-f $config) {
1979 mkfile($config);
1980 open my $fh, '>', $config or
1981 die "Can't open $config: $!\n";
1982 print $fh "; This file is used internally by ",
1983 "git-svn\n" or die
1984 "Couldn't write to $config: $!\n";
1985 print $fh "; You should not have to edit it\n" or
1986 die "Couldn't write to $config: $!\n";
1987 close $fh or die "Couldn't close $config: $!\n";
1988 }
1989 command('config', @args);
1990 };
1991 my $err = $@;
1992 if (defined $old_config) {
1993 $ENV{GIT_CONFIG} = $old_config;
1994 } else {
1995 delete $ENV{GIT_CONFIG};
1996 }
1997 die $err if $err;
1998 wantarray ? @ret : $ret[0];
1999 }
2001 sub tmp_index_do {
2002 my ($self, $sub) = @_;
2003 my $old_index = $ENV{GIT_INDEX_FILE};
2004 $ENV{GIT_INDEX_FILE} = $self->{index};
2005 $@ = undef;
2006 my @ret = eval {
2007 my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
2008 mkpath([$dir]) unless -d $dir;
2009 &$sub;
2010 };
2011 my $err = $@;
2012 if (defined $old_index) {
2013 $ENV{GIT_INDEX_FILE} = $old_index;
2014 } else {
2015 delete $ENV{GIT_INDEX_FILE};
2016 }
2017 die $err if $err;
2018 wantarray ? @ret : $ret[0];
2019 }
2021 sub assert_index_clean {
2022 my ($self, $treeish) = @_;
2024 $self->tmp_index_do(sub {
2025 command_noisy('read-tree', $treeish) unless -e $self->{index};
2026 my $x = command_oneline('write-tree');
2027 my ($y) = (command(qw/cat-file commit/, $treeish) =~
2028 /^tree ($::sha1)/mo);
2029 return if $y eq $x;
2031 warn "Index mismatch: $y != $x\nrereading $treeish\n";
2032 unlink $self->{index} or die "unlink $self->{index}: $!\n";
2033 command_noisy('read-tree', $treeish);
2034 $x = command_oneline('write-tree');
2035 if ($y ne $x) {
2036 ::fatal "trees ($treeish) $y != $x\n",
2037 "Something is seriously wrong...";
2038 }
2039 });
2040 }
2042 sub get_commit_parents {
2043 my ($self, $log_entry) = @_;
2044 my (%seen, @ret, @tmp);
2045 # legacy support for 'set-tree'; this is only used by set_tree_cb:
2046 if (my $ip = $self->{inject_parents}) {
2047 if (my $commit = delete $ip->{$log_entry->{revision}}) {
2048 push @tmp, $commit;
2049 }
2050 }
2051 if (my $cur = ::verify_ref($self->refname.'^0')) {
2052 push @tmp, $cur;
2053 }
2054 if (my $ipd = $self->{inject_parents_dcommit}) {
2055 if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2056 push @tmp, @$commit;
2057 }
2058 }
2059 push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2060 while (my $p = shift @tmp) {
2061 next if $seen{$p};
2062 $seen{$p} = 1;
2063 push @ret, $p;
2064 # MAXPARENT is defined to 16 in commit-tree.c:
2065 last if @ret >= 16;
2066 }
2067 if (@tmp) {
2068 die "r$log_entry->{revision}: No room for parents:\n\t",
2069 join("\n\t", @tmp), "\n";
2070 }
2071 @ret;
2072 }
2074 sub rewrite_root {
2075 my ($self) = @_;
2076 return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2077 my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2078 my $rwr = eval { command_oneline(qw/config --get/, $k) };
2079 if ($rwr) {
2080 $rwr =~ s#/+$##;
2081 if ($rwr !~ m#^[a-z\+]+://#) {
2082 die "$rwr is not a valid URL (key: $k)\n";
2083 }
2084 }
2085 $self->{-rewrite_root} = $rwr;
2086 }
2088 sub metadata_url {
2089 my ($self) = @_;
2090 ($self->rewrite_root || $self->{url}) .
2091 (length $self->{path} ? '/' . $self->{path} : '');
2092 }
2094 sub full_url {
2095 my ($self) = @_;
2096 $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2097 }
2100 sub set_commit_header_env {
2101 my ($log_entry) = @_;
2102 my %env;
2103 foreach my $ned (qw/NAME EMAIL DATE/) {
2104 foreach my $ac (qw/AUTHOR COMMITTER/) {
2105 $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2106 }
2107 }
2109 $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2110 $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2111 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2113 $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2114 ? $log_entry->{commit_name}
2115 : $log_entry->{name};
2116 $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2117 ? $log_entry->{commit_email}
2118 : $log_entry->{email};
2119 \%env;
2120 }
2122 sub restore_commit_header_env {
2123 my ($env) = @_;
2124 foreach my $ned (qw/NAME EMAIL DATE/) {
2125 foreach my $ac (qw/AUTHOR COMMITTER/) {
2126 my $k = "GIT_${ac}_${ned}";
2127 if (defined $env->{$k}) {
2128 $ENV{$k} = $env->{$k};
2129 } else {
2130 delete $ENV{$k};
2131 }
2132 }
2133 }
2134 }
2136 sub gc {
2137 command_noisy('gc', '--auto');
2138 };
2140 sub do_git_commit {
2141 my ($self, $log_entry) = @_;
2142 my $lr = $self->last_rev;
2143 if (defined $lr && $lr >= $log_entry->{revision}) {
2144 die "Last fetched revision of ", $self->refname,
2145 " was r$lr, but we are about to fetch: ",
2146 "r$log_entry->{revision}!\n";
2147 }
2148 if (my $c = $self->rev_map_get($log_entry->{revision})) {
2149 croak "$log_entry->{revision} = $c already exists! ",
2150 "Why are we refetching it?\n";
2151 }
2152 my $old_env = set_commit_header_env($log_entry);
2153 my $tree = $log_entry->{tree};
2154 if (!defined $tree) {
2155 $tree = $self->tmp_index_do(sub {
2156 command_oneline('write-tree') });
2157 }
2158 die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2160 my @exec = ('git-commit-tree', $tree);
2161 foreach ($self->get_commit_parents($log_entry)) {
2162 push @exec, '-p', $_;
2163 }
2164 defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2165 or croak $!;
2166 print $msg_fh $log_entry->{log} or croak $!;
2167 restore_commit_header_env($old_env);
2168 unless ($self->no_metadata) {
2169 print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2170 or croak $!;
2171 }
2172 $msg_fh->flush == 0 or croak $!;
2173 close $msg_fh or croak $!;
2174 chomp(my $commit = do { local $/; <$out_fh> });
2175 close $out_fh or croak $!;
2176 waitpid $pid, 0;
2177 croak $? if $?;
2178 if ($commit !~ /^$::sha1$/o) {
2179 die "Failed to commit, invalid sha1: $commit\n";
2180 }
2182 $self->rev_map_set($log_entry->{revision}, $commit, 1);
2184 $self->{last_rev} = $log_entry->{revision};
2185 $self->{last_commit} = $commit;
2186 print "r$log_entry->{revision}";
2187 if (defined $log_entry->{svm_revision}) {
2188 print " (\@$log_entry->{svm_revision})";
2189 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2190 0, $self->svm_uuid);
2191 }
2192 print " = $commit ($self->{ref_id})\n";
2193 if (--$_gc_nr == 0) {
2194 $_gc_nr = $_gc_period;
2195 gc();
2196 }
2197 return $commit;
2198 }
2200 sub match_paths {
2201 my ($self, $paths, $r) = @_;
2202 return 1 if $self->{path} eq '';
2203 if (my $path = $paths->{"/$self->{path}"}) {
2204 return ($path->{action} eq 'D') ? 0 : 1;
2205 }
2206 $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
2207 if (grep /$self->{path_regex}/, keys %$paths) {
2208 return 1;
2209 }
2210 my $c = '';
2211 foreach (split m#/#, $self->{path}) {
2212 $c .= "/$_";
2213 next unless ($paths->{$c} &&
2214 ($paths->{$c}->{action} =~ /^[AR]$/));
2215 if ($self->ra->check_path($self->{path}, $r) ==
2216 $SVN::Node::dir) {
2217 return 1;
2218 }
2219 }
2220 return 0;
2221 }
2223 sub find_parent_branch {
2224 my ($self, $paths, $rev) = @_;
2225 return undef unless $self->follow_parent;
2226 unless (defined $paths) {
2227 my $err_handler = $SVN::Error::handler;
2228 $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2229 $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1, sub {
2230 $paths =
2231 Git::SVN::Ra::dup_changed_paths($_[0]) });
2232 $SVN::Error::handler = $err_handler;
2233 }
2234 return undef unless defined $paths;
2236 # look for a parent from another branch:
2237 my @b_path_components = split m#/#, $self->rel_path;
2238 my @a_path_components;
2239 my $i;
2240 while (@b_path_components) {
2241 $i = $paths->{'/'.join('/', @b_path_components)};
2242 last if $i && defined $i->{copyfrom_path};
2243 unshift(@a_path_components, pop(@b_path_components));
2244 }
2245 return undef unless defined $i && defined $i->{copyfrom_path};
2246 my $branch_from = $i->{copyfrom_path};
2247 if (@a_path_components) {
2248 print STDERR "branch_from: $branch_from => ";
2249 $branch_from .= '/'.join('/', @a_path_components);
2250 print STDERR $branch_from, "\n";
2251 }
2252 my $r = $i->{copyfrom_rev};
2253 my $repos_root = $self->ra->{repos_root};
2254 my $url = $self->ra->{url};
2255 my $new_url = $repos_root . $branch_from;
2256 print STDERR "Found possible branch point: ",
2257 "$new_url => ", $self->full_url, ", $r\n";
2258 $branch_from =~ s#^/##;
2259 my $gs = Git::SVN->find_by_url($new_url, $repos_root, $branch_from);
2260 unless ($gs) {
2261 my $ref_id = $self->{ref_id};
2262 $ref_id =~ s/\@\d+$//;
2263 $ref_id .= "\@$r";
2264 # just grow a tail if we're not unique enough :x
2265 $ref_id .= '-' while find_ref($ref_id);
2266 print STDERR "Initializing parent: $ref_id\n";
2267 my ($u, $p, $repo_id) = ($new_url, '', $ref_id);
2268 if ($u =~ s#^\Q$url\E(/|$)##) {
2269 $p = $u;
2270 $u = $url;
2271 $repo_id = $self->{repo_id};
2272 }
2273 $gs = Git::SVN->init($u, $p, $repo_id, $ref_id, 1);
2274 }
2275 my ($r0, $parent) = $gs->find_rev_before($r, 1);
2276 if (!defined $r0 || !defined $parent) {
2277 my ($base, $head) = parse_revision_argument(0, $r);
2278 if ($base <= $r) {
2279 $gs->fetch($base, $r);
2280 }
2281 ($r0, $parent) = $gs->last_rev_commit;
2282 }
2283 if (defined $r0 && defined $parent) {
2284 print STDERR "Found branch parent: ($self->{ref_id}) $parent\n";
2285 my $ed;
2286 if ($self->ra->can_do_switch) {
2287 $self->assert_index_clean($parent);
2288 print STDERR "Following parent with do_switch\n";
2289 # do_switch works with svn/trunk >= r22312, but that
2290 # is not included with SVN 1.4.3 (the latest version
2291 # at the moment), so we can't rely on it
2292 $self->{last_commit} = $parent;
2293 $ed = SVN::Git::Fetcher->new($self);
2294 $gs->ra->gs_do_switch($r0, $rev, $gs,
2295 $self->full_url, $ed)
2296 or die "SVN connection failed somewhere...\n";
2297 } elsif ($self->ra->trees_match($new_url, $r0,
2298 $self->full_url, $rev)) {
2299 print STDERR "Trees match:\n",
2300 " $new_url\@$r0\n",
2301 " ${\$self->full_url}\@$rev\n",
2302 "Following parent with no changes\n";
2303 $self->tmp_index_do(sub {
2304 command_noisy('read-tree', $parent);
2305 });
2306 $self->{last_commit} = $parent;
2307 } else {
2308 print STDERR "Following parent with do_update\n";
2309 $ed = SVN::Git::Fetcher->new($self);
2310 $self->ra->gs_do_update($rev, $rev, $self, $ed)
2311 or die "SVN connection failed somewhere...\n";
2312 }
2313 print STDERR "Successfully followed parent\n";
2314 return $self->make_log_entry($rev, [$parent], $ed);
2315 }
2316 return undef;
2317 }
2319 sub do_fetch {
2320 my ($self, $paths, $rev) = @_;
2321 my $ed;
2322 my ($last_rev, @parents);
2323 if (my $lc = $self->last_commit) {
2324 # we can have a branch that was deleted, then re-added
2325 # under the same name but copied from another path, in
2326 # which case we'll have multiple parents (we don't
2327 # want to break the original ref, nor lose copypath info):
2328 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2329 push @{$log_entry->{parents}}, $lc;
2330 return $log_entry;
2331 }
2332 $ed = SVN::Git::Fetcher->new($self);
2333 $last_rev = $self->{last_rev};
2334 $ed->{c} = $lc;
2335 @parents = ($lc);
2336 } else {
2337 $last_rev = $rev;
2338 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2339 return $log_entry;
2340 }
2341 $ed = SVN::Git::Fetcher->new($self);
2342 }
2343 unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2344 die "SVN connection failed somewhere...\n";
2345 }
2346 $self->make_log_entry($rev, \@parents, $ed);
2347 }
2349 sub get_untracked {
2350 my ($self, $ed) = @_;
2351 my @out;
2352 my $h = $ed->{empty};
2353 foreach (sort keys %$h) {
2354 my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2355 push @out, " $act: " . uri_encode($_);
2356 warn "W: $act: $_\n";
2357 }
2358 foreach my $t (qw/dir_prop file_prop/) {
2359 $h = $ed->{$t} or next;
2360 foreach my $path (sort keys %$h) {
2361 my $ppath = $path eq '' ? '.' : $path;
2362 foreach my $prop (sort keys %{$h->{$path}}) {
2363 next if $SKIP_PROP{$prop};
2364 my $v = $h->{$path}->{$prop};
2365 my $t_ppath_prop = "$t: " .
2366 uri_encode($ppath) . ' ' .
2367 uri_encode($prop);
2368 if (defined $v) {
2369 push @out, " +$t_ppath_prop " .
2370 uri_encode($v);
2371 } else {
2372 push @out, " -$t_ppath_prop";
2373 }
2374 }
2375 }
2376 }
2377 foreach my $t (qw/absent_file absent_directory/) {
2378 $h = $ed->{$t} or next;
2379 foreach my $parent (sort keys %$h) {
2380 foreach my $path (sort @{$h->{$parent}}) {
2381 push @out, " $t: " .
2382 uri_encode("$parent/$path");
2383 warn "W: $t: $parent/$path ",
2384 "Insufficient permissions?\n";
2385 }
2386 }
2387 }
2388 \@out;
2389 }
2391 sub parse_svn_date {
2392 my $date = shift || return '+0000 1970-01-01 00:00:00';
2393 my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2394 (\d\d)\:(\d\d)\:(\d\d).\d+Z$/x) or
2395 croak "Unable to parse date: $date\n";
2396 "+0000 $Y-$m-$d $H:$M:$S";
2397 }
2399 sub check_author {
2400 my ($author) = @_;
2401 if (!defined $author || length $author == 0) {
2402 $author = '(no author)';
2403 } elsif (defined $::_authors && ! defined $::users{$author}) {
2404 die "Author: $author not defined in $::_authors file\n";
2405 }
2406 $author;
2407 }
2409 sub make_log_entry {
2410 my ($self, $rev, $parents, $ed) = @_;
2411 my $untracked = $self->get_untracked($ed);
2413 open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
2414 print $un "r$rev\n" or croak $!;
2415 print $un $_, "\n" foreach @$untracked;
2416 my %log_entry = ( parents => $parents || [], revision => $rev,
2417 log => '');
2419 my $headrev;
2420 my $logged = delete $self->{logged_rev_props};
2421 if (!$logged || $self->{-want_revprops}) {
2422 my $rp = $self->ra->rev_proplist($rev);
2423 foreach (sort keys %$rp) {
2424 my $v = $rp->{$_};
2425 if (/^svn:(author|date|log)$/) {
2426 $log_entry{$1} = $v;
2427 } elsif ($_ eq 'svm:headrev') {
2428 $headrev = $v;
2429 } else {
2430 print $un " rev_prop: ", uri_encode($_), ' ',
2431 uri_encode($v), "\n";
2432 }
2433 }
2434 } else {
2435 map { $log_entry{$_} = $logged->{$_} } keys %$logged;
2436 }
2437 close $un or croak $!;
2439 $log_entry{date} = parse_svn_date($log_entry{date});
2440 $log_entry{log} .= "\n";
2441 my $author = $log_entry{author} = check_author($log_entry{author});
2442 my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
2443 : ($author, undef);
2445 my ($commit_name, $commit_email) = ($name, $email);
2446 if ($_use_log_author) {
2447 my $name_field;
2448 if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
2449 $name_field = $1;
2450 } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
2451 $name_field = $1;
2452 }
2453 if (!defined $name_field) {
2454 if (!defined $email) {
2455 $email = $name;
2456 }
2457 } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
2458 ($name, $email) = ($1, $2);
2459 } elsif ($name_field =~ /(.*)@/) {
2460 ($name, $email) = ($1, $name_field);
2461 } else {
2462 ($name, $email) = ($name_field, $name_field);
2463 }
2464 }
2465 if (defined $headrev && $self->use_svm_props) {
2466 if ($self->rewrite_root) {
2467 die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
2468 "options set!\n";
2469 }
2470 my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$};
2471 # we don't want "SVM: initializing mirror for junk" ...
2472 return undef if $r == 0;
2473 my $svm = $self->svm;
2474 if ($uuid ne $svm->{uuid}) {
2475 die "UUID mismatch on SVM path:\n",
2476 "expected: $svm->{uuid}\n",
2477 " got: $uuid\n";
2478 }
2479 my $full_url = $self->full_url;
2480 $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
2481 die "Failed to replace '$svm->{replace}' with ",
2482 "'$svm->{source}' in $full_url\n";
2483 # throw away username for storing in records
2484 remove_username($full_url);
2485 $log_entry{metadata} = "$full_url\@$r $uuid";
2486 $log_entry{svm_revision} = $r;
2487 $email ||= "$author\@$uuid";
2488 $commit_email ||= "$author\@$uuid";
2489 } elsif ($self->use_svnsync_props) {
2490 my $full_url = $self->svnsync->{url};
2491 $full_url .= "/$self->{path}" if length $self->{path};
2492 remove_username($full_url);
2493 my $uuid = $self->svnsync->{uuid};
2494 $log_entry{metadata} = "$full_url\@$rev $uuid";
2495 $email ||= "$author\@$uuid";
2496 $commit_email ||= "$author\@$uuid";
2497 } else {
2498 my $url = $self->metadata_url;
2499 remove_username($url);
2500 $log_entry{metadata} = "$url\@$rev " .
2501 $self->ra->get_uuid;
2502 $email ||= "$author\@" . $self->ra->get_uuid;
2503 $commit_email ||= "$author\@" . $self->ra->get_uuid;
2504 }
2505 $log_entry{name} = $name;
2506 $log_entry{email} = $email;
2507 $log_entry{commit_name} = $commit_name;
2508 $log_entry{commit_email} = $commit_email;
2509 \%log_entry;
2510 }
2512 sub fetch {
2513 my ($self, $min_rev, $max_rev, @parents) = @_;
2514 my ($last_rev, $last_commit) = $self->last_rev_commit;
2515 my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
2516 $self->ra->gs_fetch_loop_common($base, $head, [$self]);
2517 }
2519 sub set_tree_cb {
2520 my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
2521 $self->{inject_parents} = { $rev => $tree };
2522 $self->fetch(undef, undef);
2523 }
2525 sub set_tree {
2526 my ($self, $tree) = (shift, shift);
2527 my $log_entry = ::get_commit_entry($tree);
2528 unless ($self->{last_rev}) {
2529 fatal("Must have an existing revision to commit");
2530 }
2531 my %ed_opts = ( r => $self->{last_rev},
2532 log => $log_entry->{log},
2533 ra => $self->ra,
2534 tree_a => $self->{last_commit},
2535 tree_b => $tree,
2536 editor_cb => sub {
2537 $self->set_tree_cb($log_entry, $tree, @_) },
2538 svn_path => $self->{path} );
2539 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
2540 print "No changes\nr$self->{last_rev} = $tree\n";
2541 }
2542 }
2544 sub rebuild_from_rev_db {
2545 my ($self, $path) = @_;
2546 my $r = -1;
2547 open my $fh, '<', $path or croak "open: $!";
2548 binmode $fh or croak "binmode: $!";
2549 while (<$fh>) {
2550 length($_) == 41 or croak "inconsistent size in ($_) != 41";
2551 chomp($_);
2552 ++$r;
2553 next if $_ eq ('0' x 40);
2554 $self->rev_map_set($r, $_);
2555 print "r$r = $_\n";
2556 }
2557 close $fh or croak "close: $!";
2558 unlink $path or croak "unlink: $!";
2559 }
2561 sub rebuild {
2562 my ($self) = @_;
2563 my $map_path = $self->map_path;
2564 return if (-e $map_path && ! -z $map_path);
2565 return unless ::verify_ref($self->refname.'^0');
2566 if ($self->use_svm_props || $self->no_metadata) {
2567 my $rev_db = $self->rev_db_path;
2568 $self->rebuild_from_rev_db($rev_db);
2569 if ($self->use_svm_props) {
2570 my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
2571 $self->rebuild_from_rev_db($svm_rev_db);
2572 }
2573 $self->unlink_rev_db_symlink;
2574 return;
2575 }
2576 print "Rebuilding $map_path ...\n";
2577 my ($log, $ctx) =
2578 command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
2579 $self->refname, '--');
2580 my $full_url = $self->full_url;
2581 remove_username($full_url);
2582 my $svn_uuid = $self->ra_uuid;
2583 my $c;
2584 while (<$log>) {
2585 if ( m{^commit ($::sha1)$} ) {
2586 $c = $1;
2587 next;
2588 }
2589 next unless s{^\s*(git-svn-id:)}{$1};
2590 my ($url, $rev, $uuid) = ::extract_metadata($_);
2591 remove_username($url);
2593 # ignore merges (from set-tree)
2594 next if (!defined $rev || !$uuid);
2596 # if we merged or otherwise started elsewhere, this is
2597 # how we break out of it
2598 if (($uuid ne $svn_uuid) ||
2599 ($full_url && $url && ($url ne $full_url))) {
2600 next;
2601 }
2603 $self->rev_map_set($rev, $c);
2604 print "r$rev = $c\n";
2605 }
2606 command_close_pipe($log, $ctx);
2607 print "Done rebuilding $map_path\n";
2608 my $rev_db_path = $self->rev_db_path;
2609 if (-f $self->rev_db_path) {
2610 unlink $self->rev_db_path or croak "unlink: $!";
2611 }
2612 $self->unlink_rev_db_symlink;
2613 }
2615 # rev_map:
2616 # Tie::File seems to be prone to offset errors if revisions get sparse,
2617 # it's not that fast, either. Tie::File is also not in Perl 5.6. So
2618 # one of my favorite modules is out :< Next up would be one of the DBM
2619 # modules, but I'm not sure which is most portable...
2620 #
2621 # This is the replacement for the rev_db format, which was too big
2622 # and inefficient for large repositories with a lot of sparse history
2623 # (mainly tags)
2624 #
2625 # The format is this:
2626 # - 24 bytes for every record,
2627 # * 4 bytes for the integer representing an SVN revision number
2628 # * 20 bytes representing the sha1 of a git commit
2629 # - No empty padding records like the old format
2630 # (except the last record, which can be overwritten)
2631 # - new records are written append-only since SVN revision numbers
2632 # increase monotonically
2633 # - lookups on SVN revision number are done via a binary search
2634 # - Piping the file to xxd -c24 is a good way of dumping it for
2635 # viewing or editing (piped back through xxd -r), should the need
2636 # ever arise.
2637 # - The last record can be padding revision with an all-zero sha1
2638 # This is used to optimize fetch performance when using multiple
2639 # "fetch" directives in .git/config
2640 #
2641 # These files are disposable unless noMetadata or useSvmProps is set
2643 sub _rev_map_set {
2644 my ($fh, $rev, $commit) = @_;
2646 binmode $fh or croak "binmode: $!";
2647 my $size = (stat($fh))[7];
2648 ($size % 24) == 0 or croak "inconsistent size: $size";
2650 my $wr_offset = 0;
2651 if ($size > 0) {
2652 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2653 my $read = sysread($fh, my $buf, 24) or croak "read: $!";
2654 $read == 24 or croak "read only $read bytes (!= 24)";
2655 my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
2656 if ($last_commit eq ('0' x40)) {
2657 if ($size >= 48) {
2658 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2659 $read = sysread($fh, $buf, 24) or
2660 croak "read: $!";
2661 $read == 24 or
2662 croak "read only $read bytes (!= 24)";
2663 ($last_rev, $last_commit) =
2664 unpack(rev_map_fmt, $buf);
2665 if ($last_commit eq ('0' x40)) {
2666 croak "inconsistent .rev_map\n";
2667 }
2668 }
2669 if ($last_rev >= $rev) {
2670 croak "last_rev is higher!: $last_rev >= $rev";
2671 }
2672 $wr_offset = -24;
2673 }
2674 }
2675 sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
2676 syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
2677 croak "write: $!";
2678 }
2680 sub mkfile {
2681 my ($path) = @_;
2682 unless (-e $path) {
2683 my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
2684 mkpath([$dir]) unless -d $dir;
2685 open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
2686 close $fh or die "Couldn't close (create) $path: $!\n";
2687 }
2688 }
2690 sub rev_map_set {
2691 my ($self, $rev, $commit, $update_ref, $uuid) = @_;
2692 length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
2693 my $db = $self->map_path($uuid);
2694 my $db_lock = "$db.lock";
2695 my $sig;
2696 if ($update_ref) {
2697 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2698 $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
2699 }
2700 mkfile($db);
2702 $LOCKFILES{$db_lock} = 1;
2703 my $sync;
2704 # both of these options make our .rev_db file very, very important
2705 # and we can't afford to lose it because rebuild() won't work
2706 if ($self->use_svm_props || $self->no_metadata) {
2707 $sync = 1;
2708 copy($db, $db_lock) or die "rev_map_set(@_): ",
2709 "Failed to copy: ",
2710 "$db => $db_lock ($!)\n";
2711 } else {
2712 rename $db, $db_lock or die "rev_map_set(@_): ",
2713 "Failed to rename: ",
2714 "$db => $db_lock ($!)\n";
2715 }
2717 sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
2718 or croak "Couldn't open $db_lock: $!\n";
2719 _rev_map_set($fh, $rev, $commit);
2720 if ($sync) {
2721 $fh->flush or die "Couldn't flush $db_lock: $!\n";
2722 $fh->sync or die "Couldn't sync $db_lock: $!\n";
2723 }
2724 close $fh or croak $!;
2725 if ($update_ref) {
2726 $_head = $self;
2727 command_noisy('update-ref', '-m', "r$rev",
2728 $self->refname, $commit);
2729 }
2730 rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
2731 "$db_lock => $db ($!)\n";
2732 delete $LOCKFILES{$db_lock};
2733 if ($update_ref) {
2734 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2735 $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
2736 kill $sig, $$ if defined $sig;
2737 }
2738 }
2740 # If want_commit, this will return an array of (rev, commit) where
2741 # commit _must_ be a valid commit in the archive.
2742 # Otherwise, it'll return the max revision (whether or not the
2743 # commit is valid or just a 0x40 placeholder).
2744 sub rev_map_max {
2745 my ($self, $want_commit) = @_;
2746 $self->rebuild;
2747 my $map_path = $self->map_path;
2748 stat $map_path or return $want_commit ? (0, undef) : 0;
2749 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2750 binmode $fh or croak "binmode: $!";
2751 my $size = (stat($fh))[7];
2752 ($size % 24) == 0 or croak "inconsistent size: $size";
2754 if ($size == 0) {
2755 close $fh or croak "close: $!";
2756 return $want_commit ? (0, undef) : 0;
2757 }
2759 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2760 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2761 my ($r, $c) = unpack(rev_map_fmt, $buf);
2762 if ($want_commit && $c eq ('0' x40)) {
2763 if ($size < 48) {
2764 return $want_commit ? (0, undef) : 0;
2765 }
2766 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2767 sysread($fh, $buf, 24) == 24 or croak "read: $!";
2768 ($r, $c) = unpack(rev_map_fmt, $buf);
2769 if ($c eq ('0'x40)) {
2770 croak "Penultimate record is all-zeroes in $map_path";
2771 }
2772 }
2773 close $fh or croak "close: $!";
2774 $want_commit ? ($r, $c) : $r;
2775 }
2777 sub rev_map_get {
2778 my ($self, $rev, $uuid) = @_;
2779 my $map_path = $self->map_path($uuid);
2780 return undef unless -e $map_path;
2782 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2783 binmode $fh or croak "binmode: $!";
2784 my $size = (stat($fh))[7];
2785 ($size % 24) == 0 or croak "inconsistent size: $size";
2787 if ($size == 0) {
2788 close $fh or croak "close: $fh";
2789 return undef;
2790 }
2792 my ($l, $u) = (0, $size - 24);
2793 my ($r, $c, $buf);
2795 while ($l <= $u) {
2796 my $i = int(($l/24 + $u/24) / 2) * 24;
2797 sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
2798 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2799 my ($r, $c) = unpack('NH40', $buf);
2801 if ($r < $rev) {
2802 $l = $i + 24;
2803 } elsif ($r > $rev) {
2804 $u = $i - 24;
2805 } else { # $r == $rev
2806 close($fh) or croak "close: $!";
2807 return $c eq ('0' x 40) ? undef : $c;
2808 }
2809 }
2810 close($fh) or croak "close: $!";
2811 undef;
2812 }
2814 # Finds the first svn revision that exists on (if $eq_ok is true) or
2815 # before $rev for the current branch. It will not search any lower
2816 # than $min_rev. Returns the git commit hash and svn revision number
2817 # if found, else (undef, undef).
2818 sub find_rev_before {
2819 my ($self, $rev, $eq_ok, $min_rev) = @_;
2820 --$rev unless $eq_ok;
2821 $min_rev ||= 1;
2822 while ($rev >= $min_rev) {
2823 if (my $c = $self->rev_map_get($rev)) {
2824 return ($rev, $c);
2825 }
2826 --$rev;
2827 }
2828 return (undef, undef);
2829 }
2831 # Finds the first svn revision that exists on (if $eq_ok is true) or
2832 # after $rev for the current branch. It will not search any higher
2833 # than $max_rev. Returns the git commit hash and svn revision number
2834 # if found, else (undef, undef).
2835 sub find_rev_after {
2836 my ($self, $rev, $eq_ok, $max_rev) = @_;
2837 ++$rev unless $eq_ok;
2838 $max_rev ||= $self->rev_map_max;
2839 while ($rev <= $max_rev) {
2840 if (my $c = $self->rev_map_get($rev)) {
2841 return ($rev, $c);
2842 }
2843 ++$rev;
2844 }
2845 return (undef, undef);
2846 }
2848 sub _new {
2849 my ($class, $repo_id, $ref_id, $path) = @_;
2850 unless (defined $repo_id && length $repo_id) {
2851 $repo_id = $Git::SVN::default_repo_id;
2852 }
2853 unless (defined $ref_id && length $ref_id) {
2854 $_[2] = $ref_id = $Git::SVN::default_ref_id;
2855 }
2856 $_[1] = $repo_id = sanitize_remote_name($repo_id);
2857 my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
2858 $_[3] = $path = '' unless (defined $path);
2859 mkpath(["$ENV{GIT_DIR}/svn"]);
2860 bless {
2861 ref_id => $ref_id, dir => $dir, index => "$dir/index",
2862 path => $path, config => "$ENV{GIT_DIR}/svn/config",
2863 map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
2864 }
2866 # for read-only access of old .rev_db formats
2867 sub unlink_rev_db_symlink {
2868 my ($self) = @_;
2869 my $link = $self->rev_db_path;
2870 $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
2871 if (-l $link) {
2872 unlink $link or croak "unlink: $link failed!";
2873 }
2874 }
2876 sub rev_db_path {
2877 my ($self, $uuid) = @_;
2878 my $db_path = $self->map_path($uuid);
2879 $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
2880 or croak "map_path: $db_path does not contain '/.rev_map.' !";
2881 $db_path;
2882 }
2884 # the new replacement for .rev_db
2885 sub map_path {
2886 my ($self, $uuid) = @_;
2887 $uuid ||= $self->ra_uuid;
2888 "$self->{map_root}.$uuid";
2889 }
2891 sub uri_encode {
2892 my ($f) = @_;
2893 $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
2894 $f
2895 }
2897 sub remove_username {
2898 $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
2899 }
2901 package Git::SVN::Prompt;
2902 use strict;
2903 use warnings;
2904 require SVN::Core;
2905 use vars qw/$_no_auth_cache $_username/;
2907 sub simple {
2908 my ($cred, $realm, $default_username, $may_save, $pool) = @_;
2909 $may_save = undef if $_no_auth_cache;
2910 $default_username = $_username if defined $_username;
2911 if (defined $default_username && length $default_username) {
2912 if (defined $realm && length $realm) {
2913 print STDERR "Authentication realm: $realm\n";
2914 STDERR->flush;
2915 }
2916 $cred->username($default_username);
2917 } else {
2918 username($cred, $realm, $may_save, $pool);
2919 }
2920 $cred->password(_read_password("Password for '" .
2921 $cred->username . "': ", $realm));
2922 $cred->may_save($may_save);
2923 $SVN::_Core::SVN_NO_ERROR;
2924 }
2926 sub ssl_server_trust {
2927 my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
2928 $may_save = undef if $_no_auth_cache;
2929 print STDERR "Error validating server certificate for '$realm':\n";
2930 {
2931 no warnings 'once';
2932 # All variables SVN::Auth::SSL::* are used only once,
2933 # so we're shutting up Perl warnings about this.
2934 if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
2935 print STDERR " - The certificate is not issued ",
2936 "by a trusted authority. Use the\n",
2937 " fingerprint to validate ",
2938 "the certificate manually!\n";
2939 }
2940 if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
2941 print STDERR " - The certificate hostname ",
2942 "does not match.\n";
2943 }
2944 if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
2945 print STDERR " - The certificate is not yet valid.\n";
2946 }
2947 if ($failures & $SVN::Auth::SSL::EXPIRED) {
2948 print STDERR " - The certificate has expired.\n";
2949 }
2950 if ($failures & $SVN::Auth::SSL::OTHER) {
2951 print STDERR " - The certificate has ",
2952 "an unknown error.\n";
2953 }
2954 } # no warnings 'once'
2955 printf STDERR
2956 "Certificate information:\n".
2957 " - Hostname: %s\n".
2958 " - Valid: from %s until %s\n".
2959 " - Issuer: %s\n".
2960 " - Fingerprint: %s\n",
2961 map $cert_info->$_, qw(hostname valid_from valid_until
2962 issuer_dname fingerprint);
2963 my $choice;
2964 prompt:
2965 print STDERR $may_save ?
2966 "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
2967 "(R)eject or accept (t)emporarily? ";
2968 STDERR->flush;
2969 $choice = lc(substr(<STDIN> || 'R', 0, 1));
2970 if ($choice =~ /^t$/i) {
2971 $cred->may_save(undef);
2972 } elsif ($choice =~ /^r$/i) {
2973 return -1;
2974 } elsif ($may_save && $choice =~ /^p$/i) {
2975 $cred->may_save($may_save);
2976 } else {
2977 goto prompt;
2978 }
2979 $cred->accepted_failures($failures);
2980 $SVN::_Core::SVN_NO_ERROR;
2981 }
2983 sub ssl_client_cert {
2984 my ($cred, $realm, $may_save, $pool) = @_;
2985 $may_save = undef if $_no_auth_cache;
2986 print STDERR "Client certificate filename: ";
2987 STDERR->flush;
2988 chomp(my $filename = <STDIN>);
2989 $cred->cert_file($filename);
2990 $cred->may_save($may_save);
2991 $SVN::_Core::SVN_NO_ERROR;
2992 }
2994 sub ssl_client_cert_pw {
2995 my ($cred, $realm, $may_save, $pool) = @_;
2996 $may_save = undef if $_no_auth_cache;
2997 $cred->password(_read_password("Password: ", $realm));
2998 $cred->may_save($may_save);
2999 $SVN::_Core::SVN_NO_ERROR;
3000 }
3002 sub username {
3003 my ($cred, $realm, $may_save, $pool) = @_;
3004 $may_save = undef if $_no_auth_cache;
3005 if (defined $realm && length $realm) {
3006 print STDERR "Authentication realm: $realm\n";
3007 }
3008 my $username;
3009 if (defined $_username) {
3010 $username = $_username;
3011 } else {
3012 print STDERR "Username: ";
3013 STDERR->flush;
3014 chomp($username = <STDIN>);
3015 }
3016 $cred->username($username);
3017 $cred->may_save($may_save);
3018 $SVN::_Core::SVN_NO_ERROR;
3019 }
3021 sub _read_password {
3022 my ($prompt, $realm) = @_;
3023 print STDERR $prompt;
3024 STDERR->flush;
3025 require Term::ReadKey;
3026 Term::ReadKey::ReadMode('noecho');
3027 my $password = '';
3028 while (defined(my $key = Term::ReadKey::ReadKey(0))) {
3029 last if $key =~ /[\012\015]/; # \n\r
3030 $password .= $key;
3031 }
3032 Term::ReadKey::ReadMode('restore');
3033 print STDERR "\n";
3034 STDERR->flush;
3035 $password;
3036 }
3038 package SVN::Git::Fetcher;
3039 use vars qw/@ISA/;
3040 use strict;
3041 use warnings;
3042 use Carp qw/croak/;
3043 use File::Temp qw/tempfile/;
3044 use IO::File qw//;
3046 # file baton members: path, mode_a, mode_b, pool, fh, blob, base
3047 sub new {
3048 my ($class, $git_svn) = @_;
3049 my $self = SVN::Delta::Editor->new;
3050 bless $self, $class;
3051 $self->{c} = $git_svn->{last_commit} if exists $git_svn->{last_commit};
3052 $self->{empty} = {};
3053 $self->{dir_prop} = {};
3054 $self->{file_prop} = {};
3055 $self->{absent_dir} = {};
3056 $self->{absent_file} = {};
3057 $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
3058 $self;
3059 }
3061 sub set_path_strip {
3062 my ($self, $path) = @_;
3063 $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
3064 }
3066 sub open_root {
3067 { path => '' };
3068 }
3070 sub open_directory {
3071 my ($self, $path, $pb, $rev) = @_;
3072 { path => $path };
3073 }
3075 sub git_path {
3076 my ($self, $path) = @_;
3077 if ($self->{path_strip}) {
3078 $path =~ s!$self->{path_strip}!! or
3079 die "Failed to strip path '$path' ($self->{path_strip})\n";
3080 }
3081 $path;
3082 }
3084 sub delete_entry {
3085 my ($self, $path, $rev, $pb) = @_;
3087 my $gpath = $self->git_path($path);
3088 return undef if ($gpath eq '');
3090 # remove entire directories.
3091 if (command('ls-tree', $self->{c}, '--', $gpath) =~ /^040000 tree/) {
3092 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3093 -r --name-only -z/,
3094 $self->{c}, '--', $gpath);
3095 local $/ = "\0";
3096 while (<$ls>) {
3097 chomp;
3098 $self->{gii}->remove($_);
3099 print "\tD\t$_\n" unless $::_q;
3100 }
3101 print "\tD\t$gpath/\n" unless $::_q;
3102 command_close_pipe($ls, $ctx);
3103 $self->{empty}->{$path} = 0
3104 } else {
3105 $self->{gii}->remove($gpath);
3106 print "\tD\t$gpath\n" unless $::_q;
3107 }
3108 undef;
3109 }
3111 sub open_file {
3112 my ($self, $path, $pb, $rev) = @_;
3113 my $gpath = $self->git_path($path);
3114 my ($mode, $blob) = (command('ls-tree', $self->{c}, '--', $gpath)
3115 =~ /^(\d{6}) blob ([a-f\d]{40})\t/);
3116 unless (defined $mode && defined $blob) {
3117 die "$path was not found in commit $self->{c} (r$rev)\n";
3118 }
3119 { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3120 pool => SVN::Pool->new, action => 'M' };
3121 }
3123 sub add_file {
3124 my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3125 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3126 delete $self->{empty}->{$dir};
3127 { path => $path, mode_a => 100644, mode_b => 100644,
3128 pool => SVN::Pool->new, action => 'A' };
3129 }
3131 sub add_directory {
3132 my ($self, $path, $cp_path, $cp_rev) = @_;
3133 my $gpath = $self->git_path($path);
3134 if ($gpath eq '') {
3135 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3136 -r --name-only -z/,
3137 $self->{c});
3138 local $/ = "\0";
3139 while (<$ls>) {
3140 chomp;
3141 $self->{gii}->remove($_);
3142 print "\tD\t$_\n" unless $::_q;
3143 }
3144 command_close_pipe($ls, $ctx);
3145 $self->{empty}->{$path} = 0;
3146 }
3147 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3148 delete $self->{empty}->{$dir};
3149 $self->{empty}->{$path} = 1;
3150 { path => $path };
3151 }
3153 sub change_dir_prop {
3154 my ($self, $db, $prop, $value) = @_;
3155 $self->{dir_prop}->{$db->{path}} ||= {};
3156 $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3157 undef;
3158 }
3160 sub absent_directory {
3161 my ($self, $path, $pb) = @_;
3162 $self->{absent_dir}->{$pb->{path}} ||= [];
3163 push @{$self->{absent_dir}->{$pb->{path}}}, $path;
3164 undef;
3165 }
3167 sub absent_file {
3168 my ($self, $path, $pb) = @_;
3169 $self->{absent_file}->{$pb->{path}} ||= [];
3170 push @{$self->{absent_file}->{$pb->{path}}}, $path;
3171 undef;
3172 }
3174 sub change_file_prop {
3175 my ($self, $fb, $prop, $value) = @_;
3176 if ($prop eq 'svn:executable') {
3177 if ($fb->{mode_b} != 120000) {
3178 $fb->{mode_b} = defined $value ? 100755 : 100644;
3179 }
3180 } elsif ($prop eq 'svn:special') {
3181 $fb->{mode_b} = defined $value ? 120000 : 100644;
3182 } else {
3183 $self->{file_prop}->{$fb->{path}} ||= {};
3184 $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
3185 }
3186 undef;
3187 }
3189 sub apply_textdelta {
3190 my ($self, $fb, $exp) = @_;
3191 my $fh = IO::File->new_tmpfile;
3192 $fh->autoflush(1);
3193 # $fh gets auto-closed() by SVN::TxDelta::apply(),
3194 # (but $base does not,) so dup() it for reading in close_file
3195 open my $dup, '<&', $fh or croak $!;
3196 my $base = IO::File->new_tmpfile;
3197 $base->autoflush(1);
3198 if ($fb->{blob}) {
3199 print $base 'link ' if ($fb->{mode_a} == 120000);
3200 my $size = $::_repository->cat_blob($fb->{blob}, $base);
3201 die "Failed to read object $fb->{blob}" if ($size < 0);
3203 if (defined $exp) {
3204 seek $base, 0, 0 or croak $!;
3205 my $got = ::md5sum($base);
3206 die "Checksum mismatch: $fb->{path} $fb->{blob}\n",
3207 "expected: $exp\n",
3208 " got: $got\n" if ($got ne $exp);
3209 }
3210 }
3211 seek $base, 0, 0 or croak $!;
3212 $fb->{fh} = $dup;
3213 $fb->{base} = $base;
3214 [ SVN::TxDelta::apply($base, $fh, undef, $fb->{path}, $fb->{pool}) ];
3215 }
3217 sub close_file {
3218 my ($self, $fb, $exp) = @_;
3219 my $hash;
3220 my $path = $self->git_path($fb->{path});
3221 if (my $fh = $fb->{fh}) {
3222 if (defined $exp) {
3223 seek($fh, 0, 0) or croak $!;
3224 my $got = ::md5sum($fh);
3225 if ($got ne $exp) {
3226 die "Checksum mismatch: $path\n",
3227 "expected: $exp\n got: $got\n";
3228 }
3229 }
3230 sysseek($fh, 0, 0) or croak $!;
3231 if ($fb->{mode_b} == 120000) {
3232 eval {
3233 sysread($fh, my $buf, 5) == 5 or croak $!;
3234 $buf eq 'link ' or die "$path has mode 120000",
3235 " but is not a link";
3236 };
3237 if ($@) {
3238 warn "$@\n";
3239 sysseek($fh, 0, 0) or croak $!;
3240 }
3241 }
3243 my ($tmp_fh, $tmp_filename) = File::Temp::tempfile(UNLINK => 1);
3244 my $result;
3245 while ($result = sysread($fh, my $string, 1024)) {
3246 syswrite($tmp_fh, $string, $result);
3247 }
3248 defined $result or croak $!;
3249 close $tmp_fh or croak $!;
3251 close $fh or croak $!;
3253 $hash = $::_repository->hash_and_insert_object($tmp_filename);
3254 $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
3255 close $fb->{base} or croak $!;
3256 } else {
3257 $hash = $fb->{blob} or die "no blob information\n";
3258 }
3259 $fb->{pool}->clear;
3260 $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
3261 print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
3262 undef;
3263 }
3265 sub abort_edit {
3266 my $self = shift;
3267 $self->{nr} = $self->{gii}->{nr};
3268 delete $self->{gii};
3269 $self->SUPER::abort_edit(@_);
3270 }
3272 sub close_edit {
3273 my $self = shift;
3274 $self->{git_commit_ok} = 1;
3275 $self->{nr} = $self->{gii}->{nr};
3276 delete $self->{gii};
3277 $self->SUPER::close_edit(@_);
3278 }
3280 package SVN::Git::Editor;
3281 use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
3282 use strict;
3283 use warnings;
3284 use Carp qw/croak/;
3285 use IO::File;
3287 sub new {
3288 my ($class, $opts) = @_;
3289 foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
3290 die "$_ required!\n" unless (defined $opts->{$_});
3291 }
3293 my $pool = SVN::Pool->new;
3294 my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
3295 my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
3296 $opts->{r}, $mods);
3298 # $opts->{ra} functions should not be used after this:
3299 my @ce = $opts->{ra}->get_commit_editor($opts->{log},
3300 $opts->{editor_cb}, $pool);
3301 my $self = SVN::Delta::Editor->new(@ce, $pool);
3302 bless $self, $class;
3303 foreach (qw/svn_path r tree_a tree_b/) {
3304 $self->{$_} = $opts->{$_};
3305 }
3306 $self->{url} = $opts->{ra}->{url};
3307 $self->{mods} = $mods;
3308 $self->{types} = $types;
3309 $self->{pool} = $pool;
3310 $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
3311 $self->{rm} = { };
3312 $self->{path_prefix} = length $self->{svn_path} ?
3313 "$self->{svn_path}/" : '';
3314 return $self;
3315 }
3317 sub generate_diff {
3318 my ($tree_a, $tree_b) = @_;
3319 my @diff_tree = qw(diff-tree -z -r);
3320 if ($_cp_similarity) {
3321 push @diff_tree, "-C$_cp_similarity";
3322 } else {
3323 push @diff_tree, '-C';
3324 }
3325 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
3326 push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
3327 push @diff_tree, $tree_a, $tree_b;
3328 my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
3329 local $/ = "\0";
3330 my $state = 'meta';
3331 my @mods;
3332 while (<$diff_fh>) {
3333 chomp $_; # this gets rid of the trailing "\0"
3334 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
3335 $::sha1\s($::sha1)\s
3336 ([MTCRAD])\d*$/xo) {
3337 push @mods, { mode_a => $1, mode_b => $2,
3338 sha1_b => $3, chg => $4 };
3339 if ($4 =~ /^(?:C|R)$/) {
3340 $state = 'file_a';
3341 } else {
3342 $state = 'file_b';
3343 }
3344 } elsif ($state eq 'file_a') {
3345 my $x = $mods[$#mods] or croak "Empty array\n";
3346 if ($x->{chg} !~ /^(?:C|R)$/) {
3347 croak "Error parsing $_, $x->{chg}\n";
3348 }
3349 $x->{file_a} = $_;
3350 $state = 'file_b';
3351 } elsif ($state eq 'file_b') {
3352 my $x = $mods[$#mods] or croak "Empty array\n";
3353 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
3354 croak "Error parsing $_, $x->{chg}\n";
3355 }
3356 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
3357 croak "Error parsing $_, $x->{chg}\n";
3358 }
3359 $x->{file_b} = $_;
3360 $state = 'meta';
3361 } else {
3362 croak "Error parsing $_\n";
3363 }
3364 }
3365 command_close_pipe($diff_fh, $ctx);
3366 \@mods;
3367 }
3369 sub check_diff_paths {
3370 my ($ra, $pfx, $rev, $mods) = @_;
3371 my %types;
3372 $pfx .= '/' if length $pfx;
3374 sub type_diff_paths {
3375 my ($ra, $types, $path, $rev) = @_;
3376 my @p = split m#/+#, $path;
3377 my $c = shift @p;
3378 unless (defined $types->{$c}) {
3379 $types->{$c} = $ra->check_path($c, $rev);
3380 }
3381 while (@p) {
3382 $c .= '/' . shift @p;
3383 next if defined $types->{$c};
3384 $types->{$c} = $ra->check_path($c, $rev);
3385 }
3386 }
3388 foreach my $m (@$mods) {
3389 foreach my $f (qw/file_a file_b/) {
3390 next unless defined $m->{$f};
3391 my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
3392 if (length $pfx.$dir && ! defined $types{$dir}) {
3393 type_diff_paths($ra, \%types, $pfx.$dir, $rev);
3394 }
3395 }
3396 }
3397 \%types;
3398 }
3400 sub split_path {
3401 return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
3402 }
3404 sub repo_path {
3405 my ($self, $path) = @_;
3406 $self->{path_prefix}.(defined $path ? $path : '');
3407 }
3409 sub url_path {
3410 my ($self, $path) = @_;
3411 if ($self->{url} =~ m#^https?://#) {
3412 $path =~ s/([^a-zA-Z0-9_.-])/uc sprintf("%%%02x",ord($1))/eg;
3413 }
3414 $self->{url} . '/' . $self->repo_path($path);
3415 }
3417 sub rmdirs {
3418 my ($self) = @_;
3419 my $rm = $self->{rm};
3420 delete $rm->{''}; # we never delete the url we're tracking
3421 return unless %$rm;
3423 foreach (keys %$rm) {
3424 my @d = split m#/#, $_;
3425 my $c = shift @d;
3426 $rm->{$c} = 1;
3427 while (@d) {
3428 $c .= '/' . shift @d;
3429 $rm->{$c} = 1;
3430 }
3431 }
3432 delete $rm->{$self->{svn_path}};
3433 delete $rm->{''}; # we never delete the url we're tracking
3434 return unless %$rm;
3436 my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
3437 $self->{tree_b});
3438 local $/ = "\0";
3439 while (<$fh>) {
3440 chomp;
3441 my @dn = split m#/#, $_;
3442 while (pop @dn) {
3443 delete $rm->{join '/', @dn};
3444 }
3445 unless (%$rm) {
3446 close $fh;
3447 return;
3448 }
3449 }
3450 command_close_pipe($fh, $ctx);
3452 my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
3453 foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
3454 $self->close_directory($bat->{$d}, $p);
3455 my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
3456 print "\tD+\t$d/\n" unless $::_q;
3457 $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
3458 delete $bat->{$d};
3459 }
3460 }
3462 sub open_or_add_dir {
3463 my ($self, $full_path, $baton) = @_;
3464 my $t = $self->{types}->{$full_path};
3465 if (!defined $t) {
3466 die "$full_path not known in r$self->{r} or we have a bug!\n";
3467 }
3468 {
3469 no warnings 'once';
3470 # SVN::Node::none and SVN::Node::file are used only once,
3471 # so we're shutting up Perl's warnings about them.
3472 if ($t == $SVN::Node::none) {
3473 return $self->add_directory($full_path, $baton,
3474 undef, -1, $self->{pool});
3475 } elsif ($t == $SVN::Node::dir) {
3476 return $self->open_directory($full_path, $baton,
3477 $self->{r}, $self->{pool});
3478 } # no warnings 'once'
3479 print STDERR "$full_path already exists in repository at ",
3480 "r$self->{r} and it is not a directory (",
3481 ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
3482 } # no warnings 'once'
3483 exit 1;
3484 }
3486 sub ensure_path {
3487 my ($self, $path) = @_;
3488 my $bat = $self->{bat};
3489 my $repo_path = $self->repo_path($path);
3490 return $bat->{''} unless (length $repo_path);
3491 my @p = split m#/+#, $repo_path;
3492 my $c = shift @p;
3493 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
3494 while (@p) {
3495 my $c0 = $c;
3496 $c .= '/' . shift @p;
3497 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
3498 }
3499 return $bat->{$c};
3500 }
3502 sub A {
3503 my ($self, $m) = @_;
3504 my ($dir, $file) = split_path($m->{file_b});
3505 my $pbat = $self->ensure_path($dir);
3506 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3507 undef, -1);
3508 print "\tA\t$m->{file_b}\n" unless $::_q;
3509 $self->chg_file($fbat, $m);
3510 $self->close_file($fbat,undef,$self->{pool});
3511 }
3513 sub C {
3514 my ($self, $m) = @_;
3515 my ($dir, $file) = split_path($m->{file_b});
3516 my $pbat = $self->ensure_path($dir);
3517 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3518 $self->url_path($m->{file_a}), $self->{r});
3519 print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3520 $self->chg_file($fbat, $m);
3521 $self->close_file($fbat,undef,$self->{pool});
3522 }
3524 sub delete_entry {
3525 my ($self, $path, $pbat) = @_;
3526 my $rpath = $self->repo_path($path);
3527 my ($dir, $file) = split_path($rpath);
3528 $self->{rm}->{$dir} = 1;
3529 $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
3530 }
3532 sub R {
3533 my ($self, $m) = @_;
3534 my ($dir, $file) = split_path($m->{file_b});
3535 my $pbat = $self->ensure_path($dir);
3536 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3537 $self->url_path($m->{file_a}), $self->{r});
3538 print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3539 $self->chg_file($fbat, $m);
3540 $self->close_file($fbat,undef,$self->{pool});
3542 ($dir, $file) = split_path($m->{file_a});
3543 $pbat = $self->ensure_path($dir);
3544 $self->delete_entry($m->{file_a}, $pbat);
3545 }
3547 sub M {
3548 my ($self, $m) = @_;
3549 my ($dir, $file) = split_path($m->{file_b});
3550 my $pbat = $self->ensure_path($dir);
3551 my $fbat = $self->open_file($self->repo_path($m->{file_b}),
3552 $pbat,$self->{r},$self->{pool});
3553 print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
3554 $self->chg_file($fbat, $m);
3555 $self->close_file($fbat,undef,$self->{pool});
3556 }
3558 sub T { shift->M(@_) }
3560 sub change_file_prop {
3561 my ($self, $fbat, $pname, $pval) = @_;
3562 $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
3563 }
3565 sub chg_file {
3566 my ($self, $fbat, $m) = @_;
3567 if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
3568 $self->change_file_prop($fbat,'svn:executable','*');
3569 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
3570 $self->change_file_prop($fbat,'svn:executable',undef);
3571 }
3572 my $fh = IO::File->new_tmpfile or croak $!;
3573 if ($m->{mode_b} =~ /^120/) {
3574 print $fh 'link ' or croak $!;
3575 $self->change_file_prop($fbat,'svn:special','*');
3576 } elsif ($m->{mode_a} =~ /^120/ && $m->{mode_b} !~ /^120/) {
3577 $self->change_file_prop($fbat,'svn:special',undef);
3578 }
3579 my $size = $::_repository->cat_blob($m->{sha1_b}, $fh);
3580 croak "Failed to read object $m->{sha1_b}" if ($size < 0);
3581 $fh->flush == 0 or croak $!;
3582 seek $fh, 0, 0 or croak $!;
3584 my $exp = ::md5sum($fh);
3585 seek $fh, 0, 0 or croak $!;
3587 my $pool = SVN::Pool->new;
3588 my $atd = $self->apply_textdelta($fbat, undef, $pool);
3589 my $got = SVN::TxDelta::send_stream($fh, @$atd, $pool);
3590 die "Checksum mismatch\nexpected: $exp\ngot: $got\n" if ($got ne $exp);
3591 $pool->clear;
3593 close $fh or croak $!;
3594 }
3596 sub D {
3597 my ($self, $m) = @_;
3598 my ($dir, $file) = split_path($m->{file_b});
3599 my $pbat = $self->ensure_path($dir);
3600 print "\tD\t$m->{file_b}\n" unless $::_q;
3601 $self->delete_entry($m->{file_b}, $pbat);
3602 }
3604 sub close_edit {
3605 my ($self) = @_;
3606 my ($p,$bat) = ($self->{pool}, $self->{bat});
3607 foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
3608 next if $_ eq '';
3609 $self->close_directory($bat->{$_}, $p);
3610 }
3611 $self->close_directory($bat->{''}, $p);
3612 $self->SUPER::close_edit($p);
3613 $p->clear;
3614 }
3616 sub abort_edit {
3617 my ($self) = @_;
3618 $self->SUPER::abort_edit($self->{pool});
3619 }
3621 sub DESTROY {
3622 my $self = shift;
3623 $self->SUPER::DESTROY(@_);
3624 $self->{pool}->clear;
3625 }
3627 # this drives the editor
3628 sub apply_diff {
3629 my ($self) = @_;
3630 my $mods = $self->{mods};
3631 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
3632 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
3633 my $f = $m->{chg};
3634 if (defined $o{$f}) {
3635 $self->$f($m);
3636 } else {
3637 fatal("Invalid change type: $f");
3638 }
3639 }
3640 $self->rmdirs if $_rmdir;
3641 if (@$mods == 0) {
3642 $self->abort_edit;
3643 } else {
3644 $self->close_edit;
3645 }
3646 return scalar @$mods;
3647 }
3649 package Git::SVN::Ra;
3650 use vars qw/@ISA $config_dir $_log_window_size/;
3651 use strict;
3652 use warnings;
3653 my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
3655 BEGIN {
3656 # enforce temporary pool usage for some simple functions
3657 no strict 'refs';
3658 for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root/) {
3659 my $SUPER = "SUPER::$f";
3660 *$f = sub {
3661 my $self = shift;
3662 my $pool = SVN::Pool->new;
3663 my @ret = $self->$SUPER(@_,$pool);
3664 $pool->clear;
3665 wantarray ? @ret : $ret[0];
3666 };
3667 }
3668 }
3670 sub _auth_providers () {
3671 [
3672 SVN::Client::get_simple_provider(),
3673 SVN::Client::get_ssl_server_trust_file_provider(),
3674 SVN::Client::get_simple_prompt_provider(
3675 \&Git::SVN::Prompt::simple, 2),
3676 SVN::Client::get_ssl_client_cert_file_provider(),
3677 SVN::Client::get_ssl_client_cert_prompt_provider(
3678 \&Git::SVN::Prompt::ssl_client_cert, 2),
3679 SVN::Client::get_ssl_client_cert_pw_file_provider(),
3680 SVN::Client::get_ssl_client_cert_pw_prompt_provider(
3681 \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
3682 SVN::Client::get_username_provider(),
3683 SVN::Client::get_ssl_server_trust_prompt_provider(
3684 \&Git::SVN::Prompt::ssl_server_trust),
3685 SVN::Client::get_username_prompt_provider(
3686 \&Git::SVN::Prompt::username, 2)
3687 ]
3688 }
3690 sub escape_uri_only {
3691 my ($uri) = @_;
3692 my @tmp;
3693 foreach (split m{/}, $uri) {
3694 s/([^\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
3695 push @tmp, $_;
3696 }
3697 join('/', @tmp);
3698 }
3700 sub escape_url {
3701 my ($url) = @_;
3702 if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
3703 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
3704 $url = "$scheme://$domain$uri";
3705 }
3706 $url;
3707 }
3709 sub new {
3710 my ($class, $url) = @_;
3711 $url =~ s!/+$!!;
3712 return $RA if ($RA && $RA->{url} eq $url);
3714 SVN::_Core::svn_config_ensure($config_dir, undef);
3715 my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
3716 my $config = SVN::Core::config_get_config($config_dir);
3717 $RA = undef;
3718 my $dont_store_passwords = 1;
3719 my $conf_t = ${$config}{'config'};
3720 {
3721 no warnings 'once';
3722 # The usage of $SVN::_Core::SVN_CONFIG_* variables
3723 # produces warnings that variables are used only once.
3724 # I had not found the better way to shut them up, so
3725 # the warnings of type 'once' are disabled in this block.
3726 if (SVN::_Core::svn_config_get_bool($conf_t,
3727 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3728 $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
3729 1) == 0) {
3730 SVN::_Core::svn_auth_set_parameter($baton,
3731 $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
3732 bless (\$dont_store_passwords, "_p_void"));
3733 }
3734 if (SVN::_Core::svn_config_get_bool($conf_t,
3735 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3736 $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
3737 1) == 0) {
3738 $Git::SVN::Prompt::_no_auth_cache = 1;
3739 }
3740 } # no warnings 'once'
3741 my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
3742 config => $config,
3743 pool => SVN::Pool->new,
3744 auth_provider_callbacks => $callbacks);
3745 $self->{url} = $url;
3746 $self->{svn_path} = $url;
3747 $self->{repos_root} = $self->get_repos_root;
3748 $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
3749 $self->{cache} = { check_path => { r => 0, data => {} },
3750 get_dir => { r => 0, data => {} } };
3751 $RA = bless $self, $class;
3752 }
3754 sub check_path {
3755 my ($self, $path, $r) = @_;
3756 my $cache = $self->{cache}->{check_path};
3757 if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
3758 return $cache->{data}->{$path};
3759 }
3760 my $pool = SVN::Pool->new;
3761 my $t = $self->SUPER::check_path($path, $r, $pool);
3762 $pool->clear;
3763 if ($r != $cache->{r}) {
3764 %{$cache->{data}} = ();
3765 $cache->{r} = $r;
3766 }
3767 $cache->{data}->{$path} = $t;
3768 }
3770 sub get_dir {
3771 my ($self, $dir, $r) = @_;
3772 my $cache = $self->{cache}->{get_dir};
3773 if ($r == $cache->{r}) {
3774 if (my $x = $cache->{data}->{$dir}) {
3775 return wantarray ? @$x : $x->[0];
3776 }
3777 }
3778 my $pool = SVN::Pool->new;
3779 my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
3780 my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
3781 $pool->clear;
3782 if ($r != $cache->{r}) {
3783 %{$cache->{data}} = ();
3784 $cache->{r} = $r;
3785 }
3786 $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
3787 wantarray ? (\%dirents, $r, $props) : \%dirents;
3788 }
3790 sub DESTROY {
3791 # do not call the real DESTROY since we store ourselves in $RA
3792 }
3794 sub get_log {
3795 my ($self, @args) = @_;
3796 my $pool = SVN::Pool->new;
3797 splice(@args, 3, 1) if ($SVN::Core::VERSION le '1.2.0');
3798 my $ret = $self->SUPER::get_log(@args, $pool);
3799 $pool->clear;
3800 $ret;
3801 }
3803 sub trees_match {
3804 my ($self, $url1, $rev1, $url2, $rev2) = @_;
3805 my $ctx = SVN::Client->new(auth => _auth_providers);
3806 my $out = IO::File->new_tmpfile;
3808 # older SVN (1.1.x) doesn't take $pool as the last parameter for
3809 # $ctx->diff(), so we'll create a default one
3810 my $pool = SVN::Pool->new_default_sub;
3812 $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
3813 $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
3814 $out->flush;
3815 my $ret = (($out->stat)[7] == 0);
3816 close $out or croak $!;
3818 $ret;
3819 }
3821 sub get_commit_editor {
3822 my ($self, $log, $cb, $pool) = @_;
3823 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
3824 $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
3825 }
3827 sub gs_do_update {
3828 my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
3829 my $new = ($rev_a == $rev_b);
3830 my $path = $gs->{path};
3832 if ($new && -e $gs->{index}) {
3833 unlink $gs->{index} or die
3834 "Couldn't unlink index: $gs->{index}: $!\n";
3835 }
3836 my $pool = SVN::Pool->new;
3837 $editor->set_path_strip($path);
3838 my (@pc) = split m#/#, $path;
3839 my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
3840 1, $editor, $pool);
3841 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3843 # Since we can't rely on svn_ra_reparent being available, we'll
3844 # just have to do some magic with set_path to make it so
3845 # we only want a partial path.
3846 my $sp = '';
3847 my $final = join('/', @pc);
3848 while (@pc) {
3849 $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
3850 $sp .= '/' if length $sp;
3851 $sp .= shift @pc;
3852 }
3853 die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
3855 $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
3857 $reporter->finish_report($pool);
3858 $pool->clear;
3859 $editor->{git_commit_ok};
3860 }
3862 # this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
3863 # svn_ra_reparent didn't work before 1.4)
3864 sub gs_do_switch {
3865 my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
3866 my $path = $gs->{path};
3867 my $pool = SVN::Pool->new;
3869 my $full_url = $self->{url};
3870 my $old_url = $full_url;
3871 $full_url .= '/' . escape_uri_only($path) if length $path;
3872 my ($ra, $reparented);
3873 if ($old_url ne $full_url) {
3874 if ($old_url !~ m#^svn(\+ssh)?://#) {
3875 SVN::_Ra::svn_ra_reparent($self->{session}, $full_url,
3876 $pool);
3877 $self->{url} = $full_url;
3878 $reparented = 1;
3879 } else {
3880 $_[0] = undef;
3881 $self = undef;
3882 $RA = undef;
3883 $ra = Git::SVN::Ra->new($full_url);
3884 $ra_invalid = 1;
3885 }
3886 }
3887 $ra ||= $self;
3888 my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
3889 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3890 $reporter->set_path('', $rev_a, 0, @lock, $pool);
3891 $reporter->finish_report($pool);
3893 if ($reparented) {
3894 SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
3895 $self->{url} = $old_url;
3896 }
3898 $pool->clear;
3899 $editor->{git_commit_ok};
3900 }
3902 sub longest_common_path {
3903 my ($gsv, $globs) = @_;
3904 my %common;
3905 my $common_max = scalar @$gsv;
3907 foreach my $gs (@$gsv) {
3908 my @tmp = split m#/#, $gs->{path};
3909 my $p = '';
3910 foreach (@tmp) {
3911 $p .= length($p) ? "/$_" : $_;
3912 $common{$p} ||= 0;
3913 $common{$p}++;
3914 }
3915 }
3916 $globs ||= [];
3917 $common_max += scalar @$globs;
3918 foreach my $glob (@$globs) {
3919 my @tmp = split m#/#, $glob->{path}->{left};
3920 my $p = '';
3921 foreach (@tmp) {
3922 $p .= length($p) ? "/$_" : $_;
3923 $common{$p} ||= 0;
3924 $common{$p}++;
3925 }
3926 }
3928 my $longest_path = '';
3929 foreach (sort {length $b <=> length $a} keys %common) {
3930 if ($common{$_} == $common_max) {
3931 $longest_path = $_;
3932 last;
3933 }
3934 }
3935 $longest_path;
3936 }
3938 sub gs_fetch_loop_common {
3939 my ($self, $base, $head, $gsv, $globs) = @_;
3940 return if ($base > $head);
3941 my $inc = $_log_window_size;
3942 my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
3943 my $longest_path = longest_common_path($gsv, $globs);
3944 my $ra_url = $self->{url};
3945 while (1) {
3946 my %revs;
3947 my $err;
3948 my $err_handler = $SVN::Error::handler;
3949 $SVN::Error::handler = sub {
3950 ($err) = @_;
3951 skip_unknown_revs($err);
3952 };
3953 sub _cb {
3954 my ($paths, $r, $author, $date, $log) = @_;
3955 [ dup_changed_paths($paths),
3956 { author => $author, date => $date, log => $log } ];
3957 }
3958 $self->get_log([$longest_path], $min, $max, 0, 1, 1,
3959 sub { $revs{$_[1]} = _cb(@_) });
3960 if ($err && $max >= $head) {
3961 print STDERR "Path '$longest_path' ",
3962 "was probably deleted:\n",
3963 $err->expanded_message,
3964 "\nWill attempt to follow ",
3965 "revisions r$min .. r$max ",
3966 "committed before the deletion\n";
3967 my $hi = $max;
3968 while (--$hi >= $min) {
3969 my $ok;
3970 $self->get_log([$longest_path], $min, $hi,
3971 0, 1, 1, sub {
3972 $ok ||= $_[1];
3973 $revs{$_[1]} = _cb(@_) });
3974 if ($ok) {
3975 print STDERR "r$min .. r$ok OK\n";
3976 last;
3977 }
3978 }
3979 }
3980 $SVN::Error::handler = $err_handler;
3982 my %exists = map { $_->{path} => $_ } @$gsv;
3983 foreach my $r (sort {$a <=> $b} keys %revs) {
3984 my ($paths, $logged) = @{$revs{$r}};
3986 foreach my $gs ($self->match_globs(\%exists, $paths,
3987 $globs, $r)) {
3988 if ($gs->rev_map_max >= $r) {
3989 next;
3990 }
3991 next unless $gs->match_paths($paths, $r);
3992 $gs->{logged_rev_props} = $logged;
3993 if (my $last_commit = $gs->last_commit) {
3994 $gs->assert_index_clean($last_commit);
3995 }
3996 my $log_entry = $gs->do_fetch($paths, $r);
3997 if ($log_entry) {
3998 $gs->do_git_commit($log_entry);
3999 }
4000 $INDEX_FILES{$gs->{index}} = 1;
4001 }
4002 foreach my $g (@$globs) {
4003 my $k = "svn-remote.$g->{remote}." .
4004 "$g->{t}-maxRev";
4005 Git::SVN::tmp_config($k, $r);
4006 }
4007 if ($ra_invalid) {
4008 $_[0] = undef;
4009 $self = undef;
4010 $RA = undef;
4011 $self = Git::SVN::Ra->new($ra_url);
4012 $ra_invalid = undef;
4013 }
4014 }
4015 # pre-fill the .rev_db since it'll eventually get filled in
4016 # with '0' x40 if something new gets committed
4017 foreach my $gs (@$gsv) {
4018 next if $gs->rev_map_max >= $max;
4019 next if defined $gs->rev_map_get($max);
4020 $gs->rev_map_set($max, 0 x40);
4021 }
4022 foreach my $g (@$globs) {
4023 my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
4024 Git::SVN::tmp_config($k, $max);
4025 }
4026 last if $max >= $head;
4027 $min = $max + 1;
4028 $max += $inc;
4029 $max = $head if ($max > $head);
4030 }
4031 Git::SVN::gc();
4032 }
4034 sub match_globs {
4035 my ($self, $exists, $paths, $globs, $r) = @_;
4037 sub get_dir_check {
4038 my ($self, $exists, $g, $r) = @_;
4039 my @x = eval { $self->get_dir($g->{path}->{left}, $r) };
4040 return unless scalar @x == 3;
4041 my $dirents = $x[0];
4042 foreach my $de (keys %$dirents) {
4043 next if $dirents->{$de}->{kind} != $SVN::Node::dir;
4044 my $p = $g->{path}->full_path($de);
4045 next if $exists->{$p};
4046 next if (length $g->{path}->{right} &&
4047 ($self->check_path($p, $r) !=
4048 $SVN::Node::dir));
4049 $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
4050 $g->{ref}->full_path($de), 1);
4051 }
4052 }
4053 foreach my $g (@$globs) {
4054 if (my $path = $paths->{"/$g->{path}->{left}"}) {
4055 if ($path->{action} =~ /^[AR]$/) {
4056 get_dir_check($self, $exists, $g, $r);
4057 }
4058 }
4059 foreach (keys %$paths) {
4060 if (/$g->{path}->{left_regex}/ &&
4061 !/$g->{path}->{regex}/) {
4062 next if $paths->{$_}->{action} !~ /^[AR]$/;
4063 get_dir_check($self, $exists, $g, $r);
4064 }
4065 next unless /$g->{path}->{regex}/;
4066 my $p = $1;
4067 my $pathname = $g->{path}->full_path($p);
4068 next if $exists->{$pathname};
4069 next if ($self->check_path($pathname, $r) !=
4070 $SVN::Node::dir);
4071 $exists->{$pathname} = Git::SVN->init(
4072 $self->{url}, $pathname, undef,
4073 $g->{ref}->full_path($p), 1);
4074 }
4075 my $c = '';
4076 foreach (split m#/#, $g->{path}->{left}) {
4077 $c .= "/$_";
4078 next unless ($paths->{$c} &&
4079 ($paths->{$c}->{action} =~ /^[AR]$/));
4080 get_dir_check($self, $exists, $g, $r);
4081 }
4082 }
4083 values %$exists;
4084 }
4086 sub minimize_url {
4087 my ($self) = @_;
4088 return $self->{url} if ($self->{url} eq $self->{repos_root});
4089 my $url = $self->{repos_root};
4090 my @components = split(m!/!, $self->{svn_path});
4091 my $c = '';
4092 do {
4093 $url .= "/$c" if length $c;
4094 eval { (ref $self)->new($url)->get_latest_revnum };
4095 } while ($@ && ($c = shift @components));
4096 $url;
4097 }
4099 sub can_do_switch {
4100 my $self = shift;
4101 unless (defined $can_do_switch) {
4102 my $pool = SVN::Pool->new;
4103 my $rep = eval {
4104 $self->do_switch(1, '', 0, $self->{url},
4105 SVN::Delta::Editor->new, $pool);
4106 };
4107 if ($@) {
4108 $can_do_switch = 0;
4109 } else {
4110 $rep->abort_report($pool);
4111 $can_do_switch = 1;
4112 }
4113 $pool->clear;
4114 }
4115 $can_do_switch;
4116 }
4118 sub skip_unknown_revs {
4119 my ($err) = @_;
4120 my $errno = $err->apr_err();
4121 # Maybe the branch we're tracking didn't
4122 # exist when the repo started, so it's
4123 # not an error if it doesn't, just continue
4124 #
4125 # Wonderfully consistent library, eh?
4126 # 160013 - svn:// and file://
4127 # 175002 - http(s)://
4128 # 175007 - http(s):// (this repo required authorization, too...)
4129 # More codes may be discovered later...
4130 if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
4131 my $err_key = $err->expanded_message;
4132 # revision numbers change every time, filter them out
4133 $err_key =~ s/\d+/\0/g;
4134 $err_key = "$errno\0$err_key";
4135 unless ($ignored_err{$err_key}) {
4136 warn "W: Ignoring error from SVN, path probably ",
4137 "does not exist: ($errno): ",
4138 $err->expanded_message,"\n";
4139 warn "W: Do not be alarmed at the above message ",
4140 "git-svn is just searching aggressively for ",
4141 "old history.\n",
4142 "This may take a while on large repositories\n";
4143 $ignored_err{$err_key} = 1;
4144 }
4145 return;
4146 }
4147 die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
4148 }
4150 # svn_log_changed_path_t objects passed to get_log are likely to be
4151 # overwritten even if only the refs are copied to an external variable,
4152 # so we should dup the structures in their entirety. Using an externally
4153 # passed pool (instead of our temporary and quickly cleared pool in
4154 # Git::SVN::Ra) does not help matters at all...
4155 sub dup_changed_paths {
4156 my ($paths) = @_;
4157 return undef unless $paths;
4158 my %ret;
4159 foreach my $p (keys %$paths) {
4160 my $i = $paths->{$p};
4161 my %s = map { $_ => $i->$_ }
4162 qw/copyfrom_path copyfrom_rev action/;
4163 $ret{$p} = \%s;
4164 }
4165 \%ret;
4166 }
4168 package Git::SVN::Log;
4169 use strict;
4170 use warnings;
4171 use POSIX qw/strftime/;
4172 use constant commit_log_separator => ('-' x 72) . "\n";
4173 use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
4174 %rusers $show_commit $incremental/;
4175 my $l_fmt;
4177 sub cmt_showable {
4178 my ($c) = @_;
4179 return 1 if defined $c->{r};
4181 # big commit message got truncated by the 16k pretty buffer in rev-list
4182 if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
4183 $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
4184 @{$c->{l}} = ();
4185 my @log = command(qw/cat-file commit/, $c->{c});
4187 # shift off the headers
4188 shift @log while ($log[0] ne '');
4189 shift @log;
4191 # TODO: make $c->{l} not have a trailing newline in the future
4192 @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
4194 (undef, $c->{r}, undef) = ::extract_metadata(
4195 (grep(/^git-svn-id: /, @log))[-1]);
4196 }
4197 return defined $c->{r};
4198 }
4200 sub log_use_color {
4201 return $color || Git->repository->get_colorbool('color.diff');
4202 }
4204 sub git_svn_log_cmd {
4205 my ($r_min, $r_max, @args) = @_;
4206 my $head = 'HEAD';
4207 my (@files, @log_opts);
4208 foreach my $x (@args) {
4209 if ($x eq '--' || @files) {
4210 push @files, $x;
4211 } else {
4212 if (::verify_ref("$x^0")) {
4213 $head = $x;
4214 } else {
4215 push @log_opts, $x;
4216 }
4217 }
4218 }
4220 my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
4221 $gs ||= Git::SVN->_new;
4222 my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
4223 $gs->refname);
4224 push @cmd, '-r' unless $non_recursive;
4225 push @cmd, qw/--raw --name-status/ if $verbose;
4226 push @cmd, '--color' if log_use_color();
4227 push @cmd, @log_opts;
4228 if (defined $r_max && $r_max == $r_min) {
4229 push @cmd, '--max-count=1';
4230 if (my $c = $gs->rev_map_get($r_max)) {
4231 push @cmd, $c;
4232 }
4233 } elsif (defined $r_max) {
4234 if ($r_max < $r_min) {
4235 ($r_min, $r_max) = ($r_max, $r_min);
4236 }
4237 my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
4238 my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
4239 # If there are no commits in the range, both $c_max and $c_min
4240 # will be undefined. If there is at least 1 commit in the
4241 # range, both will be defined.
4242 return () if !defined $c_min || !defined $c_max;
4243 if ($c_min eq $c_max) {
4244 push @cmd, '--max-count=1', $c_min;
4245 } else {
4246 push @cmd, '--boundary', "$c_min..$c_max";
4247 }
4248 }
4249 return (@cmd, @files);
4250 }
4252 # adapted from pager.c
4253 sub config_pager {
4254 $pager ||= $ENV{GIT_PAGER} || $ENV{PAGER};
4255 if (!defined $pager) {
4256 $pager = 'less';
4257 } elsif (length $pager == 0 || $pager eq 'cat') {
4258 $pager = undef;
4259 }
4260 $ENV{GIT_PAGER_IN_USE} = defined($pager);
4261 }
4263 sub run_pager {
4264 return unless -t *STDOUT && defined $pager;
4265 pipe my $rfd, my $wfd or return;
4266 defined(my $pid = fork) or ::fatal "Can't fork: $!";
4267 if (!$pid) {
4268 open STDOUT, '>&', $wfd or
4269 ::fatal "Can't redirect to stdout: $!";
4270 return;
4271 }
4272 open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
4273 $ENV{LESS} ||= 'FRSX';
4274 exec $pager or ::fatal "Can't run pager: $! ($pager)";
4275 }
4277 sub format_svn_date {
4278 return strftime("%Y-%m-%d %H:%M:%S %z (%a, %d %b %Y)", localtime(shift));
4279 }
4281 sub parse_git_date {
4282 my ($t, $tz) = @_;
4283 # Date::Parse isn't in the standard Perl distro :(
4284 if ($tz =~ s/^\+//) {
4285 $t += tz_to_s_offset($tz);
4286 } elsif ($tz =~ s/^\-//) {
4287 $t -= tz_to_s_offset($tz);
4288 }
4289 return $t;
4290 }
4292 sub set_local_timezone {
4293 if (defined $TZ) {
4294 $ENV{TZ} = $TZ;
4295 } else {
4296 delete $ENV{TZ};
4297 }
4298 }
4300 sub tz_to_s_offset {
4301 my ($tz) = @_;
4302 $tz =~ s/(\d\d)$//;
4303 return ($1 * 60) + ($tz * 3600);
4304 }
4306 sub get_author_info {
4307 my ($dest, $author, $t, $tz) = @_;
4308 $author =~ s/(?:^\s*|\s*$)//g;
4309 $dest->{a_raw} = $author;
4310 my $au;
4311 if ($::_authors) {
4312 $au = $rusers{$author} || undef;
4313 }
4314 if (!$au) {
4315 ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
4316 }
4317 $dest->{t} = $t;
4318 $dest->{tz} = $tz;
4319 $dest->{a} = $au;
4320 $dest->{t_utc} = parse_git_date($t, $tz);
4321 }
4323 sub process_commit {
4324 my ($c, $r_min, $r_max, $defer) = @_;
4325 if (defined $r_min && defined $r_max) {
4326 if ($r_min == $c->{r} && $r_min == $r_max) {
4327 show_commit($c);
4328 return 0;
4329 }
4330 return 1 if $r_min == $r_max;
4331 if ($r_min < $r_max) {
4332 # we need to reverse the print order
4333 return 0 if (defined $limit && --$limit < 0);
4334 push @$defer, $c;
4335 return 1;
4336 }
4337 if ($r_min != $r_max) {
4338 return 1 if ($r_min < $c->{r});
4339 return 1 if ($r_max > $c->{r});
4340 }
4341 }
4342 return 0 if (defined $limit && --$limit < 0);
4343 show_commit($c);
4344 return 1;
4345 }
4347 sub show_commit {
4348 my $c = shift;
4349 if ($oneline) {
4350 my $x = "\n";
4351 if (my $l = $c->{l}) {
4352 while ($l->[0] =~ /^\s*$/) { shift @$l }
4353 $x = $l->[0];
4354 }
4355 $l_fmt ||= 'A' . length($c->{r});
4356 print 'r',pack($l_fmt, $c->{r}),' | ';
4357 print "$c->{c} | " if $show_commit;
4358 print $x;
4359 } else {
4360 show_commit_normal($c);
4361 }
4362 }
4364 sub show_commit_changed_paths {
4365 my ($c) = @_;
4366 return unless $c->{changed};
4367 print "Changed paths:\n", @{$c->{changed}};
4368 }
4370 sub show_commit_normal {
4371 my ($c) = @_;
4372 print commit_log_separator, "r$c->{r} | ";
4373 print "$c->{c} | " if $show_commit;
4374 print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
4375 my $nr_line = 0;
4377 if (my $l = $c->{l}) {
4378 while ($l->[$#$l] eq "\n" && $#$l > 0
4379 && $l->[($#$l - 1)] eq "\n") {
4380 pop @$l;
4381 }
4382 $nr_line = scalar @$l;
4383 if (!$nr_line) {
4384 print "1 line\n\n\n";
4385 } else {
4386 if ($nr_line == 1) {
4387 $nr_line = '1 line';
4388 } else {
4389 $nr_line .= ' lines';
4390 }
4391 print $nr_line, "\n";
4392 show_commit_changed_paths($c);
4393 print "\n";
4394 print $_ foreach @$l;
4395 }
4396 } else {
4397 print "1 line\n";
4398 show_commit_changed_paths($c);
4399 print "\n";
4401 }
4402 foreach my $x (qw/raw stat diff/) {
4403 if ($c->{$x}) {
4404 print "\n";
4405 print $_ foreach @{$c->{$x}}
4406 }
4407 }
4408 }
4410 sub cmd_show_log {
4411 my (@args) = @_;
4412 my ($r_min, $r_max);
4413 my $r_last = -1; # prevent dupes
4414 set_local_timezone();
4415 if (defined $::_revision) {
4416 if ($::_revision =~ /^(\d+):(\d+)$/) {
4417 ($r_min, $r_max) = ($1, $2);
4418 } elsif ($::_revision =~ /^\d+$/) {
4419 $r_min = $r_max = $::_revision;
4420 } else {
4421 ::fatal "-r$::_revision is not supported, use ",
4422 "standard 'git log' arguments instead";
4423 }
4424 }
4426 config_pager();
4427 @args = git_svn_log_cmd($r_min, $r_max, @args);
4428 if (!@args) {
4429 print commit_log_separator unless $incremental || $oneline;
4430 return;
4431 }
4432 my $log = command_output_pipe(@args);
4433 run_pager();
4434 my (@k, $c, $d, $stat);
4435 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
4436 while (<$log>) {
4437 if (/^${esc_color}commit -?($::sha1_short)/o) {
4438 my $cmt = $1;
4439 if ($c && cmt_showable($c) && $c->{r} != $r_last) {
4440 $r_last = $c->{r};
4441 process_commit($c, $r_min, $r_max, \@k) or
4442 goto out;
4443 }
4444 $d = undef;
4445 $c = { c => $cmt };
4446 } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
4447 get_author_info($c, $1, $2, $3);
4448 } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
4449 # ignore
4450 } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
4451 push @{$c->{raw}}, $_;
4452 } elsif (/^${esc_color}[ACRMDT]\t/) {
4453 # we could add $SVN->{svn_path} here, but that requires
4454 # remote access at the moment (repo_path_split)...
4455 s#^(${esc_color})([ACRMDT])\t#$1 $2 #o;
4456 push @{$c->{changed}}, $_;
4457 } elsif (/^${esc_color}diff /o) {
4458 $d = 1;
4459 push @{$c->{diff}}, $_;
4460 } elsif ($d) {
4461 push @{$c->{diff}}, $_;
4462 } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
4463 $esc_color*[\+\-]*$esc_color$/x) {
4464 $stat = 1;
4465 push @{$c->{stat}}, $_;
4466 } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
4467 push @{$c->{stat}}, $_;
4468 $stat = undef;
4469 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
4470 ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
4471 } elsif (s/^${esc_color} //o) {
4472 push @{$c->{l}}, $_;
4473 }
4474 }
4475 if ($c && defined $c->{r} && $c->{r} != $r_last) {
4476 $r_last = $c->{r};
4477 process_commit($c, $r_min, $r_max, \@k);
4478 }
4479 if (@k) {
4480 ($r_min, $r_max) = ($r_max, $r_min);
4481 process_commit($_, $r_min, $r_max) foreach reverse @k;
4482 }
4483 out:
4484 close $log;
4485 print commit_log_separator unless $incremental || $oneline;
4486 }
4488 sub cmd_blame {
4489 my $path = pop;
4491 config_pager();
4492 run_pager();
4494 my ($fh, $ctx, $rev);
4496 if ($_git_format) {
4497 ($fh, $ctx) = command_output_pipe('blame', @_, $path);
4498 while (my $line = <$fh>) {
4499 if ($line =~ /^\^?([[:xdigit:]]+)\s/) {
4500 # Uncommitted edits show up as a rev ID of
4501 # all zeros, which we can't look up with
4502 # cmt_metadata
4503 if ($1 !~ /^0+$/) {
4504 (undef, $rev, undef) =
4505 ::cmt_metadata($1);
4506 $rev = '0' if (!$rev);
4507 } else {
4508 $rev = '0';
4509 }
4510 $rev = sprintf('%-10s', $rev);
4511 $line =~ s/^\^?[[:xdigit:]]+(\s)/$rev$1/;
4512 }
4513 print $line;
4514 }
4515 } else {
4516 ($fh, $ctx) = command_output_pipe('blame', '-p', @_, 'HEAD',
4517 '--', $path);
4518 my ($sha1);
4519 my %authors;
4520 while (my $line = <$fh>) {
4521 if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
4522 $sha1 = $1;
4523 (undef, $rev, undef) = ::cmt_metadata($1);
4524 $rev = '0' if (!$rev);
4525 }
4526 elsif ($line =~ /^author (.*)/) {
4527 $authors{$rev} = $1;
4528 $authors{$rev} =~ s/\s/_/g;
4529 }
4530 elsif ($line =~ /^\t(.*)$/) {
4531 printf("%6s %10s %s\n", $rev, $authors{$rev}, $1);
4532 }
4533 }
4534 }
4535 command_close_pipe($fh, $ctx);
4536 }
4538 package Git::SVN::Migration;
4539 # these version numbers do NOT correspond to actual version numbers
4540 # of git nor git-svn. They are just relative.
4541 #
4542 # v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
4543 #
4544 # v1 layout: .git/$id/info/url, refs/remotes/$id
4545 #
4546 # v2 layout: .git/svn/$id/info/url, refs/remotes/$id
4547 #
4548 # v3 layout: .git/svn/$id, refs/remotes/$id
4549 # - info/url may remain for backwards compatibility
4550 # - this is what we migrate up to this layout automatically,
4551 # - this will be used by git svn init on single branches
4552 # v3.1 layout (auto migrated):
4553 # - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
4554 # for backwards compatibility
4555 #
4556 # v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
4557 # - this is only created for newly multi-init-ed
4558 # repositories. Similar in spirit to the
4559 # --use-separate-remotes option in git-clone (now default)
4560 # - we do not automatically migrate to this (following
4561 # the example set by core git)
4562 #
4563 # v5 layout: .rev_db.$UUID => .rev_map.$UUID
4564 # - newer, more-efficient format that uses 24-bytes per record
4565 # with no filler space.
4566 # - use xxd -c24 < .rev_map.$UUID to view and debug
4567 # - This is a one-way migration, repositories updated to the
4568 # new format will not be able to use old git-svn without
4569 # rebuilding the .rev_db. Rebuilding the rev_db is not
4570 # possible if noMetadata or useSvmProps are set; but should
4571 # be no problem for users that use the (sensible) defaults.
4572 use strict;
4573 use warnings;
4574 use Carp qw/croak/;
4575 use File::Path qw/mkpath/;
4576 use File::Basename qw/dirname basename/;
4577 use vars qw/$_minimize/;
4579 sub migrate_from_v0 {
4580 my $git_dir = $ENV{GIT_DIR};
4581 return undef unless -d $git_dir;
4582 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4583 my $migrated = 0;
4584 while (<$fh>) {
4585 chomp;
4586 my ($id, $orig_ref) = ($_, $_);
4587 next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
4588 next unless -f "$git_dir/$id/info/url";
4589 my $new_ref = "refs/remotes/$id";
4590 if (::verify_ref("$new_ref^0")) {
4591 print STDERR "W: $orig_ref is probably an old ",
4592 "branch used by an ancient version of ",
4593 "git-svn.\n",
4594 "However, $new_ref also exists.\n",
4595 "We will not be able ",
4596 "to use this branch until this ",
4597 "ambiguity is resolved.\n";
4598 next;
4599 }
4600 print STDERR "Migrating from v0 layout...\n" if !$migrated;
4601 print STDERR "Renaming ref: $orig_ref => $new_ref\n";
4602 command_noisy('update-ref', $new_ref, $orig_ref);
4603 command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
4604 $migrated++;
4605 }
4606 command_close_pipe($fh, $ctx);
4607 print STDERR "Done migrating from v0 layout...\n" if $migrated;
4608 $migrated;
4609 }
4611 sub migrate_from_v1 {
4612 my $git_dir = $ENV{GIT_DIR};
4613 my $migrated = 0;
4614 return $migrated unless -d $git_dir;
4615 my $svn_dir = "$git_dir/svn";
4617 # just in case somebody used 'svn' as their $id at some point...
4618 return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
4620 print STDERR "Migrating from a git-svn v1 layout...\n";
4621 mkpath([$svn_dir]);
4622 print STDERR "Data from a previous version of git-svn exists, but\n\t",
4623 "$svn_dir\n\t(required for this version ",
4624 "($::VERSION) of git-svn) does not. exist\n";
4625 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4626 while (<$fh>) {
4627 my $x = $_;
4628 next unless $x =~ s#^refs/remotes/##;
4629 chomp $x;
4630 next unless -f "$git_dir/$x/info/url";
4631 my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
4632 next unless $u;
4633 my $dn = dirname("$git_dir/svn/$x");
4634 mkpath([$dn]) unless -d $dn;
4635 if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
4636 mkpath(["$git_dir/svn/svn"]);
4637 print STDERR " - $git_dir/$x/info => ",
4638 "$git_dir/svn/$x/info\n";
4639 rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
4640 croak "$!: $x";
4641 # don't worry too much about these, they probably
4642 # don't exist with repos this old (save for index,
4643 # and we can easily regenerate that)
4644 foreach my $f (qw/unhandled.log index .rev_db/) {
4645 rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
4646 }
4647 } else {
4648 print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
4649 rename "$git_dir/$x", "$git_dir/svn/$x" or
4650 croak "$!: $x";
4651 }
4652 $migrated++;
4653 }
4654 command_close_pipe($fh, $ctx);
4655 print STDERR "Done migrating from a git-svn v1 layout\n";
4656 $migrated;
4657 }
4659 sub read_old_urls {
4660 my ($l_map, $pfx, $path) = @_;
4661 my @dir;
4662 foreach (<$path/*>) {
4663 if (-r "$_/info/url") {
4664 $pfx .= '/' if $pfx && $pfx !~ m!/$!;
4665 my $ref_id = $pfx . basename $_;
4666 my $url = ::file_to_s("$_/info/url");
4667 $l_map->{$ref_id} = $url;
4668 } elsif (-d $_) {
4669 push @dir, $_;
4670 }
4671 }
4672 foreach (@dir) {
4673 my $x = $_;
4674 $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
4675 read_old_urls($l_map, $x, $_);
4676 }
4677 }
4679 sub migrate_from_v2 {
4680 my @cfg = command(qw/config -l/);
4681 return if grep /^svn-remote\..+\.url=/, @cfg;
4682 my %l_map;
4683 read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
4684 my $migrated = 0;
4686 foreach my $ref_id (sort keys %l_map) {
4687 eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
4688 if ($@) {
4689 Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
4690 }
4691 $migrated++;
4692 }
4693 $migrated;
4694 }
4696 sub minimize_connections {
4697 my $r = Git::SVN::read_all_remotes();
4698 my $new_urls = {};
4699 my $root_repos = {};
4700 foreach my $repo_id (keys %$r) {
4701 my $url = $r->{$repo_id}->{url} or next;
4702 my $fetch = $r->{$repo_id}->{fetch} or next;
4703 my $ra = Git::SVN::Ra->new($url);
4705 # skip existing cases where we already connect to the root
4706 if (($ra->{url} eq $ra->{repos_root}) ||
4707 (Git::SVN::sanitize_remote_name($ra->{repos_root}) eq
4708 $repo_id)) {
4709 $root_repos->{$ra->{url}} = $repo_id;
4710 next;
4711 }
4713 my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
4714 my $root_path = $ra->{url};
4715 $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
4716 foreach my $path (keys %$fetch) {
4717 my $ref_id = $fetch->{$path};
4718 my $gs = Git::SVN->new($ref_id, $repo_id, $path);
4720 # make sure we can read when connecting to
4721 # a higher level of a repository
4722 my ($last_rev, undef) = $gs->last_rev_commit;
4723 if (!defined $last_rev) {
4724 $last_rev = eval {
4725 $root_ra->get_latest_revnum;
4726 };
4727 next if $@;
4728 }
4729 my $new = $root_path;
4730 $new .= length $path ? "/$path" : '';
4731 eval {
4732 $root_ra->get_log([$new], $last_rev, $last_rev,
4733 0, 0, 1, sub { });
4734 };
4735 next if $@;
4736 $new_urls->{$ra->{repos_root}}->{$new} =
4737 { ref_id => $ref_id,
4738 old_repo_id => $repo_id,
4739 old_path => $path };
4740 }
4741 }
4743 my @emptied;
4744 foreach my $url (keys %$new_urls) {
4745 # see if we can re-use an existing [svn-remote "repo_id"]
4746 # instead of creating a(n ugly) new section:
4747 my $repo_id = $root_repos->{$url} ||
4748 Git::SVN::sanitize_remote_name($url);
4750 my $fetch = $new_urls->{$url};
4751 foreach my $path (keys %$fetch) {
4752 my $x = $fetch->{$path};
4753 Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
4754 my $pfx = "svn-remote.$x->{old_repo_id}";
4756 my $old_fetch = quotemeta("$x->{old_path}:".
4757 "refs/remotes/$x->{ref_id}");
4758 command_noisy(qw/config --unset/,
4759 "$pfx.fetch", '^'. $old_fetch . '$');
4760 delete $r->{$x->{old_repo_id}}->
4761 {fetch}->{$x->{old_path}};
4762 if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
4763 command_noisy(qw/config --unset/,
4764 "$pfx.url");
4765 push @emptied, $x->{old_repo_id}
4766 }
4767 }
4768 }
4769 if (@emptied) {
4770 my $file = $ENV{GIT_CONFIG} || $ENV{GIT_CONFIG_LOCAL} ||
4771 "$ENV{GIT_DIR}/config";
4772 print STDERR <<EOF;
4773 The following [svn-remote] sections in your config file ($file) are empty
4774 and can be safely removed:
4775 EOF
4776 print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
4777 }
4778 }
4780 sub migration_check {
4781 migrate_from_v0();
4782 migrate_from_v1();
4783 migrate_from_v2();
4784 minimize_connections() if $_minimize;
4785 }
4787 package Git::IndexInfo;
4788 use strict;
4789 use warnings;
4790 use Git qw/command_input_pipe command_close_pipe/;
4792 sub new {
4793 my ($class) = @_;
4794 my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
4795 bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
4796 }
4798 sub remove {
4799 my ($self, $path) = @_;
4800 if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
4801 return ++$self->{nr};
4802 }
4803 undef;
4804 }
4806 sub update {
4807 my ($self, $mode, $hash, $path) = @_;
4808 if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
4809 return ++$self->{nr};
4810 }
4811 undef;
4812 }
4814 sub DESTROY {
4815 my ($self) = @_;
4816 command_close_pipe($self->{gui}, $self->{ctx});
4817 }
4819 package Git::SVN::GlobSpec;
4820 use strict;
4821 use warnings;
4823 sub new {
4824 my ($class, $glob) = @_;
4825 my $re = $glob;
4826 $re =~ s!/+$!!g; # no need for trailing slashes
4827 my $nr = ($re =~ s!^(.*)\*(.*)$!\(\[^/\]+\)!g);
4828 my ($left, $right) = ($1, $2);
4829 if ($nr > 1) {
4830 die "Only one '*' wildcard expansion ",
4831 "is supported (got $nr): '$glob'\n";
4832 } elsif ($nr == 0) {
4833 die "One '*' is needed for glob: '$glob'\n";
4834 }
4835 $re = quotemeta($left) . $re . quotemeta($right);
4836 if (length $left && !($left =~ s!/+$!!g)) {
4837 die "Missing trailing '/' on left side of: '$glob' ($left)\n";
4838 }
4839 if (length $right && !($right =~ s!^/+!!g)) {
4840 die "Missing leading '/' on right side of: '$glob' ($right)\n";
4841 }
4842 my $left_re = qr/^\/\Q$left\E(\/|$)/;
4843 bless { left => $left, right => $right, left_regex => $left_re,
4844 regex => qr/$re/, glob => $glob }, $class;
4845 }
4847 sub full_path {
4848 my ($self, $path) = @_;
4849 return (length $self->{left} ? "$self->{left}/" : '') .
4850 $path . (length $self->{right} ? "/$self->{right}" : '');
4851 }
4853 __END__
4855 Data structures:
4858 $remotes = { # returned by read_all_remotes()
4859 'svn' => {
4860 # svn-remote.svn.url=https://svn.musicpd.org
4861 url => 'https://svn.musicpd.org',
4862 # svn-remote.svn.fetch=mpd/trunk:trunk
4863 fetch => {
4864 'mpd/trunk' => 'trunk',
4865 },
4866 # svn-remote.svn.tags=mpd/tags/*:tags/*
4867 tags => {
4868 path => {
4869 left => 'mpd/tags',
4870 right => '',
4871 regex => qr!mpd/tags/([^/]+)$!,
4872 glob => 'tags/*',
4873 },
4874 ref => {
4875 left => 'tags',
4876 right => '',
4877 regex => qr!tags/([^/]+)$!,
4878 glob => 'tags/*',
4879 },
4880 }
4881 }
4882 };
4884 $log_entry hashref as returned by libsvn_log_entry()
4885 {
4886 log => 'whitespace-formatted log entry
4887 ', # trailing newline is preserved
4888 revision => '8', # integer
4889 date => '2004-02-24T17:01:44.108345Z', # commit date
4890 author => 'committer name'
4891 };
4894 # this is generated by generate_diff();
4895 @mods = array of diff-index line hashes, each element represents one line
4896 of diff-index output
4898 diff-index line ($m hash)
4899 {
4900 mode_a => first column of diff-index output, no leading ':',
4901 mode_b => second column of diff-index output,
4902 sha1_b => sha1sum of the final blob,
4903 chg => change type [MCRADT],
4904 file_a => original file name of a file (iff chg is 'C' or 'R')
4905 file_b => new/current file name of a file (any chg)
4906 }
4907 ;
4909 # retval of read_url_paths{,_all}();
4910 $l_map = {
4911 # repository root url
4912 'https://svn.musicpd.org' => {
4913 # repository path # GIT_SVN_ID
4914 'mpd/trunk' => 'trunk',
4915 'mpd/tags/0.11.5' => 'tags/0.11.5',
4916 },
4917 }
4919 Notes:
4920 I don't trust the each() function on unless I created %hash myself
4921 because the internal iterator may not have started at base.