?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
git-svn: Add --localtime option to "fetch"
[git/gitweb.git] / git-svn.perl
blobe3e125b098f63e3c7fc434770e7281981bb8940b
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@@';
11
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 } || '';
16
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;
22
23 $Git::SVN::Log::TZ = $ENV{TZ};
24 $ENV{TZ} = 'UTC';
25 $| = 1; # unbuffer STDOUT
26
27 sub fatal (@) { print STDERR "@_\n"; exit 1 }
28 require SVN::Core; # use()-ing this causes segfaults for me... *shrug*
29 require SVN::Ra;
30 require SVN::Delta;
31 if ($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;
45
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 }
58
59 my ($SVN);
60
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, $_commit_url, $_tag);
70 $Git::SVN::_follow_parent = 1;
71 my %remote_opts = ( 'username=s' => \$Git::SVN::Prompt::_username,
72 'config-dir=s' => \$Git::SVN::Ra::config_dir,
73 'no-auth-cache' => \$Git::SVN::Prompt::_no_auth_cache );
74 my %fc_opts = ( 'follow-parent|follow!' => \$Git::SVN::_follow_parent,
75 'authors-file|A=s' => \$_authors,
76 'repack:i' => \$Git::SVN::_repack,
77 'noMetadata' => \$Git::SVN::_no_metadata,
78 'useSvmProps' => \$Git::SVN::_use_svm_props,
79 'useSvnsyncProps' => \$Git::SVN::_use_svnsync_props,
80 'log-window-size=i' => \$Git::SVN::Ra::_log_window_size,
81 'no-checkout' => \$_no_checkout,
82 'quiet|q' => \$_q,
83 'repack-flags|repack-args|repack-opts=s' =>
84 \$Git::SVN::_repack_flags,
85 'use-log-author' => \$Git::SVN::_use_log_author,
86 'add-author-from' => \$Git::SVN::_add_author_from,
87 'localtime' => \$Git::SVN::_localtime,
88 %remote_opts );
89
90 my ($_trunk, $_tags, $_branches, $_stdlayout);
91 my %icv;
92 my %init_opts = ( 'template=s' => \$_template, 'shared:s' => \$_shared,
93 'trunk|T=s' => \$_trunk, 'tags|t=s' => \$_tags,
94 'branches|b=s' => \$_branches, 'prefix=s' => \$_prefix,
95 'stdlayout|s' => \$_stdlayout,
96 'minimize-url|m' => \$Git::SVN::_minimize_url,
97 'no-metadata' => sub { $icv{noMetadata} = 1 },
98 'use-svm-props' => sub { $icv{useSvmProps} = 1 },
99 'use-svnsync-props' => sub { $icv{useSvnsyncProps} = 1 },
100 'rewrite-root=s' => sub { $icv{rewriteRoot} = $_[1] },
101 %remote_opts );
102 my %cmt_opts = ( 'edit|e' => \$_edit,
103 'rmdir' => \$SVN::Git::Editor::_rmdir,
104 'find-copies-harder' => \$SVN::Git::Editor::_find_copies_harder,
105 'l=i' => \$SVN::Git::Editor::_rename_limit,
106 'copy-similarity|C=i'=> \$SVN::Git::Editor::_cp_similarity
107 );
108
109 my %cmd = (
110 fetch => [ \&cmd_fetch, "Download new revisions from SVN",
111 { 'revision|r=s' => \$_revision,
112 'fetch-all|all' => \$_fetch_all,
113 %fc_opts } ],
114 clone => [ \&cmd_clone, "Initialize and fetch revisions",
115 { 'revision|r=s' => \$_revision,
116 %fc_opts, %init_opts } ],
117 init => [ \&cmd_init, "Initialize a repo for tracking" .
118 " (requires URL argument)",
119 \%init_opts ],
120 'multi-init' => [ \&cmd_multi_init,
121 "Deprecated alias for ".
122 "'$0 init -T<trunk> -b<branches> -t<tags>'",
123 \%init_opts ],
124 dcommit => [ \&cmd_dcommit,
125 'Commit several diffs to merge with upstream',
126 { 'merge|m|M' => \$_merge,
127 'strategy|s=s' => \$_strategy,
128 'verbose|v' => \$_verbose,
129 'dry-run|n' => \$_dry_run,
130 'fetch-all|all' => \$_fetch_all,
131 'commit-url=s' => \$_commit_url,
132 'revision|r=i' => \$_revision,
133 'no-rebase' => \$_no_rebase,
134 %cmt_opts, %fc_opts } ],
135 branch => [ \&cmd_branch,
136 'Create a branch in the SVN repository',
137 { 'message|m=s' => \$_message,
138 'dry-run|n' => \$_dry_run,
139 'tag|t' => \$_tag } ],
140 tag => [ sub { $_tag = 1; cmd_branch(@_) },
141 'Create a tag in the SVN repository',
142 { 'message|m=s' => \$_message,
143 'dry-run|n' => \$_dry_run } ],
144 'set-tree' => [ \&cmd_set_tree,
145 "Set an SVN repository to a git tree-ish",
146 { 'stdin|' => \$_stdin, %cmt_opts, %fc_opts, } ],
147 'create-ignore' => [ \&cmd_create_ignore,
148 'Create a .gitignore per svn:ignore',
149 { 'revision|r=i' => \$_revision
150 } ],
151 'propget' => [ \&cmd_propget,
152 'Print the value of a property on a file or directory',
153 { 'revision|r=i' => \$_revision } ],
154 'proplist' => [ \&cmd_proplist,
155 'List all properties of a file or directory',
156 { 'revision|r=i' => \$_revision } ],
157 'show-ignore' => [ \&cmd_show_ignore, "Show svn:ignore listings",
158 { 'revision|r=i' => \$_revision
159 } ],
160 'show-externals' => [ \&cmd_show_externals, "Show svn:externals listings",
161 { 'revision|r=i' => \$_revision
162 } ],
163 'multi-fetch' => [ \&cmd_multi_fetch,
164 "Deprecated alias for $0 fetch --all",
165 { 'revision|r=s' => \$_revision, %fc_opts } ],
166 'migrate' => [ sub { },
167 # no-op, we automatically run this anyways,
168 'Migrate configuration/metadata/layout from
169 previous versions of git-svn',
170 { 'minimize' => \$Git::SVN::Migration::_minimize,
171 %remote_opts } ],
172 'log' => [ \&Git::SVN::Log::cmd_show_log, 'Show commit logs',
173 { 'limit=i' => \$Git::SVN::Log::limit,
174 'revision|r=s' => \$_revision,
175 'verbose|v' => \$Git::SVN::Log::verbose,
176 'incremental' => \$Git::SVN::Log::incremental,
177 'oneline' => \$Git::SVN::Log::oneline,
178 'show-commit' => \$Git::SVN::Log::show_commit,
179 'non-recursive' => \$Git::SVN::Log::non_recursive,
180 'authors-file|A=s' => \$_authors,
181 'color' => \$Git::SVN::Log::color,
182 'pager=s' => \$Git::SVN::Log::pager
183 } ],
184 'find-rev' => [ \&cmd_find_rev,
185 "Translate between SVN revision numbers and tree-ish",
186 {} ],
187 'rebase' => [ \&cmd_rebase, "Fetch and rebase your working directory",
188 { 'merge|m|M' => \$_merge,
189 'verbose|v' => \$_verbose,
190 'strategy|s=s' => \$_strategy,
191 'local|l' => \$_local,
192 'fetch-all|all' => \$_fetch_all,
193 'dry-run|n' => \$_dry_run,
194 %fc_opts } ],
195 'commit-diff' => [ \&cmd_commit_diff,
196 'Commit a diff between two trees',
197 { 'message|m=s' => \$_message,
198 'file|F=s' => \$_file,
199 'revision|r=s' => \$_revision,
200 %cmt_opts } ],
201 'info' => [ \&cmd_info,
202 "Show info about the latest SVN revision
203 on the current branch",
204 { 'url' => \$_url, } ],
205 'blame' => [ \&Git::SVN::Log::cmd_blame,
206 "Show what revision and author last modified each line of a file",
207 { 'git-format' => \$_git_format } ],
208 );
209
210 my $cmd;
211 for (my $i = 0; $i < @ARGV; $i++) {
212 if (defined $cmd{$ARGV[$i]}) {
213 $cmd = $ARGV[$i];
214 splice @ARGV, $i, 1;
215 last;
216 }
217 };
218
219 # make sure we're always running at the top-level working directory
220 unless ($cmd && $cmd =~ /(?:clone|init|multi-init)$/) {
221 unless (-d $ENV{GIT_DIR}) {
222 if ($git_dir_user_set) {
223 die "GIT_DIR=$ENV{GIT_DIR} explicitly set, ",
224 "but it is not a directory\n";
225 }
226 my $git_dir = delete $ENV{GIT_DIR};
227 my $cdup = undef;
228 git_cmd_try {
229 $cdup = command_oneline(qw/rev-parse --show-cdup/);
230 $git_dir = '.' unless ($cdup);
231 chomp $cdup if ($cdup);
232 $cdup = "." unless ($cdup && length $cdup);
233 } "Already at toplevel, but $git_dir not found\n";
234 chdir $cdup or die "Unable to chdir up to '$cdup'\n";
235 unless (-d $git_dir) {
236 die "$git_dir still not found after going to ",
237 "'$cdup'\n";
238 }
239 $ENV{GIT_DIR} = $git_dir;
240 }
241 $_repository = Git->repository(Repository => $ENV{GIT_DIR});
242 }
243
244 my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
245
246 read_repo_config(\%opts);
247 if ($cmd && ($cmd eq 'log' || $cmd eq 'blame')) {
248 Getopt::Long::Configure('pass_through');
249 }
250 my $rv = GetOptions(%opts, 'help|H|h' => \$_help, 'version|V' => \$_version,
251 'minimize-connections' => \$Git::SVN::Migration::_minimize,
252 'id|i=s' => \$Git::SVN::default_ref_id,
253 'svn-remote|remote|R=s' => sub {
254 $Git::SVN::no_reuse_existing = 1;
255 $Git::SVN::default_repo_id = $_[1] });
256 exit 1 if (!$rv && $cmd && $cmd ne 'log');
257
258 usage(0) if $_help;
259 version() if $_version;
260 usage(1) unless defined $cmd;
261 load_authors() if $_authors;
262
263 unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
264 Git::SVN::Migration::migration_check();
265 }
266 Git::SVN::init_vars();
267 eval {
268 Git::SVN::verify_remotes_sanity();
269 $cmd{$cmd}->[0]->(@ARGV);
270 };
271 fatal $@ if $@;
272 post_fetch_checkout();
273 exit 0;
274
275 ####################### primary functions ######################
276 sub usage {
277 my $exit = shift || 0;
278 my $fd = $exit ? \*STDERR : \*STDOUT;
279 print $fd <<"";
280 git-svn - bidirectional operations between a single Subversion tree and git
281 Usage: git svn <command> [options] [arguments]\n
282
283 print $fd "Available commands:\n" unless $cmd;
284
285 foreach (sort keys %cmd) {
286 next if $cmd && $cmd ne $_;
287 next if /^multi-/; # don't show deprecated commands
288 print $fd ' ',pack('A17',$_),$cmd{$_}->[1],"\n";
289 foreach (sort keys %{$cmd{$_}->[2]}) {
290 # mixed-case options are for .git/config only
291 next if /[A-Z]/ && /^[a-z]+$/i;
292 # prints out arguments as they should be passed:
293 my $x = s#[:=]s$## ? '<arg>' : s#[:=]i$## ? '<num>' : '';
294 print $fd ' ' x 21, join(', ', map { length $_ > 1 ?
295 "--$_" : "-$_" }
296 split /\|/,$_)," $x\n";
297 }
298 }
299 print $fd <<"";
300 \nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
301 arbitrary identifier if you're tracking multiple SVN branches/repositories in
302 one git repository and want to keep them separate. See git-svn(1) for more
303 information.
304
305 exit $exit;
306 }
307
308 sub version {
309 print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
310 exit 0;
311 }
312
313 sub do_git_init_db {
314 unless (-d $ENV{GIT_DIR}) {
315 my @init_db = ('init');
316 push @init_db, "--template=$_template" if defined $_template;
317 if (defined $_shared) {
318 if ($_shared =~ /[a-z]/) {
319 push @init_db, "--shared=$_shared";
320 } else {
321 push @init_db, "--shared";
322 }
323 }
324 command_noisy(@init_db);
325 $_repository = Git->repository(Repository => ".git");
326 }
327 my $set;
328 my $pfx = "svn-remote.$Git::SVN::default_repo_id";
329 foreach my $i (keys %icv) {
330 die "'$set' and '$i' cannot both be set\n" if $set;
331 next unless defined $icv{$i};
332 command_noisy('config', "$pfx.$i", $icv{$i});
333 $set = $i;
334 }
335 }
336
337 sub init_subdir {
338 my $repo_path = shift or return;
339 mkpath([$repo_path]) unless -d $repo_path;
340 chdir $repo_path or die "Couldn't chdir to $repo_path: $!\n";
341 $ENV{GIT_DIR} = '.git';
342 $_repository = Git->repository(Repository => $ENV{GIT_DIR});
343 }
344
345 sub cmd_clone {
346 my ($url, $path) = @_;
347 if (!defined $path &&
348 (defined $_trunk || defined $_branches || defined $_tags ||
349 defined $_stdlayout) &&
350 $url !~ m#^[a-z\+]+://#) {
351 $path = $url;
352 }
353 $path = basename($url) if !defined $path || !length $path;
354 cmd_init($url, $path);
355 Git::SVN::fetch_all($Git::SVN::default_repo_id);
356 }
357
358 sub cmd_init {
359 if (defined $_stdlayout) {
360 $_trunk = 'trunk' if (!defined $_trunk);
361 $_tags = 'tags' if (!defined $_tags);
362 $_branches = 'branches' if (!defined $_branches);
363 }
364 if (defined $_trunk || defined $_branches || defined $_tags) {
365 return cmd_multi_init(@_);
366 }
367 my $url = shift or die "SVN repository location required ",
368 "as a command-line argument\n";
369 init_subdir(@_);
370 do_git_init_db();
371
372 Git::SVN->init($url);
373 }
374
375 sub cmd_fetch {
376 if (grep /^\d+=./, @_) {
377 die "'<rev>=<commit>' fetch arguments are ",
378 "no longer supported.\n";
379 }
380 my ($remote) = @_;
381 if (@_ > 1) {
382 die "Usage: $0 fetch [--all] [svn-remote]\n";
383 }
384 $remote ||= $Git::SVN::default_repo_id;
385 if ($_fetch_all) {
386 cmd_multi_fetch();
387 } else {
388 Git::SVN::fetch_all($remote, Git::SVN::read_all_remotes());
389 }
390 }
391
392 sub cmd_set_tree {
393 my (@commits) = @_;
394 if ($_stdin || !@commits) {
395 print "Reading from stdin...\n";
396 @commits = ();
397 while (<STDIN>) {
398 if (/\b($sha1_short)\b/o) {
399 unshift @commits, $1;
400 }
401 }
402 }
403 my @revs;
404 foreach my $c (@commits) {
405 my @tmp = command('rev-parse',$c);
406 if (scalar @tmp == 1) {
407 push @revs, $tmp[0];
408 } elsif (scalar @tmp > 1) {
409 push @revs, reverse(command('rev-list',@tmp));
410 } else {
411 fatal "Failed to rev-parse $c";
412 }
413 }
414 my $gs = Git::SVN->new;
415 my ($r_last, $cmt_last) = $gs->last_rev_commit;
416 $gs->fetch;
417 if (defined $gs->{last_rev} && $r_last != $gs->{last_rev}) {
418 fatal "There are new revisions that were fetched ",
419 "and need to be merged (or acknowledged) ",
420 "before committing.\nlast rev: $r_last\n",
421 " current: $gs->{last_rev}";
422 }
423 $gs->set_tree($_) foreach @revs;
424 print "Done committing ",scalar @revs," revisions to SVN\n";
425 unlink $gs->{index};
426 }
427
428 sub cmd_dcommit {
429 my $head = shift;
430 git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
431 'Cannot dcommit with a dirty index. Commit your changes first, '
432 . "or stash them with `git stash'.\n";
433 $head ||= 'HEAD';
434 my @refs;
435 my ($url, $rev, $uuid, $gs) = working_head_info($head, \@refs);
436 unless ($gs) {
437 die "Unable to determine upstream SVN information from ",
438 "$head history.\nPerhaps the repository is empty.";
439 }
440 $url = defined $_commit_url ? $_commit_url : $gs->full_url;
441 my $last_rev = $_revision if defined $_revision;
442 if ($url) {
443 print "Committing to $url ...\n";
444 }
445 my ($linear_refs, $parents) = linearize_history($gs, \@refs);
446 if ($_no_rebase && scalar(@$linear_refs) > 1) {
447 warn "Attempting to commit more than one change while ",
448 "--no-rebase is enabled.\n",
449 "If these changes depend on each other, re-running ",
450 "without --no-rebase may be required."
451 }
452 my $expect_url = $url;
453 Git::SVN::remove_username($expect_url);
454 while (1) {
455 my $d = shift @$linear_refs or last;
456 unless (defined $last_rev) {
457 (undef, $last_rev, undef) = cmt_metadata("$d~1");
458 unless (defined $last_rev) {
459 fatal "Unable to extract revision information ",
460 "from commit $d~1";
461 }
462 }
463 if ($_dry_run) {
464 print "diff-tree $d~1 $d\n";
465 } else {
466 my $cmt_rev;
467 my %ed_opts = ( r => $last_rev,
468 log => get_commit_entry($d)->{log},
469 ra => Git::SVN::Ra->new($url),
470 config => SVN::Core::config_get_config(
471 $Git::SVN::Ra::config_dir
472 ),
473 tree_a => "$d~1",
474 tree_b => $d,
475 editor_cb => sub {
476 print "Committed r$_[0]\n";
477 $cmt_rev = $_[0];
478 },
479 svn_path => '');
480 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
481 print "No changes\n$d~1 == $d\n";
482 } elsif ($parents->{$d} && @{$parents->{$d}}) {
483 $gs->{inject_parents_dcommit}->{$cmt_rev} =
484 $parents->{$d};
485 }
486 $_fetch_all ? $gs->fetch_all : $gs->fetch;
487 $last_rev = $cmt_rev;
488 next if $_no_rebase;
489
490 # we always want to rebase against the current HEAD,
491 # not any head that was passed to us
492 my @diff = command('diff-tree', $d,
493 $gs->refname, '--');
494 my @finish;
495 if (@diff) {
496 @finish = rebase_cmd();
497 print STDERR "W: $d and ", $gs->refname,
498 " differ, using @finish:\n",
499 join("\n", @diff), "\n";
500 } else {
501 print "No changes between current HEAD and ",
502 $gs->refname,
503 "\nResetting to the latest ",
504 $gs->refname, "\n";
505 @finish = qw/reset --mixed/;
506 }
507 command_noisy(@finish, $gs->refname);
508 if (@diff) {
509 @refs = ();
510 my ($url_, $rev_, $uuid_, $gs_) =
511 working_head_info($head, \@refs);
512 my ($linear_refs_, $parents_) =
513 linearize_history($gs_, \@refs);
514 if (scalar(@$linear_refs) !=
515 scalar(@$linear_refs_)) {
516 fatal "# of revisions changed ",
517 "\nbefore:\n",
518 join("\n", @$linear_refs),
519 "\n\nafter:\n",
520 join("\n", @$linear_refs_), "\n",
521 'If you are attempting to commit ',
522 "merges, try running:\n\t",
523 'git rebase --interactive',
524 '--preserve-merges ',
525 $gs->refname,
526 "\nBefore dcommitting";
527 }
528 if ($url_ ne $expect_url) {
529 fatal "URL mismatch after rebase: ",
530 "$url_ != $expect_url";
531 }
532 if ($uuid_ ne $uuid) {
533 fatal "uuid mismatch after rebase: ",
534 "$uuid_ != $uuid";
535 }
536 # remap parents
537 my (%p, @l, $i);
538 for ($i = 0; $i < scalar @$linear_refs; $i++) {
539 my $new = $linear_refs_->[$i] or next;
540 $p{$new} =
541 $parents->{$linear_refs->[$i]};
542 push @l, $new;
543 }
544 $parents = \%p;
545 $linear_refs = \@l;
546 }
547 }
548 }
549 unlink $gs->{index};
550 }
551
552 sub cmd_branch {
553 my ($branch_name, $head) = @_;
554
555 unless (defined $branch_name && length $branch_name) {
556 die(($_tag ? "tag" : "branch") . " name required\n");
557 }
558 $head ||= 'HEAD';
559
560 my ($src, $rev, undef, $gs) = working_head_info($head);
561
562 my $remote = Git::SVN::read_all_remotes()->{$gs->{repo_id}};
563 my $glob = $remote->{ $_tag ? 'tags' : 'branches' };
564 my ($lft, $rgt) = @{ $glob->{path} }{qw/left right/};
565 my $dst = join '/', $remote->{url}, $lft, $branch_name, ($rgt || ());
566
567 my $ctx = SVN::Client->new(
568 auth => Git::SVN::Ra::_auth_providers(),
569 log_msg => sub {
570 ${ $_[0] } = defined $_message
571 ? $_message
572 : 'Create ' . ($_tag ? 'tag ' : 'branch ' )
573 . $branch_name;
574 },
575 );
576
577 eval {
578 $ctx->ls($dst, 'HEAD', 0);
579 } and die "branch ${branch_name} already exists\n";
580
581 print "Copying ${src} at r${rev} to ${dst}...\n";
582 $ctx->copy($src, $rev, $dst)
583 unless $_dry_run;
584
585 $gs->fetch_all;
586 }
587
588 sub cmd_find_rev {
589 my $revision_or_hash = shift or die "SVN or git revision required ",
590 "as a command-line argument\n";
591 my $result;
592 if ($revision_or_hash =~ /^r\d+$/) {
593 my $head = shift;
594 $head ||= 'HEAD';
595 my @refs;
596 my (undef, undef, $uuid, $gs) = working_head_info($head, \@refs);
597 unless ($gs) {
598 die "Unable to determine upstream SVN information from ",
599 "$head history\n";
600 }
601 my $desired_revision = substr($revision_or_hash, 1);
602 $result = $gs->rev_map_get($desired_revision, $uuid);
603 } else {
604 my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
605 $result = $rev;
606 }
607 print "$result\n" if $result;
608 }
609
610 sub cmd_rebase {
611 command_noisy(qw/update-index --refresh/);
612 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
613 unless ($gs) {
614 die "Unable to determine upstream SVN information from ",
615 "working tree history\n";
616 }
617 if ($_dry_run) {
618 print "Remote Branch: " . $gs->refname . "\n";
619 print "SVN URL: " . $url . "\n";
620 return;
621 }
622 if (command(qw/diff-index HEAD --/)) {
623 print STDERR "Cannot rebase with uncommited changes:\n";
624 command_noisy('status');
625 exit 1;
626 }
627 unless ($_local) {
628 # rebase will checkout for us, so no need to do it explicitly
629 $_no_checkout = 'true';
630 $_fetch_all ? $gs->fetch_all : $gs->fetch;
631 }
632 command_noisy(rebase_cmd(), $gs->refname);
633 }
634
635 sub cmd_show_ignore {
636 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
637 $gs ||= Git::SVN->new;
638 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
639 $gs->prop_walk($gs->{path}, $r, sub {
640 my ($gs, $path, $props) = @_;
641 print STDOUT "\n# $path\n";
642 my $s = $props->{'svn:ignore'} or return;
643 $s =~ s/[\r\n]+/\n/g;
644 chomp $s;
645 $s =~ s#^#$path#gm;
646 print STDOUT "$s\n";
647 });
648 }
649
650 sub cmd_show_externals {
651 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
652 $gs ||= Git::SVN->new;
653 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
654 $gs->prop_walk($gs->{path}, $r, sub {
655 my ($gs, $path, $props) = @_;
656 print STDOUT "\n# $path\n";
657 my $s = $props->{'svn:externals'} or return;
658 $s =~ s/[\r\n]+/\n/g;
659 chomp $s;
660 $s =~ s#^#$path#gm;
661 print STDOUT "$s\n";
662 });
663 }
664
665 sub cmd_create_ignore {
666 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
667 $gs ||= Git::SVN->new;
668 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
669 $gs->prop_walk($gs->{path}, $r, sub {
670 my ($gs, $path, $props) = @_;
671 # $path is of the form /path/to/dir/
672 my $ignore = '.' . $path . '.gitignore';
673 my $s = $props->{'svn:ignore'} or return;
674 open(GITIGNORE, '>', $ignore)
675 or fatal("Failed to open `$ignore' for writing: $!");
676 $s =~ s/[\r\n]+/\n/g;
677 chomp $s;
678 # Prefix all patterns so that the ignore doesn't apply
679 # to sub-directories.
680 $s =~ s#^#/#gm;
681 print GITIGNORE "$s\n";
682 close(GITIGNORE)
683 or fatal("Failed to close `$ignore': $!");
684 command_noisy('add', '-f', $ignore);
685 });
686 }
687
688 sub canonicalize_path {
689 my ($path) = @_;
690 my $dot_slash_added = 0;
691 if (substr($path, 0, 1) ne "/") {
692 $path = "./" . $path;
693 $dot_slash_added = 1;
694 }
695 # File::Spec->canonpath doesn't collapse x/../y into y (for a
696 # good reason), so let's do this manually.
697 $path =~ s#/+#/#g;
698 $path =~ s#/\.(?:/|$)#/#g;
699 $path =~ s#/[^/]+/\.\.##g;
700 $path =~ s#/$##g;
701 $path =~ s#^\./## if $dot_slash_added;
702 $path =~ s#^/##;
703 $path =~ s#^\.$##;
704 return $path;
705 }
706
707 # get_svnprops(PATH)
708 # ------------------
709 # Helper for cmd_propget and cmd_proplist below.
710 sub get_svnprops {
711 my $path = shift;
712 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
713 $gs ||= Git::SVN->new;
714
715 # prefix THE PATH by the sub-directory from which the user
716 # invoked us.
717 $path = $cmd_dir_prefix . $path;
718 fatal("No such file or directory: $path") unless -e $path;
719 my $is_dir = -d $path ? 1 : 0;
720 $path = $gs->{path} . '/' . $path;
721
722 # canonicalize the path (otherwise libsvn will abort or fail to
723 # find the file)
724 $path = canonicalize_path($path);
725
726 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
727 my $props;
728 if ($is_dir) {
729 (undef, undef, $props) = $gs->ra->get_dir($path, $r);
730 }
731 else {
732 (undef, $props) = $gs->ra->get_file($path, $r, undef);
733 }
734 return $props;
735 }
736
737 # cmd_propget (PROP, PATH)
738 # ------------------------
739 # Print the SVN property PROP for PATH.
740 sub cmd_propget {
741 my ($prop, $path) = @_;
742 $path = '.' if not defined $path;
743 usage(1) if not defined $prop;
744 my $props = get_svnprops($path);
745 if (not defined $props->{$prop}) {
746 fatal("`$path' does not have a `$prop' SVN property.");
747 }
748 print $props->{$prop} . "\n";
749 }
750
751 # cmd_proplist (PATH)
752 # -------------------
753 # Print the list of SVN properties for PATH.
754 sub cmd_proplist {
755 my $path = shift;
756 $path = '.' if not defined $path;
757 my $props = get_svnprops($path);
758 print "Properties on '$path':\n";
759 foreach (sort keys %{$props}) {
760 print " $_\n";
761 }
762 }
763
764 sub cmd_multi_init {
765 my $url = shift;
766 unless (defined $_trunk || defined $_branches || defined $_tags) {
767 usage(1);
768 }
769
770 # there are currently some bugs that prevent multi-init/multi-fetch
771 # setups from working well without this.
772 $Git::SVN::_minimize_url = 1;
773
774 $_prefix = '' unless defined $_prefix;
775 if (defined $url) {
776 $url =~ s#/+$##;
777 init_subdir(@_);
778 }
779 do_git_init_db();
780 if (defined $_trunk) {
781 my $trunk_ref = $_prefix . 'trunk';
782 # try both old-style and new-style lookups:
783 my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
784 unless ($gs_trunk) {
785 my ($trunk_url, $trunk_path) =
786 complete_svn_url($url, $_trunk);
787 $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
788 undef, $trunk_ref);
789 }
790 }
791 return unless defined $_branches || defined $_tags;
792 my $ra = $url ? Git::SVN::Ra->new($url) : undef;
793 complete_url_ls_init($ra, $_branches, '--branches/-b', $_prefix);
794 complete_url_ls_init($ra, $_tags, '--tags/-t', $_prefix . 'tags/');
795 }
796
797 sub cmd_multi_fetch {
798 my $remotes = Git::SVN::read_all_remotes();
799 foreach my $repo_id (sort keys %$remotes) {
800 if ($remotes->{$repo_id}->{url}) {
801 Git::SVN::fetch_all($repo_id, $remotes);
802 }
803 }
804 }
805
806 # this command is special because it requires no metadata
807 sub cmd_commit_diff {
808 my ($ta, $tb, $url) = @_;
809 my $usage = "Usage: $0 commit-diff -r<revision> ".
810 "<tree-ish> <tree-ish> [<URL>]";
811 fatal($usage) if (!defined $ta || !defined $tb);
812 my $svn_path = '';
813 if (!defined $url) {
814 my $gs = eval { Git::SVN->new };
815 if (!$gs) {
816 fatal("Needed URL or usable git-svn --id in ",
817 "the command-line\n", $usage);
818 }
819 $url = $gs->{url};
820 $svn_path = $gs->{path};
821 }
822 unless (defined $_revision) {
823 fatal("-r|--revision is a required argument\n", $usage);
824 }
825 if (defined $_message && defined $_file) {
826 fatal("Both --message/-m and --file/-F specified ",
827 "for the commit message.\n",
828 "I have no idea what you mean");
829 }
830 if (defined $_file) {
831 $_message = file_to_s($_file);
832 } else {
833 $_message ||= get_commit_entry($tb)->{log};
834 }
835 my $ra ||= Git::SVN::Ra->new($url);
836 my $r = $_revision;
837 if ($r eq 'HEAD') {
838 $r = $ra->get_latest_revnum;
839 } elsif ($r !~ /^\d+$/) {
840 die "revision argument: $r not understood by git-svn\n";
841 }
842 my %ed_opts = ( r => $r,
843 log => $_message,
844 ra => $ra,
845 tree_a => $ta,
846 tree_b => $tb,
847 editor_cb => sub { print "Committed r$_[0]\n" },
848 svn_path => $svn_path );
849 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
850 print "No changes\n$ta == $tb\n";
851 }
852 }
853
854 sub escape_uri_only {
855 my ($uri) = @_;
856 my @tmp;
857 foreach (split m{/}, $uri) {
858 s/([^~\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
859 push @tmp, $_;
860 }
861 join('/', @tmp);
862 }
863
864 sub escape_url {
865 my ($url) = @_;
866 if ($url =~ m#^([^:]+)://([^/]*)(.*)$#) {
867 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
868 $url = "$scheme://$domain$uri";
869 }
870 $url;
871 }
872
873 sub cmd_info {
874 my $path = canonicalize_path(defined($_[0]) ? $_[0] : ".");
875 my $fullpath = canonicalize_path($cmd_dir_prefix . $path);
876 if (exists $_[1]) {
877 die "Too many arguments specified\n";
878 }
879
880 my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
881
882 if (!$file_type && !$diff_status) {
883 print STDERR "svn: '$path' is not under version control\n";
884 exit 1;
885 }
886
887 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
888 unless ($gs) {
889 die "Unable to determine upstream SVN information from ",
890 "working tree history\n";
891 }
892
893 # canonicalize_path() will return "" to make libsvn 1.5.x happy,
894 $path = "." if $path eq "";
895
896 my $full_url = $url . ($fullpath eq "" ? "" : "/$fullpath");
897
898 if ($_url) {
899 print escape_url($full_url), "\n";
900 return;
901 }
902
903 my $result = "Path: $path\n";
904 $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
905 $result .= "URL: " . escape_url($full_url) . "\n";
906
907 eval {
908 my $repos_root = $gs->repos_root;
909 Git::SVN::remove_username($repos_root);
910 $result .= "Repository Root: " . escape_url($repos_root) . "\n";
911 };
912 if ($@) {
913 $result .= "Repository Root: (offline)\n";
914 }
915 $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A";
916 $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
917
918 $result .= "Node Kind: " .
919 ($file_type eq "dir" ? "directory" : "file") . "\n";
920
921 my $schedule = $diff_status eq "A"
922 ? "add"
923 : ($diff_status eq "D" ? "delete" : "normal");
924 $result .= "Schedule: $schedule\n";
925
926 if ($diff_status eq "A") {
927 print $result, "\n";
928 return;
929 }
930
931 my ($lc_author, $lc_rev, $lc_date_utc);
932 my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $fullpath);
933 my $log = command_output_pipe(@args);
934 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
935 while (<$log>) {
936 if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
937 $lc_author = $1;
938 $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
939 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
940 (undef, $lc_rev, undef) = ::extract_metadata($1);
941 }
942 }
943 close $log;
944
945 Git::SVN::Log::set_local_timezone();
946
947 $result .= "Last Changed Author: $lc_author\n";
948 $result .= "Last Changed Rev: $lc_rev\n";
949 $result .= "Last Changed Date: " .
950 Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
951
952 if ($file_type ne "dir") {
953 my $text_last_updated_date =
954 ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
955 $result .=
956 "Text Last Updated: " .
957 Git::SVN::Log::format_svn_date($text_last_updated_date) .
958 "\n";
959 my $checksum;
960 if ($diff_status eq "D") {
961 my ($fh, $ctx) =
962 command_output_pipe(qw(cat-file blob), "HEAD:$path");
963 if ($file_type eq "link") {
964 my $file_name = <$fh>;
965 $checksum = md5sum("link $file_name");
966 } else {
967 $checksum = md5sum($fh);
968 }
969 command_close_pipe($fh, $ctx);
970 } elsif ($file_type eq "link") {
971 my $file_name =
972 command(qw(cat-file blob), "HEAD:$path");
973 $checksum =
974 md5sum("link " . $file_name);
975 } else {
976 open FILE, "<", $path or die $!;
977 $checksum = md5sum(\*FILE);
978 close FILE or die $!;
979 }
980 $result .= "Checksum: " . $checksum . "\n";
981 }
982
983 print $result, "\n";
984 }
985
986 ########################### utility functions #########################
987
988 sub rebase_cmd {
989 my @cmd = qw/rebase/;
990 push @cmd, '-v' if $_verbose;
991 push @cmd, qw/--merge/ if $_merge;
992 push @cmd, "--strategy=$_strategy" if $_strategy;
993 @cmd;
994 }
995
996 sub post_fetch_checkout {
997 return if $_no_checkout;
998 my $gs = $Git::SVN::_head or return;
999 return if verify_ref('refs/heads/master^0');
1000
1001 my $valid_head = verify_ref('HEAD^0');
1002 command_noisy(qw(update-ref refs/heads/master), $gs->refname);
1003 return if ($valid_head || !verify_ref('HEAD^0'));
1004
1005 return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
1006 my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
1007 return if -f $index;
1008
1009 return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
1010 return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
1011 command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
1012 print STDERR "Checked out HEAD:\n ",
1013 $gs->full_url, " r", $gs->last_rev, "\n";
1014 }
1015
1016 sub complete_svn_url {
1017 my ($url, $path) = @_;
1018 $path =~ s#/+$##;
1019 if ($path !~ m#^[a-z\+]+://#) {
1020 if (!defined $url || $url !~ m#^[a-z\+]+://#) {
1021 fatal("E: '$path' is not a complete URL ",
1022 "and a separate URL is not specified");
1023 }
1024 return ($url, $path);
1025 }
1026 return ($path, '');
1027 }
1028
1029 sub complete_url_ls_init {
1030 my ($ra, $repo_path, $switch, $pfx) = @_;
1031 unless ($repo_path) {
1032 print STDERR "W: $switch not specified\n";
1033 return;
1034 }
1035 $repo_path =~ s#/+$##;
1036 if ($repo_path =~ m#^[a-z\+]+://#) {
1037 $ra = Git::SVN::Ra->new($repo_path);
1038 $repo_path = '';
1039 } else {
1040 $repo_path =~ s#^/+##;
1041 unless ($ra) {
1042 fatal("E: '$repo_path' is not a complete URL ",
1043 "and a separate URL is not specified");
1044 }
1045 }
1046 my $url = $ra->{url};
1047 my $gs = Git::SVN->init($url, undef, undef, undef, 1);
1048 my $k = "svn-remote.$gs->{repo_id}.url";
1049 my $orig_url = eval { command_oneline(qw/config --get/, $k) };
1050 if ($orig_url && ($orig_url ne $gs->{url})) {
1051 die "$k already set: $orig_url\n",
1052 "wanted to set to: $gs->{url}\n";
1053 }
1054 command_oneline('config', $k, $gs->{url}) unless $orig_url;
1055 my $remote_path = "$ra->{svn_path}/$repo_path";
1056 $remote_path =~ s#/+#/#g;
1057 $remote_path =~ s#^/##g;
1058 $remote_path .= "/*" if $remote_path !~ /\*/;
1059 my ($n) = ($switch =~ /^--(\w+)/);
1060 if (length $pfx && $pfx !~ m#/$#) {
1061 die "--prefix='$pfx' must have a trailing slash '/'\n";
1062 }
1063 command_noisy('config',
1064 "svn-remote.$gs->{repo_id}.$n",
1065 "$remote_path:refs/remotes/$pfx*" .
1066 ('/*' x (($remote_path =~ tr/*/*/) - 1)) );
1067 }
1068
1069 sub verify_ref {
1070 my ($ref) = @_;
1071 eval { command_oneline([ 'rev-parse', '--verify', $ref ],
1072 { STDERR => 0 }); };
1073 }
1074
1075 sub get_tree_from_treeish {
1076 my ($treeish) = @_;
1077 # $treeish can be a symbolic ref, too:
1078 my $type = command_oneline(qw/cat-file -t/, $treeish);
1079 my $expected;
1080 while ($type eq 'tag') {
1081 ($treeish, $type) = command(qw/cat-file tag/, $treeish);
1082 }
1083 if ($type eq 'commit') {
1084 $expected = (grep /^tree /, command(qw/cat-file commit/,
1085 $treeish))[0];
1086 ($expected) = ($expected =~ /^tree ($sha1)$/o);
1087 die "Unable to get tree from $treeish\n" unless $expected;
1088 } elsif ($type eq 'tree') {
1089 $expected = $treeish;
1090 } else {
1091 die "$treeish is a $type, expected tree, tag or commit\n";
1092 }
1093 return $expected;
1094 }
1095
1096 sub get_commit_entry {
1097 my ($treeish) = shift;
1098 my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
1099 my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
1100 my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
1101 open my $log_fh, '>', $commit_editmsg or croak $!;
1102
1103 my $type = command_oneline(qw/cat-file -t/, $treeish);
1104 if ($type eq 'commit' || $type eq 'tag') {
1105 my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1106 $type, $treeish);
1107 my $in_msg = 0;
1108 my $author;
1109 my $saw_from = 0;
1110 my $msgbuf = "";
1111 while (<$msg_fh>) {
1112 if (!$in_msg) {
1113 $in_msg = 1 if (/^\s*$/);
1114 $author = $1 if (/^author (.*>)/);
1115 } elsif (/^git-svn-id: /) {
1116 # skip this for now, we regenerate the
1117 # correct one on re-fetch anyways
1118 # TODO: set *:merge properties or like...
1119 } else {
1120 if (/^From:/ || /^Signed-off-by:/) {
1121 $saw_from = 1;
1122 }
1123 $msgbuf .= $_;
1124 }
1125 }
1126 $msgbuf =~ s/\s+$//s;
1127 if ($Git::SVN::_add_author_from && defined($author)
1128 && !$saw_from) {
1129 $msgbuf .= "\n\nFrom: $author";
1130 }
1131 print $log_fh $msgbuf or croak $!;
1132 command_close_pipe($msg_fh, $ctx);
1133 }
1134 close $log_fh or croak $!;
1135
1136 if ($_edit || ($type eq 'tree')) {
1137 my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
1138 # TODO: strip out spaces, comments, like git-commit.sh
1139 system($editor, $commit_editmsg);
1140 }
1141 rename $commit_editmsg, $commit_msg or croak $!;
1142 {
1143 # SVN requires messages to be UTF-8 when entering the repo
1144 local $/;
1145 open $log_fh, '<', $commit_msg or croak $!;
1146 binmode $log_fh;
1147 chomp($log_entry{log} = <$log_fh>);
1148
1149 if (my $enc = Git::config('i18n.commitencoding')) {
1150 require Encode;
1151 Encode::from_to($log_entry{log}, $enc, 'UTF-8');
1152 }
1153 close $log_fh or croak $!;
1154 }
1155 unlink $commit_msg;
1156 \%log_entry;
1157 }
1158
1159 sub s_to_file {
1160 my ($str, $file, $mode) = @_;
1161 open my $fd,'>',$file or croak $!;
1162 print $fd $str,"\n" or croak $!;
1163 close $fd or croak $!;
1164 chmod ($mode &~ umask, $file) if (defined $mode);
1165 }
1166
1167 sub file_to_s {
1168 my $file = shift;
1169 open my $fd,'<',$file or croak "$!: file: $file\n";
1170 local $/;
1171 my $ret = <$fd>;
1172 close $fd or croak $!;
1173 $ret =~ s/\s*$//s;
1174 return $ret;
1175 }
1176
1177 # '<svn username> = real-name <email address>' mapping based on git-svnimport:
1178 sub load_authors {
1179 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1180 my $log = $cmd eq 'log';
1181 while (<$authors>) {
1182 chomp;
1183 next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1184 my ($user, $name, $email) = ($1, $2, $3);
1185 if ($log) {
1186 $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1187 } else {
1188 $users{$user} = [$name, $email];
1189 }
1190 }
1191 close $authors or croak $!;
1192 }
1193
1194 # convert GetOpt::Long specs for use by git-config
1195 sub read_repo_config {
1196 return unless -d $ENV{GIT_DIR};
1197 my $opts = shift;
1198 my @config_only;
1199 foreach my $o (keys %$opts) {
1200 # if we have mixedCase and a long option-only, then
1201 # it's a config-only variable that we don't need for
1202 # the command-line.
1203 push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1204 my $v = $opts->{$o};
1205 my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1206 $key =~ s/-//g;
1207 my $arg = 'git config';
1208 $arg .= ' --int' if ($o =~ /[:=]i$/);
1209 $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1210 if (ref $v eq 'ARRAY') {
1211 chomp(my @tmp = `$arg --get-all svn.$key`);
1212 @$v = @tmp if @tmp;
1213 } else {
1214 chomp(my $tmp = `$arg --get svn.$key`);
1215 if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1216 $$v = $tmp;
1217 }
1218 }
1219 }
1220 delete @$opts{@config_only} if @config_only;
1221 }
1222
1223 sub extract_metadata {
1224 my $id = shift or return (undef, undef, undef);
1225 my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1226 \s([a-f\d\-]+)$/x);
1227 if (!defined $rev || !$uuid || !$url) {
1228 # some of the original repositories I made had
1229 # identifiers like this:
1230 ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/);
1231 }
1232 return ($url, $rev, $uuid);
1233 }
1234
1235 sub cmt_metadata {
1236 return extract_metadata((grep(/^git-svn-id: /,
1237 command(qw/cat-file commit/, shift)))[-1]);
1238 }
1239
1240 sub working_head_info {
1241 my ($head, $refs) = @_;
1242 my @args = ('log', '--no-color', '--first-parent', '--pretty=medium');
1243 my ($fh, $ctx) = command_output_pipe(@args, $head);
1244 my $hash;
1245 my %max;
1246 while (<$fh>) {
1247 if ( m{^commit ($::sha1)$} ) {
1248 unshift @$refs, $hash if $hash and $refs;
1249 $hash = $1;
1250 next;
1251 }
1252 next unless s{^\s*(git-svn-id:)}{$1};
1253 my ($url, $rev, $uuid) = extract_metadata($_);
1254 if (defined $url && defined $rev) {
1255 next if $max{$url} and $max{$url} < $rev;
1256 if (my $gs = Git::SVN->find_by_url($url)) {
1257 my $c = $gs->rev_map_get($rev, $uuid);
1258 if ($c && $c eq $hash) {
1259 close $fh; # break the pipe
1260 return ($url, $rev, $uuid, $gs);
1261 } else {
1262 $max{$url} ||= $gs->rev_map_max;
1263 }
1264 }
1265 }
1266 }
1267 command_close_pipe($fh, $ctx);
1268 (undef, undef, undef, undef);
1269 }
1270
1271 sub read_commit_parents {
1272 my ($parents, $c) = @_;
1273 chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1274 $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1275 @{$parents->{$c}} = split(/ /, $p);
1276 }
1277
1278 sub linearize_history {
1279 my ($gs, $refs) = @_;
1280 my %parents;
1281 foreach my $c (@$refs) {
1282 read_commit_parents(\%parents, $c);
1283 }
1284
1285 my @linear_refs;
1286 my %skip = ();
1287 my $last_svn_commit = $gs->last_commit;
1288 foreach my $c (reverse @$refs) {
1289 next if $c eq $last_svn_commit;
1290 last if $skip{$c};
1291
1292 unshift @linear_refs, $c;
1293 $skip{$c} = 1;
1294
1295 # we only want the first parent to diff against for linear
1296 # history, we save the rest to inject when we finalize the
1297 # svn commit
1298 my $fp_a = verify_ref("$c~1");
1299 my $fp_b = shift @{$parents{$c}} if $parents{$c};
1300 if (!$fp_a || !$fp_b) {
1301 die "Commit $c\n",
1302 "has no parent commit, and therefore ",
1303 "nothing to diff against.\n",
1304 "You should be working from a repository ",
1305 "originally created by git-svn\n";
1306 }
1307 if ($fp_a ne $fp_b) {
1308 die "$c~1 = $fp_a, however parsing commit $c ",
1309 "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1310 }
1311
1312 foreach my $p (@{$parents{$c}}) {
1313 $skip{$p} = 1;
1314 }
1315 }
1316 (\@linear_refs, \%parents);
1317 }
1318
1319 sub find_file_type_and_diff_status {
1320 my ($path) = @_;
1321 return ('dir', '') if $path eq '';
1322
1323 my $diff_output =
1324 command_oneline(qw(diff --cached --name-status --), $path) || "";
1325 my $diff_status = (split(' ', $diff_output))[0] || "";
1326
1327 my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1328
1329 return (undef, undef) if !$diff_status && !$ls_tree;
1330
1331 if ($diff_status eq "A") {
1332 return ("link", $diff_status) if -l $path;
1333 return ("dir", $diff_status) if -d $path;
1334 return ("file", $diff_status);
1335 }
1336
1337 my $mode = (split(' ', $ls_tree))[0] || "";
1338
1339 return ("link", $diff_status) if $mode eq "120000";
1340 return ("dir", $diff_status) if $mode eq "040000";
1341 return ("file", $diff_status);
1342 }
1343
1344 sub md5sum {
1345 my $arg = shift;
1346 my $ref = ref $arg;
1347 my $md5 = Digest::MD5->new();
1348 if ($ref eq 'GLOB' || $ref eq 'IO::File' || $ref eq 'File::Temp') {
1349 $md5->addfile($arg) or croak $!;
1350 } elsif ($ref eq 'SCALAR') {
1351 $md5->add($$arg) or croak $!;
1352 } elsif (!$ref) {
1353 $md5->add($arg) or croak $!;
1354 } else {
1355 ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
1356 }
1357 return $md5->hexdigest();
1358 }
1359
1360 package Git::SVN;
1361 use strict;
1362 use warnings;
1363 use Fcntl qw/:DEFAULT :seek/;
1364 use constant rev_map_fmt => 'NH40';
1365 use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1366 $_repack $_repack_flags $_use_svm_props $_head
1367 $_use_svnsync_props $no_reuse_existing $_minimize_url
1368 $_use_log_author $_add_author_from $_localtime/;
1369 use Carp qw/croak/;
1370 use File::Path qw/mkpath/;
1371 use File::Copy qw/copy/;
1372 use IPC::Open3;
1373
1374 my ($_gc_nr, $_gc_period);
1375
1376 # properties that we do not log:
1377 my %SKIP_PROP;
1378 BEGIN {
1379 %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1380 svn:special svn:executable
1381 svn:entry:committed-rev
1382 svn:entry:last-author
1383 svn:entry:uuid
1384 svn:entry:committed-date/;
1385
1386 # some options are read globally, but can be overridden locally
1387 # per [svn-remote "..."] section. Command-line options will *NOT*
1388 # override options set in an [svn-remote "..."] section
1389 no strict 'refs';
1390 for my $option (qw/follow_parent no_metadata use_svm_props
1391 use_svnsync_props/) {
1392 my $key = $option;
1393 $key =~ tr/_//d;
1394 my $prop = "-$option";
1395 *$option = sub {
1396 my ($self) = @_;
1397 return $self->{$prop} if exists $self->{$prop};
1398 my $k = "svn-remote.$self->{repo_id}.$key";
1399 eval { command_oneline(qw/config --get/, $k) };
1400 if ($@) {
1401 $self->{$prop} = ${"Git::SVN::_$option"};
1402 } else {
1403 my $v = command_oneline(qw/config --bool/,$k);
1404 $self->{$prop} = $v eq 'false' ? 0 : 1;
1405 }
1406 return $self->{$prop};
1407 }
1408 }
1409 }
1410
1411
1412 my (%LOCKFILES, %INDEX_FILES);
1413 END {
1414 unlink keys %LOCKFILES if %LOCKFILES;
1415 unlink keys %INDEX_FILES if %INDEX_FILES;
1416 }
1417
1418 sub resolve_local_globs {
1419 my ($url, $fetch, $glob_spec) = @_;
1420 return unless defined $glob_spec;
1421 my $ref = $glob_spec->{ref};
1422 my $path = $glob_spec->{path};
1423 foreach (command(qw#for-each-ref --format=%(refname) refs/remotes#)) {
1424 next unless m#^refs/remotes/$ref->{regex}$#;
1425 my $p = $1;
1426 my $pathname = desanitize_refname($path->full_path($p));
1427 my $refname = desanitize_refname($ref->full_path($p));
1428 if (my $existing = $fetch->{$pathname}) {
1429 if ($existing ne $refname) {
1430 die "Refspec conflict:\n",
1431 "existing: refs/remotes/$existing\n",
1432 " globbed: refs/remotes/$refname\n";
1433 }
1434 my $u = (::cmt_metadata("refs/remotes/$refname"))[0];
1435 $u =~ s!^\Q$url\E(/|$)!! or die
1436 "refs/remotes/$refname: '$url' not found in '$u'\n";
1437 if ($pathname ne $u) {
1438 warn "W: Refspec glob conflict ",
1439 "(ref: refs/remotes/$refname):\n",
1440 "expected path: $pathname\n",
1441 " real path: $u\n",
1442 "Continuing ahead with $u\n";
1443 next;
1444 }
1445 } else {
1446 $fetch->{$pathname} = $refname;
1447 }
1448 }
1449 }
1450
1451 sub parse_revision_argument {
1452 my ($base, $head) = @_;
1453 if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1454 return ($base, $head);
1455 }
1456 return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1457 return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1458 return ($head, $head) if ($::_revision eq 'HEAD');
1459 return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1460 return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1461 die "revision argument: $::_revision not understood by git-svn\n";
1462 }
1463
1464 sub fetch_all {
1465 my ($repo_id, $remotes) = @_;
1466 if (ref $repo_id) {
1467 my $gs = $repo_id;
1468 $repo_id = undef;
1469 $repo_id = $gs->{repo_id};
1470 }
1471 $remotes ||= read_all_remotes();
1472 my $remote = $remotes->{$repo_id} or
1473 die "[svn-remote \"$repo_id\"] unknown\n";
1474 my $fetch = $remote->{fetch};
1475 my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1476 my (@gs, @globs);
1477 my $ra = Git::SVN::Ra->new($url);
1478 my $uuid = $ra->get_uuid;
1479 my $head = $ra->get_latest_revnum;
1480 my $base = defined $fetch ? $head : 0;
1481
1482 # read the max revs for wildcard expansion (branches/*, tags/*)
1483 foreach my $t (qw/branches tags/) {
1484 defined $remote->{$t} or next;
1485 push @globs, $remote->{$t};
1486 my $max_rev = eval { tmp_config(qw/--int --get/,
1487 "svn-remote.$repo_id.${t}-maxRev") };
1488 if (defined $max_rev && ($max_rev < $base)) {
1489 $base = $max_rev;
1490 } elsif (!defined $max_rev) {
1491 $base = 0;
1492 }
1493 }
1494
1495 if ($fetch) {
1496 foreach my $p (sort keys %$fetch) {
1497 my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1498 my $lr = $gs->rev_map_max;
1499 if (defined $lr) {
1500 $base = $lr if ($lr < $base);
1501 }
1502 push @gs, $gs;
1503 }
1504 }
1505
1506 ($base, $head) = parse_revision_argument($base, $head);
1507 $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1508 }
1509
1510 sub read_all_remotes {
1511 my $r = {};
1512 my $use_svm_props = eval { command_oneline(qw/config --bool
1513 svn.useSvmProps/) };
1514 $use_svm_props = $use_svm_props eq 'true' if $use_svm_props;
1515 foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1516 if (m!^(.+)\.fetch=\s*(.*)\s*:\s*(.+)\s*$!) {
1517 my ($remote, $local_ref, $_remote_ref) = ($1, $2, $3);
1518 die("svn-remote.$remote: remote ref '$_remote_ref' "
1519 . "must start with 'refs/remotes/'\n")
1520 unless $_remote_ref =~ m{^refs/remotes/(.+)};
1521 my $remote_ref = $1;
1522 $local_ref =~ s{^/}{};
1523 $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1524 $r->{$remote}->{svm} = {} if $use_svm_props;
1525 } elsif (m!^(.+)\.usesvmprops=\s*(.*)\s*$!) {
1526 $r->{$1}->{svm} = {};
1527 } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1528 $r->{$1}->{url} = $2;
1529 } elsif (m!^(.+)\.(branches|tags)=
1530 (.*):refs/remotes/(.+)\s*$/!x) {
1531 my ($p, $g) = ($3, $4);
1532 my $rs = $r->{$1}->{$2} = {
1533 t => $2,
1534 remote => $1,
1535 path => Git::SVN::GlobSpec->new($p),
1536 ref => Git::SVN::GlobSpec->new($g) };
1537 if (length($rs->{ref}->{right}) != 0) {
1538 die "The '*' glob character must be the last ",
1539 "character of '$g'\n";
1540 }
1541 }
1542 }
1543
1544 map {
1545 if (defined $r->{$_}->{svm}) {
1546 my $svm;
1547 eval {
1548 my $section = "svn-remote.$_";
1549 $svm = {
1550 source => tmp_config('--get',
1551 "$section.svm-source"),
1552 replace => tmp_config('--get',
1553 "$section.svm-replace"),
1554 }
1555 };
1556 $r->{$_}->{svm} = $svm;
1557 }
1558 } keys %$r;
1559
1560 $r;
1561 }
1562
1563 sub init_vars {
1564 $_gc_nr = $_gc_period = 1000;
1565 if (defined $_repack || defined $_repack_flags) {
1566 warn "Repack options are obsolete; they have no effect.\n";
1567 }
1568 }
1569
1570 sub verify_remotes_sanity {
1571 return unless -d $ENV{GIT_DIR};
1572 my %seen;
1573 foreach (command(qw/config -l/)) {
1574 if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1575 if ($seen{$1}) {
1576 die "Remote ref refs/remote/$1 is tracked by",
1577 "\n \"$_\"\nand\n \"$seen{$1}\"\n",
1578 "Please resolve this ambiguity in ",
1579 "your git configuration file before ",
1580 "continuing\n";
1581 }
1582 $seen{$1} = $_;
1583 }
1584 }
1585 }
1586
1587 sub find_existing_remote {
1588 my ($url, $remotes) = @_;
1589 return undef if $no_reuse_existing;
1590 my $existing;
1591 foreach my $repo_id (keys %$remotes) {
1592 my $u = $remotes->{$repo_id}->{url} or next;
1593 next if $u ne $url;
1594 $existing = $repo_id;
1595 last;
1596 }
1597 $existing;
1598 }
1599
1600 sub init_remote_config {
1601 my ($self, $url, $no_write) = @_;
1602 $url =~ s!/+$!!; # strip trailing slash
1603 my $r = read_all_remotes();
1604 my $existing = find_existing_remote($url, $r);
1605 if ($existing) {
1606 unless ($no_write) {
1607 print STDERR "Using existing ",
1608 "[svn-remote \"$existing\"]\n";
1609 }
1610 $self->{repo_id} = $existing;
1611 } elsif ($_minimize_url) {
1612 my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1613 $existing = find_existing_remote($min_url, $r);
1614 if ($existing) {
1615 unless ($no_write) {
1616 print STDERR "Using existing ",
1617 "[svn-remote \"$existing\"]\n";
1618 }
1619 $self->{repo_id} = $existing;
1620 }
1621 if ($min_url ne $url) {
1622 unless ($no_write) {
1623 print STDERR "Using higher level of URL: ",
1624 "$url => $min_url\n";
1625 }
1626 my $old_path = $self->{path};
1627 $self->{path} = $url;
1628 $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1629 if (length $old_path) {
1630 $self->{path} .= "/$old_path";
1631 }
1632 $url = $min_url;
1633 }
1634 }
1635 my $orig_url;
1636 if (!$existing) {
1637 # verify that we aren't overwriting anything:
1638 $orig_url = eval {
1639 command_oneline('config', '--get',
1640 "svn-remote.$self->{repo_id}.url")
1641 };
1642 if ($orig_url && ($orig_url ne $url)) {
1643 die "svn-remote.$self->{repo_id}.url already set: ",
1644 "$orig_url\nwanted to set to: $url\n";
1645 }
1646 }
1647 my ($xrepo_id, $xpath) = find_ref($self->refname);
1648 if (defined $xpath) {
1649 die "svn-remote.$xrepo_id.fetch already set to track ",
1650 "$xpath:refs/remotes/", $self->refname, "\n";
1651 }
1652 unless ($no_write) {
1653 command_noisy('config',
1654 "svn-remote.$self->{repo_id}.url", $url);
1655 $self->{path} =~ s{^/}{};
1656 command_noisy('config', '--add',
1657 "svn-remote.$self->{repo_id}.fetch",
1658 "$self->{path}:".$self->refname);
1659 }
1660 $self->{url} = $url;
1661 }
1662
1663 sub find_by_url { # repos_root and, path are optional
1664 my ($class, $full_url, $repos_root, $path) = @_;
1665
1666 return undef unless defined $full_url;
1667 remove_username($full_url);
1668 remove_username($repos_root) if defined $repos_root;
1669 my $remotes = read_all_remotes();
1670 if (defined $full_url && defined $repos_root && !defined $path) {
1671 $path = $full_url;
1672 $path =~ s#^\Q$repos_root\E(?:/|$)##;
1673 }
1674 foreach my $repo_id (keys %$remotes) {
1675 my $u = $remotes->{$repo_id}->{url} or next;
1676 remove_username($u);
1677 next if defined $repos_root && $repos_root ne $u;
1678
1679 my $fetch = $remotes->{$repo_id}->{fetch} || {};
1680 foreach (qw/branches tags/) {
1681 resolve_local_globs($u, $fetch,
1682 $remotes->{$repo_id}->{$_});
1683 }
1684 my $p = $path;
1685 my $rwr = rewrite_root({repo_id => $repo_id});
1686 my $svm = $remotes->{$repo_id}->{svm}
1687 if defined $remotes->{$repo_id}->{svm};
1688 unless (defined $p) {
1689 $p = $full_url;
1690 my $z = $u;
1691 my $prefix = '';
1692 if ($rwr) {
1693 $z = $rwr;
1694 } elsif (defined $svm) {
1695 $z = $svm->{source};
1696 $prefix = $svm->{replace};
1697 $prefix =~ s#^\Q$u\E(?:/|$)##;
1698 $prefix =~ s#/$##;
1699 }
1700 $p =~ s#^\Q$z\E(?:/|$)#$prefix# or next;
1701 }
1702 foreach my $f (keys %$fetch) {
1703 next if $f ne $p;
1704 return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1705 }
1706 }
1707 undef;
1708 }
1709
1710 sub init {
1711 my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1712 my $self = _new($class, $repo_id, $ref_id, $path);
1713 if (defined $url) {
1714 $self->init_remote_config($url, $no_write);
1715 }
1716 $self;
1717 }
1718
1719 sub find_ref {
1720 my ($ref_id) = @_;
1721 foreach (command(qw/config -l/)) {
1722 next unless m!^svn-remote\.(.+)\.fetch=
1723 \s*(.*)\s*:\s*refs/remotes/(.+)\s*$!x;
1724 my ($repo_id, $path, $ref) = ($1, $2, $3);
1725 if ($ref eq $ref_id) {
1726 $path = '' if ($path =~ m#^\./?#);
1727 return ($repo_id, $path);
1728 }
1729 }
1730 (undef, undef, undef);
1731 }
1732
1733 sub new {
1734 my ($class, $ref_id, $repo_id, $path) = @_;
1735 if (defined $ref_id && !defined $repo_id && !defined $path) {
1736 ($repo_id, $path) = find_ref($ref_id);
1737 if (!defined $repo_id) {
1738 die "Could not find a \"svn-remote.*.fetch\" key ",
1739 "in the repository configuration matching: ",
1740 "refs/remotes/$ref_id\n";
1741 }
1742 }
1743 my $self = _new($class, $repo_id, $ref_id, $path);
1744 if (!defined $self->{path} || !length $self->{path}) {
1745 my $fetch = command_oneline('config', '--get',
1746 "svn-remote.$repo_id.fetch",
1747 ":refs/remotes/$ref_id\$") or
1748 die "Failed to read \"svn-remote.$repo_id.fetch\" ",
1749 "\":refs/remotes/$ref_id\$\" in config\n";
1750 ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
1751 }
1752 $self->{url} = command_oneline('config', '--get',
1753 "svn-remote.$repo_id.url") or
1754 die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
1755 $self->rebuild;
1756 $self;
1757 }
1758
1759 sub refname {
1760 my ($refname) = "refs/remotes/$_[0]->{ref_id}" ;
1761
1762 # It cannot end with a slash /, we'll throw up on this because
1763 # SVN can't have directories with a slash in their name, either:
1764 if ($refname =~ m{/$}) {
1765 die "ref: '$refname' ends with a trailing slash, this is ",
1766 "not permitted by git nor Subversion\n";
1767 }
1768
1769 # It cannot have ASCII control character space, tilde ~, caret ^,
1770 # colon :, question-mark ?, asterisk *, space, or open bracket [
1771 # anywhere.
1772 #
1773 # Additionally, % must be escaped because it is used for escaping
1774 # and we want our escaped refname to be reversible
1775 $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
1776
1777 # no slash-separated component can begin with a dot .
1778 # /.* becomes /%2E*
1779 $refname =~ s{/\.}{/%2E}g;
1780
1781 # It cannot have two consecutive dots .. anywhere
1782 # .. becomes %2E%2E
1783 $refname =~ s{\.\.}{%2E%2E}g;
1784
1785 return $refname;
1786 }
1787
1788 sub desanitize_refname {
1789 my ($refname) = @_;
1790 $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
1791 return $refname;
1792 }
1793
1794 sub svm_uuid {
1795 my ($self) = @_;
1796 return $self->{svm}->{uuid} if $self->svm;
1797 $self->ra;
1798 unless ($self->{svm}) {
1799 die "SVM UUID not cached, and reading remotely failed\n";
1800 }
1801 $self->{svm}->{uuid};
1802 }
1803
1804 sub svm {
1805 my ($self) = @_;
1806 return $self->{svm} if $self->{svm};
1807 my $svm;
1808 # see if we have it in our config, first:
1809 eval {
1810 my $section = "svn-remote.$self->{repo_id}";
1811 $svm = {
1812 source => tmp_config('--get', "$section.svm-source"),
1813 uuid => tmp_config('--get', "$section.svm-uuid"),
1814 replace => tmp_config('--get', "$section.svm-replace"),
1815 }
1816 };
1817 if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
1818 $self->{svm} = $svm;
1819 }
1820 $self->{svm};
1821 }
1822
1823 sub _set_svm_vars {
1824 my ($self, $ra) = @_;
1825 return $ra if $self->svm;
1826
1827 my @err = ( "useSvmProps set, but failed to read SVM properties\n",
1828 "(svm:source, svm:uuid) ",
1829 "from the following URLs:\n" );
1830 sub read_svm_props {
1831 my ($self, $ra, $path, $r) = @_;
1832 my $props = ($ra->get_dir($path, $r))[2];
1833 my $src = $props->{'svm:source'};
1834 my $uuid = $props->{'svm:uuid'};
1835 return undef if (!$src || !$uuid);
1836
1837 chomp($src, $uuid);
1838
1839 $uuid =~ m{^[0-9a-f\-]{30,}$}
1840 or die "doesn't look right - svm:uuid is '$uuid'\n";
1841
1842 # the '!' is used to mark the repos_root!/relative/path
1843 $src =~ s{/?!/?}{/};
1844 $src =~ s{/+$}{}; # no trailing slashes please
1845 # username is of no interest
1846 $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
1847
1848 my $replace = $ra->{url};
1849 $replace .= "/$path" if length $path;
1850
1851 my $section = "svn-remote.$self->{repo_id}";
1852 tmp_config("$section.svm-source", $src);
1853 tmp_config("$section.svm-replace", $replace);
1854 tmp_config("$section.svm-uuid", $uuid);
1855 $self->{svm} = {
1856 source => $src,
1857 uuid => $uuid,
1858 replace => $replace
1859 };
1860 }
1861
1862 my $r = $ra->get_latest_revnum;
1863 my $path = $self->{path};
1864 my %tried;
1865 while (length $path) {
1866 unless ($tried{"$self->{url}/$path"}) {
1867 return $ra if $self->read_svm_props($ra, $path, $r);
1868 $tried{"$self->{url}/$path"} = 1;
1869 }
1870 $path =~ s#/?[^/]+$##;
1871 }
1872 die "Path: '$path' should be ''\n" if $path ne '';
1873 return $ra if $self->read_svm_props($ra, $path, $r);
1874 $tried{"$self->{url}/$path"} = 1;
1875
1876 if ($ra->{repos_root} eq $self->{url}) {
1877 die @err, (map { " $_\n" } keys %tried), "\n";
1878 }
1879
1880 # nope, make sure we're connected to the repository root:
1881 my $ok;
1882 my @tried_b;
1883 $path = $ra->{svn_path};
1884 $ra = Git::SVN::Ra->new($ra->{repos_root});
1885 while (length $path) {
1886 unless ($tried{"$ra->{url}/$path"}) {
1887 $ok = $self->read_svm_props($ra, $path, $r);
1888 last if $ok;
1889 $tried{"$ra->{url}/$path"} = 1;
1890 }
1891 $path =~ s#/?[^/]+$##;
1892 }
1893 die "Path: '$path' should be ''\n" if $path ne '';
1894 $ok ||= $self->read_svm_props($ra, $path, $r);
1895 $tried{"$ra->{url}/$path"} = 1;
1896 if (!$ok) {
1897 die @err, (map { " $_\n" } keys %tried), "\n";
1898 }
1899 Git::SVN::Ra->new($self->{url});
1900 }
1901
1902 sub svnsync {
1903 my ($self) = @_;
1904 return $self->{svnsync} if $self->{svnsync};
1905
1906 if ($self->no_metadata) {
1907 die "Can't have both 'noMetadata' and ",
1908 "'useSvnsyncProps' options set!\n";
1909 }
1910 if ($self->rewrite_root) {
1911 die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
1912 "options set!\n";
1913 }
1914
1915 my $svnsync;
1916 # see if we have it in our config, first:
1917 eval {
1918 my $section = "svn-remote.$self->{repo_id}";
1919
1920 my $url = tmp_config('--get', "$section.svnsync-url");
1921 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1922 die "doesn't look right - svn:sync-from-url is '$url'\n";
1923
1924 my $uuid = tmp_config('--get', "$section.svnsync-uuid");
1925 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1926 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1927
1928 $svnsync = { url => $url, uuid => $uuid }
1929 };
1930 if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
1931 return $self->{svnsync} = $svnsync;
1932 }
1933
1934 my $err = "useSvnsyncProps set, but failed to read " .
1935 "svnsync property: svn:sync-from-";
1936 my $rp = $self->ra->rev_proplist(0);
1937
1938 my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
1939 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1940 die "doesn't look right - svn:sync-from-url is '$url'\n";
1941
1942 my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
1943 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1944 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1945
1946 my $section = "svn-remote.$self->{repo_id}";
1947 tmp_config('--add', "$section.svnsync-uuid", $uuid);
1948 tmp_config('--add', "$section.svnsync-url", $url);
1949 return $self->{svnsync} = { url => $url, uuid => $uuid };
1950 }
1951
1952 # this allows us to memoize our SVN::Ra UUID locally and avoid a
1953 # remote lookup (useful for 'git svn log').
1954 sub ra_uuid {
1955 my ($self) = @_;
1956 unless ($self->{ra_uuid}) {
1957 my $key = "svn-remote.$self->{repo_id}.uuid";
1958 my $uuid = eval { tmp_config('--get', $key) };
1959 if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/) {
1960 $self->{ra_uuid} = $uuid;
1961 } else {
1962 die "ra_uuid called without URL\n" unless $self->{url};
1963 $self->{ra_uuid} = $self->ra->get_uuid;
1964 tmp_config('--add', $key, $self->{ra_uuid});
1965 }
1966 }
1967 $self->{ra_uuid};
1968 }
1969
1970 sub _set_repos_root {
1971 my ($self, $repos_root) = @_;
1972 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1973 $repos_root ||= $self->ra->{repos_root};
1974 tmp_config($k, $repos_root);
1975 $repos_root;
1976 }
1977
1978 sub repos_root {
1979 my ($self) = @_;
1980 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1981 eval { tmp_config('--get', $k) } || $self->_set_repos_root;
1982 }
1983
1984 sub ra {
1985 my ($self) = shift;
1986 my $ra = Git::SVN::Ra->new($self->{url});
1987 $self->_set_repos_root($ra->{repos_root});
1988 if ($self->use_svm_props && !$self->{svm}) {
1989 if ($self->no_metadata) {
1990 die "Can't have both 'noMetadata' and ",
1991 "'useSvmProps' options set!\n";
1992 } elsif ($self->use_svnsync_props) {
1993 die "Can't have both 'useSvnsyncProps' and ",
1994 "'useSvmProps' options set!\n";
1995 }
1996 $ra = $self->_set_svm_vars($ra);
1997 $self->{-want_revprops} = 1;
1998 }
1999 $ra;
2000 }
2001
2002 sub rel_path {
2003 my ($self) = @_;
2004 my $repos_root = $self->ra->{repos_root};
2005 return $self->{path} if ($self->{url} eq $repos_root);
2006 my $url = $self->{url} .
2007 (length $self->{path} ? "/$self->{path}" : $self->{path});
2008 $url =~ s!^\Q$repos_root\E(?:/+|$)!!g;
2009 $url;
2010 }
2011
2012 # prop_walk(PATH, REV, SUB)
2013 # -------------------------
2014 # Recursively traverse PATH at revision REV and invoke SUB for each
2015 # directory that contains a SVN property. SUB will be invoked as
2016 # follows: &SUB(gs, path, props); where `gs' is this instance of
2017 # Git::SVN, `path' the path to the directory where the properties
2018 # `props' were found. The `path' will be relative to point of checkout,
2019 # that is, if url://repo/trunk is the current Git branch, and that
2020 # directory contains a sub-directory `d', SUB will be invoked with `/d/'
2021 # as `path' (note the trailing `/').
2022 sub prop_walk {
2023 my ($self, $path, $rev, $sub) = @_;
2024
2025 $path =~ s#^/##;
2026 my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
2027 $path =~ s#^/*#/#g;
2028 my $p = $path;
2029 # Strip the irrelevant part of the path.
2030 $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
2031 # Ensure the path is terminated by a `/'.
2032 $p =~ s#/*$#/#;
2033
2034 # The properties contain all the internal SVN stuff nobody
2035 # (usually) cares about.
2036 my $interesting_props = 0;
2037 foreach (keys %{$props}) {
2038 # If it doesn't start with `svn:', it must be a
2039 # user-defined property.
2040 ++$interesting_props and next if $_ !~ /^svn:/;
2041 # FIXME: Fragile, if SVN adds new public properties,
2042 # this needs to be updated.
2043 ++$interesting_props if /^svn:(?:ignore|keywords|executable
2044 |eol-style|mime-type
2045 |externals|needs-lock)$/x;
2046 }
2047 &$sub($self, $p, $props) if $interesting_props;
2048
2049 foreach (sort keys %$dirent) {
2050 next if $dirent->{$_}->{kind} != $SVN::Node::dir;
2051 $self->prop_walk($self->{path} . $p . $_, $rev, $sub);
2052 }
2053 }
2054
2055 sub last_rev { ($_[0]->last_rev_commit)[0] }
2056 sub last_commit { ($_[0]->last_rev_commit)[1] }
2057
2058 # returns the newest SVN revision number and newest commit SHA1
2059 sub last_rev_commit {
2060 my ($self) = @_;
2061 if (defined $self->{last_rev} && defined $self->{last_commit}) {
2062 return ($self->{last_rev}, $self->{last_commit});
2063 }
2064 my $c = ::verify_ref($self->refname.'^0');
2065 if ($c && !$self->use_svm_props && !$self->no_metadata) {
2066 my $rev = (::cmt_metadata($c))[1];
2067 if (defined $rev) {
2068 ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
2069 return ($rev, $c);
2070 }
2071 }
2072 my $map_path = $self->map_path;
2073 unless (-e $map_path) {
2074 ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
2075 return (undef, undef);
2076 }
2077 my ($rev, $commit) = $self->rev_map_max(1);
2078 ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
2079 return ($rev, $commit);
2080 }
2081
2082 sub get_fetch_range {
2083 my ($self, $min, $max) = @_;
2084 $max ||= $self->ra->get_latest_revnum;
2085 $min ||= $self->rev_map_max;
2086 (++$min, $max);
2087 }
2088
2089 sub tmp_config {
2090 my (@args) = @_;
2091 my $old_def_config = "$ENV{GIT_DIR}/svn/config";
2092 my $config = "$ENV{GIT_DIR}/svn/.metadata";
2093 if (! -f $config && -f $old_def_config) {
2094 rename $old_def_config, $config or
2095 die "Failed rename $old_def_config => $config: $!\n";
2096 }
2097 my $old_config = $ENV{GIT_CONFIG};
2098 $ENV{GIT_CONFIG} = $config;
2099 $@ = undef;
2100 my @ret = eval {
2101 unless (-f $config) {
2102 mkfile($config);
2103 open my $fh, '>', $config or
2104 die "Can't open $config: $!\n";
2105 print $fh "; This file is used internally by ",
2106 "git-svn\n" or die
2107 "Couldn't write to $config: $!\n";
2108 print $fh "; You should not have to edit it\n" or
2109 die "Couldn't write to $config: $!\n";
2110 close $fh or die "Couldn't close $config: $!\n";
2111 }
2112 command('config', @args);
2113 };
2114 my $err = $@;
2115 if (defined $old_config) {
2116 $ENV{GIT_CONFIG} = $old_config;
2117 } else {
2118 delete $ENV{GIT_CONFIG};
2119 }
2120 die $err if $err;
2121 wantarray ? @ret : $ret[0];
2122 }
2123
2124 sub tmp_index_do {
2125 my ($self, $sub) = @_;
2126 my $old_index = $ENV{GIT_INDEX_FILE};
2127 $ENV{GIT_INDEX_FILE} = $self->{index};
2128 $@ = undef;
2129 my @ret = eval {
2130 my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
2131 mkpath([$dir]) unless -d $dir;
2132 &$sub;
2133 };
2134 my $err = $@;
2135 if (defined $old_index) {
2136 $ENV{GIT_INDEX_FILE} = $old_index;
2137 } else {
2138 delete $ENV{GIT_INDEX_FILE};
2139 }
2140 die $err if $err;
2141 wantarray ? @ret : $ret[0];
2142 }
2143
2144 sub assert_index_clean {
2145 my ($self, $treeish) = @_;
2146
2147 $self->tmp_index_do(sub {
2148 command_noisy('read-tree', $treeish) unless -e $self->{index};
2149 my $x = command_oneline('write-tree');
2150 my ($y) = (command(qw/cat-file commit/, $treeish) =~
2151 /^tree ($::sha1)/mo);
2152 return if $y eq $x;
2153
2154 warn "Index mismatch: $y != $x\nrereading $treeish\n";
2155 unlink $self->{index} or die "unlink $self->{index}: $!\n";
2156 command_noisy('read-tree', $treeish);
2157 $x = command_oneline('write-tree');
2158 if ($y ne $x) {
2159 ::fatal "trees ($treeish) $y != $x\n",
2160 "Something is seriously wrong...";
2161 }
2162 });
2163 }
2164
2165 sub get_commit_parents {
2166 my ($self, $log_entry) = @_;
2167 my (%seen, @ret, @tmp);
2168 # legacy support for 'set-tree'; this is only used by set_tree_cb:
2169 if (my $ip = $self->{inject_parents}) {
2170 if (my $commit = delete $ip->{$log_entry->{revision}}) {
2171 push @tmp, $commit;
2172 }
2173 }
2174 if (my $cur = ::verify_ref($self->refname.'^0')) {
2175 push @tmp, $cur;
2176 }
2177 if (my $ipd = $self->{inject_parents_dcommit}) {
2178 if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2179 push @tmp, @$commit;
2180 }
2181 }
2182 push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2183 while (my $p = shift @tmp) {
2184 next if $seen{$p};
2185 $seen{$p} = 1;
2186 push @ret, $p;
2187 # MAXPARENT is defined to 16 in commit-tree.c:
2188 last if @ret >= 16;
2189 }
2190 if (@tmp) {
2191 die "r$log_entry->{revision}: No room for parents:\n\t",
2192 join("\n\t", @tmp), "\n";
2193 }
2194 @ret;
2195 }
2196
2197 sub rewrite_root {
2198 my ($self) = @_;
2199 return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2200 my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2201 my $rwr = eval { command_oneline(qw/config --get/, $k) };
2202 if ($rwr) {
2203 $rwr =~ s#/+$##;
2204 if ($rwr !~ m#^[a-z\+]+://#) {
2205 die "$rwr is not a valid URL (key: $k)\n";
2206 }
2207 }
2208 $self->{-rewrite_root} = $rwr;
2209 }
2210
2211 sub metadata_url {
2212 my ($self) = @_;
2213 ($self->rewrite_root || $self->{url}) .
2214 (length $self->{path} ? '/' . $self->{path} : '');
2215 }
2216
2217 sub full_url {
2218 my ($self) = @_;
2219 $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2220 }
2221
2222
2223 sub set_commit_header_env {
2224 my ($log_entry) = @_;
2225 my %env;
2226 foreach my $ned (qw/NAME EMAIL DATE/) {
2227 foreach my $ac (qw/AUTHOR COMMITTER/) {
2228 $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2229 }
2230 }
2231
2232 $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2233 $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2234 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2235
2236 $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2237 ? $log_entry->{commit_name}
2238 : $log_entry->{name};
2239 $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2240 ? $log_entry->{commit_email}
2241 : $log_entry->{email};
2242 \%env;
2243 }
2244
2245 sub restore_commit_header_env {
2246 my ($env) = @_;
2247 foreach my $ned (qw/NAME EMAIL DATE/) {
2248 foreach my $ac (qw/AUTHOR COMMITTER/) {
2249 my $k = "GIT_${ac}_${ned}";
2250 if (defined $env->{$k}) {
2251 $ENV{$k} = $env->{$k};
2252 } else {
2253 delete $ENV{$k};
2254 }
2255 }
2256 }
2257 }
2258
2259 sub gc {
2260 command_noisy('gc', '--auto');
2261 };
2262
2263 sub do_git_commit {
2264 my ($self, $log_entry) = @_;
2265 my $lr = $self->last_rev;
2266 if (defined $lr && $lr >= $log_entry->{revision}) {
2267 die "Last fetched revision of ", $self->refname,
2268 " was r$lr, but we are about to fetch: ",
2269 "r$log_entry->{revision}!\n";
2270 }
2271 if (my $c = $self->rev_map_get($log_entry->{revision})) {
2272 croak "$log_entry->{revision} = $c already exists! ",
2273 "Why are we refetching it?\n";
2274 }
2275 my $old_env = set_commit_header_env($log_entry);
2276 my $tree = $log_entry->{tree};
2277 if (!defined $tree) {
2278 $tree = $self->tmp_index_do(sub {
2279 command_oneline('write-tree') });
2280 }
2281 die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2282
2283 my @exec = ('git', 'commit-tree', $tree);
2284 foreach ($self->get_commit_parents($log_entry)) {
2285 push @exec, '-p', $_;
2286 }
2287 defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2288 or croak $!;
2289 binmode $msg_fh;
2290
2291 # we always get UTF-8 from SVN, but we may want our commits in
2292 # a different encoding.
2293 if (my $enc = Git::config('i18n.commitencoding')) {
2294 require Encode;
2295 Encode::from_to($log_entry->{log}, 'UTF-8', $enc);
2296 }
2297 print $msg_fh $log_entry->{log} or croak $!;
2298 restore_commit_header_env($old_env);
2299 unless ($self->no_metadata) {
2300 print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2301 or croak $!;
2302 }
2303 $msg_fh->flush == 0 or croak $!;
2304 close $msg_fh or croak $!;
2305 chomp(my $commit = do { local $/; <$out_fh> });
2306 close $out_fh or croak $!;
2307 waitpid $pid, 0;
2308 croak $? if $?;
2309 if ($commit !~ /^$::sha1$/o) {
2310 die "Failed to commit, invalid sha1: $commit\n";
2311 }
2312
2313 $self->rev_map_set($log_entry->{revision}, $commit, 1);
2314
2315 $self->{last_rev} = $log_entry->{revision};
2316 $self->{last_commit} = $commit;
2317 print "r$log_entry->{revision}";
2318 if (defined $log_entry->{svm_revision}) {
2319 print " (\@$log_entry->{svm_revision})";
2320 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2321 0, $self->svm_uuid);
2322 }
2323 print " = $commit ($self->{ref_id})\n";
2324 if (--$_gc_nr == 0) {
2325 $_gc_nr = $_gc_period;
2326 gc();
2327 }
2328 return $commit;
2329 }
2330
2331 sub match_paths {
2332 my ($self, $paths, $r) = @_;
2333 return 1 if $self->{path} eq '';
2334 if (my $path = $paths->{"/$self->{path}"}) {
2335 return ($path->{action} eq 'D') ? 0 : 1;
2336 }
2337 $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
2338 if (grep /$self->{path_regex}/, keys %$paths) {
2339 return 1;
2340 }
2341 my $c = '';
2342 foreach (split m#/#, $self->{path}) {
2343 $c .= "/$_";
2344 next unless ($paths->{$c} &&
2345 ($paths->{$c}->{action} =~ /^[AR]$/));
2346 if ($self->ra->check_path($self->{path}, $r) ==
2347 $SVN::Node::dir) {
2348 return 1;
2349 }
2350 }
2351 return 0;
2352 }
2353
2354 sub find_parent_branch {
2355 my ($self, $paths, $rev) = @_;
2356 return undef unless $self->follow_parent;
2357 unless (defined $paths) {
2358 my $err_handler = $SVN::Error::handler;
2359 $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2360 $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1, sub {
2361 $paths =
2362 Git::SVN::Ra::dup_changed_paths($_[0]) });
2363 $SVN::Error::handler = $err_handler;
2364 }
2365 return undef unless defined $paths;
2366
2367 # look for a parent from another branch:
2368 my @b_path_components = split m#/#, $self->rel_path;
2369 my @a_path_components;
2370 my $i;
2371 while (@b_path_components) {
2372 $i = $paths->{'/'.join('/', @b_path_components)};
2373 last if $i && defined $i->{copyfrom_path};
2374 unshift(@a_path_components, pop(@b_path_components));
2375 }
2376 return undef unless defined $i && defined $i->{copyfrom_path};
2377 my $branch_from = $i->{copyfrom_path};
2378 if (@a_path_components) {
2379 print STDERR "branch_from: $branch_from => ";
2380 $branch_from .= '/'.join('/', @a_path_components);
2381 print STDERR $branch_from, "\n";
2382 }
2383 my $r = $i->{copyfrom_rev};
2384 my $repos_root = $self->ra->{repos_root};
2385 my $url = $self->ra->{url};
2386 my $new_url = $repos_root . $branch_from;
2387 print STDERR "Found possible branch point: ",
2388 "$new_url => ", $self->full_url, ", $r\n";
2389 $branch_from =~ s#^/##;
2390 my $gs = Git::SVN->find_by_url($new_url, $repos_root, $branch_from);
2391 unless ($gs) {
2392 my $ref_id = $self->{ref_id};
2393 $ref_id =~ s/\@\d+$//;
2394 $ref_id .= "\@$r";
2395 # just grow a tail if we're not unique enough :x
2396 $ref_id .= '-' while find_ref($ref_id);
2397 print STDERR "Initializing parent: $ref_id\n";
2398 my ($u, $p, $repo_id) = ($new_url, '', $ref_id);
2399 if ($u =~ s#^\Q$url\E(/|$)##) {
2400 $p = $u;
2401 $u = $url;
2402 $repo_id = $self->{repo_id};
2403 }
2404 $gs = Git::SVN->init($u, $p, $repo_id, $ref_id, 1);
2405 }
2406 my ($r0, $parent) = $gs->find_rev_before($r, 1);
2407 {
2408 my ($base, $head);
2409 if (!defined $r0 || !defined $parent) {
2410 ($base, $head) = parse_revision_argument(0, $r);
2411 } else {
2412 if ($r0 < $r) {
2413 $gs->ra->get_log([$gs->{path}], $r0 + 1, $r, 1,
2414 0, 1, sub { $base = $_[1] - 1 });
2415 }
2416 }
2417 if (defined $base && $base <= $r) {
2418 $gs->fetch($base, $r);
2419 }
2420 ($r0, $parent) = $gs->find_rev_before($r, 1);
2421 }
2422 if (defined $r0 && defined $parent) {
2423 print STDERR "Found branch parent: ($self->{ref_id}) $parent\n";
2424 my $ed;
2425 if ($self->ra->can_do_switch) {
2426 $self->assert_index_clean($parent);
2427 print STDERR "Following parent with do_switch\n";
2428 # do_switch works with svn/trunk >= r22312, but that
2429 # is not included with SVN 1.4.3 (the latest version
2430 # at the moment), so we can't rely on it
2431 $self->{last_commit} = $parent;
2432 $ed = SVN::Git::Fetcher->new($self);
2433 $gs->ra->gs_do_switch($r0, $rev, $gs,
2434 $self->full_url, $ed)
2435 or die "SVN connection failed somewhere...\n";
2436 } elsif ($self->ra->trees_match($new_url, $r0,
2437 $self->full_url, $rev)) {
2438 print STDERR "Trees match:\n",
2439 " $new_url\@$r0\n",
2440 " ${\$self->full_url}\@$rev\n",
2441 "Following parent with no changes\n";
2442 $self->tmp_index_do(sub {
2443 command_noisy('read-tree', $parent);
2444 });
2445 $self->{last_commit} = $parent;
2446 } else {
2447 print STDERR "Following parent with do_update\n";
2448 $ed = SVN::Git::Fetcher->new($self);
2449 $self->ra->gs_do_update($rev, $rev, $self, $ed)
2450 or die "SVN connection failed somewhere...\n";
2451 }
2452 print STDERR "Successfully followed parent\n";
2453 return $self->make_log_entry($rev, [$parent], $ed);
2454 }
2455 return undef;
2456 }
2457
2458 sub do_fetch {
2459 my ($self, $paths, $rev) = @_;
2460 my $ed;
2461 my ($last_rev, @parents);
2462 if (my $lc = $self->last_commit) {
2463 # we can have a branch that was deleted, then re-added
2464 # under the same name but copied from another path, in
2465 # which case we'll have multiple parents (we don't
2466 # want to break the original ref, nor lose copypath info):
2467 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2468 push @{$log_entry->{parents}}, $lc;
2469 return $log_entry;
2470 }
2471 $ed = SVN::Git::Fetcher->new($self);
2472 $last_rev = $self->{last_rev};
2473 $ed->{c} = $lc;
2474 @parents = ($lc);
2475 } else {
2476 $last_rev = $rev;
2477 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2478 return $log_entry;
2479 }
2480 $ed = SVN::Git::Fetcher->new($self);
2481 }
2482 unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2483 die "SVN connection failed somewhere...\n";
2484 }
2485 $self->make_log_entry($rev, \@parents, $ed);
2486 }
2487
2488 sub get_untracked {
2489 my ($self, $ed) = @_;
2490 my @out;
2491 my $h = $ed->{empty};
2492 foreach (sort keys %$h) {
2493 my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2494 push @out, " $act: " . uri_encode($_);
2495 warn "W: $act: $_\n";
2496 }
2497 foreach my $t (qw/dir_prop file_prop/) {
2498 $h = $ed->{$t} or next;
2499 foreach my $path (sort keys %$h) {
2500 my $ppath = $path eq '' ? '.' : $path;
2501 foreach my $prop (sort keys %{$h->{$path}}) {
2502 next if $SKIP_PROP{$prop};
2503 my $v = $h->{$path}->{$prop};
2504 my $t_ppath_prop = "$t: " .
2505 uri_encode($ppath) . ' ' .
2506 uri_encode($prop);
2507 if (defined $v) {
2508 push @out, " +$t_ppath_prop " .
2509 uri_encode($v);
2510 } else {
2511 push @out, " -$t_ppath_prop";
2512 }
2513 }
2514 }
2515 }
2516 foreach my $t (qw/absent_file absent_directory/) {
2517 $h = $ed->{$t} or next;
2518 foreach my $parent (sort keys %$h) {
2519 foreach my $path (sort @{$h->{$parent}}) {
2520 push @out, " $t: " .
2521 uri_encode("$parent/$path");
2522 warn "W: $t: $parent/$path ",
2523 "Insufficient permissions?\n";
2524 }
2525 }
2526 }
2527 \@out;
2528 }
2529
2530 # parse_svn_date(DATE)
2531 # --------------------
2532 # Given a date (in UTC) from Subversion, return a string in the format
2533 # "<TZ Offset> <local date/time>" that Git will use.
2534 #
2535 # By default the parsed date will be in UTC; if $Git::SVN::_localtime
2536 # is true we'll convert it to the local timezone instead.
2537 sub parse_svn_date {
2538 my $date = shift || return '+0000 1970-01-01 00:00:00';
2539 my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2540 (\d\d)\:(\d\d)\:(\d\d).\d+Z$/x) or
2541 croak "Unable to parse date: $date\n";
2542 my $parsed_date; # Set next.
2543
2544 if ($Git::SVN::_localtime) {
2545 # Translate the Subversion datetime to an epoch time.
2546 # Begin by switching ourselves to $date's timezone, UTC.
2547 my $old_env_TZ = $ENV{TZ};
2548 $ENV{TZ} = 'UTC';
2549
2550 my $epoch_in_UTC =
2551 POSIX::strftime('%s', $S, $M, $H, $d, $m - 1, $Y - 1900);
2552
2553 # Determine our local timezone (including DST) at the
2554 # time of $epoch_in_UTC. $Git::SVN::Log::TZ stored the
2555 # value of TZ, if any, at the time we were run.
2556 if (defined $Git::SVN::Log::TZ) {
2557 $ENV{TZ} = $Git::SVN::Log::TZ;
2558 } else {
2559 delete $ENV{TZ};
2560 }
2561
2562 my $our_TZ =
2563 POSIX::strftime('%Z', $S, $M, $H, $d, $m - 1, $Y - 1900);
2564
2565 # This converts $epoch_in_UTC into our local timezone.
2566 my ($sec, $min, $hour, $mday, $mon, $year,
2567 $wday, $yday, $isdst) = localtime($epoch_in_UTC);
2568
2569 $parsed_date = sprintf('%s %04d-%02d-%02d %02d:%02d:%02d',
2570 $our_TZ, $year + 1900, $mon + 1,
2571 $mday, $hour, $min, $sec);
2572
2573 # Reset us to the timezone in effect when we entered
2574 # this routine.
2575 if (defined $old_env_TZ) {
2576 $ENV{TZ} = $old_env_TZ;
2577 } else {
2578 delete $ENV{TZ};
2579 }
2580 } else {
2581 $parsed_date = "+0000 $Y-$m-$d $H:$M:$S";
2582 }
2583
2584 return $parsed_date;
2585 }
2586
2587 sub check_author {
2588 my ($author) = @_;
2589 if (!defined $author || length $author == 0) {
2590 $author = '(no author)';
2591 } elsif (defined $::_authors && ! defined $::users{$author}) {
2592 die "Author: $author not defined in $::_authors file\n";
2593 }
2594 $author;
2595 }
2596
2597 sub make_log_entry {
2598 my ($self, $rev, $parents, $ed) = @_;
2599 my $untracked = $self->get_untracked($ed);
2600
2601 open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
2602 print $un "r$rev\n" or croak $!;
2603 print $un $_, "\n" foreach @$untracked;
2604 my %log_entry = ( parents => $parents || [], revision => $rev,
2605 log => '');
2606
2607 my $headrev;
2608 my $logged = delete $self->{logged_rev_props};
2609 if (!$logged || $self->{-want_revprops}) {
2610 my $rp = $self->ra->rev_proplist($rev);
2611 foreach (sort keys %$rp) {
2612 my $v = $rp->{$_};
2613 if (/^svn:(author|date|log)$/) {
2614 $log_entry{$1} = $v;
2615 } elsif ($_ eq 'svm:headrev') {
2616 $headrev = $v;
2617 } else {
2618 print $un " rev_prop: ", uri_encode($_), ' ',
2619 uri_encode($v), "\n";
2620 }
2621 }
2622 } else {
2623 map { $log_entry{$_} = $logged->{$_} } keys %$logged;
2624 }
2625 close $un or croak $!;
2626
2627 $log_entry{date} = parse_svn_date($log_entry{date});
2628 $log_entry{log} .= "\n";
2629 my $author = $log_entry{author} = check_author($log_entry{author});
2630 my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
2631 : ($author, undef);
2632
2633 my ($commit_name, $commit_email) = ($name, $email);
2634 if ($_use_log_author) {
2635 my $name_field;
2636 if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
2637 $name_field = $1;
2638 } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
2639 $name_field = $1;
2640 }
2641 if (!defined $name_field) {
2642 if (!defined $email) {
2643 $email = $name;
2644 }
2645 } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
2646 ($name, $email) = ($1, $2);
2647 } elsif ($name_field =~ /(.*)@/) {
2648 ($name, $email) = ($1, $name_field);
2649 } else {
2650 ($name, $email) = ($name_field, $name_field);
2651 }
2652 }
2653 if (defined $headrev && $self->use_svm_props) {
2654 if ($self->rewrite_root) {
2655 die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
2656 "options set!\n";
2657 }
2658 my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$};
2659 # we don't want "SVM: initializing mirror for junk" ...
2660 return undef if $r == 0;
2661 my $svm = $self->svm;
2662 if ($uuid ne $svm->{uuid}) {
2663 die "UUID mismatch on SVM path:\n",
2664 "expected: $svm->{uuid}\n",
2665 " got: $uuid\n";
2666 }
2667 my $full_url = $self->full_url;
2668 $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
2669 die "Failed to replace '$svm->{replace}' with ",
2670 "'$svm->{source}' in $full_url\n";
2671 # throw away username for storing in records
2672 remove_username($full_url);
2673 $log_entry{metadata} = "$full_url\@$r $uuid";
2674 $log_entry{svm_revision} = $r;
2675 $email ||= "$author\@$uuid";
2676 $commit_email ||= "$author\@$uuid";
2677 } elsif ($self->use_svnsync_props) {
2678 my $full_url = $self->svnsync->{url};
2679 $full_url .= "/$self->{path}" if length $self->{path};
2680 remove_username($full_url);
2681 my $uuid = $self->svnsync->{uuid};
2682 $log_entry{metadata} = "$full_url\@$rev $uuid";
2683 $email ||= "$author\@$uuid";
2684 $commit_email ||= "$author\@$uuid";
2685 } else {
2686 my $url = $self->metadata_url;
2687 remove_username($url);
2688 $log_entry{metadata} = "$url\@$rev " .
2689 $self->ra->get_uuid;
2690 $email ||= "$author\@" . $self->ra->get_uuid;
2691 $commit_email ||= "$author\@" . $self->ra->get_uuid;
2692 }
2693 $log_entry{name} = $name;
2694 $log_entry{email} = $email;
2695 $log_entry{commit_name} = $commit_name;
2696 $log_entry{commit_email} = $commit_email;
2697 \%log_entry;
2698 }
2699
2700 sub fetch {
2701 my ($self, $min_rev, $max_rev, @parents) = @_;
2702 my ($last_rev, $last_commit) = $self->last_rev_commit;
2703 my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
2704 $self->ra->gs_fetch_loop_common($base, $head, [$self]);
2705 }
2706
2707 sub set_tree_cb {
2708 my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
2709 $self->{inject_parents} = { $rev => $tree };
2710 $self->fetch(undef, undef);
2711 }
2712
2713 sub set_tree {
2714 my ($self, $tree) = (shift, shift);
2715 my $log_entry = ::get_commit_entry($tree);
2716 unless ($self->{last_rev}) {
2717 ::fatal("Must have an existing revision to commit");
2718 }
2719 my %ed_opts = ( r => $self->{last_rev},
2720 log => $log_entry->{log},
2721 ra => $self->ra,
2722 tree_a => $self->{last_commit},
2723 tree_b => $tree,
2724 editor_cb => sub {
2725 $self->set_tree_cb($log_entry, $tree, @_) },
2726 svn_path => $self->{path} );
2727 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
2728 print "No changes\nr$self->{last_rev} = $tree\n";
2729 }
2730 }
2731
2732 sub rebuild_from_rev_db {
2733 my ($self, $path) = @_;
2734 my $r = -1;
2735 open my $fh, '<', $path or croak "open: $!";
2736 binmode $fh or croak "binmode: $!";
2737 while (<$fh>) {
2738 length($_) == 41 or croak "inconsistent size in ($_) != 41";
2739 chomp($_);
2740 ++$r;
2741 next if $_ eq ('0' x 40);
2742 $self->rev_map_set($r, $_);
2743 print "r$r = $_\n";
2744 }
2745 close $fh or croak "close: $!";
2746 unlink $path or croak "unlink: $!";
2747 }
2748
2749 sub rebuild {
2750 my ($self) = @_;
2751 my $map_path = $self->map_path;
2752 my $partial = (-e $map_path && ! -z $map_path);
2753 return unless ::verify_ref($self->refname.'^0');
2754 if (!$partial && ($self->use_svm_props || $self->no_metadata)) {
2755 my $rev_db = $self->rev_db_path;
2756 $self->rebuild_from_rev_db($rev_db);
2757 if ($self->use_svm_props) {
2758 my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
2759 $self->rebuild_from_rev_db($svm_rev_db);
2760 }
2761 $self->unlink_rev_db_symlink;
2762 return;
2763 }
2764 print "Rebuilding $map_path ...\n" if (!$partial);
2765 my ($base_rev, $head) = ($partial ? $self->rev_map_max_norebuild(1) :
2766 (undef, undef));
2767 my ($log, $ctx) =
2768 command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
2769 ($head ? "$head.." : "") . $self->refname,
2770 '--');
2771 my $metadata_url = $self->metadata_url;
2772 remove_username($metadata_url);
2773 my $svn_uuid = $self->ra_uuid;
2774 my $c;
2775 while (<$log>) {
2776 if ( m{^commit ($::sha1)$} ) {
2777 $c = $1;
2778 next;
2779 }
2780 next unless s{^\s*(git-svn-id:)}{$1};
2781 my ($url, $rev, $uuid) = ::extract_metadata($_);
2782 remove_username($url);
2783
2784 # ignore merges (from set-tree)
2785 next if (!defined $rev || !$uuid);
2786
2787 # if we merged or otherwise started elsewhere, this is
2788 # how we break out of it
2789 if (($uuid ne $svn_uuid) ||
2790 ($metadata_url && $url && ($url ne $metadata_url))) {
2791 next;
2792 }
2793 if ($partial && $head) {
2794 print "Partial-rebuilding $map_path ...\n";
2795 print "Currently at $base_rev = $head\n";
2796 $head = undef;
2797 }
2798
2799 $self->rev_map_set($rev, $c);
2800 print "r$rev = $c\n";
2801 }
2802 command_close_pipe($log, $ctx);
2803 print "Done rebuilding $map_path\n" if (!$partial || !$head);
2804 my $rev_db_path = $self->rev_db_path;
2805 if (-f $self->rev_db_path) {
2806 unlink $self->rev_db_path or croak "unlink: $!";
2807 }
2808 $self->unlink_rev_db_symlink;
2809 }
2810
2811 # rev_map:
2812 # Tie::File seems to be prone to offset errors if revisions get sparse,
2813 # it's not that fast, either. Tie::File is also not in Perl 5.6. So
2814 # one of my favorite modules is out :< Next up would be one of the DBM
2815 # modules, but I'm not sure which is most portable...
2816 #
2817 # This is the replacement for the rev_db format, which was too big
2818 # and inefficient for large repositories with a lot of sparse history
2819 # (mainly tags)
2820 #
2821 # The format is this:
2822 # - 24 bytes for every record,
2823 # * 4 bytes for the integer representing an SVN revision number
2824 # * 20 bytes representing the sha1 of a git commit
2825 # - No empty padding records like the old format
2826 # (except the last record, which can be overwritten)
2827 # - new records are written append-only since SVN revision numbers
2828 # increase monotonically
2829 # - lookups on SVN revision number are done via a binary search
2830 # - Piping the file to xxd -c24 is a good way of dumping it for
2831 # viewing or editing (piped back through xxd -r), should the need
2832 # ever arise.
2833 # - The last record can be padding revision with an all-zero sha1
2834 # This is used to optimize fetch performance when using multiple
2835 # "fetch" directives in .git/config
2836 #
2837 # These files are disposable unless noMetadata or useSvmProps is set
2838
2839 sub _rev_map_set {
2840 my ($fh, $rev, $commit) = @_;
2841
2842 binmode $fh or croak "binmode: $!";
2843 my $size = (stat($fh))[7];
2844 ($size % 24) == 0 or croak "inconsistent size: $size";
2845
2846 my $wr_offset = 0;
2847 if ($size > 0) {
2848 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2849 my $read = sysread($fh, my $buf, 24) or croak "read: $!";
2850 $read == 24 or croak "read only $read bytes (!= 24)";
2851 my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
2852 if ($last_commit eq ('0' x40)) {
2853 if ($size >= 48) {
2854 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2855 $read = sysread($fh, $buf, 24) or
2856 croak "read: $!";
2857 $read == 24 or
2858 croak "read only $read bytes (!= 24)";
2859 ($last_rev, $last_commit) =
2860 unpack(rev_map_fmt, $buf);
2861 if ($last_commit eq ('0' x40)) {
2862 croak "inconsistent .rev_map\n";
2863 }
2864 }
2865 if ($last_rev >= $rev) {
2866 croak "last_rev is higher!: $last_rev >= $rev";
2867 }
2868 $wr_offset = -24;
2869 }
2870 }
2871 sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
2872 syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
2873 croak "write: $!";
2874 }
2875
2876 sub mkfile {
2877 my ($path) = @_;
2878 unless (-e $path) {
2879 my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
2880 mkpath([$dir]) unless -d $dir;
2881 open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
2882 close $fh or die "Couldn't close (create) $path: $!\n";
2883 }
2884 }
2885
2886 sub rev_map_set {
2887 my ($self, $rev, $commit, $update_ref, $uuid) = @_;
2888 length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
2889 my $db = $self->map_path($uuid);
2890 my $db_lock = "$db.lock";
2891 my $sig;
2892 if ($update_ref) {
2893 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2894 $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
2895 }
2896 mkfile($db);
2897
2898 $LOCKFILES{$db_lock} = 1;
2899 my $sync;
2900 # both of these options make our .rev_db file very, very important
2901 # and we can't afford to lose it because rebuild() won't work
2902 if ($self->use_svm_props || $self->no_metadata) {
2903 $sync = 1;
2904 copy($db, $db_lock) or die "rev_map_set(@_): ",
2905 "Failed to copy: ",
2906 "$db => $db_lock ($!)\n";
2907 } else {
2908 rename $db, $db_lock or die "rev_map_set(@_): ",
2909 "Failed to rename: ",
2910 "$db => $db_lock ($!)\n";
2911 }
2912
2913 sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
2914 or croak "Couldn't open $db_lock: $!\n";
2915 _rev_map_set($fh, $rev, $commit);
2916 if ($sync) {
2917 $fh->flush or die "Couldn't flush $db_lock: $!\n";
2918 $fh->sync or die "Couldn't sync $db_lock: $!\n";
2919 }
2920 close $fh or croak $!;
2921 if ($update_ref) {
2922 $_head = $self;
2923 command_noisy('update-ref', '-m', "r$rev",
2924 $self->refname, $commit);
2925 }
2926 rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
2927 "$db_lock => $db ($!)\n";
2928 delete $LOCKFILES{$db_lock};
2929 if ($update_ref) {
2930 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2931 $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
2932 kill $sig, $$ if defined $sig;
2933 }
2934 }
2935
2936 # If want_commit, this will return an array of (rev, commit) where
2937 # commit _must_ be a valid commit in the archive.
2938 # Otherwise, it'll return the max revision (whether or not the
2939 # commit is valid or just a 0x40 placeholder).
2940 sub rev_map_max {
2941 my ($self, $want_commit) = @_;
2942 $self->rebuild;
2943 my ($r, $c) = $self->rev_map_max_norebuild($want_commit);
2944 $want_commit ? ($r, $c) : $r;
2945 }
2946
2947 sub rev_map_max_norebuild {
2948 my ($self, $want_commit) = @_;
2949 my $map_path = $self->map_path;
2950 stat $map_path or return $want_commit ? (0, undef) : 0;
2951 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2952 binmode $fh or croak "binmode: $!";
2953 my $size = (stat($fh))[7];
2954 ($size % 24) == 0 or croak "inconsistent size: $size";
2955
2956 if ($size == 0) {
2957 close $fh or croak "close: $!";
2958 return $want_commit ? (0, undef) : 0;
2959 }
2960
2961 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2962 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2963 my ($r, $c) = unpack(rev_map_fmt, $buf);
2964 if ($want_commit && $c eq ('0' x40)) {
2965 if ($size < 48) {
2966 return $want_commit ? (0, undef) : 0;
2967 }
2968 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2969 sysread($fh, $buf, 24) == 24 or croak "read: $!";
2970 ($r, $c) = unpack(rev_map_fmt, $buf);
2971 if ($c eq ('0'x40)) {
2972 croak "Penultimate record is all-zeroes in $map_path";
2973 }
2974 }
2975 close $fh or croak "close: $!";
2976 $want_commit ? ($r, $c) : $r;
2977 }
2978
2979 sub rev_map_get {
2980 my ($self, $rev, $uuid) = @_;
2981 my $map_path = $self->map_path($uuid);
2982 return undef unless -e $map_path;
2983
2984 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2985 binmode $fh or croak "binmode: $!";
2986 my $size = (stat($fh))[7];
2987 ($size % 24) == 0 or croak "inconsistent size: $size";
2988
2989 if ($size == 0) {
2990 close $fh or croak "close: $fh";
2991 return undef;
2992 }
2993
2994 my ($l, $u) = (0, $size - 24);
2995 my ($r, $c, $buf);
2996
2997 while ($l <= $u) {
2998 my $i = int(($l/24 + $u/24) / 2) * 24;
2999 sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
3000 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
3001 my ($r, $c) = unpack('NH40', $buf);
3002
3003 if ($r < $rev) {
3004 $l = $i + 24;
3005 } elsif ($r > $rev) {
3006 $u = $i - 24;
3007 } else { # $r == $rev
3008 close($fh) or croak "close: $!";
3009 return $c eq ('0' x 40) ? undef : $c;
3010 }
3011 }
3012 close($fh) or croak "close: $!";
3013 undef;
3014 }
3015
3016 # Finds the first svn revision that exists on (if $eq_ok is true) or
3017 # before $rev for the current branch. It will not search any lower
3018 # than $min_rev. Returns the git commit hash and svn revision number
3019 # if found, else (undef, undef).
3020 sub find_rev_before {
3021 my ($self, $rev, $eq_ok, $min_rev) = @_;
3022 --$rev unless $eq_ok;
3023 $min_rev ||= 1;
3024 while ($rev >= $min_rev) {
3025 if (my $c = $self->rev_map_get($rev)) {
3026 return ($rev, $c);
3027 }
3028 --$rev;
3029 }
3030 return (undef, undef);
3031 }
3032
3033 # Finds the first svn revision that exists on (if $eq_ok is true) or
3034 # after $rev for the current branch. It will not search any higher
3035 # than $max_rev. Returns the git commit hash and svn revision number
3036 # if found, else (undef, undef).
3037 sub find_rev_after {
3038 my ($self, $rev, $eq_ok, $max_rev) = @_;
3039 ++$rev unless $eq_ok;
3040 $max_rev ||= $self->rev_map_max;
3041 while ($rev <= $max_rev) {
3042 if (my $c = $self->rev_map_get($rev)) {
3043 return ($rev, $c);
3044 }
3045 ++$rev;
3046 }
3047 return (undef, undef);
3048 }
3049
3050 sub _new {
3051 my ($class, $repo_id, $ref_id, $path) = @_;
3052 unless (defined $repo_id && length $repo_id) {
3053 $repo_id = $Git::SVN::default_repo_id;
3054 }
3055 unless (defined $ref_id && length $ref_id) {
3056 $_[2] = $ref_id = $Git::SVN::default_ref_id;
3057 }
3058 $_[1] = $repo_id;
3059 my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
3060 $_[3] = $path = '' unless (defined $path);
3061 mkpath(["$ENV{GIT_DIR}/svn"]);
3062 bless {
3063 ref_id => $ref_id, dir => $dir, index => "$dir/index",
3064 path => $path, config => "$ENV{GIT_DIR}/svn/config",
3065 map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
3066 }
3067
3068 # for read-only access of old .rev_db formats
3069 sub unlink_rev_db_symlink {
3070 my ($self) = @_;
3071 my $link = $self->rev_db_path;
3072 $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
3073 if (-l $link) {
3074 unlink $link or croak "unlink: $link failed!";
3075 }
3076 }
3077
3078 sub rev_db_path {
3079 my ($self, $uuid) = @_;
3080 my $db_path = $self->map_path($uuid);
3081 $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
3082 or croak "map_path: $db_path does not contain '/.rev_map.' !";
3083 $db_path;
3084 }
3085
3086 # the new replacement for .rev_db
3087 sub map_path {
3088 my ($self, $uuid) = @_;
3089 $uuid ||= $self->ra_uuid;
3090 "$self->{map_root}.$uuid";
3091 }
3092
3093 sub uri_encode {
3094 my ($f) = @_;
3095 $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
3096 $f
3097 }
3098
3099 sub remove_username {
3100 $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
3101 }
3102
3103 package Git::SVN::Prompt;
3104 use strict;
3105 use warnings;
3106 require SVN::Core;
3107 use vars qw/$_no_auth_cache $_username/;
3108
3109 sub simple {
3110 my ($cred, $realm, $default_username, $may_save, $pool) = @_;
3111 $may_save = undef if $_no_auth_cache;
3112 $default_username = $_username if defined $_username;
3113 if (defined $default_username && length $default_username) {
3114 if (defined $realm && length $realm) {
3115 print STDERR "Authentication realm: $realm\n";
3116 STDERR->flush;
3117 }
3118 $cred->username($default_username);
3119 } else {
3120 username($cred, $realm, $may_save, $pool);
3121 }
3122 $cred->password(_read_password("Password for '" .
3123 $cred->username . "': ", $realm));
3124 $cred->may_save($may_save);
3125 $SVN::_Core::SVN_NO_ERROR;
3126 }
3127
3128 sub ssl_server_trust {
3129 my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
3130 $may_save = undef if $_no_auth_cache;
3131 print STDERR "Error validating server certificate for '$realm':\n";
3132 {
3133 no warnings 'once';
3134 # All variables SVN::Auth::SSL::* are used only once,
3135 # so we're shutting up Perl warnings about this.
3136 if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
3137 print STDERR " - The certificate is not issued ",
3138 "by a trusted authority. Use the\n",
3139 " fingerprint to validate ",
3140 "the certificate manually!\n";
3141 }
3142 if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
3143 print STDERR " - The certificate hostname ",
3144 "does not match.\n";
3145 }
3146 if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
3147 print STDERR " - The certificate is not yet valid.\n";
3148 }
3149 if ($failures & $SVN::Auth::SSL::EXPIRED) {
3150 print STDERR " - The certificate has expired.\n";
3151 }
3152 if ($failures & $SVN::Auth::SSL::OTHER) {
3153 print STDERR " - The certificate has ",
3154 "an unknown error.\n";
3155 }
3156 } # no warnings 'once'
3157 printf STDERR
3158 "Certificate information:\n".
3159 " - Hostname: %s\n".
3160 " - Valid: from %s until %s\n".
3161 " - Issuer: %s\n".
3162 " - Fingerprint: %s\n",
3163 map $cert_info->$_, qw(hostname valid_from valid_until
3164 issuer_dname fingerprint);
3165 my $choice;
3166 prompt:
3167 print STDERR $may_save ?
3168 "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
3169 "(R)eject or accept (t)emporarily? ";
3170 STDERR->flush;
3171 $choice = lc(substr(<STDIN> || 'R', 0, 1));
3172 if ($choice =~ /^t$/i) {
3173 $cred->may_save(undef);
3174 } elsif ($choice =~ /^r$/i) {
3175 return -1;
3176 } elsif ($may_save && $choice =~ /^p$/i) {
3177 $cred->may_save($may_save);
3178 } else {
3179 goto prompt;
3180 }
3181 $cred->accepted_failures($failures);
3182 $SVN::_Core::SVN_NO_ERROR;
3183 }
3184
3185 sub ssl_client_cert {
3186 my ($cred, $realm, $may_save, $pool) = @_;
3187 $may_save = undef if $_no_auth_cache;
3188 print STDERR "Client certificate filename: ";
3189 STDERR->flush;
3190 chomp(my $filename = <STDIN>);
3191 $cred->cert_file($filename);
3192 $cred->may_save($may_save);
3193 $SVN::_Core::SVN_NO_ERROR;
3194 }
3195
3196 sub ssl_client_cert_pw {
3197 my ($cred, $realm, $may_save, $pool) = @_;
3198 $may_save = undef if $_no_auth_cache;
3199 $cred->password(_read_password("Password: ", $realm));
3200 $cred->may_save($may_save);
3201 $SVN::_Core::SVN_NO_ERROR;
3202 }
3203
3204 sub username {
3205 my ($cred, $realm, $may_save, $pool) = @_;
3206 $may_save = undef if $_no_auth_cache;
3207 if (defined $realm && length $realm) {
3208 print STDERR "Authentication realm: $realm\n";
3209 }
3210 my $username;
3211 if (defined $_username) {
3212 $username = $_username;
3213 } else {
3214 print STDERR "Username: ";
3215 STDERR->flush;
3216 chomp($username = <STDIN>);
3217 }
3218 $cred->username($username);
3219 $cred->may_save($may_save);
3220 $SVN::_Core::SVN_NO_ERROR;
3221 }
3222
3223 sub _read_password {
3224 my ($prompt, $realm) = @_;
3225 print STDERR $prompt;
3226 STDERR->flush;
3227 require Term::ReadKey;
3228 Term::ReadKey::ReadMode('noecho');
3229 my $password = '';
3230 while (defined(my $key = Term::ReadKey::ReadKey(0))) {
3231 last if $key =~ /[\012\015]/; # \n\r
3232 $password .= $key;
3233 }
3234 Term::ReadKey::ReadMode('restore');
3235 print STDERR "\n";
3236 STDERR->flush;
3237 $password;
3238 }
3239
3240 package SVN::Git::Fetcher;
3241 use vars qw/@ISA/;
3242 use strict;
3243 use warnings;
3244 use Carp qw/croak/;
3245 use File::Temp qw/tempfile/;
3246 use IO::File qw//;
3247
3248 # file baton members: path, mode_a, mode_b, pool, fh, blob, base
3249 sub new {
3250 my ($class, $git_svn) = @_;
3251 my $self = SVN::Delta::Editor->new;
3252 bless $self, $class;
3253 if (exists $git_svn->{last_commit}) {
3254 $self->{c} = $git_svn->{last_commit};
3255 $self->{empty_symlinks} = _mark_empty_symlinks($git_svn);
3256 }
3257 $self->{empty} = {};
3258 $self->{dir_prop} = {};
3259 $self->{file_prop} = {};
3260 $self->{absent_dir} = {};
3261 $self->{absent_file} = {};
3262 $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
3263 $self;
3264 }
3265
3266 # this uses the Ra object, so it must be called before do_{switch,update},
3267 # not inside them (when the Git::SVN::Fetcher object is passed) to
3268 # do_{switch,update}
3269 sub _mark_empty_symlinks {
3270 my ($git_svn) = @_;
3271 my %ret;
3272 my ($rev, $cmt) = $git_svn->last_rev_commit;
3273 return {} unless ($rev && $cmt);
3274
3275 chomp(my $empty_blob = `git hash-object -t blob --stdin < /dev/null`);
3276 my ($ls, $ctx) = command_output_pipe(qw/ls-tree -r -z/, $cmt);
3277 local $/ = "\0";
3278 my $pfx = $git_svn->{path};
3279 $pfx .= '/' if length($pfx);
3280 while (<$ls>) {
3281 chomp;
3282 s/\A100644 blob $empty_blob\t//o or next;
3283 my $path = $_;
3284 my (undef, $props) =
3285 $git_svn->ra->get_file($pfx.$path, $rev, undef);
3286 if ($props->{'svn:special'}) {
3287 $ret{$path} = 1;
3288 }
3289 }
3290 command_close_pipe($ls, $ctx);
3291 \%ret;
3292 }
3293
3294 sub set_path_strip {
3295 my ($self, $path) = @_;
3296 $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
3297 }
3298
3299 sub open_root {
3300 { path => '' };
3301 }
3302
3303 sub open_directory {
3304 my ($self, $path, $pb, $rev) = @_;
3305 { path => $path };
3306 }
3307
3308 sub git_path {
3309 my ($self, $path) = @_;
3310 if ($self->{path_strip}) {
3311 $path =~ s!$self->{path_strip}!! or
3312 die "Failed to strip path '$path' ($self->{path_strip})\n";
3313 }
3314 $path;
3315 }
3316
3317 sub delete_entry {
3318 my ($self, $path, $rev, $pb) = @_;
3319
3320 my $gpath = $self->git_path($path);
3321 return undef if ($gpath eq '');
3322
3323 # remove entire directories.
3324 if (command('ls-tree', $self->{c}, '--', $gpath) =~ /^040000 tree/) {
3325 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3326 -r --name-only -z/,
3327 $self->{c}, '--', $gpath);
3328 local $/ = "\0";
3329 while (<$ls>) {
3330 chomp;
3331 $self->{gii}->remove($_);
3332 print "\tD\t$_\n" unless $::_q;
3333 }
3334 print "\tD\t$gpath/\n" unless $::_q;
3335 command_close_pipe($ls, $ctx);
3336 $self->{empty}->{$path} = 0
3337 } else {
3338 $self->{gii}->remove($gpath);
3339 print "\tD\t$gpath\n" unless $::_q;
3340 }
3341 undef;
3342 }
3343
3344 sub open_file {
3345 my ($self, $path, $pb, $rev) = @_;
3346 my $gpath = $self->git_path($path);
3347 my ($mode, $blob) = (command('ls-tree', $self->{c}, '--', $gpath)
3348 =~ /^(\d{6}) blob ([a-f\d]{40})\t/);
3349 unless (defined $mode && defined $blob) {
3350 die "$path was not found in commit $self->{c} (r$rev)\n";
3351 }
3352 if ($mode eq '100644' && $self->{empty_symlinks}->{$path}) {
3353 $mode = '120000';
3354 }
3355 { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3356 pool => SVN::Pool->new, action => 'M' };
3357 }
3358
3359 sub add_file {
3360 my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3361 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3362 delete $self->{empty}->{$dir};
3363 { path => $path, mode_a => 100644, mode_b => 100644,
3364 pool => SVN::Pool->new, action => 'A' };
3365 }
3366
3367 sub add_directory {
3368 my ($self, $path, $cp_path, $cp_rev) = @_;
3369 my $gpath = $self->git_path($path);
3370 if ($gpath eq '') {
3371 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3372 -r --name-only -z/,
3373 $self->{c});
3374 local $/ = "\0";
3375 while (<$ls>) {
3376 chomp;
3377 $self->{gii}->remove($_);
3378 print "\tD\t$_\n" unless $::_q;
3379 }
3380 command_close_pipe($ls, $ctx);
3381 $self->{empty}->{$path} = 0;
3382 }
3383 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3384 delete $self->{empty}->{$dir};
3385 $self->{empty}->{$path} = 1;
3386 { path => $path };
3387 }
3388
3389 sub change_dir_prop {
3390 my ($self, $db, $prop, $value) = @_;
3391 $self->{dir_prop}->{$db->{path}} ||= {};
3392 $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3393 undef;
3394 }
3395
3396 sub absent_directory {
3397 my ($self, $path, $pb) = @_;
3398 $self->{absent_dir}->{$pb->{path}} ||= [];
3399 push @{$self->{absent_dir}->{$pb->{path}}}, $path;
3400 undef;
3401 }
3402
3403 sub absent_file {
3404 my ($self, $path, $pb) = @_;
3405 $self->{absent_file}->{$pb->{path}} ||= [];
3406 push @{$self->{absent_file}->{$pb->{path}}}, $path;
3407 undef;
3408 }
3409
3410 sub change_file_prop {
3411 my ($self, $fb, $prop, $value) = @_;
3412 if ($prop eq 'svn:executable') {
3413 if ($fb->{mode_b} != 120000) {
3414 $fb->{mode_b} = defined $value ? 100755 : 100644;
3415 }
3416 } elsif ($prop eq 'svn:special') {
3417 $fb->{mode_b} = defined $value ? 120000 : 100644;
3418 } else {
3419 $self->{file_prop}->{$fb->{path}} ||= {};
3420 $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
3421 }
3422 undef;
3423 }
3424
3425 sub apply_textdelta {
3426 my ($self, $fb, $exp) = @_;
3427 my $fh = $::_repository->temp_acquire('svn_delta');
3428 # $fh gets auto-closed() by SVN::TxDelta::apply(),
3429 # (but $base does not,) so dup() it for reading in close_file
3430 open my $dup, '<&', $fh or croak $!;
3431 my $base = $::_repository->temp_acquire('git_blob');
3432 if ($fb->{blob}) {
3433 my ($base_is_link, $size);
3434
3435 if ($fb->{mode_a} eq '120000' &&
3436 ! $self->{empty_symlinks}->{$fb->{path}}) {
3437 print $base 'link ' or die "print $!\n";
3438 $base_is_link = 1;
3439 }
3440 retry:
3441 $size = $::_repository->cat_blob($fb->{blob}, $base);
3442 die "Failed to read object $fb->{blob}" if ($size < 0);
3443
3444 if (defined $exp) {
3445 seek $base, 0, 0 or croak $!;
3446 my $got = ::md5sum($base);
3447 if ($got ne $exp) {
3448 my $err = "Checksum mismatch: ".
3449 "$fb->{path} $fb->{blob}\n" .
3450 "expected: $exp\n" .
3451 " got: $got\n";
3452 if ($base_is_link) {
3453 warn $err,
3454 "Retrying... (possibly ",
3455 "a bad symlink from SVN)\n";
3456 $::_repository->temp_reset($base);
3457 $base_is_link = 0;
3458 goto retry;
3459 }
3460 die $err;
3461 }
3462 }
3463 }
3464 seek $base, 0, 0 or croak $!;
3465 $fb->{fh} = $fh;
3466 $fb->{base} = $base;
3467 [ SVN::TxDelta::apply($base, $dup, undef, $fb->{path}, $fb->{pool}) ];
3468 }
3469
3470 sub close_file {
3471 my ($self, $fb, $exp) = @_;
3472 my $hash;
3473 my $path = $self->git_path($fb->{path});
3474 if (my $fh = $fb->{fh}) {
3475 if (defined $exp) {
3476 seek($fh, 0, 0) or croak $!;
3477 my $got = ::md5sum($fh);
3478 if ($got ne $exp) {
3479 die "Checksum mismatch: $path\n",
3480 "expected: $exp\n got: $got\n";
3481 }
3482 }
3483 if ($fb->{mode_b} == 120000) {
3484 sysseek($fh, 0, 0) or croak $!;
3485 my $rd = sysread($fh, my $buf, 5);
3486
3487 if (!defined $rd) {
3488 croak "sysread: $!\n";
3489 } elsif ($rd == 0) {
3490 warn "$path has mode 120000",
3491 " but it points to nothing\n",
3492 "converting to an empty file with mode",
3493 " 100644\n";
3494 $fb->{mode_b} = '100644';
3495 } elsif ($buf ne 'link ') {
3496 warn "$path has mode 120000",
3497 " but is not a link\n";
3498 } else {
3499 my $tmp_fh = $::_repository->temp_acquire(
3500 'svn_hash');
3501 my $res;
3502 while ($res = sysread($fh, my $str, 1024)) {
3503 my $out = syswrite($tmp_fh, $str, $res);
3504 defined($out) && $out == $res
3505 or croak("write ",
3506 Git::temp_path($tmp_fh),
3507 ": $!\n");
3508 }
3509 defined $res or croak $!;
3510
3511 ($fh, $tmp_fh) = ($tmp_fh, $fh);
3512 Git::temp_release($tmp_fh, 1);
3513 }
3514 }
3515
3516 $hash = $::_repository->hash_and_insert_object(
3517 Git::temp_path($fh));
3518 $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
3519
3520 Git::temp_release($fb->{base}, 1);
3521 Git::temp_release($fh, 1);
3522 } else {
3523 $hash = $fb->{blob} or die "no blob information\n";
3524 }
3525 $fb->{pool}->clear;
3526 $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
3527 print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
3528 undef;
3529 }
3530
3531 sub abort_edit {
3532 my $self = shift;
3533 $self->{nr} = $self->{gii}->{nr};
3534 delete $self->{gii};
3535 $self->SUPER::abort_edit(@_);
3536 }
3537
3538 sub close_edit {
3539 my $self = shift;
3540 $self->{git_commit_ok} = 1;
3541 $self->{nr} = $self->{gii}->{nr};
3542 delete $self->{gii};
3543 $self->SUPER::close_edit(@_);
3544 }
3545
3546 package SVN::Git::Editor;
3547 use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
3548 use strict;
3549 use warnings;
3550 use Carp qw/croak/;
3551 use IO::File;
3552
3553 sub new {
3554 my ($class, $opts) = @_;
3555 foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
3556 die "$_ required!\n" unless (defined $opts->{$_});
3557 }
3558
3559 my $pool = SVN::Pool->new;
3560 my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
3561 my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
3562 $opts->{r}, $mods);
3563
3564 # $opts->{ra} functions should not be used after this:
3565 my @ce = $opts->{ra}->get_commit_editor($opts->{log},
3566 $opts->{editor_cb}, $pool);
3567 my $self = SVN::Delta::Editor->new(@ce, $pool);
3568 bless $self, $class;
3569 foreach (qw/svn_path r tree_a tree_b/) {
3570 $self->{$_} = $opts->{$_};
3571 }
3572 $self->{url} = $opts->{ra}->{url};
3573 $self->{mods} = $mods;
3574 $self->{types} = $types;
3575 $self->{pool} = $pool;
3576 $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
3577 $self->{rm} = { };
3578 $self->{path_prefix} = length $self->{svn_path} ?
3579 "$self->{svn_path}/" : '';
3580 $self->{config} = $opts->{config};
3581 return $self;
3582 }
3583
3584 sub generate_diff {
3585 my ($tree_a, $tree_b) = @_;
3586 my @diff_tree = qw(diff-tree -z -r);
3587 if ($_cp_similarity) {
3588 push @diff_tree, "-C$_cp_similarity";
3589 } else {
3590 push @diff_tree, '-C';
3591 }
3592 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
3593 push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
3594 push @diff_tree, $tree_a, $tree_b;
3595 my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
3596 local $/ = "\0";
3597 my $state = 'meta';
3598 my @mods;
3599 while (<$diff_fh>) {
3600 chomp $_; # this gets rid of the trailing "\0"
3601 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
3602 ($::sha1)\s($::sha1)\s
3603 ([MTCRAD])\d*$/xo) {
3604 push @mods, { mode_a => $1, mode_b => $2,
3605 sha1_a => $3, sha1_b => $4,
3606 chg => $5 };
3607 if ($5 =~ /^(?:C|R)$/) {
3608 $state = 'file_a';
3609 } else {
3610 $state = 'file_b';
3611 }
3612 } elsif ($state eq 'file_a') {
3613 my $x = $mods[$#mods] or croak "Empty array\n";
3614 if ($x->{chg} !~ /^(?:C|R)$/) {
3615 croak "Error parsing $_, $x->{chg}\n";
3616 }
3617 $x->{file_a} = $_;
3618 $state = 'file_b';
3619 } elsif ($state eq 'file_b') {
3620 my $x = $mods[$#mods] or croak "Empty array\n";
3621 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
3622 croak "Error parsing $_, $x->{chg}\n";
3623 }
3624 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
3625 croak "Error parsing $_, $x->{chg}\n";
3626 }
3627 $x->{file_b} = $_;
3628 $state = 'meta';
3629 } else {
3630 croak "Error parsing $_\n";
3631 }
3632 }
3633 command_close_pipe($diff_fh, $ctx);
3634 \@mods;
3635 }
3636
3637 sub check_diff_paths {
3638 my ($ra, $pfx, $rev, $mods) = @_;
3639 my %types;
3640 $pfx .= '/' if length $pfx;
3641
3642 sub type_diff_paths {
3643 my ($ra, $types, $path, $rev) = @_;
3644 my @p = split m#/+#, $path;
3645 my $c = shift @p;
3646 unless (defined $types->{$c}) {
3647 $types->{$c} = $ra->check_path($c, $rev);
3648 }
3649 while (@p) {
3650 $c .= '/' . shift @p;
3651 next if defined $types->{$c};
3652 $types->{$c} = $ra->check_path($c, $rev);
3653 }
3654 }
3655
3656 foreach my $m (@$mods) {
3657 foreach my $f (qw/file_a file_b/) {
3658 next unless defined $m->{$f};
3659 my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
3660 if (length $pfx.$dir && ! defined $types{$dir}) {
3661 type_diff_paths($ra, \%types, $pfx.$dir, $rev);
3662 }
3663 }
3664 }
3665 \%types;
3666 }
3667
3668 sub split_path {
3669 return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
3670 }
3671
3672 sub repo_path {
3673 my ($self, $path) = @_;
3674 $self->{path_prefix}.(defined $path ? $path : '');
3675 }
3676
3677 sub url_path {
3678 my ($self, $path) = @_;
3679 if ($self->{url} =~ m#^https?://#) {
3680 $path =~ s/([^~a-zA-Z0-9_.-])/uc sprintf("%%%02x",ord($1))/eg;
3681 }
3682 $self->{url} . '/' . $self->repo_path($path);
3683 }
3684
3685 sub rmdirs {
3686 my ($self) = @_;
3687 my $rm = $self->{rm};
3688 delete $rm->{''}; # we never delete the url we're tracking
3689 return unless %$rm;
3690
3691 foreach (keys %$rm) {
3692 my @d = split m#/#, $_;
3693 my $c = shift @d;
3694 $rm->{$c} = 1;
3695 while (@d) {
3696 $c .= '/' . shift @d;
3697 $rm->{$c} = 1;
3698 }
3699 }
3700 delete $rm->{$self->{svn_path}};
3701 delete $rm->{''}; # we never delete the url we're tracking
3702 return unless %$rm;
3703
3704 my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
3705 $self->{tree_b});
3706 local $/ = "\0";
3707 while (<$fh>) {
3708 chomp;
3709 my @dn = split m#/#, $_;
3710 while (pop @dn) {
3711 delete $rm->{join '/', @dn};
3712 }
3713 unless (%$rm) {
3714 close $fh;
3715 return;
3716 }
3717 }
3718 command_close_pipe($fh, $ctx);
3719
3720 my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
3721 foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
3722 $self->close_directory($bat->{$d}, $p);
3723 my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
3724 print "\tD+\t$d/\n" unless $::_q;
3725 $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
3726 delete $bat->{$d};
3727 }
3728 }
3729
3730 sub open_or_add_dir {
3731 my ($self, $full_path, $baton) = @_;
3732 my $t = $self->{types}->{$full_path};
3733 if (!defined $t) {
3734 die "$full_path not known in r$self->{r} or we have a bug!\n";
3735 }
3736 {
3737 no warnings 'once';
3738 # SVN::Node::none and SVN::Node::file are used only once,
3739 # so we're shutting up Perl's warnings about them.
3740 if ($t == $SVN::Node::none) {
3741 return $self->add_directory($full_path, $baton,
3742 undef, -1, $self->{pool});
3743 } elsif ($t == $SVN::Node::dir) {
3744 return $self->open_directory($full_path, $baton,
3745 $self->{r}, $self->{pool});
3746 } # no warnings 'once'
3747 print STDERR "$full_path already exists in repository at ",
3748 "r$self->{r} and it is not a directory (",
3749 ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
3750 } # no warnings 'once'
3751 exit 1;
3752 }
3753
3754 sub ensure_path {
3755 my ($self, $path) = @_;
3756 my $bat = $self->{bat};
3757 my $repo_path = $self->repo_path($path);
3758 return $bat->{''} unless (length $repo_path);
3759 my @p = split m#/+#, $repo_path;
3760 my $c = shift @p;
3761 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
3762 while (@p) {
3763 my $c0 = $c;
3764 $c .= '/' . shift @p;
3765 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
3766 }
3767 return $bat->{$c};
3768 }
3769
3770 # Subroutine to convert a globbing pattern to a regular expression.
3771 # From perl cookbook.
3772 sub glob2pat {
3773 my $globstr = shift;
3774 my %patmap = ('*' => '.*', '?' => '.', '[' => '[', ']' => ']');
3775 $globstr =~ s{(.)} { $patmap{$1} || "\Q$1" }ge;
3776 return '^' . $globstr . '$';
3777 }
3778
3779 sub check_autoprop {
3780 my ($self, $pattern, $properties, $file, $fbat) = @_;
3781 # Convert the globbing pattern to a regular expression.
3782 my $regex = glob2pat($pattern);
3783 # Check if the pattern matches the file name.
3784 if($file =~ m/($regex)/) {
3785 # Parse the list of properties to set.
3786 my @props = split(/;/, $properties);
3787 foreach my $prop (@props) {
3788 # Parse 'name=value' syntax and set the property.
3789 if ($prop =~ /([^=]+)=(.*)/) {
3790 my ($n,$v) = ($1,$2);
3791 for ($n, $v) {
3792 s/^\s+//; s/\s+$//;
3793 }
3794 $self->change_file_prop($fbat, $n, $v);
3795 }
3796 }
3797 }
3798 }
3799
3800 sub apply_autoprops {
3801 my ($self, $file, $fbat) = @_;
3802 my $conf_t = ${$self->{config}}{'config'};
3803 no warnings 'once';
3804 # Check [miscellany]/enable-auto-props in svn configuration.
3805 if (SVN::_Core::svn_config_get_bool(
3806 $conf_t,
3807 $SVN::_Core::SVN_CONFIG_SECTION_MISCELLANY,
3808 $SVN::_Core::SVN_CONFIG_OPTION_ENABLE_AUTO_PROPS,
3809 0)) {
3810 # Auto-props are enabled. Enumerate them to look for matches.
3811 my $callback = sub {
3812 $self->check_autoprop($_[0], $_[1], $file, $fbat);
3813 };
3814 SVN::_Core::svn_config_enumerate(
3815 $conf_t,
3816 $SVN::_Core::SVN_CONFIG_SECTION_AUTO_PROPS,
3817 $callback);
3818 }
3819 }
3820
3821 sub A {
3822 my ($self, $m) = @_;
3823 my ($dir, $file) = split_path($m->{file_b});
3824 my $pbat = $self->ensure_path($dir);
3825 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3826 undef, -1);
3827 print "\tA\t$m->{file_b}\n" unless $::_q;
3828 $self->apply_autoprops($file, $fbat);
3829 $self->chg_file($fbat, $m);
3830 $self->close_file($fbat,undef,$self->{pool});
3831 }
3832
3833 sub C {
3834 my ($self, $m) = @_;
3835 my ($dir, $file) = split_path($m->{file_b});
3836 my $pbat = $self->ensure_path($dir);
3837 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3838 $self->url_path($m->{file_a}), $self->{r});
3839 print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3840 $self->chg_file($fbat, $m);
3841 $self->close_file($fbat,undef,$self->{pool});
3842 }
3843
3844 sub delete_entry {
3845 my ($self, $path, $pbat) = @_;
3846 my $rpath = $self->repo_path($path);
3847 my ($dir, $file) = split_path($rpath);
3848 $self->{rm}->{$dir} = 1;
3849 $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
3850 }
3851
3852 sub R {
3853 my ($self, $m) = @_;
3854 my ($dir, $file) = split_path($m->{file_b});
3855 my $pbat = $self->ensure_path($dir);
3856 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3857 $self->url_path($m->{file_a}), $self->{r});
3858 print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3859 $self->apply_autoprops($file, $fbat);
3860 $self->chg_file($fbat, $m);
3861 $self->close_file($fbat,undef,$self->{pool});
3862
3863 ($dir, $file) = split_path($m->{file_a});
3864 $pbat = $self->ensure_path($dir);
3865 $self->delete_entry($m->{file_a}, $pbat);
3866 }
3867
3868 sub M {
3869 my ($self, $m) = @_;
3870 my ($dir, $file) = split_path($m->{file_b});
3871 my $pbat = $self->ensure_path($dir);
3872 my $fbat = $self->open_file($self->repo_path($m->{file_b}),
3873 $pbat,$self->{r},$self->{pool});
3874 print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
3875 $self->chg_file($fbat, $m);
3876 $self->close_file($fbat,undef,$self->{pool});
3877 }
3878
3879 sub T { shift->M(@_) }
3880
3881 sub change_file_prop {
3882 my ($self, $fbat, $pname, $pval) = @_;
3883 $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
3884 }
3885
3886 sub _chg_file_get_blob ($$$$) {
3887 my ($self, $fbat, $m, $which) = @_;
3888 my $fh = $::_repository->temp_acquire("git_blob_$which");
3889 if ($m->{"mode_$which"} =~ /^120/) {
3890 print $fh 'link ' or croak $!;
3891 $self->change_file_prop($fbat,'svn:special','*');
3892 } elsif ($m->{mode_a} =~ /^120/ && $m->{"mode_$which"} !~ /^120/) {
3893 $self->change_file_prop($fbat,'svn:special',undef);
3894 }
3895 my $blob = $m->{"sha1_$which"};
3896 return ($fh,) if ($blob =~ /^0{40}$/);
3897 my $size = $::_repository->cat_blob($blob, $fh);
3898 croak "Failed to read object $blob" if ($size < 0);
3899 $fh->flush == 0 or croak $!;
3900 seek $fh, 0, 0 or croak $!;
3901
3902 my $exp = ::md5sum($fh);
3903 seek $fh, 0, 0 or croak $!;
3904 return ($fh, $exp);
3905 }
3906
3907 sub chg_file {
3908 my ($self, $fbat, $m) = @_;
3909 if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
3910 $self->change_file_prop($fbat,'svn:executable','*');
3911 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
3912 $self->change_file_prop($fbat,'svn:executable',undef);
3913 }
3914 my ($fh_a, $exp_a) = _chg_file_get_blob $self, $fbat, $m, 'a';
3915 my ($fh_b, $exp_b) = _chg_file_get_blob $self, $fbat, $m, 'b';
3916 my $pool = SVN::Pool->new;
3917 my $atd = $self->apply_textdelta($fbat, $exp_a, $pool);
3918 if (-s $fh_a) {
3919 my $txstream = SVN::TxDelta::new ($fh_a, $fh_b, $pool);
3920 my $res = SVN::TxDelta::send_txstream($txstream, @$atd, $pool);
3921 if (defined $res) {
3922 die "Unexpected result from send_txstream: $res\n",
3923 "(SVN::Core::VERSION: $SVN::Core::VERSION)\n";
3924 }
3925 } else {
3926 my $got = SVN::TxDelta::send_stream($fh_b, @$atd, $pool);
3927 die "Checksum mismatch\nexpected: $exp_b\ngot: $got\n"
3928 if ($got ne $exp_b);
3929 }
3930 Git::temp_release($fh_b, 1);
3931 Git::temp_release($fh_a, 1);
3932 $pool->clear;
3933 }
3934
3935 sub D {
3936 my ($self, $m) = @_;
3937 my ($dir, $file) = split_path($m->{file_b});
3938 my $pbat = $self->ensure_path($dir);
3939 print "\tD\t$m->{file_b}\n" unless $::_q;
3940 $self->delete_entry($m->{file_b}, $pbat);
3941 }
3942
3943 sub close_edit {
3944 my ($self) = @_;
3945 my ($p,$bat) = ($self->{pool}, $self->{bat});
3946 foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
3947 next if $_ eq '';
3948 $self->close_directory($bat->{$_}, $p);
3949 }
3950 $self->close_directory($bat->{''}, $p);
3951 $self->SUPER::close_edit($p);
3952 $p->clear;
3953 }
3954
3955 sub abort_edit {
3956 my ($self) = @_;
3957 $self->SUPER::abort_edit($self->{pool});
3958 }
3959
3960 sub DESTROY {
3961 my $self = shift;
3962 $self->SUPER::DESTROY(@_);
3963 $self->{pool}->clear;
3964 }
3965
3966 # this drives the editor
3967 sub apply_diff {
3968 my ($self) = @_;
3969 my $mods = $self->{mods};
3970 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
3971 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
3972 my $f = $m->{chg};
3973 if (defined $o{$f}) {
3974 $self->$f($m);
3975 } else {
3976 fatal("Invalid change type: $f");
3977 }
3978 }
3979 $self->rmdirs if $_rmdir;
3980 if (@$mods == 0) {
3981 $self->abort_edit;
3982 } else {
3983 $self->close_edit;
3984 }
3985 return scalar @$mods;
3986 }
3987
3988 package Git::SVN::Ra;
3989 use vars qw/@ISA $config_dir $_log_window_size/;
3990 use strict;
3991 use warnings;
3992 my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
3993
3994 BEGIN {
3995 # enforce temporary pool usage for some simple functions
3996 no strict 'refs';
3997 for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root/) {
3998 my $SUPER = "SUPER::$f";
3999 *$f = sub {
4000 my $self = shift;
4001 my $pool = SVN::Pool->new;
4002 my @ret = $self->$SUPER(@_,$pool);
4003 $pool->clear;
4004 wantarray ? @ret : $ret[0];
4005 };
4006 }
4007 }
4008
4009 sub _auth_providers () {
4010 [
4011 SVN::Client::get_simple_provider(),
4012 SVN::Client::get_ssl_server_trust_file_provider(),
4013 SVN::Client::get_simple_prompt_provider(
4014 \&Git::SVN::Prompt::simple, 2),
4015 SVN::Client::get_ssl_client_cert_file_provider(),
4016 SVN::Client::get_ssl_client_cert_prompt_provider(
4017 \&Git::SVN::Prompt::ssl_client_cert, 2),
4018 SVN::Client::get_ssl_client_cert_pw_file_provider(),
4019 SVN::Client::get_ssl_client_cert_pw_prompt_provider(
4020 \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
4021 SVN::Client::get_username_provider(),
4022 SVN::Client::get_ssl_server_trust_prompt_provider(
4023 \&Git::SVN::Prompt::ssl_server_trust),
4024 SVN::Client::get_username_prompt_provider(
4025 \&Git::SVN::Prompt::username, 2)
4026 ]
4027 }
4028
4029 sub escape_uri_only {
4030 my ($uri) = @_;
4031 my @tmp;
4032 foreach (split m{/}, $uri) {
4033 s/([^~\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
4034 push @tmp, $_;
4035 }
4036 join('/', @tmp);
4037 }
4038
4039 sub escape_url {
4040 my ($url) = @_;
4041 if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
4042 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
4043 $url = "$scheme://$domain$uri";
4044 }
4045 $url;
4046 }
4047
4048 sub new {
4049 my ($class, $url) = @_;
4050 $url =~ s!/+$!!;
4051 return $RA if ($RA && $RA->{url} eq $url);
4052
4053 SVN::_Core::svn_config_ensure($config_dir, undef);
4054 my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
4055 my $config = SVN::Core::config_get_config($config_dir);
4056 $RA = undef;
4057 my $dont_store_passwords = 1;
4058 my $conf_t = ${$config}{'config'};
4059 {
4060 no warnings 'once';
4061 # The usage of $SVN::_Core::SVN_CONFIG_* variables
4062 # produces warnings that variables are used only once.
4063 # I had not found the better way to shut them up, so
4064 # the warnings of type 'once' are disabled in this block.
4065 if (SVN::_Core::svn_config_get_bool($conf_t,
4066 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
4067 $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
4068 1) == 0) {
4069 SVN::_Core::svn_auth_set_parameter($baton,
4070 $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
4071 bless (\$dont_store_passwords, "_p_void"));
4072 }
4073 if (SVN::_Core::svn_config_get_bool($conf_t,
4074 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
4075 $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
4076 1) == 0) {
4077 $Git::SVN::Prompt::_no_auth_cache = 1;
4078 }
4079 } # no warnings 'once'
4080 my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
4081 config => $config,
4082 pool => SVN::Pool->new,
4083 auth_provider_callbacks => $callbacks);
4084 $self->{url} = $url;
4085 $self->{svn_path} = $url;
4086 $self->{repos_root} = $self->get_repos_root;
4087 $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
4088 $self->{cache} = { check_path => { r => 0, data => {} },
4089 get_dir => { r => 0, data => {} } };
4090 $RA = bless $self, $class;
4091 }
4092
4093 sub check_path {
4094 my ($self, $path, $r) = @_;
4095 my $cache = $self->{cache}->{check_path};
4096 if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
4097 return $cache->{data}->{$path};
4098 }
4099 my $pool = SVN::Pool->new;
4100 my $t = $self->SUPER::check_path($path, $r, $pool);
4101 $pool->clear;
4102 if ($r != $cache->{r}) {
4103 %{$cache->{data}} = ();
4104 $cache->{r} = $r;
4105 }
4106 $cache->{data}->{$path} = $t;
4107 }
4108
4109 sub get_dir {
4110 my ($self, $dir, $r) = @_;
4111 my $cache = $self->{cache}->{get_dir};
4112 if ($r == $cache->{r}) {
4113 if (my $x = $cache->{data}->{$dir}) {
4114 return wantarray ? @$x : $x->[0];
4115 }
4116 }
4117 my $pool = SVN::Pool->new;
4118 my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
4119 my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
4120 $pool->clear;
4121 if ($r != $cache->{r}) {
4122 %{$cache->{data}} = ();
4123 $cache->{r} = $r;
4124 }
4125 $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
4126 wantarray ? (\%dirents, $r, $props) : \%dirents;
4127 }
4128
4129 sub DESTROY {
4130 # do not call the real DESTROY since we store ourselves in $RA
4131 }
4132
4133 sub get_log {
4134 my ($self, @args) = @_;
4135 my $pool = SVN::Pool->new;
4136 splice(@args, 3, 1) if ($SVN::Core::VERSION le '1.2.0');
4137 my $ret = $self->SUPER::get_log(@args, $pool);
4138 $pool->clear;
4139 $ret;
4140 }
4141
4142 sub trees_match {
4143 my ($self, $url1, $rev1, $url2, $rev2) = @_;
4144 my $ctx = SVN::Client->new(auth => _auth_providers);
4145 my $out = IO::File->new_tmpfile;
4146
4147 # older SVN (1.1.x) doesn't take $pool as the last parameter for
4148 # $ctx->diff(), so we'll create a default one
4149 my $pool = SVN::Pool->new_default_sub;
4150
4151 $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
4152 $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
4153 $out->flush;
4154 my $ret = (($out->stat)[7] == 0);
4155 close $out or croak $!;
4156
4157 $ret;
4158 }
4159
4160 sub get_commit_editor {
4161 my ($self, $log, $cb, $pool) = @_;
4162 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
4163 $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
4164 }
4165
4166 sub gs_do_update {
4167 my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
4168 my $new = ($rev_a == $rev_b);
4169 my $path = $gs->{path};
4170
4171 if ($new && -e $gs->{index}) {
4172 unlink $gs->{index} or die
4173 "Couldn't unlink index: $gs->{index}: $!\n";
4174 }
4175 my $pool = SVN::Pool->new;
4176 $editor->set_path_strip($path);
4177 my (@pc) = split m#/#, $path;
4178 my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
4179 1, $editor, $pool);
4180 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
4181
4182 # Since we can't rely on svn_ra_reparent being available, we'll
4183 # just have to do some magic with set_path to make it so
4184 # we only want a partial path.
4185 my $sp = '';
4186 my $final = join('/', @pc);
4187 while (@pc) {
4188 $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
4189 $sp .= '/' if length $sp;
4190 $sp .= shift @pc;
4191 }
4192 die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
4193
4194 $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
4195
4196 $reporter->finish_report($pool);
4197 $pool->clear;
4198 $editor->{git_commit_ok};
4199 }
4200
4201 # this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
4202 # svn_ra_reparent didn't work before 1.4)
4203 sub gs_do_switch {
4204 my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
4205 my $path = $gs->{path};
4206 my $pool = SVN::Pool->new;
4207
4208 my $full_url = $self->{url};
4209 my $old_url = $full_url;
4210 $full_url .= '/' . escape_uri_only($path) if length $path;
4211 my ($ra, $reparented);
4212
4213 if ($old_url =~ m#^svn(\+ssh)?://#) {
4214 $_[0] = undef;
4215 $self = undef;
4216 $RA = undef;
4217 $ra = Git::SVN::Ra->new($full_url);
4218 $ra_invalid = 1;
4219 } elsif ($old_url ne $full_url) {
4220 SVN::_Ra::svn_ra_reparent($self->{session}, $full_url, $pool);
4221 $self->{url} = $full_url;
4222 $reparented = 1;
4223 }
4224
4225 $ra ||= $self;
4226 $url_b = escape_url($url_b);
4227 my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
4228 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
4229 $reporter->set_path('', $rev_a, 0, @lock, $pool);
4230 $reporter->finish_report($pool);
4231
4232 if ($reparented) {
4233 SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
4234 $self->{url} = $old_url;
4235 }
4236
4237 $pool->clear;
4238 $editor->{git_commit_ok};
4239 }
4240
4241 sub longest_common_path {
4242 my ($gsv, $globs) = @_;
4243 my %common;
4244 my $common_max = scalar @$gsv;
4245
4246 foreach my $gs (@$gsv) {
4247 my @tmp = split m#/#, $gs->{path};
4248 my $p = '';
4249 foreach (@tmp) {
4250 $p .= length($p) ? "/$_" : $_;
4251 $common{$p} ||= 0;
4252 $common{$p}++;
4253 }
4254 }
4255 $globs ||= [];
4256 $common_max += scalar @$globs;
4257 foreach my $glob (@$globs) {
4258 my @tmp = split m#/#, $glob->{path}->{left};
4259 my $p = '';
4260 foreach (@tmp) {
4261 $p .= length($p) ? "/$_" : $_;
4262 $common{$p} ||= 0;
4263 $common{$p}++;
4264 }
4265 }
4266
4267 my $longest_path = '';
4268 foreach (sort {length $b <=> length $a} keys %common) {
4269 if ($common{$_} == $common_max) {
4270 $longest_path = $_;
4271 last;
4272 }
4273 }
4274 $longest_path;
4275 }
4276
4277 sub gs_fetch_loop_common {
4278 my ($self, $base, $head, $gsv, $globs) = @_;
4279 return if ($base > $head);
4280 my $inc = $_log_window_size;
4281 my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
4282 my $longest_path = longest_common_path($gsv, $globs);
4283 my $ra_url = $self->{url};
4284 while (1) {
4285 my %revs;
4286 my $err;
4287 my $err_handler = $SVN::Error::handler;
4288 $SVN::Error::handler = sub {
4289 ($err) = @_;
4290 skip_unknown_revs($err);
4291 };
4292 sub _cb {
4293 my ($paths, $r, $author, $date, $log) = @_;
4294 [ dup_changed_paths($paths),
4295 { author => $author, date => $date, log => $log } ];
4296 }
4297 $self->get_log([$longest_path], $min, $max, 0, 1, 1,
4298 sub { $revs{$_[1]} = _cb(@_) });
4299 if ($err && $max >= $head) {
4300 print STDERR "Path '$longest_path' ",
4301 "was probably deleted:\n",
4302 $err->expanded_message,
4303 "\nWill attempt to follow ",
4304 "revisions r$min .. r$max ",
4305 "committed before the deletion\n";
4306 my $hi = $max;
4307 while (--$hi >= $min) {
4308 my $ok;
4309 $self->get_log([$longest_path], $min, $hi,
4310 0, 1, 1, sub {
4311 $ok ||= $_[1];
4312 $revs{$_[1]} = _cb(@_) });
4313 if ($ok) {
4314 print STDERR "r$min .. r$ok OK\n";
4315 last;
4316 }
4317 }
4318 }
4319 $SVN::Error::handler = $err_handler;
4320
4321 my %exists = map { $_->{path} => $_ } @$gsv;
4322 foreach my $r (sort {$a <=> $b} keys %revs) {
4323 my ($paths, $logged) = @{$revs{$r}};
4324
4325 foreach my $gs ($self->match_globs(\%exists, $paths,
4326 $globs, $r)) {
4327 if ($gs->rev_map_max >= $r) {
4328 next;
4329 }
4330 next unless $gs->match_paths($paths, $r);
4331 $gs->{logged_rev_props} = $logged;
4332 if (my $last_commit = $gs->last_commit) {
4333 $gs->assert_index_clean($last_commit);
4334 }
4335 my $log_entry = $gs->do_fetch($paths, $r);
4336 if ($log_entry) {
4337 $gs->do_git_commit($log_entry);
4338 }
4339 $INDEX_FILES{$gs->{index}} = 1;
4340 }
4341 foreach my $g (@$globs) {
4342 my $k = "svn-remote.$g->{remote}." .
4343 "$g->{t}-maxRev";
4344 Git::SVN::tmp_config($k, $r);
4345 }
4346 if ($ra_invalid) {
4347 $_[0] = undef;
4348 $self = undef;
4349 $RA = undef;
4350 $self = Git::SVN::Ra->new($ra_url);
4351 $ra_invalid = undef;
4352 }
4353 }
4354 # pre-fill the .rev_db since it'll eventually get filled in
4355 # with '0' x40 if something new gets committed
4356 foreach my $gs (@$gsv) {
4357 next if $gs->rev_map_max >= $max;
4358 next if defined $gs->rev_map_get($max);
4359 $gs->rev_map_set($max, 0 x40);
4360 }
4361 foreach my $g (@$globs) {
4362 my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
4363 Git::SVN::tmp_config($k, $max);
4364 }
4365 last if $max >= $head;
4366 $min = $max + 1;
4367 $max += $inc;
4368 $max = $head if ($max > $head);
4369 }
4370 Git::SVN::gc();
4371 }
4372
4373 sub get_dir_globbed {
4374 my ($self, $left, $depth, $r) = @_;
4375
4376 my @x = eval { $self->get_dir($left, $r) };
4377 return unless scalar @x == 3;
4378 my $dirents = $x[0];
4379 my @finalents;
4380 foreach my $de (keys %$dirents) {
4381 next if $dirents->{$de}->{kind} != $SVN::Node::dir;
4382 if ($depth > 1) {
4383 my @args = ("$left/$de", $depth - 1, $r);
4384 foreach my $dir ($self->get_dir_globbed(@args)) {
4385 push @finalents, "$de/$dir";
4386 }
4387 } else {
4388 push @finalents, $de;
4389 }
4390 }
4391 @finalents;
4392 }
4393
4394 sub match_globs {
4395 my ($self, $exists, $paths, $globs, $r) = @_;
4396
4397 sub get_dir_check {
4398 my ($self, $exists, $g, $r) = @_;
4399
4400 my @dirs = $self->get_dir_globbed($g->{path}->{left},
4401 $g->{path}->{depth},
4402 $r);
4403
4404 foreach my $de (@dirs) {
4405 my $p = $g->{path}->full_path($de);
4406 next if $exists->{$p};
4407 next if (length $g->{path}->{right} &&
4408 ($self->check_path($p, $r) !=
4409 $SVN::Node::dir));
4410 $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
4411 $g->{ref}->full_path($de), 1);
4412 }
4413 }
4414 foreach my $g (@$globs) {
4415 if (my $path = $paths->{"/$g->{path}->{left}"}) {
4416 if ($path->{action} =~ /^[AR]$/) {
4417 get_dir_check($self, $exists, $g, $r);
4418 }
4419 }
4420 foreach (keys %$paths) {
4421 if (/$g->{path}->{left_regex}/ &&
4422 !/$g->{path}->{regex}/) {
4423 next if $paths->{$_}->{action} !~ /^[AR]$/;
4424 get_dir_check($self, $exists, $g, $r);
4425 }
4426 next unless /$g->{path}->{regex}/;
4427 my $p = $1;
4428 my $pathname = $g->{path}->full_path($p);
4429 next if $exists->{$pathname};
4430 next if ($self->check_path($pathname, $r) !=
4431 $SVN::Node::dir);
4432 $exists->{$pathname} = Git::SVN->init(
4433 $self->{url}, $pathname, undef,
4434 $g->{ref}->full_path($p), 1);
4435 }
4436 my $c = '';
4437 foreach (split m#/#, $g->{path}->{left}) {
4438 $c .= "/$_";
4439 next unless ($paths->{$c} &&
4440 ($paths->{$c}->{action} =~ /^[AR]$/));
4441 get_dir_check($self, $exists, $g, $r);
4442 }
4443 }
4444 values %$exists;
4445 }
4446
4447 sub minimize_url {
4448 my ($self) = @_;
4449 return $self->{url} if ($self->{url} eq $self->{repos_root});
4450 my $url = $self->{repos_root};
4451 my @components = split(m!/!, $self->{svn_path});
4452 my $c = '';
4453 do {
4454 $url .= "/$c" if length $c;
4455 eval { (ref $self)->new($url)->get_latest_revnum };
4456 } while ($@ && ($c = shift @components));
4457 $url;
4458 }
4459
4460 sub can_do_switch {
4461 my $self = shift;
4462 unless (defined $can_do_switch) {
4463 my $pool = SVN::Pool->new;
4464 my $rep = eval {
4465 $self->do_switch(1, '', 0, $self->{url},
4466 SVN::Delta::Editor->new, $pool);
4467 };
4468 if ($@) {
4469 $can_do_switch = 0;
4470 } else {
4471 $rep->abort_report($pool);
4472 $can_do_switch = 1;
4473 }
4474 $pool->clear;
4475 }
4476 $can_do_switch;
4477 }
4478
4479 sub skip_unknown_revs {
4480 my ($err) = @_;
4481 my $errno = $err->apr_err();
4482 # Maybe the branch we're tracking didn't
4483 # exist when the repo started, so it's
4484 # not an error if it doesn't, just continue
4485 #
4486 # Wonderfully consistent library, eh?
4487 # 160013 - svn:// and file://
4488 # 175002 - http(s)://
4489 # 175007 - http(s):// (this repo required authorization, too...)
4490 # More codes may be discovered later...
4491 if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
4492 my $err_key = $err->expanded_message;
4493 # revision numbers change every time, filter them out
4494 $err_key =~ s/\d+/\0/g;
4495 $err_key = "$errno\0$err_key";
4496 unless ($ignored_err{$err_key}) {
4497 warn "W: Ignoring error from SVN, path probably ",
4498 "does not exist: ($errno): ",
4499 $err->expanded_message,"\n";
4500 warn "W: Do not be alarmed at the above message ",
4501 "git-svn is just searching aggressively for ",
4502 "old history.\n",
4503 "This may take a while on large repositories\n";
4504 $ignored_err{$err_key} = 1;
4505 }
4506 return;
4507 }
4508 die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
4509 }
4510
4511 # svn_log_changed_path_t objects passed to get_log are likely to be
4512 # overwritten even if only the refs are copied to an external variable,
4513 # so we should dup the structures in their entirety. Using an externally
4514 # passed pool (instead of our temporary and quickly cleared pool in
4515 # Git::SVN::Ra) does not help matters at all...
4516 sub dup_changed_paths {
4517 my ($paths) = @_;
4518 return undef unless $paths;
4519 my %ret;
4520 foreach my $p (keys %$paths) {
4521 my $i = $paths->{$p};
4522 my %s = map { $_ => $i->$_ }
4523 qw/copyfrom_path copyfrom_rev action/;
4524 $ret{$p} = \%s;
4525 }
4526 \%ret;
4527 }
4528
4529 package Git::SVN::Log;
4530 use strict;
4531 use warnings;
4532 use POSIX qw/strftime/;
4533 use constant commit_log_separator => ('-' x 72) . "\n";
4534 use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
4535 %rusers $show_commit $incremental/;
4536 my $l_fmt;
4537
4538 sub cmt_showable {
4539 my ($c) = @_;
4540 return 1 if defined $c->{r};
4541
4542 # big commit message got truncated by the 16k pretty buffer in rev-list
4543 if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
4544 $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
4545 @{$c->{l}} = ();
4546 my @log = command(qw/cat-file commit/, $c->{c});
4547
4548 # shift off the headers
4549 shift @log while ($log[0] ne '');
4550 shift @log;
4551
4552 # TODO: make $c->{l} not have a trailing newline in the future
4553 @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
4554
4555 (undef, $c->{r}, undef) = ::extract_metadata(
4556 (grep(/^git-svn-id: /, @log))[-1]);
4557 }
4558 return defined $c->{r};
4559 }
4560
4561 sub log_use_color {
4562 return $color || Git->repository->get_colorbool('color.diff');
4563 }
4564
4565 sub git_svn_log_cmd {
4566 my ($r_min, $r_max, @args) = @_;
4567 my $head = 'HEAD';
4568 my (@files, @log_opts);
4569 foreach my $x (@args) {
4570 if ($x eq '--' || @files) {
4571 push @files, $x;
4572 } else {
4573 if (::verify_ref("$x^0")) {
4574 $head = $x;
4575 } else {
4576 push @log_opts, $x;
4577 }
4578 }
4579 }
4580
4581 my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
4582 $gs ||= Git::SVN->_new;
4583 my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
4584 $gs->refname);
4585 push @cmd, '-r' unless $non_recursive;
4586 push @cmd, qw/--raw --name-status/ if $verbose;
4587 push @cmd, '--color' if log_use_color();
4588 push @cmd, @log_opts;
4589 if (defined $r_max && $r_max == $r_min) {
4590 push @cmd, '--max-count=1';
4591 if (my $c = $gs->rev_map_get($r_max)) {
4592 push @cmd, $c;
4593 }
4594 } elsif (defined $r_max) {
4595 if ($r_max < $r_min) {
4596 ($r_min, $r_max) = ($r_max, $r_min);
4597 }
4598 my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
4599 my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
4600 # If there are no commits in the range, both $c_max and $c_min
4601 # will be undefined. If there is at least 1 commit in the
4602 # range, both will be defined.
4603 return () if !defined $c_min || !defined $c_max;
4604 if ($c_min eq $c_max) {
4605 push @cmd, '--max-count=1', $c_min;
4606 } else {
4607 push @cmd, '--boundary', "$c_min..$c_max";
4608 }
4609 }
4610 return (@cmd, @files);
4611 }
4612
4613 # adapted from pager.c
4614 sub config_pager {
4615 $pager ||= $ENV{GIT_PAGER} || $ENV{PAGER};
4616 if (!defined $pager) {
4617 $pager = 'less';
4618 } elsif (length $pager == 0 || $pager eq 'cat') {
4619 $pager = undef;
4620 }
4621 $ENV{GIT_PAGER_IN_USE} = defined($pager);
4622 }
4623
4624 sub run_pager {
4625 return unless -t *STDOUT && defined $pager;
4626 pipe my ($rfd, $wfd) or return;
4627 defined(my $pid = fork) or ::fatal "Can't fork: $!";
4628 if (!$pid) {
4629 open STDOUT, '>&', $wfd or
4630 ::fatal "Can't redirect to stdout: $!";
4631 return;
4632 }
4633 open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
4634 $ENV{LESS} ||= 'FRSX';
4635 exec $pager or ::fatal "Can't run pager: $! ($pager)";
4636 }
4637
4638 sub format_svn_date {
4639 return strftime("%Y-%m-%d %H:%M:%S %z (%a, %d %b %Y)", localtime(shift));
4640 }
4641
4642 sub parse_git_date {
4643 my ($t, $tz) = @_;
4644 # Date::Parse isn't in the standard Perl distro :(
4645 if ($tz =~ s/^\+//) {
4646 $t += tz_to_s_offset($tz);
4647 } elsif ($tz =~ s/^\-//) {
4648 $t -= tz_to_s_offset($tz);
4649 }
4650 return $t;
4651 }
4652
4653 sub set_local_timezone {
4654 if (defined $TZ) {
4655 $ENV{TZ} = $TZ;
4656 } else {
4657 delete $ENV{TZ};
4658 }
4659 }
4660
4661 sub tz_to_s_offset {
4662 my ($tz) = @_;
4663 $tz =~ s/(\d\d)$//;
4664 return ($1 * 60) + ($tz * 3600);
4665 }
4666
4667 sub get_author_info {
4668 my ($dest, $author, $t, $tz) = @_;
4669 $author =~ s/(?:^\s*|\s*$)//g;
4670 $dest->{a_raw} = $author;
4671 my $au;
4672 if ($::_authors) {
4673 $au = $rusers{$author} || undef;
4674 }
4675 if (!$au) {
4676 ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
4677 }
4678 $dest->{t} = $t;
4679 $dest->{tz} = $tz;
4680 $dest->{a} = $au;
4681 $dest->{t_utc} = parse_git_date($t, $tz);
4682 }
4683
4684 sub process_commit {
4685 my ($c, $r_min, $r_max, $defer) = @_;
4686 if (defined $r_min && defined $r_max) {
4687 if ($r_min == $c->{r} && $r_min == $r_max) {
4688 show_commit($c);
4689 return 0;
4690 }
4691 return 1 if $r_min == $r_max;
4692 if ($r_min < $r_max) {
4693 # we need to reverse the print order
4694 return 0 if (defined $limit && --$limit < 0);
4695 push @$defer, $c;
4696 return 1;
4697 }
4698 if ($r_min != $r_max) {
4699 return 1 if ($r_min < $c->{r});
4700 return 1 if ($r_max > $c->{r});
4701 }
4702 }
4703 return 0 if (defined $limit && --$limit < 0);
4704 show_commit($c);
4705 return 1;
4706 }
4707
4708 sub show_commit {
4709 my $c = shift;
4710 if ($oneline) {
4711 my $x = "\n";
4712 if (my $l = $c->{l}) {
4713 while ($l->[0] =~ /^\s*$/) { shift @$l }
4714 $x = $l->[0];
4715 }
4716 $l_fmt ||= 'A' . length($c->{r});
4717 print 'r',pack($l_fmt, $c->{r}),' | ';
4718 print "$c->{c} | " if $show_commit;
4719 print $x;
4720 } else {
4721 show_commit_normal($c);
4722 }
4723 }
4724
4725 sub show_commit_changed_paths {
4726 my ($c) = @_;
4727 return unless $c->{changed};
4728 print "Changed paths:\n", @{$c->{changed}};
4729 }
4730
4731 sub show_commit_normal {
4732 my ($c) = @_;
4733 print commit_log_separator, "r$c->{r} | ";
4734 print "$c->{c} | " if $show_commit;
4735 print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
4736 my $nr_line = 0;
4737
4738 if (my $l = $c->{l}) {
4739 while ($l->[$#$l] eq "\n" && $#$l > 0
4740 && $l->[($#$l - 1)] eq "\n") {
4741 pop @$l;
4742 }
4743 $nr_line = scalar @$l;
4744 if (!$nr_line) {
4745 print "1 line\n\n\n";
4746 } else {
4747 if ($nr_line == 1) {
4748 $nr_line = '1 line';
4749 } else {
4750 $nr_line .= ' lines';
4751 }
4752 print $nr_line, "\n";
4753 show_commit_changed_paths($c);
4754 print "\n";
4755 print $_ foreach @$l;
4756 }
4757 } else {
4758 print "1 line\n";
4759 show_commit_changed_paths($c);
4760 print "\n";
4761
4762 }
4763 foreach my $x (qw/raw stat diff/) {
4764 if ($c->{$x}) {
4765 print "\n";
4766 print $_ foreach @{$c->{$x}}
4767 }
4768 }
4769 }
4770
4771 sub cmd_show_log {
4772 my (@args) = @_;
4773 my ($r_min, $r_max);
4774 my $r_last = -1; # prevent dupes
4775 set_local_timezone();
4776 if (defined $::_revision) {
4777 if ($::_revision =~ /^(\d+):(\d+)$/) {
4778 ($r_min, $r_max) = ($1, $2);
4779 } elsif ($::_revision =~ /^\d+$/) {
4780 $r_min = $r_max = $::_revision;
4781 } else {
4782 ::fatal "-r$::_revision is not supported, use ",
4783 "standard 'git log' arguments instead";
4784 }
4785 }
4786
4787 config_pager();
4788 @args = git_svn_log_cmd($r_min, $r_max, @args);
4789 if (!@args) {
4790 print commit_log_separator unless $incremental || $oneline;
4791 return;
4792 }
4793 my $log = command_output_pipe(@args);
4794 run_pager();
4795 my (@k, $c, $d, $stat);
4796 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
4797 while (<$log>) {
4798 if (/^${esc_color}commit -?($::sha1_short)/o) {
4799 my $cmt = $1;
4800 if ($c && cmt_showable($c) && $c->{r} != $r_last) {
4801 $r_last = $c->{r};
4802 process_commit($c, $r_min, $r_max, \@k) or
4803 goto out;
4804 }
4805 $d = undef;
4806 $c = { c => $cmt };
4807 } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
4808 get_author_info($c, $1, $2, $3);
4809 } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
4810 # ignore
4811 } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
4812 push @{$c->{raw}}, $_;
4813 } elsif (/^${esc_color}[ACRMDT]\t/) {
4814 # we could add $SVN->{svn_path} here, but that requires
4815 # remote access at the moment (repo_path_split)...
4816 s#^(${esc_color})([ACRMDT])\t#$1 $2 #o;
4817 push @{$c->{changed}}, $_;
4818 } elsif (/^${esc_color}diff /o) {
4819 $d = 1;
4820 push @{$c->{diff}}, $_;
4821 } elsif ($d) {
4822 push @{$c->{diff}}, $_;
4823 } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
4824 $esc_color*[\+\-]*$esc_color$/x) {
4825 $stat = 1;
4826 push @{$c->{stat}}, $_;
4827 } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
4828 push @{$c->{stat}}, $_;
4829 $stat = undef;
4830 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
4831 ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
4832 } elsif (s/^${esc_color} //o) {
4833 push @{$c->{l}}, $_;
4834 }
4835 }
4836 if ($c && defined $c->{r} && $c->{r} != $r_last) {
4837 $r_last = $c->{r};
4838 process_commit($c, $r_min, $r_max, \@k);
4839 }
4840 if (@k) {
4841 ($r_min, $r_max) = ($r_max, $r_min);
4842 process_commit($_, $r_min, $r_max) foreach reverse @k;
4843 }
4844 out:
4845 close $log;
4846 print commit_log_separator unless $incremental || $oneline;
4847 }
4848
4849 sub cmd_blame {
4850 my $path = pop;
4851
4852 config_pager();
4853 run_pager();
4854
4855 my ($fh, $ctx, $rev);
4856
4857 if ($_git_format) {
4858 ($fh, $ctx) = command_output_pipe('blame', @_, $path);
4859 while (my $line = <$fh>) {
4860 if ($line =~ /^\^?([[:xdigit:]]+)\s/) {
4861 # Uncommitted edits show up as a rev ID of
4862 # all zeros, which we can't look up with
4863 # cmt_metadata
4864 if ($1 !~ /^0+$/) {
4865 (undef, $rev, undef) =
4866 ::cmt_metadata($1);
4867 $rev = '0' if (!$rev);
4868 } else {
4869 $rev = '0';
4870 }
4871 $rev = sprintf('%-10s', $rev);
4872 $line =~ s/^\^?[[:xdigit:]]+(\s)/$rev$1/;
4873 }
4874 print $line;
4875 }
4876 } else {
4877 ($fh, $ctx) = command_output_pipe('blame', '-p', @_, 'HEAD',
4878 '--', $path);
4879 my ($sha1);
4880 my %authors;
4881 while (my $line = <$fh>) {
4882 if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
4883 $sha1 = $1;
4884 (undef, $rev, undef) = ::cmt_metadata($1);
4885 $rev = '0' if (!$rev);
4886 }
4887 elsif ($line =~ /^author (.*)/) {
4888 $authors{$rev} = $1;
4889 $authors{$rev} =~ s/\s/_/g;
4890 }
4891 elsif ($line =~ /^\t(.*)$/) {
4892 printf("%6s %10s %s\n", $rev, $authors{$rev}, $1);
4893 }
4894 }
4895 }
4896 command_close_pipe($fh, $ctx);
4897 }
4898
4899 package Git::SVN::Migration;
4900 # these version numbers do NOT correspond to actual version numbers
4901 # of git nor git-svn. They are just relative.
4902 #
4903 # v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
4904 #
4905 # v1 layout: .git/$id/info/url, refs/remotes/$id
4906 #
4907 # v2 layout: .git/svn/$id/info/url, refs/remotes/$id
4908 #
4909 # v3 layout: .git/svn/$id, refs/remotes/$id
4910 # - info/url may remain for backwards compatibility
4911 # - this is what we migrate up to this layout automatically,
4912 # - this will be used by git svn init on single branches
4913 # v3.1 layout (auto migrated):
4914 # - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
4915 # for backwards compatibility
4916 #
4917 # v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
4918 # - this is only created for newly multi-init-ed
4919 # repositories. Similar in spirit to the
4920 # --use-separate-remotes option in git-clone (now default)
4921 # - we do not automatically migrate to this (following
4922 # the example set by core git)
4923 #
4924 # v5 layout: .rev_db.$UUID => .rev_map.$UUID
4925 # - newer, more-efficient format that uses 24-bytes per record
4926 # with no filler space.
4927 # - use xxd -c24 < .rev_map.$UUID to view and debug
4928 # - This is a one-way migration, repositories updated to the
4929 # new format will not be able to use old git-svn without
4930 # rebuilding the .rev_db. Rebuilding the rev_db is not
4931 # possible if noMetadata or useSvmProps are set; but should
4932 # be no problem for users that use the (sensible) defaults.
4933 use strict;
4934 use warnings;
4935 use Carp qw/croak/;
4936 use File::Path qw/mkpath/;
4937 use File::Basename qw/dirname basename/;
4938 use vars qw/$_minimize/;
4939
4940 sub migrate_from_v0 {
4941 my $git_dir = $ENV{GIT_DIR};
4942 return undef unless -d $git_dir;
4943 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4944 my $migrated = 0;
4945 while (<$fh>) {
4946 chomp;
4947 my ($id, $orig_ref) = ($_, $_);
4948 next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
4949 next unless -f "$git_dir/$id/info/url";
4950 my $new_ref = "refs/remotes/$id";
4951 if (::verify_ref("$new_ref^0")) {
4952 print STDERR "W: $orig_ref is probably an old ",
4953 "branch used by an ancient version of ",
4954 "git-svn.\n",
4955 "However, $new_ref also exists.\n",
4956 "We will not be able ",
4957 "to use this branch until this ",
4958 "ambiguity is resolved.\n";
4959 next;
4960 }
4961 print STDERR "Migrating from v0 layout...\n" if !$migrated;
4962 print STDERR "Renaming ref: $orig_ref => $new_ref\n";
4963 command_noisy('update-ref', $new_ref, $orig_ref);
4964 command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
4965 $migrated++;
4966 }
4967 command_close_pipe($fh, $ctx);
4968 print STDERR "Done migrating from v0 layout...\n" if $migrated;
4969 $migrated;
4970 }
4971
4972 sub migrate_from_v1 {
4973 my $git_dir = $ENV{GIT_DIR};
4974 my $migrated = 0;
4975 return $migrated unless -d $git_dir;
4976 my $svn_dir = "$git_dir/svn";
4977
4978 # just in case somebody used 'svn' as their $id at some point...
4979 return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
4980
4981 print STDERR "Migrating from a git-svn v1 layout...\n";
4982 mkpath([$svn_dir]);
4983 print STDERR "Data from a previous version of git-svn exists, but\n\t",
4984 "$svn_dir\n\t(required for this version ",
4985 "($::VERSION) of git-svn) does not exist.\n";
4986 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4987 while (<$fh>) {
4988 my $x = $_;
4989 next unless $x =~ s#^refs/remotes/##;
4990 chomp $x;
4991 next unless -f "$git_dir/$x/info/url";
4992 my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
4993 next unless $u;
4994 my $dn = dirname("$git_dir/svn/$x");
4995 mkpath([$dn]) unless -d $dn;
4996 if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
4997 mkpath(["$git_dir/svn/svn"]);
4998 print STDERR " - $git_dir/$x/info => ",
4999 "$git_dir/svn/$x/info\n";
5000 rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
5001 croak "$!: $x";
5002 # don't worry too much about these, they probably
5003 # don't exist with repos this old (save for index,
5004 # and we can easily regenerate that)
5005 foreach my $f (qw/unhandled.log index .rev_db/) {
5006 rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
5007 }
5008 } else {
5009 print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
5010 rename "$git_dir/$x", "$git_dir/svn/$x" or
5011 croak "$!: $x";
5012 }
5013 $migrated++;
5014 }
5015 command_close_pipe($fh, $ctx);
5016 print STDERR "Done migrating from a git-svn v1 layout\n";
5017 $migrated;
5018 }
5019
5020 sub read_old_urls {
5021 my ($l_map, $pfx, $path) = @_;
5022 my @dir;
5023 foreach (<$path/*>) {
5024 if (-r "$_/info/url") {
5025 $pfx .= '/' if $pfx && $pfx !~ m!/$!;
5026 my $ref_id = $pfx . basename $_;
5027 my $url = ::file_to_s("$_/info/url");
5028 $l_map->{$ref_id} = $url;
5029 } elsif (-d $_) {
5030 push @dir, $_;
5031 }
5032 }
5033 foreach (@dir) {
5034 my $x = $_;
5035 $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
5036 read_old_urls($l_map, $x, $_);
5037 }
5038 }
5039
5040 sub migrate_from_v2 {
5041 my @cfg = command(qw/config -l/);
5042 return if grep /^svn-remote\..+\.url=/, @cfg;
5043 my %l_map;
5044 read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
5045 my $migrated = 0;
5046
5047 foreach my $ref_id (sort keys %l_map) {
5048 eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
5049 if ($@) {
5050 Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
5051 }
5052 $migrated++;
5053 }
5054 $migrated;
5055 }
5056
5057 sub minimize_connections {
5058 my $r = Git::SVN::read_all_remotes();
5059 my $new_urls = {};
5060 my $root_repos = {};
5061 foreach my $repo_id (keys %$r) {
5062 my $url = $r->{$repo_id}->{url} or next;
5063 my $fetch = $r->{$repo_id}->{fetch} or next;
5064 my $ra = Git::SVN::Ra->new($url);
5065
5066 # skip existing cases where we already connect to the root
5067 if (($ra->{url} eq $ra->{repos_root}) ||
5068 ($ra->{repos_root} eq $repo_id)) {
5069 $root_repos->{$ra->{url}} = $repo_id;
5070 next;
5071 }
5072
5073 my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
5074 my $root_path = $ra->{url};
5075 $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
5076 foreach my $path (keys %$fetch) {
5077 my $ref_id = $fetch->{$path};
5078 my $gs = Git::SVN->new($ref_id, $repo_id, $path);
5079
5080 # make sure we can read when connecting to
5081 # a higher level of a repository
5082 my ($last_rev, undef) = $gs->last_rev_commit;
5083 if (!defined $last_rev) {
5084 $last_rev = eval {
5085 $root_ra->get_latest_revnum;
5086 };
5087 next if $@;
5088 }
5089 my $new = $root_path;
5090 $new .= length $path ? "/$path" : '';
5091 eval {
5092 $root_ra->get_log([$new], $last_rev, $last_rev,
5093 0, 0, 1, sub { });
5094 };
5095 next if $@;
5096 $new_urls->{$ra->{repos_root}}->{$new} =
5097 { ref_id => $ref_id,
5098 old_repo_id => $repo_id,
5099 old_path => $path };
5100 }
5101 }
5102
5103 my @emptied;
5104 foreach my $url (keys %$new_urls) {
5105 # see if we can re-use an existing [svn-remote "repo_id"]
5106 # instead of creating a(n ugly) new section:
5107 my $repo_id = $root_repos->{$url} || $url;
5108
5109 my $fetch = $new_urls->{$url};
5110 foreach my $path (keys %$fetch) {
5111 my $x = $fetch->{$path};
5112 Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
5113 my $pfx = "svn-remote.$x->{old_repo_id}";
5114
5115 my $old_fetch = quotemeta("$x->{old_path}:".
5116 "refs/remotes/$x->{ref_id}");
5117 command_noisy(qw/config --unset/,
5118 "$pfx.fetch", '^'. $old_fetch . '$');
5119 delete $r->{$x->{old_repo_id}}->
5120 {fetch}->{$x->{old_path}};
5121 if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
5122 command_noisy(qw/config --unset/,
5123 "$pfx.url");
5124 push @emptied, $x->{old_repo_id}
5125 }
5126 }
5127 }
5128 if (@emptied) {
5129 my $file = $ENV{GIT_CONFIG} || "$ENV{GIT_DIR}/config";
5130 print STDERR <<EOF;
5131 The following [svn-remote] sections in your config file ($file) are empty
5132 and can be safely removed:
5133 EOF
5134 print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
5135 }
5136 }
5137
5138 sub migration_check {
5139 migrate_from_v0();
5140 migrate_from_v1();
5141 migrate_from_v2();
5142 minimize_connections() if $_minimize;
5143 }
5144
5145 package Git::IndexInfo;
5146 use strict;
5147 use warnings;
5148 use Git qw/command_input_pipe command_close_pipe/;
5149
5150 sub new {
5151 my ($class) = @_;
5152 my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
5153 bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
5154 }
5155
5156 sub remove {
5157 my ($self, $path) = @_;
5158 if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
5159 return ++$self->{nr};
5160 }
5161 undef;
5162 }
5163
5164 sub update {
5165 my ($self, $mode, $hash, $path) = @_;
5166 if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
5167 return ++$self->{nr};
5168 }
5169 undef;
5170 }
5171
5172 sub DESTROY {
5173 my ($self) = @_;
5174 command_close_pipe($self->{gui}, $self->{ctx});
5175 }
5176
5177 package Git::SVN::GlobSpec;
5178 use strict;
5179 use warnings;
5180
5181 sub new {
5182 my ($class, $glob) = @_;
5183 my $re = $glob;
5184 $re =~ s!/+$!!g; # no need for trailing slashes
5185 $re =~ m!^([^*]*)(\*(?:/\*)*)([^*]*)$!;
5186 my $temp = $re;
5187 my ($left, $right) = ($1, $3);
5188 $re = $2;
5189 my $depth = $re =~ tr/*/*/;
5190 if ($depth != $temp =~ tr/*/*/) {
5191 die "Only one set of wildcard directories " .
5192 "(e.g. '*' or '*/*/*') is supported: '$glob'\n";
5193 }
5194 if ($depth == 0) {
5195 die "One '*' is needed for glob: '$glob'\n";
5196 }
5197 $re =~ s!\*!\[^/\]*!g;
5198 $re = quotemeta($left) . "($re)" . quotemeta($right);
5199 if (length $left && !($left =~ s!/+$!!g)) {
5200 die "Missing trailing '/' on left side of: '$glob' ($left)\n";
5201 }
5202 if (length $right && !($right =~ s!^/+!!g)) {
5203 die "Missing leading '/' on right side of: '$glob' ($right)\n";
5204 }
5205 my $left_re = qr/^\/\Q$left\E(\/|$)/;
5206 bless { left => $left, right => $right, left_regex => $left_re,
5207 regex => qr/$re/, glob => $glob, depth => $depth }, $class;
5208 }
5209
5210 sub full_path {
5211 my ($self, $path) = @_;
5212 return (length $self->{left} ? "$self->{left}/" : '') .
5213 $path . (length $self->{right} ? "/$self->{right}" : '');
5214 }
5215
5216 __END__
5217
5218 Data structures:
5219
5220
5221 $remotes = { # returned by read_all_remotes()
5222 'svn' => {
5223 # svn-remote.svn.url=https://svn.musicpd.org
5224 url => 'https://svn.musicpd.org',
5225 # svn-remote.svn.fetch=mpd/trunk:trunk
5226 fetch => {
5227 'mpd/trunk' => 'trunk',
5228 },
5229 # svn-remote.svn.tags=mpd/tags/*:tags/*
5230 tags => {
5231 path => {
5232 left => 'mpd/tags',
5233 right => '',
5234 regex => qr!mpd/tags/([^/]+)$!,
5235 glob => 'tags/*',
5236 },
5237 ref => {
5238 left => 'tags',
5239 right => '',
5240 regex => qr!tags/([^/]+)$!,
5241 glob => 'tags/*',
5242 },
5243 }
5244 }
5245 };
5246
5247 $log_entry hashref as returned by libsvn_log_entry()
5248 {
5249 log => 'whitespace-formatted log entry
5250 ', # trailing newline is preserved
5251 revision => '8', # integer
5252 date => '2004-02-24T17:01:44.108345Z', # commit date
5253 author => 'committer name'
5254 };
5255
5256
5257 # this is generated by generate_diff();
5258 @mods = array of diff-index line hashes, each element represents one line
5259 of diff-index output
5260
5261 diff-index line ($m hash)
5262 {
5263 mode_a => first column of diff-index output, no leading ':',
5264 mode_b => second column of diff-index output,
5265 sha1_b => sha1sum of the final blob,
5266 chg => change type [MCRADT],
5267 file_a => original file name of a file (iff chg is 'C' or 'R')
5268 file_b => new/current file name of a file (any chg)
5269 }
5270 ;
5271
5272 # retval of read_url_paths{,_all}();
5273 $l_map = {
5274 # repository root url
5275 'https://svn.musicpd.org' => {
5276 # repository path # GIT_SVN_ID
5277 'mpd/trunk' => 'trunk',
5278 'mpd/tags/0.11.5' => 'tags/0.11.5',
5279 },
5280 }
5281
5282 Notes:
5283 I don't trust the each() function on unless I created %hash myself
5284 because the internal iterator may not have started at base.
Gitweb development nest