?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Merge branch 'maint'
[git/jnareb-git.git] / git-svn.perl
blob6cc3157e229f9cc2e4604a872efad16ce27f0323
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
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 $Git::SVN::_follow_parent = 1;
70 my %remote_opts = ( 'username=s' => \$Git::SVN::Prompt::_username,
71 'config-dir=s' => \$Git::SVN::Ra::config_dir,
72 'no-auth-cache' => \$Git::SVN::Prompt::_no_auth_cache );
73 my %fc_opts = ( 'follow-parent|follow!' => \$Git::SVN::_follow_parent,
74 'authors-file|A=s' => \$_authors,
75 'repack:i' => \$Git::SVN::_repack,
76 'noMetadata' => \$Git::SVN::_no_metadata,
77 'useSvmProps' => \$Git::SVN::_use_svm_props,
78 'useSvnsyncProps' => \$Git::SVN::_use_svnsync_props,
79 'log-window-size=i' => \$Git::SVN::Ra::_log_window_size,
80 'no-checkout' => \$_no_checkout,
81 'quiet|q' => \$_q,
82 'repack-flags|repack-args|repack-opts=s' =>
83 \$Git::SVN::_repack_flags,
84 'use-log-author' => \$Git::SVN::_use_log_author,
85 %remote_opts );
86
87 my ($_trunk, $_tags, $_branches, $_stdlayout);
88 my %icv;
89 my %init_opts = ( 'template=s' => \$_template, 'shared:s' => \$_shared,
90 'trunk|T=s' => \$_trunk, 'tags|t=s' => \$_tags,
91 'branches|b=s' => \$_branches, 'prefix=s' => \$_prefix,
92 'stdlayout|s' => \$_stdlayout,
93 'minimize-url|m' => \$Git::SVN::_minimize_url,
94 'no-metadata' => sub { $icv{noMetadata} = 1 },
95 'use-svm-props' => sub { $icv{useSvmProps} = 1 },
96 'use-svnsync-props' => sub { $icv{useSvnsyncProps} = 1 },
97 'rewrite-root=s' => sub { $icv{rewriteRoot} = $_[1] },
98 %remote_opts );
99 my %cmt_opts = ( 'edit|e' => \$_edit,
100 'rmdir' => \$SVN::Git::Editor::_rmdir,
101 'find-copies-harder' => \$SVN::Git::Editor::_find_copies_harder,
102 'l=i' => \$SVN::Git::Editor::_rename_limit,
103 'copy-similarity|C=i'=> \$SVN::Git::Editor::_cp_similarity
104 );
105
106 my %cmd = (
107 fetch => [ \&cmd_fetch, "Download new revisions from SVN",
108 { 'revision|r=s' => \$_revision,
109 'fetch-all|all' => \$_fetch_all,
110 %fc_opts } ],
111 clone => [ \&cmd_clone, "Initialize and fetch revisions",
112 { 'revision|r=s' => \$_revision,
113 %fc_opts, %init_opts } ],
114 init => [ \&cmd_init, "Initialize a repo for tracking" .
115 " (requires URL argument)",
116 \%init_opts ],
117 'multi-init' => [ \&cmd_multi_init,
118 "Deprecated alias for ".
119 "'$0 init -T<trunk> -b<branches> -t<tags>'",
120 \%init_opts ],
121 dcommit => [ \&cmd_dcommit,
122 'Commit several diffs to merge with upstream',
123 { 'merge|m|M' => \$_merge,
124 'strategy|s=s' => \$_strategy,
125 'verbose|v' => \$_verbose,
126 'dry-run|n' => \$_dry_run,
127 'fetch-all|all' => \$_fetch_all,
128 'no-rebase' => \$_no_rebase,
129 %cmt_opts, %fc_opts } ],
130 'set-tree' => [ \&cmd_set_tree,
131 "Set an SVN repository to a git tree-ish",
132 { 'stdin|' => \$_stdin, %cmt_opts, %fc_opts, } ],
133 'create-ignore' => [ \&cmd_create_ignore,
134 'Create a .gitignore per svn:ignore',
135 { 'revision|r=i' => \$_revision
136 } ],
137 'propget' => [ \&cmd_propget,
138 'Print the value of a property on a file or directory',
139 { 'revision|r=i' => \$_revision } ],
140 'proplist' => [ \&cmd_proplist,
141 'List all properties of a file or directory',
142 { 'revision|r=i' => \$_revision } ],
143 'show-ignore' => [ \&cmd_show_ignore, "Show svn:ignore listings",
144 { 'revision|r=i' => \$_revision
145 } ],
146 'show-externals' => [ \&cmd_show_externals, "Show svn:externals listings",
147 { 'revision|r=i' => \$_revision
148 } ],
149 'multi-fetch' => [ \&cmd_multi_fetch,
150 "Deprecated alias for $0 fetch --all",
151 { 'revision|r=s' => \$_revision, %fc_opts } ],
152 'migrate' => [ sub { },
153 # no-op, we automatically run this anyways,
154 'Migrate configuration/metadata/layout from
155 previous versions of git-svn',
156 { 'minimize' => \$Git::SVN::Migration::_minimize,
157 %remote_opts } ],
158 'log' => [ \&Git::SVN::Log::cmd_show_log, 'Show commit logs',
159 { 'limit=i' => \$Git::SVN::Log::limit,
160 'revision|r=s' => \$_revision,
161 'verbose|v' => \$Git::SVN::Log::verbose,
162 'incremental' => \$Git::SVN::Log::incremental,
163 'oneline' => \$Git::SVN::Log::oneline,
164 'show-commit' => \$Git::SVN::Log::show_commit,
165 'non-recursive' => \$Git::SVN::Log::non_recursive,
166 'authors-file|A=s' => \$_authors,
167 'color' => \$Git::SVN::Log::color,
168 'pager=s' => \$Git::SVN::Log::pager
169 } ],
170 'find-rev' => [ \&cmd_find_rev, "Translate between SVN revision numbers and tree-ish",
171 {} ],
172 'rebase' => [ \&cmd_rebase, "Fetch and rebase your working directory",
173 { 'merge|m|M' => \$_merge,
174 'verbose|v' => \$_verbose,
175 'strategy|s=s' => \$_strategy,
176 'local|l' => \$_local,
177 'fetch-all|all' => \$_fetch_all,
178 %fc_opts } ],
179 'commit-diff' => [ \&cmd_commit_diff,
180 'Commit a diff between two trees',
181 { 'message|m=s' => \$_message,
182 'file|F=s' => \$_file,
183 'revision|r=s' => \$_revision,
184 %cmt_opts } ],
185 'info' => [ \&cmd_info,
186 "Show info about the latest SVN revision
187 on the current branch",
188 { 'url' => \$_url, } ],
189 );
190
191 my $cmd;
192 for (my $i = 0; $i < @ARGV; $i++) {
193 if (defined $cmd{$ARGV[$i]}) {
194 $cmd = $ARGV[$i];
195 splice @ARGV, $i, 1;
196 last;
197 }
198 };
199
200 # make sure we're always running at the top-level working directory
201 unless ($cmd && $cmd =~ /(?:clone|init|multi-init)$/) {
202 unless (-d $ENV{GIT_DIR}) {
203 if ($git_dir_user_set) {
204 die "GIT_DIR=$ENV{GIT_DIR} explicitly set, ",
205 "but it is not a directory\n";
206 }
207 my $git_dir = delete $ENV{GIT_DIR};
208 chomp(my $cdup = command_oneline(qw/rev-parse --show-cdup/));
209 unless (length $cdup) {
210 die "Already at toplevel, but $git_dir ",
211 "not found '$cdup'\n";
212 }
213 chdir $cdup or die "Unable to chdir up to '$cdup'\n";
214 unless (-d $git_dir) {
215 die "$git_dir still not found after going to ",
216 "'$cdup'\n";
217 }
218 $ENV{GIT_DIR} = $git_dir;
219 }
220 }
221
222 my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
223
224 read_repo_config(\%opts);
225 Getopt::Long::Configure('pass_through') if ($cmd && $cmd eq 'log');
226 my $rv = GetOptions(%opts, 'help|H|h' => \$_help, 'version|V' => \$_version,
227 'minimize-connections' => \$Git::SVN::Migration::_minimize,
228 'id|i=s' => \$Git::SVN::default_ref_id,
229 'svn-remote|remote|R=s' => sub {
230 $Git::SVN::no_reuse_existing = 1;
231 $Git::SVN::default_repo_id = $_[1] });
232 exit 1 if (!$rv && $cmd && $cmd ne 'log');
233
234 usage(0) if $_help;
235 version() if $_version;
236 usage(1) unless defined $cmd;
237 load_authors() if $_authors;
238
239 unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
240 Git::SVN::Migration::migration_check();
241 }
242 Git::SVN::init_vars();
243 eval {
244 Git::SVN::verify_remotes_sanity();
245 $cmd{$cmd}->[0]->(@ARGV);
246 };
247 fatal $@ if $@;
248 post_fetch_checkout();
249 exit 0;
250
251 ####################### primary functions ######################
252 sub usage {
253 my $exit = shift || 0;
254 my $fd = $exit ? \*STDERR : \*STDOUT;
255 print $fd <<"";
256 git-svn - bidirectional operations between a single Subversion tree and git
257 Usage: $0 <command> [options] [arguments]\n
258
259 print $fd "Available commands:\n" unless $cmd;
260
261 foreach (sort keys %cmd) {
262 next if $cmd && $cmd ne $_;
263 next if /^multi-/; # don't show deprecated commands
264 print $fd ' ',pack('A17',$_),$cmd{$_}->[1],"\n";
265 foreach (sort keys %{$cmd{$_}->[2]}) {
266 # mixed-case options are for .git/config only
267 next if /[A-Z]/ && /^[a-z]+$/i;
268 # prints out arguments as they should be passed:
269 my $x = s#[:=]s$## ? '<arg>' : s#[:=]i$## ? '<num>' : '';
270 print $fd ' ' x 21, join(', ', map { length $_ > 1 ?
271 "--$_" : "-$_" }
272 split /\|/,$_)," $x\n";
273 }
274 }
275 print $fd <<"";
276 \nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
277 arbitrary identifier if you're tracking multiple SVN branches/repositories in
278 one git repository and want to keep them separate. See git-svn(1) for more
279 information.
280
281 exit $exit;
282 }
283
284 sub version {
285 print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
286 exit 0;
287 }
288
289 sub do_git_init_db {
290 unless (-d $ENV{GIT_DIR}) {
291 my @init_db = ('init');
292 push @init_db, "--template=$_template" if defined $_template;
293 if (defined $_shared) {
294 if ($_shared =~ /[a-z]/) {
295 push @init_db, "--shared=$_shared";
296 } else {
297 push @init_db, "--shared";
298 }
299 }
300 command_noisy(@init_db);
301 }
302 my $set;
303 my $pfx = "svn-remote.$Git::SVN::default_repo_id";
304 foreach my $i (keys %icv) {
305 die "'$set' and '$i' cannot both be set\n" if $set;
306 next unless defined $icv{$i};
307 command_noisy('config', "$pfx.$i", $icv{$i});
308 $set = $i;
309 }
310 }
311
312 sub init_subdir {
313 my $repo_path = shift or return;
314 mkpath([$repo_path]) unless -d $repo_path;
315 chdir $repo_path or die "Couldn't chdir to $repo_path: $!\n";
316 $ENV{GIT_DIR} = '.git';
317 }
318
319 sub cmd_clone {
320 my ($url, $path) = @_;
321 if (!defined $path &&
322 (defined $_trunk || defined $_branches || defined $_tags ||
323 defined $_stdlayout) &&
324 $url !~ m#^[a-z\+]+://#) {
325 $path = $url;
326 }
327 $path = basename($url) if !defined $path || !length $path;
328 cmd_init($url, $path);
329 Git::SVN::fetch_all($Git::SVN::default_repo_id);
330 }
331
332 sub cmd_init {
333 if (defined $_stdlayout) {
334 $_trunk = 'trunk' if (!defined $_trunk);
335 $_tags = 'tags' if (!defined $_tags);
336 $_branches = 'branches' if (!defined $_branches);
337 }
338 if (defined $_trunk || defined $_branches || defined $_tags) {
339 return cmd_multi_init(@_);
340 }
341 my $url = shift or die "SVN repository location required ",
342 "as a command-line argument\n";
343 init_subdir(@_);
344 do_git_init_db();
345
346 Git::SVN->init($url);
347 }
348
349 sub cmd_fetch {
350 if (grep /^\d+=./, @_) {
351 die "'<rev>=<commit>' fetch arguments are ",
352 "no longer supported.\n";
353 }
354 my ($remote) = @_;
355 if (@_ > 1) {
356 die "Usage: $0 fetch [--all] [svn-remote]\n";
357 }
358 $remote ||= $Git::SVN::default_repo_id;
359 if ($_fetch_all) {
360 cmd_multi_fetch();
361 } else {
362 Git::SVN::fetch_all($remote, Git::SVN::read_all_remotes());
363 }
364 }
365
366 sub cmd_set_tree {
367 my (@commits) = @_;
368 if ($_stdin || !@commits) {
369 print "Reading from stdin...\n";
370 @commits = ();
371 while (<STDIN>) {
372 if (/\b($sha1_short)\b/o) {
373 unshift @commits, $1;
374 }
375 }
376 }
377 my @revs;
378 foreach my $c (@commits) {
379 my @tmp = command('rev-parse',$c);
380 if (scalar @tmp == 1) {
381 push @revs, $tmp[0];
382 } elsif (scalar @tmp > 1) {
383 push @revs, reverse(command('rev-list',@tmp));
384 } else {
385 fatal "Failed to rev-parse $c";
386 }
387 }
388 my $gs = Git::SVN->new;
389 my ($r_last, $cmt_last) = $gs->last_rev_commit;
390 $gs->fetch;
391 if (defined $gs->{last_rev} && $r_last != $gs->{last_rev}) {
392 fatal "There are new revisions that were fetched ",
393 "and need to be merged (or acknowledged) ",
394 "before committing.\nlast rev: $r_last\n",
395 " current: $gs->{last_rev}";
396 }
397 $gs->set_tree($_) foreach @revs;
398 print "Done committing ",scalar @revs," revisions to SVN\n";
399 unlink $gs->{index};
400 }
401
402 sub cmd_dcommit {
403 my $head = shift;
404 git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
405 'Cannot dcommit with a dirty index. Commit your changes first, '
406 . "or stash them with `git stash'.\n";
407 $head ||= 'HEAD';
408 my @refs;
409 my ($url, $rev, $uuid, $gs) = working_head_info($head, \@refs);
410 print "Committing to $url ...\n";
411 unless ($gs) {
412 die "Unable to determine upstream SVN information from ",
413 "$head history\n";
414 }
415 my $last_rev;
416 my ($linear_refs, $parents) = linearize_history($gs, \@refs);
417 if ($_no_rebase && scalar(@$linear_refs) > 1) {
418 warn "Attempting to commit more than one change while ",
419 "--no-rebase is enabled.\n",
420 "If these changes depend on each other, re-running ",
421 "without --no-rebase may be required."
422 }
423 while (1) {
424 my $d = shift @$linear_refs or last;
425 unless (defined $last_rev) {
426 (undef, $last_rev, undef) = cmt_metadata("$d~1");
427 unless (defined $last_rev) {
428 fatal "Unable to extract revision information ",
429 "from commit $d~1";
430 }
431 }
432 if ($_dry_run) {
433 print "diff-tree $d~1 $d\n";
434 } else {
435 my $cmt_rev;
436 my %ed_opts = ( r => $last_rev,
437 log => get_commit_entry($d)->{log},
438 ra => Git::SVN::Ra->new($gs->full_url),
439 config => SVN::Core::config_get_config(
440 $Git::SVN::Ra::config_dir
441 ),
442 tree_a => "$d~1",
443 tree_b => $d,
444 editor_cb => sub {
445 print "Committed r$_[0]\n";
446 $cmt_rev = $_[0];
447 },
448 svn_path => '');
449 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
450 print "No changes\n$d~1 == $d\n";
451 } elsif ($parents->{$d} && @{$parents->{$d}}) {
452 $gs->{inject_parents_dcommit}->{$cmt_rev} =
453 $parents->{$d};
454 }
455 $_fetch_all ? $gs->fetch_all : $gs->fetch;
456 $last_rev = $cmt_rev;
457 next if $_no_rebase;
458
459 # we always want to rebase against the current HEAD,
460 # not any head that was passed to us
461 my @diff = command('diff-tree', $d,
462 $gs->refname, '--');
463 my @finish;
464 if (@diff) {
465 @finish = rebase_cmd();
466 print STDERR "W: $d and ", $gs->refname,
467 " differ, using @finish:\n",
468 join("\n", @diff), "\n";
469 } else {
470 print "No changes between current HEAD and ",
471 $gs->refname,
472 "\nResetting to the latest ",
473 $gs->refname, "\n";
474 @finish = qw/reset --mixed/;
475 }
476 command_noisy(@finish, $gs->refname);
477 if (@diff) {
478 @refs = ();
479 my ($url_, $rev_, $uuid_, $gs_) =
480 working_head_info($head, \@refs);
481 my ($linear_refs_, $parents_) =
482 linearize_history($gs_, \@refs);
483 if (scalar(@$linear_refs) !=
484 scalar(@$linear_refs_)) {
485 fatal "# of revisions changed ",
486 "\nbefore:\n",
487 join("\n", @$linear_refs),
488 "\n\nafter:\n",
489 join("\n", @$linear_refs_), "\n",
490 'If you are attempting to commit ',
491 "merges, try running:\n\t",
492 'git rebase --interactive',
493 '--preserve-merges ',
494 $gs->refname,
495 "\nBefore dcommitting";
496 }
497 if ($url_ ne $url) {
498 fatal "URL mismatch after rebase: ",
499 "$url_ != $url";
500 }
501 if ($uuid_ ne $uuid) {
502 fatal "uuid mismatch after rebase: ",
503 "$uuid_ != $uuid";
504 }
505 # remap parents
506 my (%p, @l, $i);
507 for ($i = 0; $i < scalar @$linear_refs; $i++) {
508 my $new = $linear_refs_->[$i] or next;
509 $p{$new} =
510 $parents->{$linear_refs->[$i]};
511 push @l, $new;
512 }
513 $parents = \%p;
514 $linear_refs = \@l;
515 }
516 }
517 }
518 unlink $gs->{index};
519 }
520
521 sub cmd_find_rev {
522 my $revision_or_hash = shift;
523 my $result;
524 if ($revision_or_hash =~ /^r\d+$/) {
525 my $head = shift;
526 $head ||= 'HEAD';
527 my @refs;
528 my (undef, undef, undef, $gs) = working_head_info($head, \@refs);
529 unless ($gs) {
530 die "Unable to determine upstream SVN information from ",
531 "$head history\n";
532 }
533 my $desired_revision = substr($revision_or_hash, 1);
534 $result = $gs->rev_map_get($desired_revision);
535 } else {
536 my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
537 $result = $rev;
538 }
539 print "$result\n" if $result;
540 }
541
542 sub cmd_rebase {
543 command_noisy(qw/update-index --refresh/);
544 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
545 unless ($gs) {
546 die "Unable to determine upstream SVN information from ",
547 "working tree history\n";
548 }
549 if (command(qw/diff-index HEAD --/)) {
550 print STDERR "Cannot rebase with uncommited changes:\n";
551 command_noisy('status');
552 exit 1;
553 }
554 unless ($_local) {
555 # rebase will checkout for us, so no need to do it explicitly
556 $_no_checkout = 'true';
557 $_fetch_all ? $gs->fetch_all : $gs->fetch;
558 }
559 command_noisy(rebase_cmd(), $gs->refname);
560 }
561
562 sub cmd_show_ignore {
563 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
564 $gs ||= Git::SVN->new;
565 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
566 $gs->prop_walk($gs->{path}, $r, sub {
567 my ($gs, $path, $props) = @_;
568 print STDOUT "\n# $path\n";
569 my $s = $props->{'svn:ignore'} or return;
570 $s =~ s/[\r\n]+/\n/g;
571 chomp $s;
572 $s =~ s#^#$path#gm;
573 print STDOUT "$s\n";
574 });
575 }
576
577 sub cmd_show_externals {
578 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
579 $gs ||= Git::SVN->new;
580 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
581 $gs->prop_walk($gs->{path}, $r, sub {
582 my ($gs, $path, $props) = @_;
583 print STDOUT "\n# $path\n";
584 my $s = $props->{'svn:externals'} or return;
585 $s =~ s/[\r\n]+/\n/g;
586 chomp $s;
587 $s =~ s#^#$path#gm;
588 print STDOUT "$s\n";
589 });
590 }
591
592 sub cmd_create_ignore {
593 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
594 $gs ||= Git::SVN->new;
595 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
596 $gs->prop_walk($gs->{path}, $r, sub {
597 my ($gs, $path, $props) = @_;
598 # $path is of the form /path/to/dir/
599 my $ignore = '.' . $path . '.gitignore';
600 my $s = $props->{'svn:ignore'} or return;
601 open(GITIGNORE, '>', $ignore)
602 or fatal("Failed to open `$ignore' for writing: $!");
603 $s =~ s/[\r\n]+/\n/g;
604 chomp $s;
605 # Prefix all patterns so that the ignore doesn't apply
606 # to sub-directories.
607 $s =~ s#^#/#gm;
608 print GITIGNORE "$s\n";
609 close(GITIGNORE)
610 or fatal("Failed to close `$ignore': $!");
611 command_noisy('add', $ignore);
612 });
613 }
614
615 sub canonicalize_path {
616 my ($path) = @_;
617 my $dot_slash_added = 0;
618 if (substr($path, 0, 1) ne "/") {
619 $path = "./" . $path;
620 $dot_slash_added = 1;
621 }
622 # File::Spec->canonpath doesn't collapse x/../y into y (for a
623 # good reason), so let's do this manually.
624 $path =~ s#/+#/#g;
625 $path =~ s#/\.(?:/|$)#/#g;
626 $path =~ s#/[^/]+/\.\.##g;
627 $path =~ s#/$##g;
628 $path =~ s#^\./## if $dot_slash_added;
629 return $path;
630 }
631
632 # get_svnprops(PATH)
633 # ------------------
634 # Helper for cmd_propget and cmd_proplist below.
635 sub get_svnprops {
636 my $path = shift;
637 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
638 $gs ||= Git::SVN->new;
639
640 # prefix THE PATH by the sub-directory from which the user
641 # invoked us.
642 $path = $cmd_dir_prefix . $path;
643 fatal("No such file or directory: $path") unless -e $path;
644 my $is_dir = -d $path ? 1 : 0;
645 $path = $gs->{path} . '/' . $path;
646
647 # canonicalize the path (otherwise libsvn will abort or fail to
648 # find the file)
649 $path = canonicalize_path($path);
650
651 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
652 my $props;
653 if ($is_dir) {
654 (undef, undef, $props) = $gs->ra->get_dir($path, $r);
655 }
656 else {
657 (undef, $props) = $gs->ra->get_file($path, $r, undef);
658 }
659 return $props;
660 }
661
662 # cmd_propget (PROP, PATH)
663 # ------------------------
664 # Print the SVN property PROP for PATH.
665 sub cmd_propget {
666 my ($prop, $path) = @_;
667 $path = '.' if not defined $path;
668 usage(1) if not defined $prop;
669 my $props = get_svnprops($path);
670 if (not defined $props->{$prop}) {
671 fatal("`$path' does not have a `$prop' SVN property.");
672 }
673 print $props->{$prop} . "\n";
674 }
675
676 # cmd_proplist (PATH)
677 # -------------------
678 # Print the list of SVN properties for PATH.
679 sub cmd_proplist {
680 my $path = shift;
681 $path = '.' if not defined $path;
682 my $props = get_svnprops($path);
683 print "Properties on '$path':\n";
684 foreach (sort keys %{$props}) {
685 print " $_\n";
686 }
687 }
688
689 sub cmd_multi_init {
690 my $url = shift;
691 unless (defined $_trunk || defined $_branches || defined $_tags) {
692 usage(1);
693 }
694
695 # there are currently some bugs that prevent multi-init/multi-fetch
696 # setups from working well without this.
697 $Git::SVN::_minimize_url = 1;
698
699 $_prefix = '' unless defined $_prefix;
700 if (defined $url) {
701 $url =~ s#/+$##;
702 init_subdir(@_);
703 }
704 do_git_init_db();
705 if (defined $_trunk) {
706 my $trunk_ref = $_prefix . 'trunk';
707 # try both old-style and new-style lookups:
708 my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
709 unless ($gs_trunk) {
710 my ($trunk_url, $trunk_path) =
711 complete_svn_url($url, $_trunk);
712 $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
713 undef, $trunk_ref);
714 }
715 }
716 return unless defined $_branches || defined $_tags;
717 my $ra = $url ? Git::SVN::Ra->new($url) : undef;
718 complete_url_ls_init($ra, $_branches, '--branches/-b', $_prefix);
719 complete_url_ls_init($ra, $_tags, '--tags/-t', $_prefix . 'tags/');
720 }
721
722 sub cmd_multi_fetch {
723 my $remotes = Git::SVN::read_all_remotes();
724 foreach my $repo_id (sort keys %$remotes) {
725 if ($remotes->{$repo_id}->{url}) {
726 Git::SVN::fetch_all($repo_id, $remotes);
727 }
728 }
729 }
730
731 # this command is special because it requires no metadata
732 sub cmd_commit_diff {
733 my ($ta, $tb, $url) = @_;
734 my $usage = "Usage: $0 commit-diff -r<revision> ".
735 "<tree-ish> <tree-ish> [<URL>]";
736 fatal($usage) if (!defined $ta || !defined $tb);
737 my $svn_path;
738 if (!defined $url) {
739 my $gs = eval { Git::SVN->new };
740 if (!$gs) {
741 fatal("Needed URL or usable git-svn --id in ",
742 "the command-line\n", $usage);
743 }
744 $url = $gs->{url};
745 $svn_path = $gs->{path};
746 }
747 unless (defined $_revision) {
748 fatal("-r|--revision is a required argument\n", $usage);
749 }
750 if (defined $_message && defined $_file) {
751 fatal("Both --message/-m and --file/-F specified ",
752 "for the commit message.\n",
753 "I have no idea what you mean");
754 }
755 if (defined $_file) {
756 $_message = file_to_s($_file);
757 } else {
758 $_message ||= get_commit_entry($tb)->{log};
759 }
760 my $ra ||= Git::SVN::Ra->new($url);
761 $svn_path ||= $ra->{svn_path};
762 my $r = $_revision;
763 if ($r eq 'HEAD') {
764 $r = $ra->get_latest_revnum;
765 } elsif ($r !~ /^\d+$/) {
766 die "revision argument: $r not understood by git-svn\n";
767 }
768 my %ed_opts = ( r => $r,
769 log => $_message,
770 ra => $ra,
771 tree_a => $ta,
772 tree_b => $tb,
773 editor_cb => sub { print "Committed r$_[0]\n" },
774 svn_path => $svn_path );
775 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
776 print "No changes\n$ta == $tb\n";
777 }
778 }
779
780 sub cmd_info {
781 my $path = canonicalize_path(shift or ".");
782 unless (scalar(@_) == 0) {
783 die "Too many arguments specified\n";
784 }
785
786 my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
787
788 if (!$file_type && !$diff_status) {
789 print STDERR "$path: (Not a versioned resource)\n\n";
790 return;
791 }
792
793 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
794 unless ($gs) {
795 die "Unable to determine upstream SVN information from ",
796 "working tree history\n";
797 }
798 my $full_url = $url . ($path eq "." ? "" : "/$path");
799
800 if ($_url) {
801 print $full_url, "\n";
802 return;
803 }
804
805 my $result = "Path: $path\n";
806 $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
807 $result .= "URL: " . $full_url . "\n";
808
809 eval {
810 my $repos_root = $gs->repos_root;
811 Git::SVN::remove_username($repos_root);
812 $result .= "Repository Root: $repos_root\n";
813 };
814 if ($@) {
815 $result .= "Repository Root: (offline)\n";
816 }
817 $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A";
818 $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
819
820 $result .= "Node Kind: " .
821 ($file_type eq "dir" ? "directory" : "file") . "\n";
822
823 my $schedule = $diff_status eq "A"
824 ? "add"
825 : ($diff_status eq "D" ? "delete" : "normal");
826 $result .= "Schedule: $schedule\n";
827
828 if ($diff_status eq "A") {
829 print $result, "\n";
830 return;
831 }
832
833 my ($lc_author, $lc_rev, $lc_date_utc);
834 my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $path);
835 my $log = command_output_pipe(@args);
836 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
837 while (<$log>) {
838 if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
839 $lc_author = $1;
840 $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
841 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
842 (undef, $lc_rev, undef) = ::extract_metadata($1);
843 }
844 }
845 close $log;
846
847 Git::SVN::Log::set_local_timezone();
848
849 $result .= "Last Changed Author: $lc_author\n";
850 $result .= "Last Changed Rev: $lc_rev\n";
851 $result .= "Last Changed Date: " .
852 Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
853
854 if ($file_type ne "dir") {
855 my $text_last_updated_date =
856 ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
857 $result .=
858 "Text Last Updated: " .
859 Git::SVN::Log::format_svn_date($text_last_updated_date) .
860 "\n";
861 my $checksum;
862 if ($diff_status eq "D") {
863 my ($fh, $ctx) =
864 command_output_pipe(qw(cat-file blob), "HEAD:$path");
865 if ($file_type eq "link") {
866 my $file_name = <$fh>;
867 $checksum = md5sum("link $file_name");
868 } else {
869 $checksum = md5sum($fh);
870 }
871 command_close_pipe($fh, $ctx);
872 } elsif ($file_type eq "link") {
873 my $file_name =
874 command(qw(cat-file blob), "HEAD:$path");
875 $checksum =
876 md5sum("link " . $file_name);
877 } else {
878 open FILE, "<", $path or die $!;
879 $checksum = md5sum(\*FILE);
880 close FILE or die $!;
881 }
882 $result .= "Checksum: " . $checksum . "\n";
883 }
884
885 print $result, "\n";
886 }
887
888 ########################### utility functions #########################
889
890 sub rebase_cmd {
891 my @cmd = qw/rebase/;
892 push @cmd, '-v' if $_verbose;
893 push @cmd, qw/--merge/ if $_merge;
894 push @cmd, "--strategy=$_strategy" if $_strategy;
895 @cmd;
896 }
897
898 sub post_fetch_checkout {
899 return if $_no_checkout;
900 my $gs = $Git::SVN::_head or return;
901 return if verify_ref('refs/heads/master^0');
902
903 my $valid_head = verify_ref('HEAD^0');
904 command_noisy(qw(update-ref refs/heads/master), $gs->refname);
905 return if ($valid_head || !verify_ref('HEAD^0'));
906
907 return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
908 my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
909 return if -f $index;
910
911 return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
912 return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
913 command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
914 print STDERR "Checked out HEAD:\n ",
915 $gs->full_url, " r", $gs->last_rev, "\n";
916 }
917
918 sub complete_svn_url {
919 my ($url, $path) = @_;
920 $path =~ s#/+$##;
921 if ($path !~ m#^[a-z\+]+://#) {
922 if (!defined $url || $url !~ m#^[a-z\+]+://#) {
923 fatal("E: '$path' is not a complete URL ",
924 "and a separate URL is not specified");
925 }
926 return ($url, $path);
927 }
928 return ($path, '');
929 }
930
931 sub complete_url_ls_init {
932 my ($ra, $repo_path, $switch, $pfx) = @_;
933 unless ($repo_path) {
934 print STDERR "W: $switch not specified\n";
935 return;
936 }
937 $repo_path =~ s#/+$##;
938 if ($repo_path =~ m#^[a-z\+]+://#) {
939 $ra = Git::SVN::Ra->new($repo_path);
940 $repo_path = '';
941 } else {
942 $repo_path =~ s#^/+##;
943 unless ($ra) {
944 fatal("E: '$repo_path' is not a complete URL ",
945 "and a separate URL is not specified");
946 }
947 }
948 my $url = $ra->{url};
949 my $gs = Git::SVN->init($url, undef, undef, undef, 1);
950 my $k = "svn-remote.$gs->{repo_id}.url";
951 my $orig_url = eval { command_oneline(qw/config --get/, $k) };
952 if ($orig_url && ($orig_url ne $gs->{url})) {
953 die "$k already set: $orig_url\n",
954 "wanted to set to: $gs->{url}\n";
955 }
956 command_oneline('config', $k, $gs->{url}) unless $orig_url;
957 my $remote_path = "$ra->{svn_path}/$repo_path/*";
958 $remote_path =~ s#/+#/#g;
959 $remote_path =~ s#^/##g;
960 my ($n) = ($switch =~ /^--(\w+)/);
961 if (length $pfx && $pfx !~ m#/$#) {
962 die "--prefix='$pfx' must have a trailing slash '/'\n";
963 }
964 command_noisy('config', "svn-remote.$gs->{repo_id}.$n",
965 "$remote_path:refs/remotes/$pfx*");
966 }
967
968 sub verify_ref {
969 my ($ref) = @_;
970 eval { command_oneline([ 'rev-parse', '--verify', $ref ],
971 { STDERR => 0 }); };
972 }
973
974 sub get_tree_from_treeish {
975 my ($treeish) = @_;
976 # $treeish can be a symbolic ref, too:
977 my $type = command_oneline(qw/cat-file -t/, $treeish);
978 my $expected;
979 while ($type eq 'tag') {
980 ($treeish, $type) = command(qw/cat-file tag/, $treeish);
981 }
982 if ($type eq 'commit') {
983 $expected = (grep /^tree /, command(qw/cat-file commit/,
984 $treeish))[0];
985 ($expected) = ($expected =~ /^tree ($sha1)$/o);
986 die "Unable to get tree from $treeish\n" unless $expected;
987 } elsif ($type eq 'tree') {
988 $expected = $treeish;
989 } else {
990 die "$treeish is a $type, expected tree, tag or commit\n";
991 }
992 return $expected;
993 }
994
995 sub get_commit_entry {
996 my ($treeish) = shift;
997 my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
998 my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
999 my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
1000 open my $log_fh, '>', $commit_editmsg or croak $!;
1001
1002 my $type = command_oneline(qw/cat-file -t/, $treeish);
1003 if ($type eq 'commit' || $type eq 'tag') {
1004 my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1005 $type, $treeish);
1006 my $in_msg = 0;
1007 while (<$msg_fh>) {
1008 if (!$in_msg) {
1009 $in_msg = 1 if (/^\s*$/);
1010 } elsif (/^git-svn-id: /) {
1011 # skip this for now, we regenerate the
1012 # correct one on re-fetch anyways
1013 # TODO: set *:merge properties or like...
1014 } else {
1015 print $log_fh $_ or croak $!;
1016 }
1017 }
1018 command_close_pipe($msg_fh, $ctx);
1019 }
1020 close $log_fh or croak $!;
1021
1022 if ($_edit || ($type eq 'tree')) {
1023 my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
1024 # TODO: strip out spaces, comments, like git-commit.sh
1025 system($editor, $commit_editmsg);
1026 }
1027 rename $commit_editmsg, $commit_msg or croak $!;
1028 open $log_fh, '<', $commit_msg or croak $!;
1029 { local $/; chomp($log_entry{log} = <$log_fh>); }
1030 close $log_fh or croak $!;
1031 unlink $commit_msg;
1032 \%log_entry;
1033 }
1034
1035 sub s_to_file {
1036 my ($str, $file, $mode) = @_;
1037 open my $fd,'>',$file or croak $!;
1038 print $fd $str,"\n" or croak $!;
1039 close $fd or croak $!;
1040 chmod ($mode &~ umask, $file) if (defined $mode);
1041 }
1042
1043 sub file_to_s {
1044 my $file = shift;
1045 open my $fd,'<',$file or croak "$!: file: $file\n";
1046 local $/;
1047 my $ret = <$fd>;
1048 close $fd or croak $!;
1049 $ret =~ s/\s*$//s;
1050 return $ret;
1051 }
1052
1053 # '<svn username> = real-name <email address>' mapping based on git-svnimport:
1054 sub load_authors {
1055 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1056 my $log = $cmd eq 'log';
1057 while (<$authors>) {
1058 chomp;
1059 next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1060 my ($user, $name, $email) = ($1, $2, $3);
1061 if ($log) {
1062 $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1063 } else {
1064 $users{$user} = [$name, $email];
1065 }
1066 }
1067 close $authors or croak $!;
1068 }
1069
1070 # convert GetOpt::Long specs for use by git-config
1071 sub read_repo_config {
1072 return unless -d $ENV{GIT_DIR};
1073 my $opts = shift;
1074 my @config_only;
1075 foreach my $o (keys %$opts) {
1076 # if we have mixedCase and a long option-only, then
1077 # it's a config-only variable that we don't need for
1078 # the command-line.
1079 push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1080 my $v = $opts->{$o};
1081 my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1082 $key =~ s/-//g;
1083 my $arg = 'git-config';
1084 $arg .= ' --int' if ($o =~ /[:=]i$/);
1085 $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1086 if (ref $v eq 'ARRAY') {
1087 chomp(my @tmp = `$arg --get-all svn.$key`);
1088 @$v = @tmp if @tmp;
1089 } else {
1090 chomp(my $tmp = `$arg --get svn.$key`);
1091 if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1092 $$v = $tmp;
1093 }
1094 }
1095 }
1096 delete @$opts{@config_only} if @config_only;
1097 }
1098
1099 sub extract_metadata {
1100 my $id = shift or return (undef, undef, undef);
1101 my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1102 \s([a-f\d\-]+)$/x);
1103 if (!defined $rev || !$uuid || !$url) {
1104 # some of the original repositories I made had
1105 # identifiers like this:
1106 ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/);
1107 }
1108 return ($url, $rev, $uuid);
1109 }
1110
1111 sub cmt_metadata {
1112 return extract_metadata((grep(/^git-svn-id: /,
1113 command(qw/cat-file commit/, shift)))[-1]);
1114 }
1115
1116 sub working_head_info {
1117 my ($head, $refs) = @_;
1118 my @args = ('log', '--no-color', '--first-parent');
1119 my ($fh, $ctx) = command_output_pipe(@args, $head);
1120 my $hash;
1121 my %max;
1122 while (<$fh>) {
1123 if ( m{^commit ($::sha1)$} ) {
1124 unshift @$refs, $hash if $hash and $refs;
1125 $hash = $1;
1126 next;
1127 }
1128 next unless s{^\s*(git-svn-id:)}{$1};
1129 my ($url, $rev, $uuid) = extract_metadata($_);
1130 if (defined $url && defined $rev) {
1131 next if $max{$url} and $max{$url} < $rev;
1132 if (my $gs = Git::SVN->find_by_url($url)) {
1133 my $c = $gs->rev_map_get($rev);
1134 if ($c && $c eq $hash) {
1135 close $fh; # break the pipe
1136 return ($url, $rev, $uuid, $gs);
1137 } else {
1138 $max{$url} ||= $gs->rev_map_max;
1139 }
1140 }
1141 }
1142 }
1143 command_close_pipe($fh, $ctx);
1144 (undef, undef, undef, undef);
1145 }
1146
1147 sub read_commit_parents {
1148 my ($parents, $c) = @_;
1149 chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1150 $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1151 @{$parents->{$c}} = split(/ /, $p);
1152 }
1153
1154 sub linearize_history {
1155 my ($gs, $refs) = @_;
1156 my %parents;
1157 foreach my $c (@$refs) {
1158 read_commit_parents(\%parents, $c);
1159 }
1160
1161 my @linear_refs;
1162 my %skip = ();
1163 my $last_svn_commit = $gs->last_commit;
1164 foreach my $c (reverse @$refs) {
1165 next if $c eq $last_svn_commit;
1166 last if $skip{$c};
1167
1168 unshift @linear_refs, $c;
1169 $skip{$c} = 1;
1170
1171 # we only want the first parent to diff against for linear
1172 # history, we save the rest to inject when we finalize the
1173 # svn commit
1174 my $fp_a = verify_ref("$c~1");
1175 my $fp_b = shift @{$parents{$c}} if $parents{$c};
1176 if (!$fp_a || !$fp_b) {
1177 die "Commit $c\n",
1178 "has no parent commit, and therefore ",
1179 "nothing to diff against.\n",
1180 "You should be working from a repository ",
1181 "originally created by git-svn\n";
1182 }
1183 if ($fp_a ne $fp_b) {
1184 die "$c~1 = $fp_a, however parsing commit $c ",
1185 "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1186 }
1187
1188 foreach my $p (@{$parents{$c}}) {
1189 $skip{$p} = 1;
1190 }
1191 }
1192 (\@linear_refs, \%parents);
1193 }
1194
1195 sub find_file_type_and_diff_status {
1196 my ($path) = @_;
1197 return ('dir', '') if $path eq '.';
1198
1199 my $diff_output =
1200 command_oneline(qw(diff --cached --name-status --), $path) || "";
1201 my $diff_status = (split(' ', $diff_output))[0] || "";
1202
1203 my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1204
1205 return (undef, undef) if !$diff_status && !$ls_tree;
1206
1207 if ($diff_status eq "A") {
1208 return ("link", $diff_status) if -l $path;
1209 return ("dir", $diff_status) if -d $path;
1210 return ("file", $diff_status);
1211 }
1212
1213 my $mode = (split(' ', $ls_tree))[0] || "";
1214
1215 return ("link", $diff_status) if $mode eq "120000";
1216 return ("dir", $diff_status) if $mode eq "040000";
1217 return ("file", $diff_status);
1218 }
1219
1220 sub md5sum {
1221 my $arg = shift;
1222 my $ref = ref $arg;
1223 my $md5 = Digest::MD5->new();
1224 if ($ref eq 'GLOB' || $ref eq 'IO::File') {
1225 $md5->addfile($arg) or croak $!;
1226 } elsif ($ref eq 'SCALAR') {
1227 $md5->add($$arg) or croak $!;
1228 } elsif (!$ref) {
1229 $md5->add($arg) or croak $!;
1230 } else {
1231 ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
1232 }
1233 return $md5->hexdigest();
1234 }
1235
1236 package Git::SVN;
1237 use strict;
1238 use warnings;
1239 use Fcntl qw/:DEFAULT :seek/;
1240 use constant rev_map_fmt => 'NH40';
1241 use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1242 $_repack $_repack_flags $_use_svm_props $_head
1243 $_use_svnsync_props $no_reuse_existing $_minimize_url
1244 $_use_log_author/;
1245 use Carp qw/croak/;
1246 use File::Path qw/mkpath/;
1247 use File::Copy qw/copy/;
1248 use IPC::Open3;
1249
1250 my ($_gc_nr, $_gc_period);
1251
1252 # properties that we do not log:
1253 my %SKIP_PROP;
1254 BEGIN {
1255 %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1256 svn:special svn:executable
1257 svn:entry:committed-rev
1258 svn:entry:last-author
1259 svn:entry:uuid
1260 svn:entry:committed-date/;
1261
1262 # some options are read globally, but can be overridden locally
1263 # per [svn-remote "..."] section. Command-line options will *NOT*
1264 # override options set in an [svn-remote "..."] section
1265 no strict 'refs';
1266 for my $option (qw/follow_parent no_metadata use_svm_props
1267 use_svnsync_props/) {
1268 my $key = $option;
1269 $key =~ tr/_//d;
1270 my $prop = "-$option";
1271 *$option = sub {
1272 my ($self) = @_;
1273 return $self->{$prop} if exists $self->{$prop};
1274 my $k = "svn-remote.$self->{repo_id}.$key";
1275 eval { command_oneline(qw/config --get/, $k) };
1276 if ($@) {
1277 $self->{$prop} = ${"Git::SVN::_$option"};
1278 } else {
1279 my $v = command_oneline(qw/config --bool/,$k);
1280 $self->{$prop} = $v eq 'false' ? 0 : 1;
1281 }
1282 return $self->{$prop};
1283 }
1284 }
1285 }
1286
1287 my (%LOCKFILES, %INDEX_FILES);
1288 END {
1289 unlink keys %LOCKFILES if %LOCKFILES;
1290 unlink keys %INDEX_FILES if %INDEX_FILES;
1291 }
1292
1293 sub resolve_local_globs {
1294 my ($url, $fetch, $glob_spec) = @_;
1295 return unless defined $glob_spec;
1296 my $ref = $glob_spec->{ref};
1297 my $path = $glob_spec->{path};
1298 foreach (command(qw#for-each-ref --format=%(refname) refs/remotes#)) {
1299 next unless m#^refs/remotes/$ref->{regex}$#;
1300 my $p = $1;
1301 my $pathname = desanitize_refname($path->full_path($p));
1302 my $refname = desanitize_refname($ref->full_path($p));
1303 if (my $existing = $fetch->{$pathname}) {
1304 if ($existing ne $refname) {
1305 die "Refspec conflict:\n",
1306 "existing: refs/remotes/$existing\n",
1307 " globbed: refs/remotes/$refname\n";
1308 }
1309 my $u = (::cmt_metadata("refs/remotes/$refname"))[0];
1310 $u =~ s!^\Q$url\E(/|$)!! or die
1311 "refs/remotes/$refname: '$url' not found in '$u'\n";
1312 if ($pathname ne $u) {
1313 warn "W: Refspec glob conflict ",
1314 "(ref: refs/remotes/$refname):\n",
1315 "expected path: $pathname\n",
1316 " real path: $u\n",
1317 "Continuing ahead with $u\n";
1318 next;
1319 }
1320 } else {
1321 $fetch->{$pathname} = $refname;
1322 }
1323 }
1324 }
1325
1326 sub parse_revision_argument {
1327 my ($base, $head) = @_;
1328 if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1329 return ($base, $head);
1330 }
1331 return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1332 return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1333 return ($head, $head) if ($::_revision eq 'HEAD');
1334 return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1335 return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1336 die "revision argument: $::_revision not understood by git-svn\n";
1337 }
1338
1339 sub fetch_all {
1340 my ($repo_id, $remotes) = @_;
1341 if (ref $repo_id) {
1342 my $gs = $repo_id;
1343 $repo_id = undef;
1344 $repo_id = $gs->{repo_id};
1345 }
1346 $remotes ||= read_all_remotes();
1347 my $remote = $remotes->{$repo_id} or
1348 die "[svn-remote \"$repo_id\"] unknown\n";
1349 my $fetch = $remote->{fetch};
1350 my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1351 my (@gs, @globs);
1352 my $ra = Git::SVN::Ra->new($url);
1353 my $uuid = $ra->get_uuid;
1354 my $head = $ra->get_latest_revnum;
1355 my $base = defined $fetch ? $head : 0;
1356
1357 # read the max revs for wildcard expansion (branches/*, tags/*)
1358 foreach my $t (qw/branches tags/) {
1359 defined $remote->{$t} or next;
1360 push @globs, $remote->{$t};
1361 my $max_rev = eval { tmp_config(qw/--int --get/,
1362 "svn-remote.$repo_id.${t}-maxRev") };
1363 if (defined $max_rev && ($max_rev < $base)) {
1364 $base = $max_rev;
1365 } elsif (!defined $max_rev) {
1366 $base = 0;
1367 }
1368 }
1369
1370 if ($fetch) {
1371 foreach my $p (sort keys %$fetch) {
1372 my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1373 my $lr = $gs->rev_map_max;
1374 if (defined $lr) {
1375 $base = $lr if ($lr < $base);
1376 }
1377 push @gs, $gs;
1378 }
1379 }
1380
1381 ($base, $head) = parse_revision_argument($base, $head);
1382 $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1383 }
1384
1385 sub read_all_remotes {
1386 my $r = {};
1387 foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1388 if (m!^(.+)\.fetch=\s*(.*)\s*:\s*refs/remotes/(.+)\s*$!) {
1389 my ($remote, $local_ref, $remote_ref) = ($1, $2, $3);
1390 $local_ref =~ s{^/}{};
1391 $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1392 } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1393 $r->{$1}->{url} = $2;
1394 } elsif (m!^(.+)\.(branches|tags)=
1395 (.*):refs/remotes/(.+)\s*$/!x) {
1396 my ($p, $g) = ($3, $4);
1397 my $rs = $r->{$1}->{$2} = {
1398 t => $2,
1399 remote => $1,
1400 path => Git::SVN::GlobSpec->new($p),
1401 ref => Git::SVN::GlobSpec->new($g) };
1402 if (length($rs->{ref}->{right}) != 0) {
1403 die "The '*' glob character must be the last ",
1404 "character of '$g'\n";
1405 }
1406 }
1407 }
1408 $r;
1409 }
1410
1411 sub init_vars {
1412 $_gc_nr = $_gc_period = 1000;
1413 if (defined $_repack || defined $_repack_flags) {
1414 warn "Repack options are obsolete; they have no effect.\n";
1415 }
1416 }
1417
1418 sub verify_remotes_sanity {
1419 return unless -d $ENV{GIT_DIR};
1420 my %seen;
1421 foreach (command(qw/config -l/)) {
1422 if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1423 if ($seen{$1}) {
1424 die "Remote ref refs/remote/$1 is tracked by",
1425 "\n \"$_\"\nand\n \"$seen{$1}\"\n",
1426 "Please resolve this ambiguity in ",
1427 "your git configuration file before ",
1428 "continuing\n";
1429 }
1430 $seen{$1} = $_;
1431 }
1432 }
1433 }
1434
1435 # we allow more chars than remotes2config.sh...
1436 sub sanitize_remote_name {
1437 my ($name) = @_;
1438 $name =~ tr{A-Za-z0-9:,/+-}{.}c;
1439 $name;
1440 }
1441
1442 sub find_existing_remote {
1443 my ($url, $remotes) = @_;
1444 return undef if $no_reuse_existing;
1445 my $existing;
1446 foreach my $repo_id (keys %$remotes) {
1447 my $u = $remotes->{$repo_id}->{url} or next;
1448 next if $u ne $url;
1449 $existing = $repo_id;
1450 last;
1451 }
1452 $existing;
1453 }
1454
1455 sub init_remote_config {
1456 my ($self, $url, $no_write) = @_;
1457 $url =~ s!/+$!!; # strip trailing slash
1458 my $r = read_all_remotes();
1459 my $existing = find_existing_remote($url, $r);
1460 if ($existing) {
1461 unless ($no_write) {
1462 print STDERR "Using existing ",
1463 "[svn-remote \"$existing\"]\n";
1464 }
1465 $self->{repo_id} = $existing;
1466 } elsif ($_minimize_url) {
1467 my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1468 $existing = find_existing_remote($min_url, $r);
1469 if ($existing) {
1470 unless ($no_write) {
1471 print STDERR "Using existing ",
1472 "[svn-remote \"$existing\"]\n";
1473 }
1474 $self->{repo_id} = $existing;
1475 }
1476 if ($min_url ne $url) {
1477 unless ($no_write) {
1478 print STDERR "Using higher level of URL: ",
1479 "$url => $min_url\n";
1480 }
1481 my $old_path = $self->{path};
1482 $self->{path} = $url;
1483 $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1484 if (length $old_path) {
1485 $self->{path} .= "/$old_path";
1486 }
1487 $url = $min_url;
1488 }
1489 }
1490 my $orig_url;
1491 if (!$existing) {
1492 # verify that we aren't overwriting anything:
1493 $orig_url = eval {
1494 command_oneline('config', '--get',
1495 "svn-remote.$self->{repo_id}.url")
1496 };
1497 if ($orig_url && ($orig_url ne $url)) {
1498 die "svn-remote.$self->{repo_id}.url already set: ",
1499 "$orig_url\nwanted to set to: $url\n";
1500 }
1501 }
1502 my ($xrepo_id, $xpath) = find_ref($self->refname);
1503 if (defined $xpath) {
1504 die "svn-remote.$xrepo_id.fetch already set to track ",
1505 "$xpath:refs/remotes/", $self->refname, "\n";
1506 }
1507 unless ($no_write) {
1508 command_noisy('config',
1509 "svn-remote.$self->{repo_id}.url", $url);
1510 $self->{path} =~ s{^/}{};
1511 command_noisy('config', '--add',
1512 "svn-remote.$self->{repo_id}.fetch",
1513 "$self->{path}:".$self->refname);
1514 }
1515 $self->{url} = $url;
1516 }
1517
1518 sub find_by_url { # repos_root and, path are optional
1519 my ($class, $full_url, $repos_root, $path) = @_;
1520
1521 return undef unless defined $full_url;
1522 remove_username($full_url);
1523 remove_username($repos_root) if defined $repos_root;
1524 my $remotes = read_all_remotes();
1525 if (defined $full_url && defined $repos_root && !defined $path) {
1526 $path = $full_url;
1527 $path =~ s#^\Q$repos_root\E(?:/|$)##;
1528 }
1529 foreach my $repo_id (keys %$remotes) {
1530 my $u = $remotes->{$repo_id}->{url} or next;
1531 remove_username($u);
1532 next if defined $repos_root && $repos_root ne $u;
1533
1534 my $fetch = $remotes->{$repo_id}->{fetch} || {};
1535 foreach (qw/branches tags/) {
1536 resolve_local_globs($u, $fetch,
1537 $remotes->{$repo_id}->{$_});
1538 }
1539 my $p = $path;
1540 unless (defined $p) {
1541 $p = $full_url;
1542 $p =~ s#^\Q$u\E(?:/|$)## or next;
1543 }
1544 foreach my $f (keys %$fetch) {
1545 next if $f ne $p;
1546 return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1547 }
1548 }
1549 undef;
1550 }
1551
1552 sub init {
1553 my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1554 my $self = _new($class, $repo_id, $ref_id, $path);
1555 if (defined $url) {
1556 $self->init_remote_config($url, $no_write);
1557 }
1558 $self;
1559 }
1560
1561 sub find_ref {
1562 my ($ref_id) = @_;
1563 foreach (command(qw/config -l/)) {
1564 next unless m!^svn-remote\.(.+)\.fetch=
1565 \s*(.*)\s*:\s*refs/remotes/(.+)\s*$!x;
1566 my ($repo_id, $path, $ref) = ($1, $2, $3);
1567 if ($ref eq $ref_id) {
1568 $path = '' if ($path =~ m#^\./?#);
1569 return ($repo_id, $path);
1570 }
1571 }
1572 (undef, undef, undef);
1573 }
1574
1575 sub new {
1576 my ($class, $ref_id, $repo_id, $path) = @_;
1577 if (defined $ref_id && !defined $repo_id && !defined $path) {
1578 ($repo_id, $path) = find_ref($ref_id);
1579 if (!defined $repo_id) {
1580 die "Could not find a \"svn-remote.*.fetch\" key ",
1581 "in the repository configuration matching: ",
1582 "refs/remotes/$ref_id\n";
1583 }
1584 }
1585 my $self = _new($class, $repo_id, $ref_id, $path);
1586 if (!defined $self->{path} || !length $self->{path}) {
1587 my $fetch = command_oneline('config', '--get',
1588 "svn-remote.$repo_id.fetch",
1589 ":refs/remotes/$ref_id\$") or
1590 die "Failed to read \"svn-remote.$repo_id.fetch\" ",
1591 "\":refs/remotes/$ref_id\$\" in config\n";
1592 ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
1593 }
1594 $self->{url} = command_oneline('config', '--get',
1595 "svn-remote.$repo_id.url") or
1596 die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
1597 $self->rebuild;
1598 $self;
1599 }
1600
1601 sub refname {
1602 my ($refname) = "refs/remotes/$_[0]->{ref_id}" ;
1603
1604 # It cannot end with a slash /, we'll throw up on this because
1605 # SVN can't have directories with a slash in their name, either:
1606 if ($refname =~ m{/$}) {
1607 die "ref: '$refname' ends with a trailing slash, this is ",
1608 "not permitted by git nor Subversion\n";
1609 }
1610
1611 # It cannot have ASCII control character space, tilde ~, caret ^,
1612 # colon :, question-mark ?, asterisk *, space, or open bracket [
1613 # anywhere.
1614 #
1615 # Additionally, % must be escaped because it is used for escaping
1616 # and we want our escaped refname to be reversible
1617 $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
1618
1619 # no slash-separated component can begin with a dot .
1620 # /.* becomes /%2E*
1621 $refname =~ s{/\.}{/%2E}g;
1622
1623 # It cannot have two consecutive dots .. anywhere
1624 # .. becomes %2E%2E
1625 $refname =~ s{\.\.}{%2E%2E}g;
1626
1627 return $refname;
1628 }
1629
1630 sub desanitize_refname {
1631 my ($refname) = @_;
1632 $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
1633 return $refname;
1634 }
1635
1636 sub svm_uuid {
1637 my ($self) = @_;
1638 return $self->{svm}->{uuid} if $self->svm;
1639 $self->ra;
1640 unless ($self->{svm}) {
1641 die "SVM UUID not cached, and reading remotely failed\n";
1642 }
1643 $self->{svm}->{uuid};
1644 }
1645
1646 sub svm {
1647 my ($self) = @_;
1648 return $self->{svm} if $self->{svm};
1649 my $svm;
1650 # see if we have it in our config, first:
1651 eval {
1652 my $section = "svn-remote.$self->{repo_id}";
1653 $svm = {
1654 source => tmp_config('--get', "$section.svm-source"),
1655 uuid => tmp_config('--get', "$section.svm-uuid"),
1656 replace => tmp_config('--get', "$section.svm-replace"),
1657 }
1658 };
1659 if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
1660 $self->{svm} = $svm;
1661 }
1662 $self->{svm};
1663 }
1664
1665 sub _set_svm_vars {
1666 my ($self, $ra) = @_;
1667 return $ra if $self->svm;
1668
1669 my @err = ( "useSvmProps set, but failed to read SVM properties\n",
1670 "(svm:source, svm:uuid) ",
1671 "from the following URLs:\n" );
1672 sub read_svm_props {
1673 my ($self, $ra, $path, $r) = @_;
1674 my $props = ($ra->get_dir($path, $r))[2];
1675 my $src = $props->{'svm:source'};
1676 my $uuid = $props->{'svm:uuid'};
1677 return undef if (!$src || !$uuid);
1678
1679 chomp($src, $uuid);
1680
1681 $uuid =~ m{^[0-9a-f\-]{30,}$}
1682 or die "doesn't look right - svm:uuid is '$uuid'\n";
1683
1684 # the '!' is used to mark the repos_root!/relative/path
1685 $src =~ s{/?!/?}{/};
1686 $src =~ s{/+$}{}; # no trailing slashes please
1687 # username is of no interest
1688 $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
1689
1690 my $replace = $ra->{url};
1691 $replace .= "/$path" if length $path;
1692
1693 my $section = "svn-remote.$self->{repo_id}";
1694 tmp_config("$section.svm-source", $src);
1695 tmp_config("$section.svm-replace", $replace);
1696 tmp_config("$section.svm-uuid", $uuid);
1697 $self->{svm} = {
1698 source => $src,
1699 uuid => $uuid,
1700 replace => $replace
1701 };
1702 }
1703
1704 my $r = $ra->get_latest_revnum;
1705 my $path = $self->{path};
1706 my %tried;
1707 while (length $path) {
1708 unless ($tried{"$self->{url}/$path"}) {
1709 return $ra if $self->read_svm_props($ra, $path, $r);
1710 $tried{"$self->{url}/$path"} = 1;
1711 }
1712 $path =~ s#/?[^/]+$##;
1713 }
1714 die "Path: '$path' should be ''\n" if $path ne '';
1715 return $ra if $self->read_svm_props($ra, $path, $r);
1716 $tried{"$self->{url}/$path"} = 1;
1717
1718 if ($ra->{repos_root} eq $self->{url}) {
1719 die @err, (map { " $_\n" } keys %tried), "\n";
1720 }
1721
1722 # nope, make sure we're connected to the repository root:
1723 my $ok;
1724 my @tried_b;
1725 $path = $ra->{svn_path};
1726 $ra = Git::SVN::Ra->new($ra->{repos_root});
1727 while (length $path) {
1728 unless ($tried{"$ra->{url}/$path"}) {
1729 $ok = $self->read_svm_props($ra, $path, $r);
1730 last if $ok;
1731 $tried{"$ra->{url}/$path"} = 1;
1732 }
1733 $path =~ s#/?[^/]+$##;
1734 }
1735 die "Path: '$path' should be ''\n" if $path ne '';
1736 $ok ||= $self->read_svm_props($ra, $path, $r);
1737 $tried{"$ra->{url}/$path"} = 1;
1738 if (!$ok) {
1739 die @err, (map { " $_\n" } keys %tried), "\n";
1740 }
1741 Git::SVN::Ra->new($self->{url});
1742 }
1743
1744 sub svnsync {
1745 my ($self) = @_;
1746 return $self->{svnsync} if $self->{svnsync};
1747
1748 if ($self->no_metadata) {
1749 die "Can't have both 'noMetadata' and ",
1750 "'useSvnsyncProps' options set!\n";
1751 }
1752 if ($self->rewrite_root) {
1753 die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
1754 "options set!\n";
1755 }
1756
1757 my $svnsync;
1758 # see if we have it in our config, first:
1759 eval {
1760 my $section = "svn-remote.$self->{repo_id}";
1761
1762 my $url = tmp_config('--get', "$section.svnsync-url");
1763 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1764 die "doesn't look right - svn:sync-from-url is '$url'\n";
1765
1766 my $uuid = tmp_config('--get', "$section.svnsync-uuid");
1767 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1768 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1769
1770 $svnsync = { url => $url, uuid => $uuid }
1771 };
1772 if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
1773 return $self->{svnsync} = $svnsync;
1774 }
1775
1776 my $err = "useSvnsyncProps set, but failed to read " .
1777 "svnsync property: svn:sync-from-";
1778 my $rp = $self->ra->rev_proplist(0);
1779
1780 my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
1781 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1782 die "doesn't look right - svn:sync-from-url is '$url'\n";
1783
1784 my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
1785 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1786 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1787
1788 my $section = "svn-remote.$self->{repo_id}";
1789 tmp_config('--add', "$section.svnsync-uuid", $uuid);
1790 tmp_config('--add', "$section.svnsync-url", $url);
1791 return $self->{svnsync} = { url => $url, uuid => $uuid };
1792 }
1793
1794 # this allows us to memoize our SVN::Ra UUID locally and avoid a
1795 # remote lookup (useful for 'git svn log').
1796 sub ra_uuid {
1797 my ($self) = @_;
1798 unless ($self->{ra_uuid}) {
1799 my $key = "svn-remote.$self->{repo_id}.uuid";
1800 my $uuid = eval { tmp_config('--get', $key) };
1801 if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/) {
1802 $self->{ra_uuid} = $uuid;
1803 } else {
1804 die "ra_uuid called without URL\n" unless $self->{url};
1805 $self->{ra_uuid} = $self->ra->get_uuid;
1806 tmp_config('--add', $key, $self->{ra_uuid});
1807 }
1808 }
1809 $self->{ra_uuid};
1810 }
1811
1812 sub _set_repos_root {
1813 my ($self, $repos_root) = @_;
1814 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1815 $repos_root ||= $self->ra->{repos_root};
1816 tmp_config($k, $repos_root);
1817 $repos_root;
1818 }
1819
1820 sub repos_root {
1821 my ($self) = @_;
1822 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1823 eval { tmp_config('--get', $k) } || $self->_set_repos_root;
1824 }
1825
1826 sub ra {
1827 my ($self) = shift;
1828 my $ra = Git::SVN::Ra->new($self->{url});
1829 $self->_set_repos_root($ra->{repos_root});
1830 if ($self->use_svm_props && !$self->{svm}) {
1831 if ($self->no_metadata) {
1832 die "Can't have both 'noMetadata' and ",
1833 "'useSvmProps' options set!\n";
1834 } elsif ($self->use_svnsync_props) {
1835 die "Can't have both 'useSvnsyncProps' and ",
1836 "'useSvmProps' options set!\n";
1837 }
1838 $ra = $self->_set_svm_vars($ra);
1839 $self->{-want_revprops} = 1;
1840 }
1841 $ra;
1842 }
1843
1844 sub rel_path {
1845 my ($self) = @_;
1846 my $repos_root = $self->ra->{repos_root};
1847 return $self->{path} if ($self->{url} eq $repos_root);
1848 my $url = $self->{url} .
1849 (length $self->{path} ? "/$self->{path}" : $self->{path});
1850 $url =~ s!^\Q$repos_root\E(?:/+|$)!!g;
1851 $url;
1852 }
1853
1854 # prop_walk(PATH, REV, SUB)
1855 # -------------------------
1856 # Recursively traverse PATH at revision REV and invoke SUB for each
1857 # directory that contains a SVN property. SUB will be invoked as
1858 # follows: &SUB(gs, path, props); where `gs' is this instance of
1859 # Git::SVN, `path' the path to the directory where the properties
1860 # `props' were found. The `path' will be relative to point of checkout,
1861 # that is, if url://repo/trunk is the current Git branch, and that
1862 # directory contains a sub-directory `d', SUB will be invoked with `/d/'
1863 # as `path' (note the trailing `/').
1864 sub prop_walk {
1865 my ($self, $path, $rev, $sub) = @_;
1866
1867 $path =~ s#^/##;
1868 my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
1869 $path =~ s#^/*#/#g;
1870 my $p = $path;
1871 # Strip the irrelevant part of the path.
1872 $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
1873 # Ensure the path is terminated by a `/'.
1874 $p =~ s#/*$#/#;
1875
1876 # The properties contain all the internal SVN stuff nobody
1877 # (usually) cares about.
1878 my $interesting_props = 0;
1879 foreach (keys %{$props}) {
1880 # If it doesn't start with `svn:', it must be a
1881 # user-defined property.
1882 ++$interesting_props and next if $_ !~ /^svn:/;
1883 # FIXME: Fragile, if SVN adds new public properties,
1884 # this needs to be updated.
1885 ++$interesting_props if /^svn:(?:ignore|keywords|executable
1886 |eol-style|mime-type
1887 |externals|needs-lock)$/x;
1888 }
1889 &$sub($self, $p, $props) if $interesting_props;
1890
1891 foreach (sort keys %$dirent) {
1892 next if $dirent->{$_}->{kind} != $SVN::Node::dir;
1893 $self->prop_walk($path . '/' . $_, $rev, $sub);
1894 }
1895 }
1896
1897 sub last_rev { ($_[0]->last_rev_commit)[0] }
1898 sub last_commit { ($_[0]->last_rev_commit)[1] }
1899
1900 # returns the newest SVN revision number and newest commit SHA1
1901 sub last_rev_commit {
1902 my ($self) = @_;
1903 if (defined $self->{last_rev} && defined $self->{last_commit}) {
1904 return ($self->{last_rev}, $self->{last_commit});
1905 }
1906 my $c = ::verify_ref($self->refname.'^0');
1907 if ($c && !$self->use_svm_props && !$self->no_metadata) {
1908 my $rev = (::cmt_metadata($c))[1];
1909 if (defined $rev) {
1910 ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
1911 return ($rev, $c);
1912 }
1913 }
1914 my $map_path = $self->map_path;
1915 unless (-e $map_path) {
1916 ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
1917 return (undef, undef);
1918 }
1919 my ($rev, $commit) = $self->rev_map_max(1);
1920 ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
1921 return ($rev, $commit);
1922 }
1923
1924 sub get_fetch_range {
1925 my ($self, $min, $max) = @_;
1926 $max ||= $self->ra->get_latest_revnum;
1927 $min ||= $self->rev_map_max;
1928 (++$min, $max);
1929 }
1930
1931 sub tmp_config {
1932 my (@args) = @_;
1933 my $old_def_config = "$ENV{GIT_DIR}/svn/config";
1934 my $config = "$ENV{GIT_DIR}/svn/.metadata";
1935 if (! -f $config && -f $old_def_config) {
1936 rename $old_def_config, $config or
1937 die "Failed rename $old_def_config => $config: $!\n";
1938 }
1939 my $old_config = $ENV{GIT_CONFIG};
1940 $ENV{GIT_CONFIG} = $config;
1941 $@ = undef;
1942 my @ret = eval {
1943 unless (-f $config) {
1944 mkfile($config);
1945 open my $fh, '>', $config or
1946 die "Can't open $config: $!\n";
1947 print $fh "; This file is used internally by ",
1948 "git-svn\n" or die
1949 "Couldn't write to $config: $!\n";
1950 print $fh "; You should not have to edit it\n" or
1951 die "Couldn't write to $config: $!\n";
1952 close $fh or die "Couldn't close $config: $!\n";
1953 }
1954 command('config', @args);
1955 };
1956 my $err = $@;
1957 if (defined $old_config) {
1958 $ENV{GIT_CONFIG} = $old_config;
1959 } else {
1960 delete $ENV{GIT_CONFIG};
1961 }
1962 die $err if $err;
1963 wantarray ? @ret : $ret[0];
1964 }
1965
1966 sub tmp_index_do {
1967 my ($self, $sub) = @_;
1968 my $old_index = $ENV{GIT_INDEX_FILE};
1969 $ENV{GIT_INDEX_FILE} = $self->{index};
1970 $@ = undef;
1971 my @ret = eval {
1972 my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
1973 mkpath([$dir]) unless -d $dir;
1974 &$sub;
1975 };
1976 my $err = $@;
1977 if (defined $old_index) {
1978 $ENV{GIT_INDEX_FILE} = $old_index;
1979 } else {
1980 delete $ENV{GIT_INDEX_FILE};
1981 }
1982 die $err if $err;
1983 wantarray ? @ret : $ret[0];
1984 }
1985
1986 sub assert_index_clean {
1987 my ($self, $treeish) = @_;
1988
1989 $self->tmp_index_do(sub {
1990 command_noisy('read-tree', $treeish) unless -e $self->{index};
1991 my $x = command_oneline('write-tree');
1992 my ($y) = (command(qw/cat-file commit/, $treeish) =~
1993 /^tree ($::sha1)/mo);
1994 return if $y eq $x;
1995
1996 warn "Index mismatch: $y != $x\nrereading $treeish\n";
1997 unlink $self->{index} or die "unlink $self->{index}: $!\n";
1998 command_noisy('read-tree', $treeish);
1999 $x = command_oneline('write-tree');
2000 if ($y ne $x) {
2001 ::fatal "trees ($treeish) $y != $x\n",
2002 "Something is seriously wrong...";
2003 }
2004 });
2005 }
2006
2007 sub get_commit_parents {
2008 my ($self, $log_entry) = @_;
2009 my (%seen, @ret, @tmp);
2010 # legacy support for 'set-tree'; this is only used by set_tree_cb:
2011 if (my $ip = $self->{inject_parents}) {
2012 if (my $commit = delete $ip->{$log_entry->{revision}}) {
2013 push @tmp, $commit;
2014 }
2015 }
2016 if (my $cur = ::verify_ref($self->refname.'^0')) {
2017 push @tmp, $cur;
2018 }
2019 if (my $ipd = $self->{inject_parents_dcommit}) {
2020 if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2021 push @tmp, @$commit;
2022 }
2023 }
2024 push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2025 while (my $p = shift @tmp) {
2026 next if $seen{$p};
2027 $seen{$p} = 1;
2028 push @ret, $p;
2029 # MAXPARENT is defined to 16 in commit-tree.c:
2030 last if @ret >= 16;
2031 }
2032 if (@tmp) {
2033 die "r$log_entry->{revision}: No room for parents:\n\t",
2034 join("\n\t", @tmp), "\n";
2035 }
2036 @ret;
2037 }
2038
2039 sub rewrite_root {
2040 my ($self) = @_;
2041 return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2042 my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2043 my $rwr = eval { command_oneline(qw/config --get/, $k) };
2044 if ($rwr) {
2045 $rwr =~ s#/+$##;
2046 if ($rwr !~ m#^[a-z\+]+://#) {
2047 die "$rwr is not a valid URL (key: $k)\n";
2048 }
2049 }
2050 $self->{-rewrite_root} = $rwr;
2051 }
2052
2053 sub metadata_url {
2054 my ($self) = @_;
2055 ($self->rewrite_root || $self->{url}) .
2056 (length $self->{path} ? '/' . $self->{path} : '');
2057 }
2058
2059 sub full_url {
2060 my ($self) = @_;
2061 $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2062 }
2063
2064
2065 sub set_commit_header_env {
2066 my ($log_entry) = @_;
2067 my %env;
2068 foreach my $ned (qw/NAME EMAIL DATE/) {
2069 foreach my $ac (qw/AUTHOR COMMITTER/) {
2070 $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2071 }
2072 }
2073
2074 $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2075 $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2076 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2077
2078 $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2079 ? $log_entry->{commit_name}
2080 : $log_entry->{name};
2081 $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2082 ? $log_entry->{commit_email}
2083 : $log_entry->{email};
2084 \%env;
2085 }
2086
2087 sub restore_commit_header_env {
2088 my ($env) = @_;
2089 foreach my $ned (qw/NAME EMAIL DATE/) {
2090 foreach my $ac (qw/AUTHOR COMMITTER/) {
2091 my $k = "GIT_${ac}_${ned}";
2092 if (defined $env->{$k}) {
2093 $ENV{$k} = $env->{$k};
2094 } else {
2095 delete $ENV{$k};
2096 }
2097 }
2098 }
2099 }
2100
2101 sub gc {
2102 command_noisy('gc', '--auto');
2103 };
2104
2105 sub do_git_commit {
2106 my ($self, $log_entry) = @_;
2107 my $lr = $self->last_rev;
2108 if (defined $lr && $lr >= $log_entry->{revision}) {
2109 die "Last fetched revision of ", $self->refname,
2110 " was r$lr, but we are about to fetch: ",
2111 "r$log_entry->{revision}!\n";
2112 }
2113 if (my $c = $self->rev_map_get($log_entry->{revision})) {
2114 croak "$log_entry->{revision} = $c already exists! ",
2115 "Why are we refetching it?\n";
2116 }
2117 my $old_env = set_commit_header_env($log_entry);
2118 my $tree = $log_entry->{tree};
2119 if (!defined $tree) {
2120 $tree = $self->tmp_index_do(sub {
2121 command_oneline('write-tree') });
2122 }
2123 die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2124
2125 my @exec = ('git-commit-tree', $tree);
2126 foreach ($self->get_commit_parents($log_entry)) {
2127 push @exec, '-p', $_;
2128 }
2129 defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2130 or croak $!;
2131 print $msg_fh $log_entry->{log} or croak $!;
2132 restore_commit_header_env($old_env);
2133 unless ($self->no_metadata) {
2134 print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2135 or croak $!;
2136 }
2137 $msg_fh->flush == 0 or croak $!;
2138 close $msg_fh or croak $!;
2139 chomp(my $commit = do { local $/; <$out_fh> });
2140 close $out_fh or croak $!;
2141 waitpid $pid, 0;
2142 croak $? if $?;
2143 if ($commit !~ /^$::sha1$/o) {
2144 die "Failed to commit, invalid sha1: $commit\n";
2145 }
2146
2147 $self->rev_map_set($log_entry->{revision}, $commit, 1);
2148
2149 $self->{last_rev} = $log_entry->{revision};
2150 $self->{last_commit} = $commit;
2151 print "r$log_entry->{revision}";
2152 if (defined $log_entry->{svm_revision}) {
2153 print " (\@$log_entry->{svm_revision})";
2154 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2155 0, $self->svm_uuid);
2156 }
2157 print " = $commit ($self->{ref_id})\n";
2158 if (--$_gc_nr == 0) {
2159 $_gc_nr = $_gc_period;
2160 gc();
2161 }
2162 return $commit;
2163 }
2164
2165 sub match_paths {
2166 my ($self, $paths, $r) = @_;
2167 return 1 if $self->{path} eq '';
2168 if (my $path = $paths->{"/$self->{path}"}) {
2169 return ($path->{action} eq 'D') ? 0 : 1;
2170 }
2171 $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
2172 if (grep /$self->{path_regex}/, keys %$paths) {
2173 return 1;
2174 }
2175 my $c = '';
2176 foreach (split m#/#, $self->{path}) {
2177 $c .= "/$_";
2178 next unless ($paths->{$c} &&
2179 ($paths->{$c}->{action} =~ /^[AR]$/));
2180 if ($self->ra->check_path($self->{path}, $r) ==
2181 $SVN::Node::dir) {
2182 return 1;
2183 }
2184 }
2185 return 0;
2186 }
2187
2188 sub find_parent_branch {
2189 my ($self, $paths, $rev) = @_;
2190 return undef unless $self->follow_parent;
2191 unless (defined $paths) {
2192 my $err_handler = $SVN::Error::handler;
2193 $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2194 $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1, sub {
2195 $paths =
2196 Git::SVN::Ra::dup_changed_paths($_[0]) });
2197 $SVN::Error::handler = $err_handler;
2198 }
2199 return undef unless defined $paths;
2200
2201 # look for a parent from another branch:
2202 my @b_path_components = split m#/#, $self->rel_path;
2203 my @a_path_components;
2204 my $i;
2205 while (@b_path_components) {
2206 $i = $paths->{'/'.join('/', @b_path_components)};
2207 last if $i && defined $i->{copyfrom_path};
2208 unshift(@a_path_components, pop(@b_path_components));
2209 }
2210 return undef unless defined $i && defined $i->{copyfrom_path};
2211 my $branch_from = $i->{copyfrom_path};
2212 if (@a_path_components) {
2213 print STDERR "branch_from: $branch_from => ";
2214 $branch_from .= '/'.join('/', @a_path_components);
2215 print STDERR $branch_from, "\n";
2216 }
2217 my $r = $i->{copyfrom_rev};
2218 my $repos_root = $self->ra->{repos_root};
2219 my $url = $self->ra->{url};
2220 my $new_url = $repos_root . $branch_from;
2221 print STDERR "Found possible branch point: ",
2222 "$new_url => ", $self->full_url, ", $r\n";
2223 $branch_from =~ s#^/##;
2224 my $gs = Git::SVN->find_by_url($new_url, $repos_root, $branch_from);
2225 unless ($gs) {
2226 my $ref_id = $self->{ref_id};
2227 $ref_id =~ s/\@\d+$//;
2228 $ref_id .= "\@$r";
2229 # just grow a tail if we're not unique enough :x
2230 $ref_id .= '-' while find_ref($ref_id);
2231 print STDERR "Initializing parent: $ref_id\n";
2232 $gs = Git::SVN->init($new_url, '', $ref_id, $ref_id, 1);
2233 }
2234 my ($r0, $parent) = $gs->find_rev_before($r, 1);
2235 if (!defined $r0 || !defined $parent) {
2236 my ($base, $head) = parse_revision_argument(0, $r);
2237 if ($base <= $r) {
2238 $gs->fetch($base, $r);
2239 }
2240 ($r0, $parent) = $gs->last_rev_commit;
2241 }
2242 if (defined $r0 && defined $parent) {
2243 print STDERR "Found branch parent: ($self->{ref_id}) $parent\n";
2244 my $ed;
2245 if ($self->ra->can_do_switch) {
2246 $self->assert_index_clean($parent);
2247 print STDERR "Following parent with do_switch\n";
2248 # do_switch works with svn/trunk >= r22312, but that
2249 # is not included with SVN 1.4.3 (the latest version
2250 # at the moment), so we can't rely on it
2251 $self->{last_commit} = $parent;
2252 $ed = SVN::Git::Fetcher->new($self);
2253 $gs->ra->gs_do_switch($r0, $rev, $gs,
2254 $self->full_url, $ed)
2255 or die "SVN connection failed somewhere...\n";
2256 } elsif ($self->ra->trees_match($new_url, $r0,
2257 $self->full_url, $rev)) {
2258 print STDERR "Trees match:\n",
2259 " $new_url\@$r0\n",
2260 " ${\$self->full_url}\@$rev\n",
2261 "Following parent with no changes\n";
2262 $self->tmp_index_do(sub {
2263 command_noisy('read-tree', $parent);
2264 });
2265 $self->{last_commit} = $parent;
2266 } else {
2267 print STDERR "Following parent with do_update\n";
2268 $ed = SVN::Git::Fetcher->new($self);
2269 $self->ra->gs_do_update($rev, $rev, $self, $ed)
2270 or die "SVN connection failed somewhere...\n";
2271 }
2272 print STDERR "Successfully followed parent\n";
2273 return $self->make_log_entry($rev, [$parent], $ed);
2274 }
2275 return undef;
2276 }
2277
2278 sub do_fetch {
2279 my ($self, $paths, $rev) = @_;
2280 my $ed;
2281 my ($last_rev, @parents);
2282 if (my $lc = $self->last_commit) {
2283 # we can have a branch that was deleted, then re-added
2284 # under the same name but copied from another path, in
2285 # which case we'll have multiple parents (we don't
2286 # want to break the original ref, nor lose copypath info):
2287 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2288 push @{$log_entry->{parents}}, $lc;
2289 return $log_entry;
2290 }
2291 $ed = SVN::Git::Fetcher->new($self);
2292 $last_rev = $self->{last_rev};
2293 $ed->{c} = $lc;
2294 @parents = ($lc);
2295 } else {
2296 $last_rev = $rev;
2297 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2298 return $log_entry;
2299 }
2300 $ed = SVN::Git::Fetcher->new($self);
2301 }
2302 unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2303 die "SVN connection failed somewhere...\n";
2304 }
2305 $self->make_log_entry($rev, \@parents, $ed);
2306 }
2307
2308 sub get_untracked {
2309 my ($self, $ed) = @_;
2310 my @out;
2311 my $h = $ed->{empty};
2312 foreach (sort keys %$h) {
2313 my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2314 push @out, " $act: " . uri_encode($_);
2315 warn "W: $act: $_\n";
2316 }
2317 foreach my $t (qw/dir_prop file_prop/) {
2318 $h = $ed->{$t} or next;
2319 foreach my $path (sort keys %$h) {
2320 my $ppath = $path eq '' ? '.' : $path;
2321 foreach my $prop (sort keys %{$h->{$path}}) {
2322 next if $SKIP_PROP{$prop};
2323 my $v = $h->{$path}->{$prop};
2324 my $t_ppath_prop = "$t: " .
2325 uri_encode($ppath) . ' ' .
2326 uri_encode($prop);
2327 if (defined $v) {
2328 push @out, " +$t_ppath_prop " .
2329 uri_encode($v);
2330 } else {
2331 push @out, " -$t_ppath_prop";
2332 }
2333 }
2334 }
2335 }
2336 foreach my $t (qw/absent_file absent_directory/) {
2337 $h = $ed->{$t} or next;
2338 foreach my $parent (sort keys %$h) {
2339 foreach my $path (sort @{$h->{$parent}}) {
2340 push @out, " $t: " .
2341 uri_encode("$parent/$path");
2342 warn "W: $t: $parent/$path ",
2343 "Insufficient permissions?\n";
2344 }
2345 }
2346 }
2347 \@out;
2348 }
2349
2350 sub parse_svn_date {
2351 my $date = shift || return '+0000 1970-01-01 00:00:00';
2352 my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2353 (\d\d)\:(\d\d)\:(\d\d).\d+Z$/x) or
2354 croak "Unable to parse date: $date\n";
2355 "+0000 $Y-$m-$d $H:$M:$S";
2356 }
2357
2358 sub check_author {
2359 my ($author) = @_;
2360 if (!defined $author || length $author == 0) {
2361 $author = '(no author)';
2362 }
2363 if (defined $::_authors && ! defined $::users{$author}) {
2364 die "Author: $author not defined in $::_authors file\n";
2365 }
2366 $author;
2367 }
2368
2369 sub make_log_entry {
2370 my ($self, $rev, $parents, $ed) = @_;
2371 my $untracked = $self->get_untracked($ed);
2372
2373 open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
2374 print $un "r$rev\n" or croak $!;
2375 print $un $_, "\n" foreach @$untracked;
2376 my %log_entry = ( parents => $parents || [], revision => $rev,
2377 log => '');
2378
2379 my $headrev;
2380 my $logged = delete $self->{logged_rev_props};
2381 if (!$logged || $self->{-want_revprops}) {
2382 my $rp = $self->ra->rev_proplist($rev);
2383 foreach (sort keys %$rp) {
2384 my $v = $rp->{$_};
2385 if (/^svn:(author|date|log)$/) {
2386 $log_entry{$1} = $v;
2387 } elsif ($_ eq 'svm:headrev') {
2388 $headrev = $v;
2389 } else {
2390 print $un " rev_prop: ", uri_encode($_), ' ',
2391 uri_encode($v), "\n";
2392 }
2393 }
2394 } else {
2395 map { $log_entry{$_} = $logged->{$_} } keys %$logged;
2396 }
2397 close $un or croak $!;
2398
2399 $log_entry{date} = parse_svn_date($log_entry{date});
2400 $log_entry{log} .= "\n";
2401 my $author = $log_entry{author} = check_author($log_entry{author});
2402 my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
2403 : ($author, undef);
2404
2405 my ($commit_name, $commit_email) = ($name, $email);
2406 if ($_use_log_author) {
2407 my $name_field;
2408 if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
2409 $name_field = $1;
2410 } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
2411 $name_field = $1;
2412 }
2413 if (!defined $name_field) {
2414 #
2415 } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
2416 ($name, $email) = ($1, $2);
2417 } elsif ($name_field =~ /(.*)@/) {
2418 ($name, $email) = ($1, $name_field);
2419 } else {
2420 ($name, $email) = ($name_field, 'unknown');
2421 }
2422 }
2423 if (defined $headrev && $self->use_svm_props) {
2424 if ($self->rewrite_root) {
2425 die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
2426 "options set!\n";
2427 }
2428 my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$};
2429 # we don't want "SVM: initializing mirror for junk" ...
2430 return undef if $r == 0;
2431 my $svm = $self->svm;
2432 if ($uuid ne $svm->{uuid}) {
2433 die "UUID mismatch on SVM path:\n",
2434 "expected: $svm->{uuid}\n",
2435 " got: $uuid\n";
2436 }
2437 my $full_url = $self->full_url;
2438 $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
2439 die "Failed to replace '$svm->{replace}' with ",
2440 "'$svm->{source}' in $full_url\n";
2441 # throw away username for storing in records
2442 remove_username($full_url);
2443 $log_entry{metadata} = "$full_url\@$r $uuid";
2444 $log_entry{svm_revision} = $r;
2445 $email ||= "$author\@$uuid";
2446 $commit_email ||= "$author\@$uuid";
2447 } elsif ($self->use_svnsync_props) {
2448 my $full_url = $self->svnsync->{url};
2449 $full_url .= "/$self->{path}" if length $self->{path};
2450 remove_username($full_url);
2451 my $uuid = $self->svnsync->{uuid};
2452 $log_entry{metadata} = "$full_url\@$rev $uuid";
2453 $email ||= "$author\@$uuid";
2454 $commit_email ||= "$author\@$uuid";
2455 } else {
2456 my $url = $self->metadata_url;
2457 remove_username($url);
2458 $log_entry{metadata} = "$url\@$rev " .
2459 $self->ra->get_uuid;
2460 $email ||= "$author\@" . $self->ra->get_uuid;
2461 $commit_email ||= "$author\@" . $self->ra->get_uuid;
2462 }
2463 $log_entry{name} = $name;
2464 $log_entry{email} = $email;
2465 $log_entry{commit_name} = $commit_name;
2466 $log_entry{commit_email} = $commit_email;
2467 \%log_entry;
2468 }
2469
2470 sub fetch {
2471 my ($self, $min_rev, $max_rev, @parents) = @_;
2472 my ($last_rev, $last_commit) = $self->last_rev_commit;
2473 my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
2474 $self->ra->gs_fetch_loop_common($base, $head, [$self]);
2475 }
2476
2477 sub set_tree_cb {
2478 my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
2479 $self->{inject_parents} = { $rev => $tree };
2480 $self->fetch(undef, undef);
2481 }
2482
2483 sub set_tree {
2484 my ($self, $tree) = (shift, shift);
2485 my $log_entry = ::get_commit_entry($tree);
2486 unless ($self->{last_rev}) {
2487 fatal("Must have an existing revision to commit");
2488 }
2489 my %ed_opts = ( r => $self->{last_rev},
2490 log => $log_entry->{log},
2491 ra => $self->ra,
2492 tree_a => $self->{last_commit},
2493 tree_b => $tree,
2494 editor_cb => sub {
2495 $self->set_tree_cb($log_entry, $tree, @_) },
2496 svn_path => $self->{path} );
2497 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
2498 print "No changes\nr$self->{last_rev} = $tree\n";
2499 }
2500 }
2501
2502 sub rebuild_from_rev_db {
2503 my ($self, $path) = @_;
2504 my $r = -1;
2505 open my $fh, '<', $path or croak "open: $!";
2506 while (<$fh>) {
2507 length($_) == 41 or croak "inconsistent size in ($_) != 41";
2508 chomp($_);
2509 ++$r;
2510 next if $_ eq ('0' x 40);
2511 $self->rev_map_set($r, $_);
2512 print "r$r = $_\n";
2513 }
2514 close $fh or croak "close: $!";
2515 unlink $path or croak "unlink: $!";
2516 }
2517
2518 sub rebuild {
2519 my ($self) = @_;
2520 my $map_path = $self->map_path;
2521 return if (-e $map_path && ! -z $map_path);
2522 return unless ::verify_ref($self->refname.'^0');
2523 if ($self->use_svm_props || $self->no_metadata) {
2524 my $rev_db = $self->rev_db_path;
2525 $self->rebuild_from_rev_db($rev_db);
2526 if ($self->use_svm_props) {
2527 my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
2528 $self->rebuild_from_rev_db($svm_rev_db);
2529 }
2530 $self->unlink_rev_db_symlink;
2531 return;
2532 }
2533 print "Rebuilding $map_path ...\n";
2534 my ($log, $ctx) =
2535 command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
2536 $self->refname, '--');
2537 my $full_url = $self->full_url;
2538 remove_username($full_url);
2539 my $svn_uuid = $self->ra_uuid;
2540 my $c;
2541 while (<$log>) {
2542 if ( m{^commit ($::sha1)$} ) {
2543 $c = $1;
2544 next;
2545 }
2546 next unless s{^\s*(git-svn-id:)}{$1};
2547 my ($url, $rev, $uuid) = ::extract_metadata($_);
2548 remove_username($url);
2549
2550 # ignore merges (from set-tree)
2551 next if (!defined $rev || !$uuid);
2552
2553 # if we merged or otherwise started elsewhere, this is
2554 # how we break out of it
2555 if (($uuid ne $svn_uuid) ||
2556 ($full_url && $url && ($url ne $full_url))) {
2557 next;
2558 }
2559
2560 $self->rev_map_set($rev, $c);
2561 print "r$rev = $c\n";
2562 }
2563 command_close_pipe($log, $ctx);
2564 print "Done rebuilding $map_path\n";
2565 my $rev_db_path = $self->rev_db_path;
2566 if (-f $self->rev_db_path) {
2567 unlink $self->rev_db_path or croak "unlink: $!";
2568 }
2569 $self->unlink_rev_db_symlink;
2570 }
2571
2572 # rev_map:
2573 # Tie::File seems to be prone to offset errors if revisions get sparse,
2574 # it's not that fast, either. Tie::File is also not in Perl 5.6. So
2575 # one of my favorite modules is out :< Next up would be one of the DBM
2576 # modules, but I'm not sure which is most portable...
2577 #
2578 # This is the replacement for the rev_db format, which was too big
2579 # and inefficient for large repositories with a lot of sparse history
2580 # (mainly tags)
2581 #
2582 # The format is this:
2583 # - 24 bytes for every record,
2584 # * 4 bytes for the integer representing an SVN revision number
2585 # * 20 bytes representing the sha1 of a git commit
2586 # - No empty padding records like the old format
2587 # (except the last record, which can be overwritten)
2588 # - new records are written append-only since SVN revision numbers
2589 # increase monotonically
2590 # - lookups on SVN revision number are done via a binary search
2591 # - Piping the file to xxd -c24 is a good way of dumping it for
2592 # viewing or editing (piped back through xxd -r), should the need
2593 # ever arise.
2594 # - The last record can be padding revision with an all-zero sha1
2595 # This is used to optimize fetch performance when using multiple
2596 # "fetch" directives in .git/config
2597 #
2598 # These files are disposable unless noMetadata or useSvmProps is set
2599
2600 sub _rev_map_set {
2601 my ($fh, $rev, $commit) = @_;
2602
2603 my $size = (stat($fh))[7];
2604 ($size % 24) == 0 or croak "inconsistent size: $size";
2605
2606 my $wr_offset = 0;
2607 if ($size > 0) {
2608 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2609 my $read = sysread($fh, my $buf, 24) or croak "read: $!";
2610 $read == 24 or croak "read only $read bytes (!= 24)";
2611 my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
2612 if ($last_commit eq ('0' x40)) {
2613 if ($size >= 48) {
2614 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2615 $read = sysread($fh, $buf, 24) or
2616 croak "read: $!";
2617 $read == 24 or
2618 croak "read only $read bytes (!= 24)";
2619 ($last_rev, $last_commit) =
2620 unpack(rev_map_fmt, $buf);
2621 if ($last_commit eq ('0' x40)) {
2622 croak "inconsistent .rev_map\n";
2623 }
2624 }
2625 if ($last_rev >= $rev) {
2626 croak "last_rev is higher!: $last_rev >= $rev";
2627 }
2628 $wr_offset = -24;
2629 }
2630 }
2631 sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
2632 syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
2633 croak "write: $!";
2634 }
2635
2636 sub mkfile {
2637 my ($path) = @_;
2638 unless (-e $path) {
2639 my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
2640 mkpath([$dir]) unless -d $dir;
2641 open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
2642 close $fh or die "Couldn't close (create) $path: $!\n";
2643 }
2644 }
2645
2646 sub rev_map_set {
2647 my ($self, $rev, $commit, $update_ref, $uuid) = @_;
2648 length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
2649 my $db = $self->map_path($uuid);
2650 my $db_lock = "$db.lock";
2651 my $sig;
2652 if ($update_ref) {
2653 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2654 $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
2655 }
2656 mkfile($db);
2657
2658 $LOCKFILES{$db_lock} = 1;
2659 my $sync;
2660 # both of these options make our .rev_db file very, very important
2661 # and we can't afford to lose it because rebuild() won't work
2662 if ($self->use_svm_props || $self->no_metadata) {
2663 $sync = 1;
2664 copy($db, $db_lock) or die "rev_map_set(@_): ",
2665 "Failed to copy: ",
2666 "$db => $db_lock ($!)\n";
2667 } else {
2668 rename $db, $db_lock or die "rev_map_set(@_): ",
2669 "Failed to rename: ",
2670 "$db => $db_lock ($!)\n";
2671 }
2672
2673 sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
2674 or croak "Couldn't open $db_lock: $!\n";
2675 _rev_map_set($fh, $rev, $commit);
2676 if ($sync) {
2677 $fh->flush or die "Couldn't flush $db_lock: $!\n";
2678 $fh->sync or die "Couldn't sync $db_lock: $!\n";
2679 }
2680 close $fh or croak $!;
2681 if ($update_ref) {
2682 $_head = $self;
2683 command_noisy('update-ref', '-m', "r$rev",
2684 $self->refname, $commit);
2685 }
2686 rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
2687 "$db_lock => $db ($!)\n";
2688 delete $LOCKFILES{$db_lock};
2689 if ($update_ref) {
2690 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2691 $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
2692 kill $sig, $$ if defined $sig;
2693 }
2694 }
2695
2696 # If want_commit, this will return an array of (rev, commit) where
2697 # commit _must_ be a valid commit in the archive.
2698 # Otherwise, it'll return the max revision (whether or not the
2699 # commit is valid or just a 0x40 placeholder).
2700 sub rev_map_max {
2701 my ($self, $want_commit) = @_;
2702 $self->rebuild;
2703 my $map_path = $self->map_path;
2704 stat $map_path or return $want_commit ? (0, undef) : 0;
2705 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2706 my $size = (stat($fh))[7];
2707 ($size % 24) == 0 or croak "inconsistent size: $size";
2708
2709 if ($size == 0) {
2710 close $fh or croak "close: $!";
2711 return $want_commit ? (0, undef) : 0;
2712 }
2713
2714 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2715 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2716 my ($r, $c) = unpack(rev_map_fmt, $buf);
2717 if ($want_commit && $c eq ('0' x40)) {
2718 if ($size < 48) {
2719 return $want_commit ? (0, undef) : 0;
2720 }
2721 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2722 sysread($fh, $buf, 24) == 24 or croak "read: $!";
2723 ($r, $c) = unpack(rev_map_fmt, $buf);
2724 if ($c eq ('0'x40)) {
2725 croak "Penultimate record is all-zeroes in $map_path";
2726 }
2727 }
2728 close $fh or croak "close: $!";
2729 $want_commit ? ($r, $c) : $r;
2730 }
2731
2732 sub rev_map_get {
2733 my ($self, $rev, $uuid) = @_;
2734 my $map_path = $self->map_path($uuid);
2735 return undef unless -e $map_path;
2736
2737 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2738 my $size = (stat($fh))[7];
2739 ($size % 24) == 0 or croak "inconsistent size: $size";
2740
2741 if ($size == 0) {
2742 close $fh or croak "close: $fh";
2743 return undef;
2744 }
2745
2746 my ($l, $u) = (0, $size - 24);
2747 my ($r, $c, $buf);
2748
2749 while ($l <= $u) {
2750 my $i = int(($l/24 + $u/24) / 2) * 24;
2751 sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
2752 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2753 my ($r, $c) = unpack('NH40', $buf);
2754
2755 if ($r < $rev) {
2756 $l = $i + 24;
2757 } elsif ($r > $rev) {
2758 $u = $i - 24;
2759 } else { # $r == $rev
2760 close($fh) or croak "close: $!";
2761 return $c eq ('0' x 40) ? undef : $c;
2762 }
2763 }
2764 close($fh) or croak "close: $!";
2765 undef;
2766 }
2767
2768 # Finds the first svn revision that exists on (if $eq_ok is true) or
2769 # before $rev for the current branch. It will not search any lower
2770 # than $min_rev. Returns the git commit hash and svn revision number
2771 # if found, else (undef, undef).
2772 sub find_rev_before {
2773 my ($self, $rev, $eq_ok, $min_rev) = @_;
2774 --$rev unless $eq_ok;
2775 $min_rev ||= 1;
2776 while ($rev >= $min_rev) {
2777 if (my $c = $self->rev_map_get($rev)) {
2778 return ($rev, $c);
2779 }
2780 --$rev;
2781 }
2782 return (undef, undef);
2783 }
2784
2785 # Finds the first svn revision that exists on (if $eq_ok is true) or
2786 # after $rev for the current branch. It will not search any higher
2787 # than $max_rev. Returns the git commit hash and svn revision number
2788 # if found, else (undef, undef).
2789 sub find_rev_after {
2790 my ($self, $rev, $eq_ok, $max_rev) = @_;
2791 ++$rev unless $eq_ok;
2792 $max_rev ||= $self->rev_map_max;
2793 while ($rev <= $max_rev) {
2794 if (my $c = $self->rev_map_get($rev)) {
2795 return ($rev, $c);
2796 }
2797 ++$rev;
2798 }
2799 return (undef, undef);
2800 }
2801
2802 sub _new {
2803 my ($class, $repo_id, $ref_id, $path) = @_;
2804 unless (defined $repo_id && length $repo_id) {
2805 $repo_id = $Git::SVN::default_repo_id;
2806 }
2807 unless (defined $ref_id && length $ref_id) {
2808 $_[2] = $ref_id = $Git::SVN::default_ref_id;
2809 }
2810 $_[1] = $repo_id = sanitize_remote_name($repo_id);
2811 my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
2812 $_[3] = $path = '' unless (defined $path);
2813 mkpath(["$ENV{GIT_DIR}/svn"]);
2814 bless {
2815 ref_id => $ref_id, dir => $dir, index => "$dir/index",
2816 path => $path, config => "$ENV{GIT_DIR}/svn/config",
2817 map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
2818 }
2819
2820 # for read-only access of old .rev_db formats
2821 sub unlink_rev_db_symlink {
2822 my ($self) = @_;
2823 my $link = $self->rev_db_path;
2824 $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
2825 if (-l $link) {
2826 unlink $link or croak "unlink: $link failed!";
2827 }
2828 }
2829
2830 sub rev_db_path {
2831 my ($self, $uuid) = @_;
2832 my $db_path = $self->map_path($uuid);
2833 $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
2834 or croak "map_path: $db_path does not contain '/.rev_map.' !";
2835 $db_path;
2836 }
2837
2838 # the new replacement for .rev_db
2839 sub map_path {
2840 my ($self, $uuid) = @_;
2841 $uuid ||= $self->ra_uuid;
2842 "$self->{map_root}.$uuid";
2843 }
2844
2845 sub uri_encode {
2846 my ($f) = @_;
2847 $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
2848 $f
2849 }
2850
2851 sub remove_username {
2852 $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
2853 }
2854
2855 package Git::SVN::Prompt;
2856 use strict;
2857 use warnings;
2858 require SVN::Core;
2859 use vars qw/$_no_auth_cache $_username/;
2860
2861 sub simple {
2862 my ($cred, $realm, $default_username, $may_save, $pool) = @_;
2863 $may_save = undef if $_no_auth_cache;
2864 $default_username = $_username if defined $_username;
2865 if (defined $default_username && length $default_username) {
2866 if (defined $realm && length $realm) {
2867 print STDERR "Authentication realm: $realm\n";
2868 STDERR->flush;
2869 }
2870 $cred->username($default_username);
2871 } else {
2872 username($cred, $realm, $may_save, $pool);
2873 }
2874 $cred->password(_read_password("Password for '" .
2875 $cred->username . "': ", $realm));
2876 $cred->may_save($may_save);
2877 $SVN::_Core::SVN_NO_ERROR;
2878 }
2879
2880 sub ssl_server_trust {
2881 my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
2882 $may_save = undef if $_no_auth_cache;
2883 print STDERR "Error validating server certificate for '$realm':\n";
2884 {
2885 no warnings 'once';
2886 # All variables SVN::Auth::SSL::* are used only once,
2887 # so we're shutting up Perl warnings about this.
2888 if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
2889 print STDERR " - The certificate is not issued ",
2890 "by a trusted authority. Use the\n",
2891 " fingerprint to validate ",
2892 "the certificate manually!\n";
2893 }
2894 if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
2895 print STDERR " - The certificate hostname ",
2896 "does not match.\n";
2897 }
2898 if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
2899 print STDERR " - The certificate is not yet valid.\n";
2900 }
2901 if ($failures & $SVN::Auth::SSL::EXPIRED) {
2902 print STDERR " - The certificate has expired.\n";
2903 }
2904 if ($failures & $SVN::Auth::SSL::OTHER) {
2905 print STDERR " - The certificate has ",
2906 "an unknown error.\n";
2907 }
2908 } # no warnings 'once'
2909 printf STDERR
2910 "Certificate information:\n".
2911 " - Hostname: %s\n".
2912 " - Valid: from %s until %s\n".
2913 " - Issuer: %s\n".
2914 " - Fingerprint: %s\n",
2915 map $cert_info->$_, qw(hostname valid_from valid_until
2916 issuer_dname fingerprint);
2917 my $choice;
2918 prompt:
2919 print STDERR $may_save ?
2920 "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
2921 "(R)eject or accept (t)emporarily? ";
2922 STDERR->flush;
2923 $choice = lc(substr(<STDIN> || 'R', 0, 1));
2924 if ($choice =~ /^t$/i) {
2925 $cred->may_save(undef);
2926 } elsif ($choice =~ /^r$/i) {
2927 return -1;
2928 } elsif ($may_save && $choice =~ /^p$/i) {
2929 $cred->may_save($may_save);
2930 } else {
2931 goto prompt;
2932 }
2933 $cred->accepted_failures($failures);
2934 $SVN::_Core::SVN_NO_ERROR;
2935 }
2936
2937 sub ssl_client_cert {
2938 my ($cred, $realm, $may_save, $pool) = @_;
2939 $may_save = undef if $_no_auth_cache;
2940 print STDERR "Client certificate filename: ";
2941 STDERR->flush;
2942 chomp(my $filename = <STDIN>);
2943 $cred->cert_file($filename);
2944 $cred->may_save($may_save);
2945 $SVN::_Core::SVN_NO_ERROR;
2946 }
2947
2948 sub ssl_client_cert_pw {
2949 my ($cred, $realm, $may_save, $pool) = @_;
2950 $may_save = undef if $_no_auth_cache;
2951 $cred->password(_read_password("Password: ", $realm));
2952 $cred->may_save($may_save);
2953 $SVN::_Core::SVN_NO_ERROR;
2954 }
2955
2956 sub username {
2957 my ($cred, $realm, $may_save, $pool) = @_;
2958 $may_save = undef if $_no_auth_cache;
2959 if (defined $realm && length $realm) {
2960 print STDERR "Authentication realm: $realm\n";
2961 }
2962 my $username;
2963 if (defined $_username) {
2964 $username = $_username;
2965 } else {
2966 print STDERR "Username: ";
2967 STDERR->flush;
2968 chomp($username = <STDIN>);
2969 }
2970 $cred->username($username);
2971 $cred->may_save($may_save);
2972 $SVN::_Core::SVN_NO_ERROR;
2973 }
2974
2975 sub _read_password {
2976 my ($prompt, $realm) = @_;
2977 print STDERR $prompt;
2978 STDERR->flush;
2979 require Term::ReadKey;
2980 Term::ReadKey::ReadMode('noecho');
2981 my $password = '';
2982 while (defined(my $key = Term::ReadKey::ReadKey(0))) {
2983 last if $key =~ /[\012\015]/; # \n\r
2984 $password .= $key;
2985 }
2986 Term::ReadKey::ReadMode('restore');
2987 print STDERR "\n";
2988 STDERR->flush;
2989 $password;
2990 }
2991
2992 package SVN::Git::Fetcher;
2993 use vars qw/@ISA/;
2994 use strict;
2995 use warnings;
2996 use Carp qw/croak/;
2997 use IO::File qw//;
2998
2999 # file baton members: path, mode_a, mode_b, pool, fh, blob, base
3000 sub new {
3001 my ($class, $git_svn) = @_;
3002 my $self = SVN::Delta::Editor->new;
3003 bless $self, $class;
3004 $self->{c} = $git_svn->{last_commit} if exists $git_svn->{last_commit};
3005 $self->{empty} = {};
3006 $self->{dir_prop} = {};
3007 $self->{file_prop} = {};
3008 $self->{absent_dir} = {};
3009 $self->{absent_file} = {};
3010 $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
3011 $self;
3012 }
3013
3014 sub set_path_strip {
3015 my ($self, $path) = @_;
3016 $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
3017 }
3018
3019 sub open_root {
3020 { path => '' };
3021 }
3022
3023 sub open_directory {
3024 my ($self, $path, $pb, $rev) = @_;
3025 { path => $path };
3026 }
3027
3028 sub git_path {
3029 my ($self, $path) = @_;
3030 if ($self->{path_strip}) {
3031 $path =~ s!$self->{path_strip}!! or
3032 die "Failed to strip path '$path' ($self->{path_strip})\n";
3033 }
3034 $path;
3035 }
3036
3037 sub delete_entry {
3038 my ($self, $path, $rev, $pb) = @_;
3039
3040 my $gpath = $self->git_path($path);
3041 return undef if ($gpath eq '');
3042
3043 # remove entire directories.
3044 if (command('ls-tree', $self->{c}, '--', $gpath) =~ /^040000 tree/) {
3045 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3046 -r --name-only -z/,
3047 $self->{c}, '--', $gpath);
3048 local $/ = "\0";
3049 while (<$ls>) {
3050 chomp;
3051 $self->{gii}->remove($_);
3052 print "\tD\t$_\n" unless $::_q;
3053 }
3054 print "\tD\t$gpath/\n" unless $::_q;
3055 command_close_pipe($ls, $ctx);
3056 $self->{empty}->{$path} = 0
3057 } else {
3058 $self->{gii}->remove($gpath);
3059 print "\tD\t$gpath\n" unless $::_q;
3060 }
3061 undef;
3062 }
3063
3064 sub open_file {
3065 my ($self, $path, $pb, $rev) = @_;
3066 my $gpath = $self->git_path($path);
3067 my ($mode, $blob) = (command('ls-tree', $self->{c}, '--', $gpath)
3068 =~ /^(\d{6}) blob ([a-f\d]{40})\t/);
3069 unless (defined $mode && defined $blob) {
3070 die "$path was not found in commit $self->{c} (r$rev)\n";
3071 }
3072 { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3073 pool => SVN::Pool->new, action => 'M' };
3074 }
3075
3076 sub add_file {
3077 my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3078 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3079 delete $self->{empty}->{$dir};
3080 { path => $path, mode_a => 100644, mode_b => 100644,
3081 pool => SVN::Pool->new, action => 'A' };
3082 }
3083
3084 sub add_directory {
3085 my ($self, $path, $cp_path, $cp_rev) = @_;
3086 my $gpath = $self->git_path($path);
3087 if ($gpath eq '') {
3088 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3089 -r --name-only -z/,
3090 $self->{c});
3091 local $/ = "\0";
3092 while (<$ls>) {
3093 chomp;
3094 $self->{gii}->remove($_);
3095 print "\tD\t$_\n" unless $::_q;
3096 }
3097 command_close_pipe($ls, $ctx);
3098 $self->{empty}->{$path} = 0;
3099 }
3100 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3101 delete $self->{empty}->{$dir};
3102 $self->{empty}->{$path} = 1;
3103 { path => $path };
3104 }
3105
3106 sub change_dir_prop {
3107 my ($self, $db, $prop, $value) = @_;
3108 $self->{dir_prop}->{$db->{path}} ||= {};
3109 $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3110 undef;
3111 }
3112
3113 sub absent_directory {
3114 my ($self, $path, $pb) = @_;
3115 $self->{absent_dir}->{$pb->{path}} ||= [];
3116 push @{$self->{absent_dir}->{$pb->{path}}}, $path;
3117 undef;
3118 }
3119
3120 sub absent_file {
3121 my ($self, $path, $pb) = @_;
3122 $self->{absent_file}->{$pb->{path}} ||= [];
3123 push @{$self->{absent_file}->{$pb->{path}}}, $path;
3124 undef;
3125 }
3126
3127 sub change_file_prop {
3128 my ($self, $fb, $prop, $value) = @_;
3129 if ($prop eq 'svn:executable') {
3130 if ($fb->{mode_b} != 120000) {
3131 $fb->{mode_b} = defined $value ? 100755 : 100644;
3132 }
3133 } elsif ($prop eq 'svn:special') {
3134 $fb->{mode_b} = defined $value ? 120000 : 100644;
3135 } else {
3136 $self->{file_prop}->{$fb->{path}} ||= {};
3137 $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
3138 }
3139 undef;
3140 }
3141
3142 sub apply_textdelta {
3143 my ($self, $fb, $exp) = @_;
3144 my $fh = IO::File->new_tmpfile;
3145 $fh->autoflush(1);
3146 # $fh gets auto-closed() by SVN::TxDelta::apply(),
3147 # (but $base does not,) so dup() it for reading in close_file
3148 open my $dup, '<&', $fh or croak $!;
3149 my $base = IO::File->new_tmpfile;
3150 $base->autoflush(1);
3151 if ($fb->{blob}) {
3152 defined (my $pid = fork) or croak $!;
3153 if (!$pid) {
3154 open STDOUT, '>&', $base or croak $!;
3155 print STDOUT 'link ' if ($fb->{mode_a} == 120000);
3156 exec qw/git-cat-file blob/, $fb->{blob} or croak $!;
3157 }
3158 waitpid $pid, 0;
3159 croak $? if $?;
3160
3161 if (defined $exp) {
3162 seek $base, 0, 0 or croak $!;
3163 my $got = ::md5sum($base);
3164 die "Checksum mismatch: $fb->{path} $fb->{blob}\n",
3165 "expected: $exp\n",
3166 " got: $got\n" if ($got ne $exp);
3167 }
3168 }
3169 seek $base, 0, 0 or croak $!;
3170 $fb->{fh} = $dup;
3171 $fb->{base} = $base;
3172 [ SVN::TxDelta::apply($base, $fh, undef, $fb->{path}, $fb->{pool}) ];
3173 }
3174
3175 sub close_file {
3176 my ($self, $fb, $exp) = @_;
3177 my $hash;
3178 my $path = $self->git_path($fb->{path});
3179 if (my $fh = $fb->{fh}) {
3180 if (defined $exp) {
3181 seek($fh, 0, 0) or croak $!;
3182 my $got = ::md5sum($fh);
3183 if ($got ne $exp) {
3184 die "Checksum mismatch: $path\n",
3185 "expected: $exp\n got: $got\n";
3186 }
3187 }
3188 sysseek($fh, 0, 0) or croak $!;
3189 if ($fb->{mode_b} == 120000) {
3190 eval {
3191 sysread($fh, my $buf, 5) == 5 or croak $!;
3192 $buf eq 'link ' or die "$path has mode 120000",
3193 " but is not a link";
3194 };
3195 if ($@) {
3196 warn "$@\n";
3197 sysseek($fh, 0, 0) or croak $!;
3198 }
3199 }
3200 defined(my $pid = open my $out,'-|') or die "Can't fork: $!\n";
3201 if (!$pid) {
3202 open STDIN, '<&', $fh or croak $!;
3203 exec qw/git-hash-object -w --stdin/ or croak $!;
3204 }
3205 chomp($hash = do { local $/; <$out> });
3206 close $out or croak $!;
3207 close $fh or croak $!;
3208 $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
3209 close $fb->{base} or croak $!;
3210 } else {
3211 $hash = $fb->{blob} or die "no blob information\n";
3212 }
3213 $fb->{pool}->clear;
3214 $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
3215 print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
3216 undef;
3217 }
3218
3219 sub abort_edit {
3220 my $self = shift;
3221 $self->{nr} = $self->{gii}->{nr};
3222 delete $self->{gii};
3223 $self->SUPER::abort_edit(@_);
3224 }
3225
3226 sub close_edit {
3227 my $self = shift;
3228 $self->{git_commit_ok} = 1;
3229 $self->{nr} = $self->{gii}->{nr};
3230 delete $self->{gii};
3231 $self->SUPER::close_edit(@_);
3232 }
3233
3234 package SVN::Git::Editor;
3235 use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
3236 use strict;
3237 use warnings;
3238 use Carp qw/croak/;
3239 use IO::File;
3240
3241 sub new {
3242 my ($class, $opts) = @_;
3243 foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
3244 die "$_ required!\n" unless (defined $opts->{$_});
3245 }
3246
3247 my $pool = SVN::Pool->new;
3248 my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
3249 my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
3250 $opts->{r}, $mods);
3251
3252 # $opts->{ra} functions should not be used after this:
3253 my @ce = $opts->{ra}->get_commit_editor($opts->{log},
3254 $opts->{editor_cb}, $pool);
3255 my $self = SVN::Delta::Editor->new(@ce, $pool);
3256 bless $self, $class;
3257 foreach (qw/svn_path r tree_a tree_b/) {
3258 $self->{$_} = $opts->{$_};
3259 }
3260 $self->{url} = $opts->{ra}->{url};
3261 $self->{mods} = $mods;
3262 $self->{types} = $types;
3263 $self->{pool} = $pool;
3264 $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
3265 $self->{rm} = { };
3266 $self->{path_prefix} = length $self->{svn_path} ?
3267 "$self->{svn_path}/" : '';
3268 return $self;
3269 }
3270
3271 sub generate_diff {
3272 my ($tree_a, $tree_b) = @_;
3273 my @diff_tree = qw(diff-tree -z -r);
3274 if ($_cp_similarity) {
3275 push @diff_tree, "-C$_cp_similarity";
3276 } else {
3277 push @diff_tree, '-C';
3278 }
3279 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
3280 push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
3281 push @diff_tree, $tree_a, $tree_b;
3282 my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
3283 local $/ = "\0";
3284 my $state = 'meta';
3285 my @mods;
3286 while (<$diff_fh>) {
3287 chomp $_; # this gets rid of the trailing "\0"
3288 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
3289 $::sha1\s($::sha1)\s
3290 ([MTCRAD])\d*$/xo) {
3291 push @mods, { mode_a => $1, mode_b => $2,
3292 sha1_b => $3, chg => $4 };
3293 if ($4 =~ /^(?:C|R)$/) {
3294 $state = 'file_a';
3295 } else {
3296 $state = 'file_b';
3297 }
3298 } elsif ($state eq 'file_a') {
3299 my $x = $mods[$#mods] or croak "Empty array\n";
3300 if ($x->{chg} !~ /^(?:C|R)$/) {
3301 croak "Error parsing $_, $x->{chg}\n";
3302 }
3303 $x->{file_a} = $_;
3304 $state = 'file_b';
3305 } elsif ($state eq 'file_b') {
3306 my $x = $mods[$#mods] or croak "Empty array\n";
3307 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
3308 croak "Error parsing $_, $x->{chg}\n";
3309 }
3310 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
3311 croak "Error parsing $_, $x->{chg}\n";
3312 }
3313 $x->{file_b} = $_;
3314 $state = 'meta';
3315 } else {
3316 croak "Error parsing $_\n";
3317 }
3318 }
3319 command_close_pipe($diff_fh, $ctx);
3320 \@mods;
3321 }
3322
3323 sub check_diff_paths {
3324 my ($ra, $pfx, $rev, $mods) = @_;
3325 my %types;
3326 $pfx .= '/' if length $pfx;
3327
3328 sub type_diff_paths {
3329 my ($ra, $types, $path, $rev) = @_;
3330 my @p = split m#/+#, $path;
3331 my $c = shift @p;
3332 unless (defined $types->{$c}) {
3333 $types->{$c} = $ra->check_path($c, $rev);
3334 }
3335 while (@p) {
3336 $c .= '/' . shift @p;
3337 next if defined $types->{$c};
3338 $types->{$c} = $ra->check_path($c, $rev);
3339 }
3340 }
3341
3342 foreach my $m (@$mods) {
3343 foreach my $f (qw/file_a file_b/) {
3344 next unless defined $m->{$f};
3345 my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
3346 if (length $pfx.$dir && ! defined $types{$dir}) {
3347 type_diff_paths($ra, \%types, $pfx.$dir, $rev);
3348 }
3349 }
3350 }
3351 \%types;
3352 }
3353
3354 sub split_path {
3355 return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
3356 }
3357
3358 sub repo_path {
3359 my ($self, $path) = @_;
3360 $self->{path_prefix}.(defined $path ? $path : '');
3361 }
3362
3363 sub url_path {
3364 my ($self, $path) = @_;
3365 if ($self->{url} =~ m#^https?://#) {
3366 $path =~ s/([^a-zA-Z0-9_.-])/uc sprintf("%%%02x",ord($1))/eg;
3367 }
3368 $self->{url} . '/' . $self->repo_path($path);
3369 }
3370
3371 sub rmdirs {
3372 my ($self) = @_;
3373 my $rm = $self->{rm};
3374 delete $rm->{''}; # we never delete the url we're tracking
3375 return unless %$rm;
3376
3377 foreach (keys %$rm) {
3378 my @d = split m#/#, $_;
3379 my $c = shift @d;
3380 $rm->{$c} = 1;
3381 while (@d) {
3382 $c .= '/' . shift @d;
3383 $rm->{$c} = 1;
3384 }
3385 }
3386 delete $rm->{$self->{svn_path}};
3387 delete $rm->{''}; # we never delete the url we're tracking
3388 return unless %$rm;
3389
3390 my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
3391 $self->{tree_b});
3392 local $/ = "\0";
3393 while (<$fh>) {
3394 chomp;
3395 my @dn = split m#/#, $_;
3396 while (pop @dn) {
3397 delete $rm->{join '/', @dn};
3398 }
3399 unless (%$rm) {
3400 close $fh;
3401 return;
3402 }
3403 }
3404 command_close_pipe($fh, $ctx);
3405
3406 my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
3407 foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
3408 $self->close_directory($bat->{$d}, $p);
3409 my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
3410 print "\tD+\t$d/\n" unless $::_q;
3411 $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
3412 delete $bat->{$d};
3413 }
3414 }
3415
3416 sub open_or_add_dir {
3417 my ($self, $full_path, $baton) = @_;
3418 my $t = $self->{types}->{$full_path};
3419 if (!defined $t) {
3420 die "$full_path not known in r$self->{r} or we have a bug!\n";
3421 }
3422 {
3423 no warnings 'once';
3424 # SVN::Node::none and SVN::Node::file are used only once,
3425 # so we're shutting up Perl's warnings about them.
3426 if ($t == $SVN::Node::none) {
3427 return $self->add_directory($full_path, $baton,
3428 undef, -1, $self->{pool});
3429 } elsif ($t == $SVN::Node::dir) {
3430 return $self->open_directory($full_path, $baton,
3431 $self->{r}, $self->{pool});
3432 } # no warnings 'once'
3433 print STDERR "$full_path already exists in repository at ",
3434 "r$self->{r} and it is not a directory (",
3435 ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
3436 } # no warnings 'once'
3437 exit 1;
3438 }
3439
3440 sub ensure_path {
3441 my ($self, $path) = @_;
3442 my $bat = $self->{bat};
3443 my $repo_path = $self->repo_path($path);
3444 return $bat->{''} unless (length $repo_path);
3445 my @p = split m#/+#, $repo_path;
3446 my $c = shift @p;
3447 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
3448 while (@p) {
3449 my $c0 = $c;
3450 $c .= '/' . shift @p;
3451 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
3452 }
3453 return $bat->{$c};
3454 }
3455
3456 sub A {
3457 my ($self, $m) = @_;
3458 my ($dir, $file) = split_path($m->{file_b});
3459 my $pbat = $self->ensure_path($dir);
3460 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3461 undef, -1);
3462 print "\tA\t$m->{file_b}\n" unless $::_q;
3463 $self->chg_file($fbat, $m);
3464 $self->close_file($fbat,undef,$self->{pool});
3465 }
3466
3467 sub C {
3468 my ($self, $m) = @_;
3469 my ($dir, $file) = split_path($m->{file_b});
3470 my $pbat = $self->ensure_path($dir);
3471 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3472 $self->url_path($m->{file_a}), $self->{r});
3473 print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3474 $self->chg_file($fbat, $m);
3475 $self->close_file($fbat,undef,$self->{pool});
3476 }
3477
3478 sub delete_entry {
3479 my ($self, $path, $pbat) = @_;
3480 my $rpath = $self->repo_path($path);
3481 my ($dir, $file) = split_path($rpath);
3482 $self->{rm}->{$dir} = 1;
3483 $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
3484 }
3485
3486 sub R {
3487 my ($self, $m) = @_;
3488 my ($dir, $file) = split_path($m->{file_b});
3489 my $pbat = $self->ensure_path($dir);
3490 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3491 $self->url_path($m->{file_a}), $self->{r});
3492 print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3493 $self->chg_file($fbat, $m);
3494 $self->close_file($fbat,undef,$self->{pool});
3495
3496 ($dir, $file) = split_path($m->{file_a});
3497 $pbat = $self->ensure_path($dir);
3498 $self->delete_entry($m->{file_a}, $pbat);
3499 }
3500
3501 sub M {
3502 my ($self, $m) = @_;
3503 my ($dir, $file) = split_path($m->{file_b});
3504 my $pbat = $self->ensure_path($dir);
3505 my $fbat = $self->open_file($self->repo_path($m->{file_b}),
3506 $pbat,$self->{r},$self->{pool});
3507 print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
3508 $self->chg_file($fbat, $m);
3509 $self->close_file($fbat,undef,$self->{pool});
3510 }
3511
3512 sub T { shift->M(@_) }
3513
3514 sub change_file_prop {
3515 my ($self, $fbat, $pname, $pval) = @_;
3516 $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
3517 }
3518
3519 sub chg_file {
3520 my ($self, $fbat, $m) = @_;
3521 if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
3522 $self->change_file_prop($fbat,'svn:executable','*');
3523 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
3524 $self->change_file_prop($fbat,'svn:executable',undef);
3525 }
3526 my $fh = IO::File->new_tmpfile or croak $!;
3527 if ($m->{mode_b} =~ /^120/) {
3528 print $fh 'link ' or croak $!;
3529 $self->change_file_prop($fbat,'svn:special','*');
3530 } elsif ($m->{mode_a} =~ /^120/ && $m->{mode_b} !~ /^120/) {
3531 $self->change_file_prop($fbat,'svn:special',undef);
3532 }
3533 defined(my $pid = fork) or croak $!;
3534 if (!$pid) {
3535 open STDOUT, '>&', $fh or croak $!;
3536 exec qw/git-cat-file blob/, $m->{sha1_b} or croak $!;
3537 }
3538 waitpid $pid, 0;
3539 croak $? if $?;
3540 $fh->flush == 0 or croak $!;
3541 seek $fh, 0, 0 or croak $!;
3542
3543 my $exp = ::md5sum($fh);
3544 seek $fh, 0, 0 or croak $!;
3545
3546 my $pool = SVN::Pool->new;
3547 my $atd = $self->apply_textdelta($fbat, undef, $pool);
3548 my $got = SVN::TxDelta::send_stream($fh, @$atd, $pool);
3549 die "Checksum mismatch\nexpected: $exp\ngot: $got\n" if ($got ne $exp);
3550 $pool->clear;
3551
3552 close $fh or croak $!;
3553 }
3554
3555 sub D {
3556 my ($self, $m) = @_;
3557 my ($dir, $file) = split_path($m->{file_b});
3558 my $pbat = $self->ensure_path($dir);
3559 print "\tD\t$m->{file_b}\n" unless $::_q;
3560 $self->delete_entry($m->{file_b}, $pbat);
3561 }
3562
3563 sub close_edit {
3564 my ($self) = @_;
3565 my ($p,$bat) = ($self->{pool}, $self->{bat});
3566 foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
3567 next if $_ eq '';
3568 $self->close_directory($bat->{$_}, $p);
3569 }
3570 $self->close_directory($bat->{''}, $p);
3571 $self->SUPER::close_edit($p);
3572 $p->clear;
3573 }
3574
3575 sub abort_edit {
3576 my ($self) = @_;
3577 $self->SUPER::abort_edit($self->{pool});
3578 }
3579
3580 sub DESTROY {
3581 my $self = shift;
3582 $self->SUPER::DESTROY(@_);
3583 $self->{pool}->clear;
3584 }
3585
3586 # this drives the editor
3587 sub apply_diff {
3588 my ($self) = @_;
3589 my $mods = $self->{mods};
3590 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
3591 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
3592 my $f = $m->{chg};
3593 if (defined $o{$f}) {
3594 $self->$f($m);
3595 } else {
3596 fatal("Invalid change type: $f");
3597 }
3598 }
3599 $self->rmdirs if $_rmdir;
3600 if (@$mods == 0) {
3601 $self->abort_edit;
3602 } else {
3603 $self->close_edit;
3604 }
3605 return scalar @$mods;
3606 }
3607
3608 package Git::SVN::Ra;
3609 use vars qw/@ISA $config_dir $_log_window_size/;
3610 use strict;
3611 use warnings;
3612 my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
3613
3614 BEGIN {
3615 # enforce temporary pool usage for some simple functions
3616 no strict 'refs';
3617 for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root/) {
3618 my $SUPER = "SUPER::$f";
3619 *$f = sub {
3620 my $self = shift;
3621 my $pool = SVN::Pool->new;
3622 my @ret = $self->$SUPER(@_,$pool);
3623 $pool->clear;
3624 wantarray ? @ret : $ret[0];
3625 };
3626 }
3627 }
3628
3629 sub _auth_providers () {
3630 [
3631 SVN::Client::get_simple_provider(),
3632 SVN::Client::get_ssl_server_trust_file_provider(),
3633 SVN::Client::get_simple_prompt_provider(
3634 \&Git::SVN::Prompt::simple, 2),
3635 SVN::Client::get_ssl_client_cert_file_provider(),
3636 SVN::Client::get_ssl_client_cert_prompt_provider(
3637 \&Git::SVN::Prompt::ssl_client_cert, 2),
3638 SVN::Client::get_ssl_client_cert_pw_prompt_provider(
3639 \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
3640 SVN::Client::get_username_provider(),
3641 SVN::Client::get_ssl_server_trust_prompt_provider(
3642 \&Git::SVN::Prompt::ssl_server_trust),
3643 SVN::Client::get_username_prompt_provider(
3644 \&Git::SVN::Prompt::username, 2)
3645 ]
3646 }
3647
3648 sub escape_uri_only {
3649 my ($uri) = @_;
3650 my @tmp;
3651 foreach (split m{/}, $uri) {
3652 s/([^\w.-])/sprintf("%%%02X",ord($1))/eg;
3653 push @tmp, $_;
3654 }
3655 join('/', @tmp);
3656 }
3657
3658 sub escape_url {
3659 my ($url) = @_;
3660 if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
3661 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
3662 $url = "$scheme://$domain$uri";
3663 }
3664 $url;
3665 }
3666
3667 sub new {
3668 my ($class, $url) = @_;
3669 $url =~ s!/+$!!;
3670 return $RA if ($RA && $RA->{url} eq $url);
3671
3672 SVN::_Core::svn_config_ensure($config_dir, undef);
3673 my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
3674 my $config = SVN::Core::config_get_config($config_dir);
3675 $RA = undef;
3676 my $dont_store_passwords = 1;
3677 my $conf_t = ${$config}{'config'};
3678 {
3679 no warnings 'once';
3680 # The usage of $SVN::_Core::SVN_CONFIG_* variables
3681 # produces warnings that variables are used only once.
3682 # I had not found the better way to shut them up, so
3683 # the warnings of type 'once' are disabled in this block.
3684 if (SVN::_Core::svn_config_get_bool($conf_t,
3685 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3686 $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
3687 1) == 0) {
3688 SVN::_Core::svn_auth_set_parameter($baton,
3689 $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
3690 bless (\$dont_store_passwords, "_p_void"));
3691 }
3692 if (SVN::_Core::svn_config_get_bool($conf_t,
3693 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3694 $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
3695 1) == 0) {
3696 $Git::SVN::Prompt::_no_auth_cache = 1;
3697 }
3698 } # no warnings 'once'
3699 my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
3700 config => $config,
3701 pool => SVN::Pool->new,
3702 auth_provider_callbacks => $callbacks);
3703 $self->{url} = $url;
3704 $self->{svn_path} = $url;
3705 $self->{repos_root} = $self->get_repos_root;
3706 $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
3707 $self->{cache} = { check_path => { r => 0, data => {} },
3708 get_dir => { r => 0, data => {} } };
3709 $RA = bless $self, $class;
3710 }
3711
3712 sub check_path {
3713 my ($self, $path, $r) = @_;
3714 my $cache = $self->{cache}->{check_path};
3715 if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
3716 return $cache->{data}->{$path};
3717 }
3718 my $pool = SVN::Pool->new;
3719 my $t = $self->SUPER::check_path($path, $r, $pool);
3720 $pool->clear;
3721 if ($r != $cache->{r}) {
3722 %{$cache->{data}} = ();
3723 $cache->{r} = $r;
3724 }
3725 $cache->{data}->{$path} = $t;
3726 }
3727
3728 sub get_dir {
3729 my ($self, $dir, $r) = @_;
3730 my $cache = $self->{cache}->{get_dir};
3731 if ($r == $cache->{r}) {
3732 if (my $x = $cache->{data}->{$dir}) {
3733 return wantarray ? @$x : $x->[0];
3734 }
3735 }
3736 my $pool = SVN::Pool->new;
3737 my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
3738 my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
3739 $pool->clear;
3740 if ($r != $cache->{r}) {
3741 %{$cache->{data}} = ();
3742 $cache->{r} = $r;
3743 }
3744 $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
3745 wantarray ? (\%dirents, $r, $props) : \%dirents;
3746 }
3747
3748 sub DESTROY {
3749 # do not call the real DESTROY since we store ourselves in $RA
3750 }
3751
3752 sub get_log {
3753 my ($self, @args) = @_;
3754 my $pool = SVN::Pool->new;
3755 splice(@args, 3, 1) if ($SVN::Core::VERSION le '1.2.0');
3756 my $ret = $self->SUPER::get_log(@args, $pool);
3757 $pool->clear;
3758 $ret;
3759 }
3760
3761 sub trees_match {
3762 my ($self, $url1, $rev1, $url2, $rev2) = @_;
3763 my $ctx = SVN::Client->new(auth => _auth_providers);
3764 my $out = IO::File->new_tmpfile;
3765
3766 # older SVN (1.1.x) doesn't take $pool as the last parameter for
3767 # $ctx->diff(), so we'll create a default one
3768 my $pool = SVN::Pool->new_default_sub;
3769
3770 $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
3771 $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
3772 $out->flush;
3773 my $ret = (($out->stat)[7] == 0);
3774 close $out or croak $!;
3775
3776 $ret;
3777 }
3778
3779 sub get_commit_editor {
3780 my ($self, $log, $cb, $pool) = @_;
3781 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
3782 $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
3783 }
3784
3785 sub gs_do_update {
3786 my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
3787 my $new = ($rev_a == $rev_b);
3788 my $path = $gs->{path};
3789
3790 if ($new && -e $gs->{index}) {
3791 unlink $gs->{index} or die
3792 "Couldn't unlink index: $gs->{index}: $!\n";
3793 }
3794 my $pool = SVN::Pool->new;
3795 $editor->set_path_strip($path);
3796 my (@pc) = split m#/#, $path;
3797 my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
3798 1, $editor, $pool);
3799 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3800
3801 # Since we can't rely on svn_ra_reparent being available, we'll
3802 # just have to do some magic with set_path to make it so
3803 # we only want a partial path.
3804 my $sp = '';
3805 my $final = join('/', @pc);
3806 while (@pc) {
3807 $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
3808 $sp .= '/' if length $sp;
3809 $sp .= shift @pc;
3810 }
3811 die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
3812
3813 $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
3814
3815 $reporter->finish_report($pool);
3816 $pool->clear;
3817 $editor->{git_commit_ok};
3818 }
3819
3820 # this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
3821 # svn_ra_reparent didn't work before 1.4)
3822 sub gs_do_switch {
3823 my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
3824 my $path = $gs->{path};
3825 my $pool = SVN::Pool->new;
3826
3827 my $full_url = $self->{url};
3828 my $old_url = $full_url;
3829 $full_url .= '/' . escape_uri_only($path) if length $path;
3830 my ($ra, $reparented);
3831 if ($old_url ne $full_url) {
3832 if ($old_url !~ m#^svn(\+ssh)?://#) {
3833 SVN::_Ra::svn_ra_reparent($self->{session}, $full_url,
3834 $pool);
3835 $self->{url} = $full_url;
3836 $reparented = 1;
3837 } else {
3838 $_[0] = undef;
3839 $self = undef;
3840 $RA = undef;
3841 $ra = Git::SVN::Ra->new($full_url);
3842 $ra_invalid = 1;
3843 }
3844 }
3845 $ra ||= $self;
3846 my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
3847 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3848 $reporter->set_path('', $rev_a, 0, @lock, $pool);
3849 $reporter->finish_report($pool);
3850
3851 if ($reparented) {
3852 SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
3853 $self->{url} = $old_url;
3854 }
3855
3856 $pool->clear;
3857 $editor->{git_commit_ok};
3858 }
3859
3860 sub longest_common_path {
3861 my ($gsv, $globs) = @_;
3862 my %common;
3863 my $common_max = scalar @$gsv;
3864
3865 foreach my $gs (@$gsv) {
3866 my @tmp = split m#/#, $gs->{path};
3867 my $p = '';
3868 foreach (@tmp) {
3869 $p .= length($p) ? "/$_" : $_;
3870 $common{$p} ||= 0;
3871 $common{$p}++;
3872 }
3873 }
3874 $globs ||= [];
3875 $common_max += scalar @$globs;
3876 foreach my $glob (@$globs) {
3877 my @tmp = split m#/#, $glob->{path}->{left};
3878 my $p = '';
3879 foreach (@tmp) {
3880 $p .= length($p) ? "/$_" : $_;
3881 $common{$p} ||= 0;
3882 $common{$p}++;
3883 }
3884 }
3885
3886 my $longest_path = '';
3887 foreach (sort {length $b <=> length $a} keys %common) {
3888 if ($common{$_} == $common_max) {
3889 $longest_path = $_;
3890 last;
3891 }
3892 }
3893 $longest_path;
3894 }
3895
3896 sub gs_fetch_loop_common {
3897 my ($self, $base, $head, $gsv, $globs) = @_;
3898 return if ($base > $head);
3899 my $inc = $_log_window_size;
3900 my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
3901 my $longest_path = longest_common_path($gsv, $globs);
3902 my $ra_url = $self->{url};
3903 while (1) {
3904 my %revs;
3905 my $err;
3906 my $err_handler = $SVN::Error::handler;
3907 $SVN::Error::handler = sub {
3908 ($err) = @_;
3909 skip_unknown_revs($err);
3910 };
3911 sub _cb {
3912 my ($paths, $r, $author, $date, $log) = @_;
3913 [ dup_changed_paths($paths),
3914 { author => $author, date => $date, log => $log } ];
3915 }
3916 $self->get_log([$longest_path], $min, $max, 0, 1, 1,
3917 sub { $revs{$_[1]} = _cb(@_) });
3918 if ($err && $max >= $head) {
3919 print STDERR "Path '$longest_path' ",
3920 "was probably deleted:\n",
3921 $err->expanded_message,
3922 "\nWill attempt to follow ",
3923 "revisions r$min .. r$max ",
3924 "committed before the deletion\n";
3925 my $hi = $max;
3926 while (--$hi >= $min) {
3927 my $ok;
3928 $self->get_log([$longest_path], $min, $hi,
3929 0, 1, 1, sub {
3930 $ok ||= $_[1];
3931 $revs{$_[1]} = _cb(@_) });
3932 if ($ok) {
3933 print STDERR "r$min .. r$ok OK\n";
3934 last;
3935 }
3936 }
3937 }
3938 $SVN::Error::handler = $err_handler;
3939
3940 my %exists = map { $_->{path} => $_ } @$gsv;
3941 foreach my $r (sort {$a <=> $b} keys %revs) {
3942 my ($paths, $logged) = @{$revs{$r}};
3943
3944 foreach my $gs ($self->match_globs(\%exists, $paths,
3945 $globs, $r)) {
3946 if ($gs->rev_map_max >= $r) {
3947 next;
3948 }
3949 next unless $gs->match_paths($paths, $r);
3950 $gs->{logged_rev_props} = $logged;
3951 if (my $last_commit = $gs->last_commit) {
3952 $gs->assert_index_clean($last_commit);
3953 }
3954 my $log_entry = $gs->do_fetch($paths, $r);
3955 if ($log_entry) {
3956 $gs->do_git_commit($log_entry);
3957 }
3958 $INDEX_FILES{$gs->{index}} = 1;
3959 }
3960 foreach my $g (@$globs) {
3961 my $k = "svn-remote.$g->{remote}." .
3962 "$g->{t}-maxRev";
3963 Git::SVN::tmp_config($k, $r);
3964 }
3965 if ($ra_invalid) {
3966 $_[0] = undef;
3967 $self = undef;
3968 $RA = undef;
3969 $self = Git::SVN::Ra->new($ra_url);
3970 $ra_invalid = undef;
3971 }
3972 }
3973 # pre-fill the .rev_db since it'll eventually get filled in
3974 # with '0' x40 if something new gets committed
3975 foreach my $gs (@$gsv) {
3976 next if $gs->rev_map_max >= $max;
3977 next if defined $gs->rev_map_get($max);
3978 $gs->rev_map_set($max, 0 x40);
3979 }
3980 foreach my $g (@$globs) {
3981 my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
3982 Git::SVN::tmp_config($k, $max);
3983 }
3984 last if $max >= $head;
3985 $min = $max + 1;
3986 $max += $inc;
3987 $max = $head if ($max > $head);
3988 }
3989 Git::SVN::gc();
3990 }
3991
3992 sub match_globs {
3993 my ($self, $exists, $paths, $globs, $r) = @_;
3994
3995 sub get_dir_check {
3996 my ($self, $exists, $g, $r) = @_;
3997 my @x = eval { $self->get_dir($g->{path}->{left}, $r) };
3998 return unless scalar @x == 3;
3999 my $dirents = $x[0];
4000 foreach my $de (keys %$dirents) {
4001 next if $dirents->{$de}->{kind} != $SVN::Node::dir;
4002 my $p = $g->{path}->full_path($de);
4003 next if $exists->{$p};
4004 next if (length $g->{path}->{right} &&
4005 ($self->check_path($p, $r) !=
4006 $SVN::Node::dir));
4007 $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
4008 $g->{ref}->full_path($de), 1);
4009 }
4010 }
4011 foreach my $g (@$globs) {
4012 if (my $path = $paths->{"/$g->{path}->{left}"}) {
4013 if ($path->{action} =~ /^[AR]$/) {
4014 get_dir_check($self, $exists, $g, $r);
4015 }
4016 }
4017 foreach (keys %$paths) {
4018 if (/$g->{path}->{left_regex}/ &&
4019 !/$g->{path}->{regex}/) {
4020 next if $paths->{$_}->{action} !~ /^[AR]$/;
4021 get_dir_check($self, $exists, $g, $r);
4022 }
4023 next unless /$g->{path}->{regex}/;
4024 my $p = $1;
4025 my $pathname = $g->{path}->full_path($p);
4026 next if $exists->{$pathname};
4027 next if ($self->check_path($pathname, $r) !=
4028 $SVN::Node::dir);
4029 $exists->{$pathname} = Git::SVN->init(
4030 $self->{url}, $pathname, undef,
4031 $g->{ref}->full_path($p), 1);
4032 }
4033 my $c = '';
4034 foreach (split m#/#, $g->{path}->{left}) {
4035 $c .= "/$_";
4036 next unless ($paths->{$c} &&
4037 ($paths->{$c}->{action} =~ /^[AR]$/));
4038 get_dir_check($self, $exists, $g, $r);
4039 }
4040 }
4041 values %$exists;
4042 }
4043
4044 sub minimize_url {
4045 my ($self) = @_;
4046 return $self->{url} if ($self->{url} eq $self->{repos_root});
4047 my $url = $self->{repos_root};
4048 my @components = split(m!/!, $self->{svn_path});
4049 my $c = '';
4050 do {
4051 $url .= "/$c" if length $c;
4052 eval { (ref $self)->new($url)->get_latest_revnum };
4053 } while ($@ && ($c = shift @components));
4054 $url;
4055 }
4056
4057 sub can_do_switch {
4058 my $self = shift;
4059 unless (defined $can_do_switch) {
4060 my $pool = SVN::Pool->new;
4061 my $rep = eval {
4062 $self->do_switch(1, '', 0, $self->{url},
4063 SVN::Delta::Editor->new, $pool);
4064 };
4065 if ($@) {
4066 $can_do_switch = 0;
4067 } else {
4068 $rep->abort_report($pool);
4069 $can_do_switch = 1;
4070 }
4071 $pool->clear;
4072 }
4073 $can_do_switch;
4074 }
4075
4076 sub skip_unknown_revs {
4077 my ($err) = @_;
4078 my $errno = $err->apr_err();
4079 # Maybe the branch we're tracking didn't
4080 # exist when the repo started, so it's
4081 # not an error if it doesn't, just continue
4082 #
4083 # Wonderfully consistent library, eh?
4084 # 160013 - svn:// and file://
4085 # 175002 - http(s)://
4086 # 175007 - http(s):// (this repo required authorization, too...)
4087 # More codes may be discovered later...
4088 if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
4089 my $err_key = $err->expanded_message;
4090 # revision numbers change every time, filter them out
4091 $err_key =~ s/\d+/\0/g;
4092 $err_key = "$errno\0$err_key";
4093 unless ($ignored_err{$err_key}) {
4094 warn "W: Ignoring error from SVN, path probably ",
4095 "does not exist: ($errno): ",
4096 $err->expanded_message,"\n";
4097 warn "W: Do not be alarmed at the above message ",
4098 "git-svn is just searching aggressively for ",
4099 "old history.\n",
4100 "This may take a while on large repositories\n";
4101 $ignored_err{$err_key} = 1;
4102 }
4103 return;
4104 }
4105 die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
4106 }
4107
4108 # svn_log_changed_path_t objects passed to get_log are likely to be
4109 # overwritten even if only the refs are copied to an external variable,
4110 # so we should dup the structures in their entirety. Using an externally
4111 # passed pool (instead of our temporary and quickly cleared pool in
4112 # Git::SVN::Ra) does not help matters at all...
4113 sub dup_changed_paths {
4114 my ($paths) = @_;
4115 return undef unless $paths;
4116 my %ret;
4117 foreach my $p (keys %$paths) {
4118 my $i = $paths->{$p};
4119 my %s = map { $_ => $i->$_ }
4120 qw/copyfrom_path copyfrom_rev action/;
4121 $ret{$p} = \%s;
4122 }
4123 \%ret;
4124 }
4125
4126 package Git::SVN::Log;
4127 use strict;
4128 use warnings;
4129 use POSIX qw/strftime/;
4130 use constant commit_log_separator => ('-' x 72) . "\n";
4131 use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
4132 %rusers $show_commit $incremental/;
4133 my $l_fmt;
4134
4135 sub cmt_showable {
4136 my ($c) = @_;
4137 return 1 if defined $c->{r};
4138
4139 # big commit message got truncated by the 16k pretty buffer in rev-list
4140 if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
4141 $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
4142 @{$c->{l}} = ();
4143 my @log = command(qw/cat-file commit/, $c->{c});
4144
4145 # shift off the headers
4146 shift @log while ($log[0] ne '');
4147 shift @log;
4148
4149 # TODO: make $c->{l} not have a trailing newline in the future
4150 @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
4151
4152 (undef, $c->{r}, undef) = ::extract_metadata(
4153 (grep(/^git-svn-id: /, @log))[-1]);
4154 }
4155 return defined $c->{r};
4156 }
4157
4158 sub log_use_color {
4159 return $color || Git->repository->get_colorbool('color.diff');
4160 }
4161
4162 sub git_svn_log_cmd {
4163 my ($r_min, $r_max, @args) = @_;
4164 my $head = 'HEAD';
4165 my (@files, @log_opts);
4166 foreach my $x (@args) {
4167 if ($x eq '--' || @files) {
4168 push @files, $x;
4169 } else {
4170 if (::verify_ref("$x^0")) {
4171 $head = $x;
4172 } else {
4173 push @log_opts, $x;
4174 }
4175 }
4176 }
4177
4178 my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
4179 $gs ||= Git::SVN->_new;
4180 my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
4181 $gs->refname);
4182 push @cmd, '-r' unless $non_recursive;
4183 push @cmd, qw/--raw --name-status/ if $verbose;
4184 push @cmd, '--color' if log_use_color();
4185 push @cmd, @log_opts;
4186 if (defined $r_max && $r_max == $r_min) {
4187 push @cmd, '--max-count=1';
4188 if (my $c = $gs->rev_map_get($r_max)) {
4189 push @cmd, $c;
4190 }
4191 } elsif (defined $r_max) {
4192 if ($r_max < $r_min) {
4193 ($r_min, $r_max) = ($r_max, $r_min);
4194 }
4195 my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
4196 my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
4197 # If there are no commits in the range, both $c_max and $c_min
4198 # will be undefined. If there is at least 1 commit in the
4199 # range, both will be defined.
4200 return () if !defined $c_min || !defined $c_max;
4201 if ($c_min eq $c_max) {
4202 push @cmd, '--max-count=1', $c_min;
4203 } else {
4204 push @cmd, '--boundary', "$c_min..$c_max";
4205 }
4206 }
4207 return (@cmd, @files);
4208 }
4209
4210 # adapted from pager.c
4211 sub config_pager {
4212 $pager ||= $ENV{GIT_PAGER} || $ENV{PAGER};
4213 if (!defined $pager) {
4214 $pager = 'less';
4215 } elsif (length $pager == 0 || $pager eq 'cat') {
4216 $pager = undef;
4217 }
4218 $ENV{GIT_PAGER_IN_USE} = defined($pager);
4219 }
4220
4221 sub run_pager {
4222 return unless -t *STDOUT && defined $pager;
4223 pipe my $rfd, my $wfd or return;
4224 defined(my $pid = fork) or ::fatal "Can't fork: $!";
4225 if (!$pid) {
4226 open STDOUT, '>&', $wfd or
4227 ::fatal "Can't redirect to stdout: $!";
4228 return;
4229 }
4230 open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
4231 $ENV{LESS} ||= 'FRSX';
4232 exec $pager or ::fatal "Can't run pager: $! ($pager)";
4233 }
4234
4235 sub format_svn_date {
4236 return strftime("%Y-%m-%d %H:%M:%S %z (%a, %d %b %Y)", localtime(shift));
4237 }
4238
4239 sub parse_git_date {
4240 my ($t, $tz) = @_;
4241 # Date::Parse isn't in the standard Perl distro :(
4242 if ($tz =~ s/^\+//) {
4243 $t += tz_to_s_offset($tz);
4244 } elsif ($tz =~ s/^\-//) {
4245 $t -= tz_to_s_offset($tz);
4246 }
4247 return $t;
4248 }
4249
4250 sub set_local_timezone {
4251 if (defined $TZ) {
4252 $ENV{TZ} = $TZ;
4253 } else {
4254 delete $ENV{TZ};
4255 }
4256 }
4257
4258 sub tz_to_s_offset {
4259 my ($tz) = @_;
4260 $tz =~ s/(\d\d)$//;
4261 return ($1 * 60) + ($tz * 3600);
4262 }
4263
4264 sub get_author_info {
4265 my ($dest, $author, $t, $tz) = @_;
4266 $author =~ s/(?:^\s*|\s*$)//g;
4267 $dest->{a_raw} = $author;
4268 my $au;
4269 if ($::_authors) {
4270 $au = $rusers{$author} || undef;
4271 }
4272 if (!$au) {
4273 ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
4274 }
4275 $dest->{t} = $t;
4276 $dest->{tz} = $tz;
4277 $dest->{a} = $au;
4278 $dest->{t_utc} = parse_git_date($t, $tz);
4279 }
4280
4281 sub process_commit {
4282 my ($c, $r_min, $r_max, $defer) = @_;
4283 if (defined $r_min && defined $r_max) {
4284 if ($r_min == $c->{r} && $r_min == $r_max) {
4285 show_commit($c);
4286 return 0;
4287 }
4288 return 1 if $r_min == $r_max;
4289 if ($r_min < $r_max) {
4290 # we need to reverse the print order
4291 return 0 if (defined $limit && --$limit < 0);
4292 push @$defer, $c;
4293 return 1;
4294 }
4295 if ($r_min != $r_max) {
4296 return 1 if ($r_min < $c->{r});
4297 return 1 if ($r_max > $c->{r});
4298 }
4299 }
4300 return 0 if (defined $limit && --$limit < 0);
4301 show_commit($c);
4302 return 1;
4303 }
4304
4305 sub show_commit {
4306 my $c = shift;
4307 if ($oneline) {
4308 my $x = "\n";
4309 if (my $l = $c->{l}) {
4310 while ($l->[0] =~ /^\s*$/) { shift @$l }
4311 $x = $l->[0];
4312 }
4313 $l_fmt ||= 'A' . length($c->{r});
4314 print 'r',pack($l_fmt, $c->{r}),' | ';
4315 print "$c->{c} | " if $show_commit;
4316 print $x;
4317 } else {
4318 show_commit_normal($c);
4319 }
4320 }
4321
4322 sub show_commit_changed_paths {
4323 my ($c) = @_;
4324 return unless $c->{changed};
4325 print "Changed paths:\n", @{$c->{changed}};
4326 }
4327
4328 sub show_commit_normal {
4329 my ($c) = @_;
4330 print commit_log_separator, "r$c->{r} | ";
4331 print "$c->{c} | " if $show_commit;
4332 print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
4333 my $nr_line = 0;
4334
4335 if (my $l = $c->{l}) {
4336 while ($l->[$#$l] eq "\n" && $#$l > 0
4337 && $l->[($#$l - 1)] eq "\n") {
4338 pop @$l;
4339 }
4340 $nr_line = scalar @$l;
4341 if (!$nr_line) {
4342 print "1 line\n\n\n";
4343 } else {
4344 if ($nr_line == 1) {
4345 $nr_line = '1 line';
4346 } else {
4347 $nr_line .= ' lines';
4348 }
4349 print $nr_line, "\n";
4350 show_commit_changed_paths($c);
4351 print "\n";
4352 print $_ foreach @$l;
4353 }
4354 } else {
4355 print "1 line\n";
4356 show_commit_changed_paths($c);
4357 print "\n";
4358
4359 }
4360 foreach my $x (qw/raw stat diff/) {
4361 if ($c->{$x}) {
4362 print "\n";
4363 print $_ foreach @{$c->{$x}}
4364 }
4365 }
4366 }
4367
4368 sub cmd_show_log {
4369 my (@args) = @_;
4370 my ($r_min, $r_max);
4371 my $r_last = -1; # prevent dupes
4372 set_local_timezone();
4373 if (defined $::_revision) {
4374 if ($::_revision =~ /^(\d+):(\d+)$/) {
4375 ($r_min, $r_max) = ($1, $2);
4376 } elsif ($::_revision =~ /^\d+$/) {
4377 $r_min = $r_max = $::_revision;
4378 } else {
4379 ::fatal "-r$::_revision is not supported, use ",
4380 "standard 'git log' arguments instead";
4381 }
4382 }
4383
4384 config_pager();
4385 @args = git_svn_log_cmd($r_min, $r_max, @args);
4386 if (!@args) {
4387 print commit_log_separator unless $incremental || $oneline;
4388 return;
4389 }
4390 my $log = command_output_pipe(@args);
4391 run_pager();
4392 my (@k, $c, $d, $stat);
4393 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
4394 while (<$log>) {
4395 if (/^${esc_color}commit -?($::sha1_short)/o) {
4396 my $cmt = $1;
4397 if ($c && cmt_showable($c) && $c->{r} != $r_last) {
4398 $r_last = $c->{r};
4399 process_commit($c, $r_min, $r_max, \@k) or
4400 goto out;
4401 }
4402 $d = undef;
4403 $c = { c => $cmt };
4404 } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
4405 get_author_info($c, $1, $2, $3);
4406 } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
4407 # ignore
4408 } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
4409 push @{$c->{raw}}, $_;
4410 } elsif (/^${esc_color}[ACRMDT]\t/) {
4411 # we could add $SVN->{svn_path} here, but that requires
4412 # remote access at the moment (repo_path_split)...
4413 s#^(${esc_color})([ACRMDT])\t#$1 $2 #o;
4414 push @{$c->{changed}}, $_;
4415 } elsif (/^${esc_color}diff /o) {
4416 $d = 1;
4417 push @{$c->{diff}}, $_;
4418 } elsif ($d) {
4419 push @{$c->{diff}}, $_;
4420 } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
4421 $esc_color*[\+\-]*$esc_color$/x) {
4422 $stat = 1;
4423 push @{$c->{stat}}, $_;
4424 } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
4425 push @{$c->{stat}}, $_;
4426 $stat = undef;
4427 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
4428 ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
4429 } elsif (s/^${esc_color} //o) {
4430 push @{$c->{l}}, $_;
4431 }
4432 }
4433 if ($c && defined $c->{r} && $c->{r} != $r_last) {
4434 $r_last = $c->{r};
4435 process_commit($c, $r_min, $r_max, \@k);
4436 }
4437 if (@k) {
4438 ($r_min, $r_max) = ($r_max, $r_min);
4439 process_commit($_, $r_min, $r_max) foreach reverse @k;
4440 }
4441 out:
4442 close $log;
4443 print commit_log_separator unless $incremental || $oneline;
4444 }
4445
4446 package Git::SVN::Migration;
4447 # these version numbers do NOT correspond to actual version numbers
4448 # of git nor git-svn. They are just relative.
4449 #
4450 # v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
4451 #
4452 # v1 layout: .git/$id/info/url, refs/remotes/$id
4453 #
4454 # v2 layout: .git/svn/$id/info/url, refs/remotes/$id
4455 #
4456 # v3 layout: .git/svn/$id, refs/remotes/$id
4457 # - info/url may remain for backwards compatibility
4458 # - this is what we migrate up to this layout automatically,
4459 # - this will be used by git svn init on single branches
4460 # v3.1 layout (auto migrated):
4461 # - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
4462 # for backwards compatibility
4463 #
4464 # v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
4465 # - this is only created for newly multi-init-ed
4466 # repositories. Similar in spirit to the
4467 # --use-separate-remotes option in git-clone (now default)
4468 # - we do not automatically migrate to this (following
4469 # the example set by core git)
4470 #
4471 # v5 layout: .rev_db.$UUID => .rev_map.$UUID
4472 # - newer, more-efficient format that uses 24-bytes per record
4473 # with no filler space.
4474 # - use xxd -c24 < .rev_map.$UUID to view and debug
4475 # - This is a one-way migration, repositories updated to the
4476 # new format will not be able to use old git-svn without
4477 # rebuilding the .rev_db. Rebuilding the rev_db is not
4478 # possible if noMetadata or useSvmProps are set; but should
4479 # be no problem for users that use the (sensible) defaults.
4480 use strict;
4481 use warnings;
4482 use Carp qw/croak/;
4483 use File::Path qw/mkpath/;
4484 use File::Basename qw/dirname basename/;
4485 use vars qw/$_minimize/;
4486
4487 sub migrate_from_v0 {
4488 my $git_dir = $ENV{GIT_DIR};
4489 return undef unless -d $git_dir;
4490 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4491 my $migrated = 0;
4492 while (<$fh>) {
4493 chomp;
4494 my ($id, $orig_ref) = ($_, $_);
4495 next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
4496 next unless -f "$git_dir/$id/info/url";
4497 my $new_ref = "refs/remotes/$id";
4498 if (::verify_ref("$new_ref^0")) {
4499 print STDERR "W: $orig_ref is probably an old ",
4500 "branch used by an ancient version of ",
4501 "git-svn.\n",
4502 "However, $new_ref also exists.\n",
4503 "We will not be able ",
4504 "to use this branch until this ",
4505 "ambiguity is resolved.\n";
4506 next;
4507 }
4508 print STDERR "Migrating from v0 layout...\n" if !$migrated;
4509 print STDERR "Renaming ref: $orig_ref => $new_ref\n";
4510 command_noisy('update-ref', $new_ref, $orig_ref);
4511 command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
4512 $migrated++;
4513 }
4514 command_close_pipe($fh, $ctx);
4515 print STDERR "Done migrating from v0 layout...\n" if $migrated;
4516 $migrated;
4517 }
4518
4519 sub migrate_from_v1 {
4520 my $git_dir = $ENV{GIT_DIR};
4521 my $migrated = 0;
4522 return $migrated unless -d $git_dir;
4523 my $svn_dir = "$git_dir/svn";
4524
4525 # just in case somebody used 'svn' as their $id at some point...
4526 return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
4527
4528 print STDERR "Migrating from a git-svn v1 layout...\n";
4529 mkpath([$svn_dir]);
4530 print STDERR "Data from a previous version of git-svn exists, but\n\t",
4531 "$svn_dir\n\t(required for this version ",
4532 "($::VERSION) of git-svn) does not. exist\n";
4533 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4534 while (<$fh>) {
4535 my $x = $_;
4536 next unless $x =~ s#^refs/remotes/##;
4537 chomp $x;
4538 next unless -f "$git_dir/$x/info/url";
4539 my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
4540 next unless $u;
4541 my $dn = dirname("$git_dir/svn/$x");
4542 mkpath([$dn]) unless -d $dn;
4543 if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
4544 mkpath(["$git_dir/svn/svn"]);
4545 print STDERR " - $git_dir/$x/info => ",
4546 "$git_dir/svn/$x/info\n";
4547 rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
4548 croak "$!: $x";
4549 # don't worry too much about these, they probably
4550 # don't exist with repos this old (save for index,
4551 # and we can easily regenerate that)
4552 foreach my $f (qw/unhandled.log index .rev_db/) {
4553 rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
4554 }
4555 } else {
4556 print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
4557 rename "$git_dir/$x", "$git_dir/svn/$x" or
4558 croak "$!: $x";
4559 }
4560 $migrated++;
4561 }
4562 command_close_pipe($fh, $ctx);
4563 print STDERR "Done migrating from a git-svn v1 layout\n";
4564 $migrated;
4565 }
4566
4567 sub read_old_urls {
4568 my ($l_map, $pfx, $path) = @_;
4569 my @dir;
4570 foreach (<$path/*>) {
4571 if (-r "$_/info/url") {
4572 $pfx .= '/' if $pfx && $pfx !~ m!/$!;
4573 my $ref_id = $pfx . basename $_;
4574 my $url = ::file_to_s("$_/info/url");
4575 $l_map->{$ref_id} = $url;
4576 } elsif (-d $_) {
4577 push @dir, $_;
4578 }
4579 }
4580 foreach (@dir) {
4581 my $x = $_;
4582 $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
4583 read_old_urls($l_map, $x, $_);
4584 }
4585 }
4586
4587 sub migrate_from_v2 {
4588 my @cfg = command(qw/config -l/);
4589 return if grep /^svn-remote\..+\.url=/, @cfg;
4590 my %l_map;
4591 read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
4592 my $migrated = 0;
4593
4594 foreach my $ref_id (sort keys %l_map) {
4595 eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
4596 if ($@) {
4597 Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
4598 }
4599 $migrated++;
4600 }
4601 $migrated;
4602 }
4603
4604 sub minimize_connections {
4605 my $r = Git::SVN::read_all_remotes();
4606 my $new_urls = {};
4607 my $root_repos = {};
4608 foreach my $repo_id (keys %$r) {
4609 my $url = $r->{$repo_id}->{url} or next;
4610 my $fetch = $r->{$repo_id}->{fetch} or next;
4611 my $ra = Git::SVN::Ra->new($url);
4612
4613 # skip existing cases where we already connect to the root
4614 if (($ra->{url} eq $ra->{repos_root}) ||
4615 (Git::SVN::sanitize_remote_name($ra->{repos_root}) eq
4616 $repo_id)) {
4617 $root_repos->{$ra->{url}} = $repo_id;
4618 next;
4619 }
4620
4621 my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
4622 my $root_path = $ra->{url};
4623 $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
4624 foreach my $path (keys %$fetch) {
4625 my $ref_id = $fetch->{$path};
4626 my $gs = Git::SVN->new($ref_id, $repo_id, $path);
4627
4628 # make sure we can read when connecting to
4629 # a higher level of a repository
4630 my ($last_rev, undef) = $gs->last_rev_commit;
4631 if (!defined $last_rev) {
4632 $last_rev = eval {
4633 $root_ra->get_latest_revnum;
4634 };
4635 next if $@;
4636 }
4637 my $new = $root_path;
4638 $new .= length $path ? "/$path" : '';
4639 eval {
4640 $root_ra->get_log([$new], $last_rev, $last_rev,
4641 0, 0, 1, sub { });
4642 };
4643 next if $@;
4644 $new_urls->{$ra->{repos_root}}->{$new} =
4645 { ref_id => $ref_id,
4646 old_repo_id => $repo_id,
4647 old_path => $path };
4648 }
4649 }
4650
4651 my @emptied;
4652 foreach my $url (keys %$new_urls) {
4653 # see if we can re-use an existing [svn-remote "repo_id"]
4654 # instead of creating a(n ugly) new section:
4655 my $repo_id = $root_repos->{$url} ||
4656 Git::SVN::sanitize_remote_name($url);
4657
4658 my $fetch = $new_urls->{$url};
4659 foreach my $path (keys %$fetch) {
4660 my $x = $fetch->{$path};
4661 Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
4662 my $pfx = "svn-remote.$x->{old_repo_id}";
4663
4664 my $old_fetch = quotemeta("$x->{old_path}:".
4665 "refs/remotes/$x->{ref_id}");
4666 command_noisy(qw/config --unset/,
4667 "$pfx.fetch", '^'. $old_fetch . '$');
4668 delete $r->{$x->{old_repo_id}}->
4669 {fetch}->{$x->{old_path}};
4670 if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
4671 command_noisy(qw/config --unset/,
4672 "$pfx.url");
4673 push @emptied, $x->{old_repo_id}
4674 }
4675 }
4676 }
4677 if (@emptied) {
4678 my $file = $ENV{GIT_CONFIG} || $ENV{GIT_CONFIG_LOCAL} ||
4679 "$ENV{GIT_DIR}/config";
4680 print STDERR <<EOF;
4681 The following [svn-remote] sections in your config file ($file) are empty
4682 and can be safely removed:
4683 EOF
4684 print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
4685 }
4686 }
4687
4688 sub migration_check {
4689 migrate_from_v0();
4690 migrate_from_v1();
4691 migrate_from_v2();
4692 minimize_connections() if $_minimize;
4693 }
4694
4695 package Git::IndexInfo;
4696 use strict;
4697 use warnings;
4698 use Git qw/command_input_pipe command_close_pipe/;
4699
4700 sub new {
4701 my ($class) = @_;
4702 my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
4703 bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
4704 }
4705
4706 sub remove {
4707 my ($self, $path) = @_;
4708 if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
4709 return ++$self->{nr};
4710 }
4711 undef;
4712 }
4713
4714 sub update {
4715 my ($self, $mode, $hash, $path) = @_;
4716 if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
4717 return ++$self->{nr};
4718 }
4719 undef;
4720 }
4721
4722 sub DESTROY {
4723 my ($self) = @_;
4724 command_close_pipe($self->{gui}, $self->{ctx});
4725 }
4726
4727 package Git::SVN::GlobSpec;
4728 use strict;
4729 use warnings;
4730
4731 sub new {
4732 my ($class, $glob) = @_;
4733 my $re = $glob;
4734 $re =~ s!/+$!!g; # no need for trailing slashes
4735 my $nr = ($re =~ s!^(.*)\*(.*)$!\(\[^/\]+\)!g);
4736 my ($left, $right) = ($1, $2);
4737 if ($nr > 1) {
4738 die "Only one '*' wildcard expansion ",
4739 "is supported (got $nr): '$glob'\n";
4740 } elsif ($nr == 0) {
4741 die "One '*' is needed for glob: '$glob'\n";
4742 }
4743 $re = quotemeta($left) . $re . quotemeta($right);
4744 if (length $left && !($left =~ s!/+$!!g)) {
4745 die "Missing trailing '/' on left side of: '$glob' ($left)\n";
4746 }
4747 if (length $right && !($right =~ s!^/+!!g)) {
4748 die "Missing leading '/' on right side of: '$glob' ($right)\n";
4749 }
4750 my $left_re = qr/^\/\Q$left\E(\/|$)/;
4751 bless { left => $left, right => $right, left_regex => $left_re,
4752 regex => qr/$re/, glob => $glob }, $class;
4753 }
4754
4755 sub full_path {
4756 my ($self, $path) = @_;
4757 return (length $self->{left} ? "$self->{left}/" : '') .
4758 $path . (length $self->{right} ? "/$self->{right}" : '');
4759 }
4760
4761 __END__
4762
4763 Data structures:
4764
4765
4766 $remotes = { # returned by read_all_remotes()
4767 'svn' => {
4768 # svn-remote.svn.url=https://svn.musicpd.org
4769 url => 'https://svn.musicpd.org',
4770 # svn-remote.svn.fetch=mpd/trunk:trunk
4771 fetch => {
4772 'mpd/trunk' => 'trunk',
4773 },
4774 # svn-remote.svn.tags=mpd/tags/*:tags/*
4775 tags => {
4776 path => {
4777 left => 'mpd/tags',
4778 right => '',
4779 regex => qr!mpd/tags/([^/]+)$!,
4780 glob => 'tags/*',
4781 },
4782 ref => {
4783 left => 'tags',
4784 right => '',
4785 regex => qr!tags/([^/]+)$!,
4786 glob => 'tags/*',
4787 },
4788 }
4789 }
4790 };
4791
4792 $log_entry hashref as returned by libsvn_log_entry()
4793 {
4794 log => 'whitespace-formatted log entry
4795 ', # trailing newline is preserved
4796 revision => '8', # integer
4797 date => '2004-02-24T17:01:44.108345Z', # commit date
4798 author => 'committer name'
4799 };
4800
4801
4802 # this is generated by generate_diff();
4803 @mods = array of diff-index line hashes, each element represents one line
4804 of diff-index output
4805
4806 diff-index line ($m hash)
4807 {
4808 mode_a => first column of diff-index output, no leading ':',
4809 mode_b => second column of diff-index output,
4810 sha1_b => sha1sum of the final blob,
4811 chg => change type [MCRADT],
4812 file_a => original file name of a file (iff chg is 'C' or 'R')
4813 file_b => new/current file name of a file (any chg)
4814 }
4815 ;
4816
4817 # retval of read_url_paths{,_all}();
4818 $l_map = {
4819 # repository root url
4820 'https://svn.musicpd.org' => {
4821 # repository path # GIT_SVN_ID
4822 'mpd/trunk' => 'trunk',
4823 'mpd/tags/0.11.5' => 'tags/0.11.5',
4824 },
4825 }
4826
4827 Notes:
4828 I don't trust the each() function on unless I created %hash myself
4829 because the internal iterator may not have started at base.
git with Jakub Narębski modifications