?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Use the modern syntax of git-diff-files in t2002-checkout-cache-u.sh
[git/jrn.git] / git-svn.perl
blob711e7b7eb9274c9de292d6f36fca9a9f0baf6318
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 'blame' => [ \&Git::SVN::Log::cmd_blame,
190 "Show what revision and author last modified each line of a file",
191 {} ],
192 );
193
194 my $cmd;
195 for (my $i = 0; $i < @ARGV; $i++) {
196 if (defined $cmd{$ARGV[$i]}) {
197 $cmd = $ARGV[$i];
198 splice @ARGV, $i, 1;
199 last;
200 }
201 };
202
203 # make sure we're always running at the top-level working directory
204 unless ($cmd && $cmd =~ /(?:clone|init|multi-init)$/) {
205 unless (-d $ENV{GIT_DIR}) {
206 if ($git_dir_user_set) {
207 die "GIT_DIR=$ENV{GIT_DIR} explicitly set, ",
208 "but it is not a directory\n";
209 }
210 my $git_dir = delete $ENV{GIT_DIR};
211 chomp(my $cdup = command_oneline(qw/rev-parse --show-cdup/));
212 unless (length $cdup) {
213 die "Already at toplevel, but $git_dir ",
214 "not found '$cdup'\n";
215 }
216 chdir $cdup or die "Unable to chdir up to '$cdup'\n";
217 unless (-d $git_dir) {
218 die "$git_dir still not found after going to ",
219 "'$cdup'\n";
220 }
221 $ENV{GIT_DIR} = $git_dir;
222 }
223 }
224
225 my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
226
227 read_repo_config(\%opts);
228 Getopt::Long::Configure('pass_through') if ($cmd && $cmd eq 'log');
229 my $rv = GetOptions(%opts, 'help|H|h' => \$_help, 'version|V' => \$_version,
230 'minimize-connections' => \$Git::SVN::Migration::_minimize,
231 'id|i=s' => \$Git::SVN::default_ref_id,
232 'svn-remote|remote|R=s' => sub {
233 $Git::SVN::no_reuse_existing = 1;
234 $Git::SVN::default_repo_id = $_[1] });
235 exit 1 if (!$rv && $cmd && $cmd ne 'log');
236
237 usage(0) if $_help;
238 version() if $_version;
239 usage(1) unless defined $cmd;
240 load_authors() if $_authors;
241
242 unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
243 Git::SVN::Migration::migration_check();
244 }
245 Git::SVN::init_vars();
246 eval {
247 Git::SVN::verify_remotes_sanity();
248 $cmd{$cmd}->[0]->(@ARGV);
249 };
250 fatal $@ if $@;
251 post_fetch_checkout();
252 exit 0;
253
254 ####################### primary functions ######################
255 sub usage {
256 my $exit = shift || 0;
257 my $fd = $exit ? \*STDERR : \*STDOUT;
258 print $fd <<"";
259 git-svn - bidirectional operations between a single Subversion tree and git
260 Usage: $0 <command> [options] [arguments]\n
261
262 print $fd "Available commands:\n" unless $cmd;
263
264 foreach (sort keys %cmd) {
265 next if $cmd && $cmd ne $_;
266 next if /^multi-/; # don't show deprecated commands
267 print $fd ' ',pack('A17',$_),$cmd{$_}->[1],"\n";
268 foreach (sort keys %{$cmd{$_}->[2]}) {
269 # mixed-case options are for .git/config only
270 next if /[A-Z]/ && /^[a-z]+$/i;
271 # prints out arguments as they should be passed:
272 my $x = s#[:=]s$## ? '<arg>' : s#[:=]i$## ? '<num>' : '';
273 print $fd ' ' x 21, join(', ', map { length $_ > 1 ?
274 "--$_" : "-$_" }
275 split /\|/,$_)," $x\n";
276 }
277 }
278 print $fd <<"";
279 \nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
280 arbitrary identifier if you're tracking multiple SVN branches/repositories in
281 one git repository and want to keep them separate. See git-svn(1) for more
282 information.
283
284 exit $exit;
285 }
286
287 sub version {
288 print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
289 exit 0;
290 }
291
292 sub do_git_init_db {
293 unless (-d $ENV{GIT_DIR}) {
294 my @init_db = ('init');
295 push @init_db, "--template=$_template" if defined $_template;
296 if (defined $_shared) {
297 if ($_shared =~ /[a-z]/) {
298 push @init_db, "--shared=$_shared";
299 } else {
300 push @init_db, "--shared";
301 }
302 }
303 command_noisy(@init_db);
304 }
305 my $set;
306 my $pfx = "svn-remote.$Git::SVN::default_repo_id";
307 foreach my $i (keys %icv) {
308 die "'$set' and '$i' cannot both be set\n" if $set;
309 next unless defined $icv{$i};
310 command_noisy('config', "$pfx.$i", $icv{$i});
311 $set = $i;
312 }
313 }
314
315 sub init_subdir {
316 my $repo_path = shift or return;
317 mkpath([$repo_path]) unless -d $repo_path;
318 chdir $repo_path or die "Couldn't chdir to $repo_path: $!\n";
319 $ENV{GIT_DIR} = '.git';
320 }
321
322 sub cmd_clone {
323 my ($url, $path) = @_;
324 if (!defined $path &&
325 (defined $_trunk || defined $_branches || defined $_tags ||
326 defined $_stdlayout) &&
327 $url !~ m#^[a-z\+]+://#) {
328 $path = $url;
329 }
330 $path = basename($url) if !defined $path || !length $path;
331 cmd_init($url, $path);
332 Git::SVN::fetch_all($Git::SVN::default_repo_id);
333 }
334
335 sub cmd_init {
336 if (defined $_stdlayout) {
337 $_trunk = 'trunk' if (!defined $_trunk);
338 $_tags = 'tags' if (!defined $_tags);
339 $_branches = 'branches' if (!defined $_branches);
340 }
341 if (defined $_trunk || defined $_branches || defined $_tags) {
342 return cmd_multi_init(@_);
343 }
344 my $url = shift or die "SVN repository location required ",
345 "as a command-line argument\n";
346 init_subdir(@_);
347 do_git_init_db();
348
349 Git::SVN->init($url);
350 }
351
352 sub cmd_fetch {
353 if (grep /^\d+=./, @_) {
354 die "'<rev>=<commit>' fetch arguments are ",
355 "no longer supported.\n";
356 }
357 my ($remote) = @_;
358 if (@_ > 1) {
359 die "Usage: $0 fetch [--all] [svn-remote]\n";
360 }
361 $remote ||= $Git::SVN::default_repo_id;
362 if ($_fetch_all) {
363 cmd_multi_fetch();
364 } else {
365 Git::SVN::fetch_all($remote, Git::SVN::read_all_remotes());
366 }
367 }
368
369 sub cmd_set_tree {
370 my (@commits) = @_;
371 if ($_stdin || !@commits) {
372 print "Reading from stdin...\n";
373 @commits = ();
374 while (<STDIN>) {
375 if (/\b($sha1_short)\b/o) {
376 unshift @commits, $1;
377 }
378 }
379 }
380 my @revs;
381 foreach my $c (@commits) {
382 my @tmp = command('rev-parse',$c);
383 if (scalar @tmp == 1) {
384 push @revs, $tmp[0];
385 } elsif (scalar @tmp > 1) {
386 push @revs, reverse(command('rev-list',@tmp));
387 } else {
388 fatal "Failed to rev-parse $c";
389 }
390 }
391 my $gs = Git::SVN->new;
392 my ($r_last, $cmt_last) = $gs->last_rev_commit;
393 $gs->fetch;
394 if (defined $gs->{last_rev} && $r_last != $gs->{last_rev}) {
395 fatal "There are new revisions that were fetched ",
396 "and need to be merged (or acknowledged) ",
397 "before committing.\nlast rev: $r_last\n",
398 " current: $gs->{last_rev}";
399 }
400 $gs->set_tree($_) foreach @revs;
401 print "Done committing ",scalar @revs," revisions to SVN\n";
402 unlink $gs->{index};
403 }
404
405 sub cmd_dcommit {
406 my $head = shift;
407 git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
408 'Cannot dcommit with a dirty index. Commit your changes first, '
409 . "or stash them with `git stash'.\n";
410 $head ||= 'HEAD';
411 my @refs;
412 my ($url, $rev, $uuid, $gs) = working_head_info($head, \@refs);
413 if ($url) {
414 print "Committing to $url ...\n";
415 }
416 unless ($gs) {
417 die "Unable to determine upstream SVN information from ",
418 "$head history.\nPerhaps the repository is empty.";
419 }
420 my $last_rev;
421 my ($linear_refs, $parents) = linearize_history($gs, \@refs);
422 if ($_no_rebase && scalar(@$linear_refs) > 1) {
423 warn "Attempting to commit more than one change while ",
424 "--no-rebase is enabled.\n",
425 "If these changes depend on each other, re-running ",
426 "without --no-rebase may be required."
427 }
428 while (1) {
429 my $d = shift @$linear_refs or last;
430 unless (defined $last_rev) {
431 (undef, $last_rev, undef) = cmt_metadata("$d~1");
432 unless (defined $last_rev) {
433 fatal "Unable to extract revision information ",
434 "from commit $d~1";
435 }
436 }
437 if ($_dry_run) {
438 print "diff-tree $d~1 $d\n";
439 } else {
440 my $cmt_rev;
441 my %ed_opts = ( r => $last_rev,
442 log => get_commit_entry($d)->{log},
443 ra => Git::SVN::Ra->new($gs->full_url),
444 config => SVN::Core::config_get_config(
445 $Git::SVN::Ra::config_dir
446 ),
447 tree_a => "$d~1",
448 tree_b => $d,
449 editor_cb => sub {
450 print "Committed r$_[0]\n";
451 $cmt_rev = $_[0];
452 },
453 svn_path => '');
454 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
455 print "No changes\n$d~1 == $d\n";
456 } elsif ($parents->{$d} && @{$parents->{$d}}) {
457 $gs->{inject_parents_dcommit}->{$cmt_rev} =
458 $parents->{$d};
459 }
460 $_fetch_all ? $gs->fetch_all : $gs->fetch;
461 $last_rev = $cmt_rev;
462 next if $_no_rebase;
463
464 # we always want to rebase against the current HEAD,
465 # not any head that was passed to us
466 my @diff = command('diff-tree', $d,
467 $gs->refname, '--');
468 my @finish;
469 if (@diff) {
470 @finish = rebase_cmd();
471 print STDERR "W: $d and ", $gs->refname,
472 " differ, using @finish:\n",
473 join("\n", @diff), "\n";
474 } else {
475 print "No changes between current HEAD and ",
476 $gs->refname,
477 "\nResetting to the latest ",
478 $gs->refname, "\n";
479 @finish = qw/reset --mixed/;
480 }
481 command_noisy(@finish, $gs->refname);
482 if (@diff) {
483 @refs = ();
484 my ($url_, $rev_, $uuid_, $gs_) =
485 working_head_info($head, \@refs);
486 my ($linear_refs_, $parents_) =
487 linearize_history($gs_, \@refs);
488 if (scalar(@$linear_refs) !=
489 scalar(@$linear_refs_)) {
490 fatal "# of revisions changed ",
491 "\nbefore:\n",
492 join("\n", @$linear_refs),
493 "\n\nafter:\n",
494 join("\n", @$linear_refs_), "\n",
495 'If you are attempting to commit ',
496 "merges, try running:\n\t",
497 'git rebase --interactive',
498 '--preserve-merges ',
499 $gs->refname,
500 "\nBefore dcommitting";
501 }
502 if ($url_ ne $url) {
503 fatal "URL mismatch after rebase: ",
504 "$url_ != $url";
505 }
506 if ($uuid_ ne $uuid) {
507 fatal "uuid mismatch after rebase: ",
508 "$uuid_ != $uuid";
509 }
510 # remap parents
511 my (%p, @l, $i);
512 for ($i = 0; $i < scalar @$linear_refs; $i++) {
513 my $new = $linear_refs_->[$i] or next;
514 $p{$new} =
515 $parents->{$linear_refs->[$i]};
516 push @l, $new;
517 }
518 $parents = \%p;
519 $linear_refs = \@l;
520 }
521 }
522 }
523 unlink $gs->{index};
524 }
525
526 sub cmd_find_rev {
527 my $revision_or_hash = shift or die "SVN or git revision required ",
528 "as a command-line argument\n";
529 my $result;
530 if ($revision_or_hash =~ /^r\d+$/) {
531 my $head = shift;
532 $head ||= 'HEAD';
533 my @refs;
534 my (undef, undef, undef, $gs) = working_head_info($head, \@refs);
535 unless ($gs) {
536 die "Unable to determine upstream SVN information from ",
537 "$head history\n";
538 }
539 my $desired_revision = substr($revision_or_hash, 1);
540 $result = $gs->rev_map_get($desired_revision);
541 } else {
542 my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
543 $result = $rev;
544 }
545 print "$result\n" if $result;
546 }
547
548 sub cmd_rebase {
549 command_noisy(qw/update-index --refresh/);
550 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
551 unless ($gs) {
552 die "Unable to determine upstream SVN information from ",
553 "working tree history\n";
554 }
555 if (command(qw/diff-index HEAD --/)) {
556 print STDERR "Cannot rebase with uncommited changes:\n";
557 command_noisy('status');
558 exit 1;
559 }
560 unless ($_local) {
561 # rebase will checkout for us, so no need to do it explicitly
562 $_no_checkout = 'true';
563 $_fetch_all ? $gs->fetch_all : $gs->fetch;
564 }
565 command_noisy(rebase_cmd(), $gs->refname);
566 }
567
568 sub cmd_show_ignore {
569 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
570 $gs ||= Git::SVN->new;
571 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
572 $gs->prop_walk($gs->{path}, $r, sub {
573 my ($gs, $path, $props) = @_;
574 print STDOUT "\n# $path\n";
575 my $s = $props->{'svn:ignore'} or return;
576 $s =~ s/[\r\n]+/\n/g;
577 chomp $s;
578 $s =~ s#^#$path#gm;
579 print STDOUT "$s\n";
580 });
581 }
582
583 sub cmd_show_externals {
584 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
585 $gs ||= Git::SVN->new;
586 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
587 $gs->prop_walk($gs->{path}, $r, sub {
588 my ($gs, $path, $props) = @_;
589 print STDOUT "\n# $path\n";
590 my $s = $props->{'svn:externals'} or return;
591 $s =~ s/[\r\n]+/\n/g;
592 chomp $s;
593 $s =~ s#^#$path#gm;
594 print STDOUT "$s\n";
595 });
596 }
597
598 sub cmd_create_ignore {
599 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
600 $gs ||= Git::SVN->new;
601 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
602 $gs->prop_walk($gs->{path}, $r, sub {
603 my ($gs, $path, $props) = @_;
604 # $path is of the form /path/to/dir/
605 my $ignore = '.' . $path . '.gitignore';
606 my $s = $props->{'svn:ignore'} or return;
607 open(GITIGNORE, '>', $ignore)
608 or fatal("Failed to open `$ignore' for writing: $!");
609 $s =~ s/[\r\n]+/\n/g;
610 chomp $s;
611 # Prefix all patterns so that the ignore doesn't apply
612 # to sub-directories.
613 $s =~ s#^#/#gm;
614 print GITIGNORE "$s\n";
615 close(GITIGNORE)
616 or fatal("Failed to close `$ignore': $!");
617 command_noisy('add', $ignore);
618 });
619 }
620
621 sub canonicalize_path {
622 my ($path) = @_;
623 my $dot_slash_added = 0;
624 if (substr($path, 0, 1) ne "/") {
625 $path = "./" . $path;
626 $dot_slash_added = 1;
627 }
628 # File::Spec->canonpath doesn't collapse x/../y into y (for a
629 # good reason), so let's do this manually.
630 $path =~ s#/+#/#g;
631 $path =~ s#/\.(?:/|$)#/#g;
632 $path =~ s#/[^/]+/\.\.##g;
633 $path =~ s#/$##g;
634 $path =~ s#^\./## if $dot_slash_added;
635 return $path;
636 }
637
638 # get_svnprops(PATH)
639 # ------------------
640 # Helper for cmd_propget and cmd_proplist below.
641 sub get_svnprops {
642 my $path = shift;
643 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
644 $gs ||= Git::SVN->new;
645
646 # prefix THE PATH by the sub-directory from which the user
647 # invoked us.
648 $path = $cmd_dir_prefix . $path;
649 fatal("No such file or directory: $path") unless -e $path;
650 my $is_dir = -d $path ? 1 : 0;
651 $path = $gs->{path} . '/' . $path;
652
653 # canonicalize the path (otherwise libsvn will abort or fail to
654 # find the file)
655 $path = canonicalize_path($path);
656
657 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
658 my $props;
659 if ($is_dir) {
660 (undef, undef, $props) = $gs->ra->get_dir($path, $r);
661 }
662 else {
663 (undef, $props) = $gs->ra->get_file($path, $r, undef);
664 }
665 return $props;
666 }
667
668 # cmd_propget (PROP, PATH)
669 # ------------------------
670 # Print the SVN property PROP for PATH.
671 sub cmd_propget {
672 my ($prop, $path) = @_;
673 $path = '.' if not defined $path;
674 usage(1) if not defined $prop;
675 my $props = get_svnprops($path);
676 if (not defined $props->{$prop}) {
677 fatal("`$path' does not have a `$prop' SVN property.");
678 }
679 print $props->{$prop} . "\n";
680 }
681
682 # cmd_proplist (PATH)
683 # -------------------
684 # Print the list of SVN properties for PATH.
685 sub cmd_proplist {
686 my $path = shift;
687 $path = '.' if not defined $path;
688 my $props = get_svnprops($path);
689 print "Properties on '$path':\n";
690 foreach (sort keys %{$props}) {
691 print " $_\n";
692 }
693 }
694
695 sub cmd_multi_init {
696 my $url = shift;
697 unless (defined $_trunk || defined $_branches || defined $_tags) {
698 usage(1);
699 }
700
701 # there are currently some bugs that prevent multi-init/multi-fetch
702 # setups from working well without this.
703 $Git::SVN::_minimize_url = 1;
704
705 $_prefix = '' unless defined $_prefix;
706 if (defined $url) {
707 $url =~ s#/+$##;
708 init_subdir(@_);
709 }
710 do_git_init_db();
711 if (defined $_trunk) {
712 my $trunk_ref = $_prefix . 'trunk';
713 # try both old-style and new-style lookups:
714 my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
715 unless ($gs_trunk) {
716 my ($trunk_url, $trunk_path) =
717 complete_svn_url($url, $_trunk);
718 $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
719 undef, $trunk_ref);
720 }
721 }
722 return unless defined $_branches || defined $_tags;
723 my $ra = $url ? Git::SVN::Ra->new($url) : undef;
724 complete_url_ls_init($ra, $_branches, '--branches/-b', $_prefix);
725 complete_url_ls_init($ra, $_tags, '--tags/-t', $_prefix . 'tags/');
726 }
727
728 sub cmd_multi_fetch {
729 my $remotes = Git::SVN::read_all_remotes();
730 foreach my $repo_id (sort keys %$remotes) {
731 if ($remotes->{$repo_id}->{url}) {
732 Git::SVN::fetch_all($repo_id, $remotes);
733 }
734 }
735 }
736
737 # this command is special because it requires no metadata
738 sub cmd_commit_diff {
739 my ($ta, $tb, $url) = @_;
740 my $usage = "Usage: $0 commit-diff -r<revision> ".
741 "<tree-ish> <tree-ish> [<URL>]";
742 fatal($usage) if (!defined $ta || !defined $tb);
743 my $svn_path;
744 if (!defined $url) {
745 my $gs = eval { Git::SVN->new };
746 if (!$gs) {
747 fatal("Needed URL or usable git-svn --id in ",
748 "the command-line\n", $usage);
749 }
750 $url = $gs->{url};
751 $svn_path = $gs->{path};
752 }
753 unless (defined $_revision) {
754 fatal("-r|--revision is a required argument\n", $usage);
755 }
756 if (defined $_message && defined $_file) {
757 fatal("Both --message/-m and --file/-F specified ",
758 "for the commit message.\n",
759 "I have no idea what you mean");
760 }
761 if (defined $_file) {
762 $_message = file_to_s($_file);
763 } else {
764 $_message ||= get_commit_entry($tb)->{log};
765 }
766 my $ra ||= Git::SVN::Ra->new($url);
767 $svn_path ||= $ra->{svn_path};
768 my $r = $_revision;
769 if ($r eq 'HEAD') {
770 $r = $ra->get_latest_revnum;
771 } elsif ($r !~ /^\d+$/) {
772 die "revision argument: $r not understood by git-svn\n";
773 }
774 my %ed_opts = ( r => $r,
775 log => $_message,
776 ra => $ra,
777 tree_a => $ta,
778 tree_b => $tb,
779 editor_cb => sub { print "Committed r$_[0]\n" },
780 svn_path => $svn_path );
781 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
782 print "No changes\n$ta == $tb\n";
783 }
784 }
785
786 sub cmd_info {
787 my $path = canonicalize_path(shift or ".");
788 unless (scalar(@_) == 0) {
789 die "Too many arguments specified\n";
790 }
791
792 my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
793
794 if (!$file_type && !$diff_status) {
795 print STDERR "$path: (Not a versioned resource)\n\n";
796 return;
797 }
798
799 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
800 unless ($gs) {
801 die "Unable to determine upstream SVN information from ",
802 "working tree history\n";
803 }
804 my $full_url = $url . ($path eq "." ? "" : "/$path");
805
806 if ($_url) {
807 print $full_url, "\n";
808 return;
809 }
810
811 my $result = "Path: $path\n";
812 $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
813 $result .= "URL: " . $full_url . "\n";
814
815 eval {
816 my $repos_root = $gs->repos_root;
817 Git::SVN::remove_username($repos_root);
818 $result .= "Repository Root: $repos_root\n";
819 };
820 if ($@) {
821 $result .= "Repository Root: (offline)\n";
822 }
823 $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A";
824 $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
825
826 $result .= "Node Kind: " .
827 ($file_type eq "dir" ? "directory" : "file") . "\n";
828
829 my $schedule = $diff_status eq "A"
830 ? "add"
831 : ($diff_status eq "D" ? "delete" : "normal");
832 $result .= "Schedule: $schedule\n";
833
834 if ($diff_status eq "A") {
835 print $result, "\n";
836 return;
837 }
838
839 my ($lc_author, $lc_rev, $lc_date_utc);
840 my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $path);
841 my $log = command_output_pipe(@args);
842 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
843 while (<$log>) {
844 if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
845 $lc_author = $1;
846 $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
847 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
848 (undef, $lc_rev, undef) = ::extract_metadata($1);
849 }
850 }
851 close $log;
852
853 Git::SVN::Log::set_local_timezone();
854
855 $result .= "Last Changed Author: $lc_author\n";
856 $result .= "Last Changed Rev: $lc_rev\n";
857 $result .= "Last Changed Date: " .
858 Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
859
860 if ($file_type ne "dir") {
861 my $text_last_updated_date =
862 ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
863 $result .=
864 "Text Last Updated: " .
865 Git::SVN::Log::format_svn_date($text_last_updated_date) .
866 "\n";
867 my $checksum;
868 if ($diff_status eq "D") {
869 my ($fh, $ctx) =
870 command_output_pipe(qw(cat-file blob), "HEAD:$path");
871 if ($file_type eq "link") {
872 my $file_name = <$fh>;
873 $checksum = md5sum("link $file_name");
874 } else {
875 $checksum = md5sum($fh);
876 }
877 command_close_pipe($fh, $ctx);
878 } elsif ($file_type eq "link") {
879 my $file_name =
880 command(qw(cat-file blob), "HEAD:$path");
881 $checksum =
882 md5sum("link " . $file_name);
883 } else {
884 open FILE, "<", $path or die $!;
885 $checksum = md5sum(\*FILE);
886 close FILE or die $!;
887 }
888 $result .= "Checksum: " . $checksum . "\n";
889 }
890
891 print $result, "\n";
892 }
893
894 ########################### utility functions #########################
895
896 sub rebase_cmd {
897 my @cmd = qw/rebase/;
898 push @cmd, '-v' if $_verbose;
899 push @cmd, qw/--merge/ if $_merge;
900 push @cmd, "--strategy=$_strategy" if $_strategy;
901 @cmd;
902 }
903
904 sub post_fetch_checkout {
905 return if $_no_checkout;
906 my $gs = $Git::SVN::_head or return;
907 return if verify_ref('refs/heads/master^0');
908
909 my $valid_head = verify_ref('HEAD^0');
910 command_noisy(qw(update-ref refs/heads/master), $gs->refname);
911 return if ($valid_head || !verify_ref('HEAD^0'));
912
913 return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
914 my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
915 return if -f $index;
916
917 return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
918 return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
919 command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
920 print STDERR "Checked out HEAD:\n ",
921 $gs->full_url, " r", $gs->last_rev, "\n";
922 }
923
924 sub complete_svn_url {
925 my ($url, $path) = @_;
926 $path =~ s#/+$##;
927 if ($path !~ m#^[a-z\+]+://#) {
928 if (!defined $url || $url !~ m#^[a-z\+]+://#) {
929 fatal("E: '$path' is not a complete URL ",
930 "and a separate URL is not specified");
931 }
932 return ($url, $path);
933 }
934 return ($path, '');
935 }
936
937 sub complete_url_ls_init {
938 my ($ra, $repo_path, $switch, $pfx) = @_;
939 unless ($repo_path) {
940 print STDERR "W: $switch not specified\n";
941 return;
942 }
943 $repo_path =~ s#/+$##;
944 if ($repo_path =~ m#^[a-z\+]+://#) {
945 $ra = Git::SVN::Ra->new($repo_path);
946 $repo_path = '';
947 } else {
948 $repo_path =~ s#^/+##;
949 unless ($ra) {
950 fatal("E: '$repo_path' is not a complete URL ",
951 "and a separate URL is not specified");
952 }
953 }
954 my $url = $ra->{url};
955 my $gs = Git::SVN->init($url, undef, undef, undef, 1);
956 my $k = "svn-remote.$gs->{repo_id}.url";
957 my $orig_url = eval { command_oneline(qw/config --get/, $k) };
958 if ($orig_url && ($orig_url ne $gs->{url})) {
959 die "$k already set: $orig_url\n",
960 "wanted to set to: $gs->{url}\n";
961 }
962 command_oneline('config', $k, $gs->{url}) unless $orig_url;
963 my $remote_path = "$ra->{svn_path}/$repo_path";
964 $remote_path =~ s#/+#/#g;
965 $remote_path =~ s#^/##g;
966 $remote_path .= "/*" if $remote_path !~ /\*/;
967 my ($n) = ($switch =~ /^--(\w+)/);
968 if (length $pfx && $pfx !~ m#/$#) {
969 die "--prefix='$pfx' must have a trailing slash '/'\n";
970 }
971 command_noisy('config', "svn-remote.$gs->{repo_id}.$n",
972 "$remote_path:refs/remotes/$pfx*");
973 }
974
975 sub verify_ref {
976 my ($ref) = @_;
977 eval { command_oneline([ 'rev-parse', '--verify', $ref ],
978 { STDERR => 0 }); };
979 }
980
981 sub get_tree_from_treeish {
982 my ($treeish) = @_;
983 # $treeish can be a symbolic ref, too:
984 my $type = command_oneline(qw/cat-file -t/, $treeish);
985 my $expected;
986 while ($type eq 'tag') {
987 ($treeish, $type) = command(qw/cat-file tag/, $treeish);
988 }
989 if ($type eq 'commit') {
990 $expected = (grep /^tree /, command(qw/cat-file commit/,
991 $treeish))[0];
992 ($expected) = ($expected =~ /^tree ($sha1)$/o);
993 die "Unable to get tree from $treeish\n" unless $expected;
994 } elsif ($type eq 'tree') {
995 $expected = $treeish;
996 } else {
997 die "$treeish is a $type, expected tree, tag or commit\n";
998 }
999 return $expected;
1000 }
1001
1002 sub get_commit_entry {
1003 my ($treeish) = shift;
1004 my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
1005 my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
1006 my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
1007 open my $log_fh, '>', $commit_editmsg or croak $!;
1008
1009 my $type = command_oneline(qw/cat-file -t/, $treeish);
1010 if ($type eq 'commit' || $type eq 'tag') {
1011 my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1012 $type, $treeish);
1013 my $in_msg = 0;
1014 while (<$msg_fh>) {
1015 if (!$in_msg) {
1016 $in_msg = 1 if (/^\s*$/);
1017 } elsif (/^git-svn-id: /) {
1018 # skip this for now, we regenerate the
1019 # correct one on re-fetch anyways
1020 # TODO: set *:merge properties or like...
1021 } else {
1022 print $log_fh $_ or croak $!;
1023 }
1024 }
1025 command_close_pipe($msg_fh, $ctx);
1026 }
1027 close $log_fh or croak $!;
1028
1029 if ($_edit || ($type eq 'tree')) {
1030 my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
1031 # TODO: strip out spaces, comments, like git-commit.sh
1032 system($editor, $commit_editmsg);
1033 }
1034 rename $commit_editmsg, $commit_msg or croak $!;
1035 open $log_fh, '<', $commit_msg or croak $!;
1036 { local $/; chomp($log_entry{log} = <$log_fh>); }
1037 close $log_fh or croak $!;
1038 unlink $commit_msg;
1039 \%log_entry;
1040 }
1041
1042 sub s_to_file {
1043 my ($str, $file, $mode) = @_;
1044 open my $fd,'>',$file or croak $!;
1045 print $fd $str,"\n" or croak $!;
1046 close $fd or croak $!;
1047 chmod ($mode &~ umask, $file) if (defined $mode);
1048 }
1049
1050 sub file_to_s {
1051 my $file = shift;
1052 open my $fd,'<',$file or croak "$!: file: $file\n";
1053 local $/;
1054 my $ret = <$fd>;
1055 close $fd or croak $!;
1056 $ret =~ s/\s*$//s;
1057 return $ret;
1058 }
1059
1060 # '<svn username> = real-name <email address>' mapping based on git-svnimport:
1061 sub load_authors {
1062 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1063 my $log = $cmd eq 'log';
1064 while (<$authors>) {
1065 chomp;
1066 next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1067 my ($user, $name, $email) = ($1, $2, $3);
1068 if ($log) {
1069 $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1070 } else {
1071 $users{$user} = [$name, $email];
1072 }
1073 }
1074 close $authors or croak $!;
1075 }
1076
1077 # convert GetOpt::Long specs for use by git-config
1078 sub read_repo_config {
1079 return unless -d $ENV{GIT_DIR};
1080 my $opts = shift;
1081 my @config_only;
1082 foreach my $o (keys %$opts) {
1083 # if we have mixedCase and a long option-only, then
1084 # it's a config-only variable that we don't need for
1085 # the command-line.
1086 push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1087 my $v = $opts->{$o};
1088 my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1089 $key =~ s/-//g;
1090 my $arg = 'git-config';
1091 $arg .= ' --int' if ($o =~ /[:=]i$/);
1092 $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1093 if (ref $v eq 'ARRAY') {
1094 chomp(my @tmp = `$arg --get-all svn.$key`);
1095 @$v = @tmp if @tmp;
1096 } else {
1097 chomp(my $tmp = `$arg --get svn.$key`);
1098 if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1099 $$v = $tmp;
1100 }
1101 }
1102 }
1103 delete @$opts{@config_only} if @config_only;
1104 }
1105
1106 sub extract_metadata {
1107 my $id = shift or return (undef, undef, undef);
1108 my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1109 \s([a-f\d\-]+)$/x);
1110 if (!defined $rev || !$uuid || !$url) {
1111 # some of the original repositories I made had
1112 # identifiers like this:
1113 ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/);
1114 }
1115 return ($url, $rev, $uuid);
1116 }
1117
1118 sub cmt_metadata {
1119 return extract_metadata((grep(/^git-svn-id: /,
1120 command(qw/cat-file commit/, shift)))[-1]);
1121 }
1122
1123 sub working_head_info {
1124 my ($head, $refs) = @_;
1125 my @args = ('log', '--no-color', '--first-parent', '--pretty=medium');
1126 my ($fh, $ctx) = command_output_pipe(@args, $head);
1127 my $hash;
1128 my %max;
1129 while (<$fh>) {
1130 if ( m{^commit ($::sha1)$} ) {
1131 unshift @$refs, $hash if $hash and $refs;
1132 $hash = $1;
1133 next;
1134 }
1135 next unless s{^\s*(git-svn-id:)}{$1};
1136 my ($url, $rev, $uuid) = extract_metadata($_);
1137 if (defined $url && defined $rev) {
1138 next if $max{$url} and $max{$url} < $rev;
1139 if (my $gs = Git::SVN->find_by_url($url)) {
1140 my $c = $gs->rev_map_get($rev);
1141 if ($c && $c eq $hash) {
1142 close $fh; # break the pipe
1143 return ($url, $rev, $uuid, $gs);
1144 } else {
1145 $max{$url} ||= $gs->rev_map_max;
1146 }
1147 }
1148 }
1149 }
1150 command_close_pipe($fh, $ctx);
1151 (undef, undef, undef, undef);
1152 }
1153
1154 sub read_commit_parents {
1155 my ($parents, $c) = @_;
1156 chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1157 $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1158 @{$parents->{$c}} = split(/ /, $p);
1159 }
1160
1161 sub linearize_history {
1162 my ($gs, $refs) = @_;
1163 my %parents;
1164 foreach my $c (@$refs) {
1165 read_commit_parents(\%parents, $c);
1166 }
1167
1168 my @linear_refs;
1169 my %skip = ();
1170 my $last_svn_commit = $gs->last_commit;
1171 foreach my $c (reverse @$refs) {
1172 next if $c eq $last_svn_commit;
1173 last if $skip{$c};
1174
1175 unshift @linear_refs, $c;
1176 $skip{$c} = 1;
1177
1178 # we only want the first parent to diff against for linear
1179 # history, we save the rest to inject when we finalize the
1180 # svn commit
1181 my $fp_a = verify_ref("$c~1");
1182 my $fp_b = shift @{$parents{$c}} if $parents{$c};
1183 if (!$fp_a || !$fp_b) {
1184 die "Commit $c\n",
1185 "has no parent commit, and therefore ",
1186 "nothing to diff against.\n",
1187 "You should be working from a repository ",
1188 "originally created by git-svn\n";
1189 }
1190 if ($fp_a ne $fp_b) {
1191 die "$c~1 = $fp_a, however parsing commit $c ",
1192 "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1193 }
1194
1195 foreach my $p (@{$parents{$c}}) {
1196 $skip{$p} = 1;
1197 }
1198 }
1199 (\@linear_refs, \%parents);
1200 }
1201
1202 sub find_file_type_and_diff_status {
1203 my ($path) = @_;
1204 return ('dir', '') if $path eq '.';
1205
1206 my $diff_output =
1207 command_oneline(qw(diff --cached --name-status --), $path) || "";
1208 my $diff_status = (split(' ', $diff_output))[0] || "";
1209
1210 my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1211
1212 return (undef, undef) if !$diff_status && !$ls_tree;
1213
1214 if ($diff_status eq "A") {
1215 return ("link", $diff_status) if -l $path;
1216 return ("dir", $diff_status) if -d $path;
1217 return ("file", $diff_status);
1218 }
1219
1220 my $mode = (split(' ', $ls_tree))[0] || "";
1221
1222 return ("link", $diff_status) if $mode eq "120000";
1223 return ("dir", $diff_status) if $mode eq "040000";
1224 return ("file", $diff_status);
1225 }
1226
1227 sub md5sum {
1228 my $arg = shift;
1229 my $ref = ref $arg;
1230 my $md5 = Digest::MD5->new();
1231 if ($ref eq 'GLOB' || $ref eq 'IO::File') {
1232 $md5->addfile($arg) or croak $!;
1233 } elsif ($ref eq 'SCALAR') {
1234 $md5->add($$arg) or croak $!;
1235 } elsif (!$ref) {
1236 $md5->add($arg) or croak $!;
1237 } else {
1238 ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
1239 }
1240 return $md5->hexdigest();
1241 }
1242
1243 package Git::SVN;
1244 use strict;
1245 use warnings;
1246 use Fcntl qw/:DEFAULT :seek/;
1247 use constant rev_map_fmt => 'NH40';
1248 use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1249 $_repack $_repack_flags $_use_svm_props $_head
1250 $_use_svnsync_props $no_reuse_existing $_minimize_url
1251 $_use_log_author/;
1252 use Carp qw/croak/;
1253 use File::Path qw/mkpath/;
1254 use File::Copy qw/copy/;
1255 use IPC::Open3;
1256
1257 my ($_gc_nr, $_gc_period);
1258
1259 # properties that we do not log:
1260 my %SKIP_PROP;
1261 BEGIN {
1262 %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1263 svn:special svn:executable
1264 svn:entry:committed-rev
1265 svn:entry:last-author
1266 svn:entry:uuid
1267 svn:entry:committed-date/;
1268
1269 # some options are read globally, but can be overridden locally
1270 # per [svn-remote "..."] section. Command-line options will *NOT*
1271 # override options set in an [svn-remote "..."] section
1272 no strict 'refs';
1273 for my $option (qw/follow_parent no_metadata use_svm_props
1274 use_svnsync_props/) {
1275 my $key = $option;
1276 $key =~ tr/_//d;
1277 my $prop = "-$option";
1278 *$option = sub {
1279 my ($self) = @_;
1280 return $self->{$prop} if exists $self->{$prop};
1281 my $k = "svn-remote.$self->{repo_id}.$key";
1282 eval { command_oneline(qw/config --get/, $k) };
1283 if ($@) {
1284 $self->{$prop} = ${"Git::SVN::_$option"};
1285 } else {
1286 my $v = command_oneline(qw/config --bool/,$k);
1287 $self->{$prop} = $v eq 'false' ? 0 : 1;
1288 }
1289 return $self->{$prop};
1290 }
1291 }
1292 }
1293
1294 my (%LOCKFILES, %INDEX_FILES);
1295 END {
1296 unlink keys %LOCKFILES if %LOCKFILES;
1297 unlink keys %INDEX_FILES if %INDEX_FILES;
1298 }
1299
1300 sub resolve_local_globs {
1301 my ($url, $fetch, $glob_spec) = @_;
1302 return unless defined $glob_spec;
1303 my $ref = $glob_spec->{ref};
1304 my $path = $glob_spec->{path};
1305 foreach (command(qw#for-each-ref --format=%(refname) refs/remotes#)) {
1306 next unless m#^refs/remotes/$ref->{regex}$#;
1307 my $p = $1;
1308 my $pathname = desanitize_refname($path->full_path($p));
1309 my $refname = desanitize_refname($ref->full_path($p));
1310 if (my $existing = $fetch->{$pathname}) {
1311 if ($existing ne $refname) {
1312 die "Refspec conflict:\n",
1313 "existing: refs/remotes/$existing\n",
1314 " globbed: refs/remotes/$refname\n";
1315 }
1316 my $u = (::cmt_metadata("refs/remotes/$refname"))[0];
1317 $u =~ s!^\Q$url\E(/|$)!! or die
1318 "refs/remotes/$refname: '$url' not found in '$u'\n";
1319 if ($pathname ne $u) {
1320 warn "W: Refspec glob conflict ",
1321 "(ref: refs/remotes/$refname):\n",
1322 "expected path: $pathname\n",
1323 " real path: $u\n",
1324 "Continuing ahead with $u\n";
1325 next;
1326 }
1327 } else {
1328 $fetch->{$pathname} = $refname;
1329 }
1330 }
1331 }
1332
1333 sub parse_revision_argument {
1334 my ($base, $head) = @_;
1335 if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1336 return ($base, $head);
1337 }
1338 return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1339 return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1340 return ($head, $head) if ($::_revision eq 'HEAD');
1341 return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1342 return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1343 die "revision argument: $::_revision not understood by git-svn\n";
1344 }
1345
1346 sub fetch_all {
1347 my ($repo_id, $remotes) = @_;
1348 if (ref $repo_id) {
1349 my $gs = $repo_id;
1350 $repo_id = undef;
1351 $repo_id = $gs->{repo_id};
1352 }
1353 $remotes ||= read_all_remotes();
1354 my $remote = $remotes->{$repo_id} or
1355 die "[svn-remote \"$repo_id\"] unknown\n";
1356 my $fetch = $remote->{fetch};
1357 my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1358 my (@gs, @globs);
1359 my $ra = Git::SVN::Ra->new($url);
1360 my $uuid = $ra->get_uuid;
1361 my $head = $ra->get_latest_revnum;
1362 my $base = defined $fetch ? $head : 0;
1363
1364 # read the max revs for wildcard expansion (branches/*, tags/*)
1365 foreach my $t (qw/branches tags/) {
1366 defined $remote->{$t} or next;
1367 push @globs, $remote->{$t};
1368 my $max_rev = eval { tmp_config(qw/--int --get/,
1369 "svn-remote.$repo_id.${t}-maxRev") };
1370 if (defined $max_rev && ($max_rev < $base)) {
1371 $base = $max_rev;
1372 } elsif (!defined $max_rev) {
1373 $base = 0;
1374 }
1375 }
1376
1377 if ($fetch) {
1378 foreach my $p (sort keys %$fetch) {
1379 my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1380 my $lr = $gs->rev_map_max;
1381 if (defined $lr) {
1382 $base = $lr if ($lr < $base);
1383 }
1384 push @gs, $gs;
1385 }
1386 }
1387
1388 ($base, $head) = parse_revision_argument($base, $head);
1389 $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1390 }
1391
1392 sub read_all_remotes {
1393 my $r = {};
1394 foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1395 if (m!^(.+)\.fetch=\s*(.*)\s*:\s*refs/remotes/(.+)\s*$!) {
1396 my ($remote, $local_ref, $remote_ref) = ($1, $2, $3);
1397 $local_ref =~ s{^/}{};
1398 $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1399 } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1400 $r->{$1}->{url} = $2;
1401 } elsif (m!^(.+)\.(branches|tags)=
1402 (.*):refs/remotes/(.+)\s*$/!x) {
1403 my ($p, $g) = ($3, $4);
1404 my $rs = $r->{$1}->{$2} = {
1405 t => $2,
1406 remote => $1,
1407 path => Git::SVN::GlobSpec->new($p),
1408 ref => Git::SVN::GlobSpec->new($g) };
1409 if (length($rs->{ref}->{right}) != 0) {
1410 die "The '*' glob character must be the last ",
1411 "character of '$g'\n";
1412 }
1413 }
1414 }
1415 $r;
1416 }
1417
1418 sub init_vars {
1419 $_gc_nr = $_gc_period = 1000;
1420 if (defined $_repack || defined $_repack_flags) {
1421 warn "Repack options are obsolete; they have no effect.\n";
1422 }
1423 }
1424
1425 sub verify_remotes_sanity {
1426 return unless -d $ENV{GIT_DIR};
1427 my %seen;
1428 foreach (command(qw/config -l/)) {
1429 if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1430 if ($seen{$1}) {
1431 die "Remote ref refs/remote/$1 is tracked by",
1432 "\n \"$_\"\nand\n \"$seen{$1}\"\n",
1433 "Please resolve this ambiguity in ",
1434 "your git configuration file before ",
1435 "continuing\n";
1436 }
1437 $seen{$1} = $_;
1438 }
1439 }
1440 }
1441
1442 # we allow more chars than remotes2config.sh...
1443 sub sanitize_remote_name {
1444 my ($name) = @_;
1445 $name =~ tr{A-Za-z0-9:,/+-}{.}c;
1446 $name;
1447 }
1448
1449 sub find_existing_remote {
1450 my ($url, $remotes) = @_;
1451 return undef if $no_reuse_existing;
1452 my $existing;
1453 foreach my $repo_id (keys %$remotes) {
1454 my $u = $remotes->{$repo_id}->{url} or next;
1455 next if $u ne $url;
1456 $existing = $repo_id;
1457 last;
1458 }
1459 $existing;
1460 }
1461
1462 sub init_remote_config {
1463 my ($self, $url, $no_write) = @_;
1464 $url =~ s!/+$!!; # strip trailing slash
1465 my $r = read_all_remotes();
1466 my $existing = find_existing_remote($url, $r);
1467 if ($existing) {
1468 unless ($no_write) {
1469 print STDERR "Using existing ",
1470 "[svn-remote \"$existing\"]\n";
1471 }
1472 $self->{repo_id} = $existing;
1473 } elsif ($_minimize_url) {
1474 my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1475 $existing = find_existing_remote($min_url, $r);
1476 if ($existing) {
1477 unless ($no_write) {
1478 print STDERR "Using existing ",
1479 "[svn-remote \"$existing\"]\n";
1480 }
1481 $self->{repo_id} = $existing;
1482 }
1483 if ($min_url ne $url) {
1484 unless ($no_write) {
1485 print STDERR "Using higher level of URL: ",
1486 "$url => $min_url\n";
1487 }
1488 my $old_path = $self->{path};
1489 $self->{path} = $url;
1490 $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1491 if (length $old_path) {
1492 $self->{path} .= "/$old_path";
1493 }
1494 $url = $min_url;
1495 }
1496 }
1497 my $orig_url;
1498 if (!$existing) {
1499 # verify that we aren't overwriting anything:
1500 $orig_url = eval {
1501 command_oneline('config', '--get',
1502 "svn-remote.$self->{repo_id}.url")
1503 };
1504 if ($orig_url && ($orig_url ne $url)) {
1505 die "svn-remote.$self->{repo_id}.url already set: ",
1506 "$orig_url\nwanted to set to: $url\n";
1507 }
1508 }
1509 my ($xrepo_id, $xpath) = find_ref($self->refname);
1510 if (defined $xpath) {
1511 die "svn-remote.$xrepo_id.fetch already set to track ",
1512 "$xpath:refs/remotes/", $self->refname, "\n";
1513 }
1514 unless ($no_write) {
1515 command_noisy('config',
1516 "svn-remote.$self->{repo_id}.url", $url);
1517 $self->{path} =~ s{^/}{};
1518 command_noisy('config', '--add',
1519 "svn-remote.$self->{repo_id}.fetch",
1520 "$self->{path}:".$self->refname);
1521 }
1522 $self->{url} = $url;
1523 }
1524
1525 sub find_by_url { # repos_root and, path are optional
1526 my ($class, $full_url, $repos_root, $path) = @_;
1527
1528 return undef unless defined $full_url;
1529 remove_username($full_url);
1530 remove_username($repos_root) if defined $repos_root;
1531 my $remotes = read_all_remotes();
1532 if (defined $full_url && defined $repos_root && !defined $path) {
1533 $path = $full_url;
1534 $path =~ s#^\Q$repos_root\E(?:/|$)##;
1535 }
1536 foreach my $repo_id (keys %$remotes) {
1537 my $u = $remotes->{$repo_id}->{url} or next;
1538 remove_username($u);
1539 next if defined $repos_root && $repos_root ne $u;
1540
1541 my $fetch = $remotes->{$repo_id}->{fetch} || {};
1542 foreach (qw/branches tags/) {
1543 resolve_local_globs($u, $fetch,
1544 $remotes->{$repo_id}->{$_});
1545 }
1546 my $p = $path;
1547 my $rwr = rewrite_root({repo_id => $repo_id});
1548 unless (defined $p) {
1549 $p = $full_url;
1550 my $z = $u;
1551 if ($rwr) {
1552 $z = $rwr;
1553 }
1554 $p =~ s#^\Q$z\E(?:/|$)## or next;
1555 }
1556 foreach my $f (keys %$fetch) {
1557 next if $f ne $p;
1558 return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1559 }
1560 }
1561 undef;
1562 }
1563
1564 sub init {
1565 my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1566 my $self = _new($class, $repo_id, $ref_id, $path);
1567 if (defined $url) {
1568 $self->init_remote_config($url, $no_write);
1569 }
1570 $self;
1571 }
1572
1573 sub find_ref {
1574 my ($ref_id) = @_;
1575 foreach (command(qw/config -l/)) {
1576 next unless m!^svn-remote\.(.+)\.fetch=
1577 \s*(.*)\s*:\s*refs/remotes/(.+)\s*$!x;
1578 my ($repo_id, $path, $ref) = ($1, $2, $3);
1579 if ($ref eq $ref_id) {
1580 $path = '' if ($path =~ m#^\./?#);
1581 return ($repo_id, $path);
1582 }
1583 }
1584 (undef, undef, undef);
1585 }
1586
1587 sub new {
1588 my ($class, $ref_id, $repo_id, $path) = @_;
1589 if (defined $ref_id && !defined $repo_id && !defined $path) {
1590 ($repo_id, $path) = find_ref($ref_id);
1591 if (!defined $repo_id) {
1592 die "Could not find a \"svn-remote.*.fetch\" key ",
1593 "in the repository configuration matching: ",
1594 "refs/remotes/$ref_id\n";
1595 }
1596 }
1597 my $self = _new($class, $repo_id, $ref_id, $path);
1598 if (!defined $self->{path} || !length $self->{path}) {
1599 my $fetch = command_oneline('config', '--get',
1600 "svn-remote.$repo_id.fetch",
1601 ":refs/remotes/$ref_id\$") or
1602 die "Failed to read \"svn-remote.$repo_id.fetch\" ",
1603 "\":refs/remotes/$ref_id\$\" in config\n";
1604 ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
1605 }
1606 $self->{url} = command_oneline('config', '--get',
1607 "svn-remote.$repo_id.url") or
1608 die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
1609 $self->rebuild;
1610 $self;
1611 }
1612
1613 sub refname {
1614 my ($refname) = "refs/remotes/$_[0]->{ref_id}" ;
1615
1616 # It cannot end with a slash /, we'll throw up on this because
1617 # SVN can't have directories with a slash in their name, either:
1618 if ($refname =~ m{/$}) {
1619 die "ref: '$refname' ends with a trailing slash, this is ",
1620 "not permitted by git nor Subversion\n";
1621 }
1622
1623 # It cannot have ASCII control character space, tilde ~, caret ^,
1624 # colon :, question-mark ?, asterisk *, space, or open bracket [
1625 # anywhere.
1626 #
1627 # Additionally, % must be escaped because it is used for escaping
1628 # and we want our escaped refname to be reversible
1629 $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
1630
1631 # no slash-separated component can begin with a dot .
1632 # /.* becomes /%2E*
1633 $refname =~ s{/\.}{/%2E}g;
1634
1635 # It cannot have two consecutive dots .. anywhere
1636 # .. becomes %2E%2E
1637 $refname =~ s{\.\.}{%2E%2E}g;
1638
1639 return $refname;
1640 }
1641
1642 sub desanitize_refname {
1643 my ($refname) = @_;
1644 $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
1645 return $refname;
1646 }
1647
1648 sub svm_uuid {
1649 my ($self) = @_;
1650 return $self->{svm}->{uuid} if $self->svm;
1651 $self->ra;
1652 unless ($self->{svm}) {
1653 die "SVM UUID not cached, and reading remotely failed\n";
1654 }
1655 $self->{svm}->{uuid};
1656 }
1657
1658 sub svm {
1659 my ($self) = @_;
1660 return $self->{svm} if $self->{svm};
1661 my $svm;
1662 # see if we have it in our config, first:
1663 eval {
1664 my $section = "svn-remote.$self->{repo_id}";
1665 $svm = {
1666 source => tmp_config('--get', "$section.svm-source"),
1667 uuid => tmp_config('--get', "$section.svm-uuid"),
1668 replace => tmp_config('--get', "$section.svm-replace"),
1669 }
1670 };
1671 if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
1672 $self->{svm} = $svm;
1673 }
1674 $self->{svm};
1675 }
1676
1677 sub _set_svm_vars {
1678 my ($self, $ra) = @_;
1679 return $ra if $self->svm;
1680
1681 my @err = ( "useSvmProps set, but failed to read SVM properties\n",
1682 "(svm:source, svm:uuid) ",
1683 "from the following URLs:\n" );
1684 sub read_svm_props {
1685 my ($self, $ra, $path, $r) = @_;
1686 my $props = ($ra->get_dir($path, $r))[2];
1687 my $src = $props->{'svm:source'};
1688 my $uuid = $props->{'svm:uuid'};
1689 return undef if (!$src || !$uuid);
1690
1691 chomp($src, $uuid);
1692
1693 $uuid =~ m{^[0-9a-f\-]{30,}$}
1694 or die "doesn't look right - svm:uuid is '$uuid'\n";
1695
1696 # the '!' is used to mark the repos_root!/relative/path
1697 $src =~ s{/?!/?}{/};
1698 $src =~ s{/+$}{}; # no trailing slashes please
1699 # username is of no interest
1700 $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
1701
1702 my $replace = $ra->{url};
1703 $replace .= "/$path" if length $path;
1704
1705 my $section = "svn-remote.$self->{repo_id}";
1706 tmp_config("$section.svm-source", $src);
1707 tmp_config("$section.svm-replace", $replace);
1708 tmp_config("$section.svm-uuid", $uuid);
1709 $self->{svm} = {
1710 source => $src,
1711 uuid => $uuid,
1712 replace => $replace
1713 };
1714 }
1715
1716 my $r = $ra->get_latest_revnum;
1717 my $path = $self->{path};
1718 my %tried;
1719 while (length $path) {
1720 unless ($tried{"$self->{url}/$path"}) {
1721 return $ra if $self->read_svm_props($ra, $path, $r);
1722 $tried{"$self->{url}/$path"} = 1;
1723 }
1724 $path =~ s#/?[^/]+$##;
1725 }
1726 die "Path: '$path' should be ''\n" if $path ne '';
1727 return $ra if $self->read_svm_props($ra, $path, $r);
1728 $tried{"$self->{url}/$path"} = 1;
1729
1730 if ($ra->{repos_root} eq $self->{url}) {
1731 die @err, (map { " $_\n" } keys %tried), "\n";
1732 }
1733
1734 # nope, make sure we're connected to the repository root:
1735 my $ok;
1736 my @tried_b;
1737 $path = $ra->{svn_path};
1738 $ra = Git::SVN::Ra->new($ra->{repos_root});
1739 while (length $path) {
1740 unless ($tried{"$ra->{url}/$path"}) {
1741 $ok = $self->read_svm_props($ra, $path, $r);
1742 last if $ok;
1743 $tried{"$ra->{url}/$path"} = 1;
1744 }
1745 $path =~ s#/?[^/]+$##;
1746 }
1747 die "Path: '$path' should be ''\n" if $path ne '';
1748 $ok ||= $self->read_svm_props($ra, $path, $r);
1749 $tried{"$ra->{url}/$path"} = 1;
1750 if (!$ok) {
1751 die @err, (map { " $_\n" } keys %tried), "\n";
1752 }
1753 Git::SVN::Ra->new($self->{url});
1754 }
1755
1756 sub svnsync {
1757 my ($self) = @_;
1758 return $self->{svnsync} if $self->{svnsync};
1759
1760 if ($self->no_metadata) {
1761 die "Can't have both 'noMetadata' and ",
1762 "'useSvnsyncProps' options set!\n";
1763 }
1764 if ($self->rewrite_root) {
1765 die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
1766 "options set!\n";
1767 }
1768
1769 my $svnsync;
1770 # see if we have it in our config, first:
1771 eval {
1772 my $section = "svn-remote.$self->{repo_id}";
1773
1774 my $url = tmp_config('--get', "$section.svnsync-url");
1775 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1776 die "doesn't look right - svn:sync-from-url is '$url'\n";
1777
1778 my $uuid = tmp_config('--get', "$section.svnsync-uuid");
1779 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1780 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1781
1782 $svnsync = { url => $url, uuid => $uuid }
1783 };
1784 if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
1785 return $self->{svnsync} = $svnsync;
1786 }
1787
1788 my $err = "useSvnsyncProps set, but failed to read " .
1789 "svnsync property: svn:sync-from-";
1790 my $rp = $self->ra->rev_proplist(0);
1791
1792 my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
1793 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1794 die "doesn't look right - svn:sync-from-url is '$url'\n";
1795
1796 my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
1797 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1798 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1799
1800 my $section = "svn-remote.$self->{repo_id}";
1801 tmp_config('--add', "$section.svnsync-uuid", $uuid);
1802 tmp_config('--add', "$section.svnsync-url", $url);
1803 return $self->{svnsync} = { url => $url, uuid => $uuid };
1804 }
1805
1806 # this allows us to memoize our SVN::Ra UUID locally and avoid a
1807 # remote lookup (useful for 'git svn log').
1808 sub ra_uuid {
1809 my ($self) = @_;
1810 unless ($self->{ra_uuid}) {
1811 my $key = "svn-remote.$self->{repo_id}.uuid";
1812 my $uuid = eval { tmp_config('--get', $key) };
1813 if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/) {
1814 $self->{ra_uuid} = $uuid;
1815 } else {
1816 die "ra_uuid called without URL\n" unless $self->{url};
1817 $self->{ra_uuid} = $self->ra->get_uuid;
1818 tmp_config('--add', $key, $self->{ra_uuid});
1819 }
1820 }
1821 $self->{ra_uuid};
1822 }
1823
1824 sub _set_repos_root {
1825 my ($self, $repos_root) = @_;
1826 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1827 $repos_root ||= $self->ra->{repos_root};
1828 tmp_config($k, $repos_root);
1829 $repos_root;
1830 }
1831
1832 sub repos_root {
1833 my ($self) = @_;
1834 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1835 eval { tmp_config('--get', $k) } || $self->_set_repos_root;
1836 }
1837
1838 sub ra {
1839 my ($self) = shift;
1840 my $ra = Git::SVN::Ra->new($self->{url});
1841 $self->_set_repos_root($ra->{repos_root});
1842 if ($self->use_svm_props && !$self->{svm}) {
1843 if ($self->no_metadata) {
1844 die "Can't have both 'noMetadata' and ",
1845 "'useSvmProps' options set!\n";
1846 } elsif ($self->use_svnsync_props) {
1847 die "Can't have both 'useSvnsyncProps' and ",
1848 "'useSvmProps' options set!\n";
1849 }
1850 $ra = $self->_set_svm_vars($ra);
1851 $self->{-want_revprops} = 1;
1852 }
1853 $ra;
1854 }
1855
1856 sub rel_path {
1857 my ($self) = @_;
1858 my $repos_root = $self->ra->{repos_root};
1859 return $self->{path} if ($self->{url} eq $repos_root);
1860 my $url = $self->{url} .
1861 (length $self->{path} ? "/$self->{path}" : $self->{path});
1862 $url =~ s!^\Q$repos_root\E(?:/+|$)!!g;
1863 $url;
1864 }
1865
1866 # prop_walk(PATH, REV, SUB)
1867 # -------------------------
1868 # Recursively traverse PATH at revision REV and invoke SUB for each
1869 # directory that contains a SVN property. SUB will be invoked as
1870 # follows: &SUB(gs, path, props); where `gs' is this instance of
1871 # Git::SVN, `path' the path to the directory where the properties
1872 # `props' were found. The `path' will be relative to point of checkout,
1873 # that is, if url://repo/trunk is the current Git branch, and that
1874 # directory contains a sub-directory `d', SUB will be invoked with `/d/'
1875 # as `path' (note the trailing `/').
1876 sub prop_walk {
1877 my ($self, $path, $rev, $sub) = @_;
1878
1879 $path =~ s#^/##;
1880 my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
1881 $path =~ s#^/*#/#g;
1882 my $p = $path;
1883 # Strip the irrelevant part of the path.
1884 $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
1885 # Ensure the path is terminated by a `/'.
1886 $p =~ s#/*$#/#;
1887
1888 # The properties contain all the internal SVN stuff nobody
1889 # (usually) cares about.
1890 my $interesting_props = 0;
1891 foreach (keys %{$props}) {
1892 # If it doesn't start with `svn:', it must be a
1893 # user-defined property.
1894 ++$interesting_props and next if $_ !~ /^svn:/;
1895 # FIXME: Fragile, if SVN adds new public properties,
1896 # this needs to be updated.
1897 ++$interesting_props if /^svn:(?:ignore|keywords|executable
1898 |eol-style|mime-type
1899 |externals|needs-lock)$/x;
1900 }
1901 &$sub($self, $p, $props) if $interesting_props;
1902
1903 foreach (sort keys %$dirent) {
1904 next if $dirent->{$_}->{kind} != $SVN::Node::dir;
1905 $self->prop_walk($p . $_, $rev, $sub);
1906 }
1907 }
1908
1909 sub last_rev { ($_[0]->last_rev_commit)[0] }
1910 sub last_commit { ($_[0]->last_rev_commit)[1] }
1911
1912 # returns the newest SVN revision number and newest commit SHA1
1913 sub last_rev_commit {
1914 my ($self) = @_;
1915 if (defined $self->{last_rev} && defined $self->{last_commit}) {
1916 return ($self->{last_rev}, $self->{last_commit});
1917 }
1918 my $c = ::verify_ref($self->refname.'^0');
1919 if ($c && !$self->use_svm_props && !$self->no_metadata) {
1920 my $rev = (::cmt_metadata($c))[1];
1921 if (defined $rev) {
1922 ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
1923 return ($rev, $c);
1924 }
1925 }
1926 my $map_path = $self->map_path;
1927 unless (-e $map_path) {
1928 ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
1929 return (undef, undef);
1930 }
1931 my ($rev, $commit) = $self->rev_map_max(1);
1932 ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
1933 return ($rev, $commit);
1934 }
1935
1936 sub get_fetch_range {
1937 my ($self, $min, $max) = @_;
1938 $max ||= $self->ra->get_latest_revnum;
1939 $min ||= $self->rev_map_max;
1940 (++$min, $max);
1941 }
1942
1943 sub tmp_config {
1944 my (@args) = @_;
1945 my $old_def_config = "$ENV{GIT_DIR}/svn/config";
1946 my $config = "$ENV{GIT_DIR}/svn/.metadata";
1947 if (! -f $config && -f $old_def_config) {
1948 rename $old_def_config, $config or
1949 die "Failed rename $old_def_config => $config: $!\n";
1950 }
1951 my $old_config = $ENV{GIT_CONFIG};
1952 $ENV{GIT_CONFIG} = $config;
1953 $@ = undef;
1954 my @ret = eval {
1955 unless (-f $config) {
1956 mkfile($config);
1957 open my $fh, '>', $config or
1958 die "Can't open $config: $!\n";
1959 print $fh "; This file is used internally by ",
1960 "git-svn\n" or die
1961 "Couldn't write to $config: $!\n";
1962 print $fh "; You should not have to edit it\n" or
1963 die "Couldn't write to $config: $!\n";
1964 close $fh or die "Couldn't close $config: $!\n";
1965 }
1966 command('config', @args);
1967 };
1968 my $err = $@;
1969 if (defined $old_config) {
1970 $ENV{GIT_CONFIG} = $old_config;
1971 } else {
1972 delete $ENV{GIT_CONFIG};
1973 }
1974 die $err if $err;
1975 wantarray ? @ret : $ret[0];
1976 }
1977
1978 sub tmp_index_do {
1979 my ($self, $sub) = @_;
1980 my $old_index = $ENV{GIT_INDEX_FILE};
1981 $ENV{GIT_INDEX_FILE} = $self->{index};
1982 $@ = undef;
1983 my @ret = eval {
1984 my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
1985 mkpath([$dir]) unless -d $dir;
1986 &$sub;
1987 };
1988 my $err = $@;
1989 if (defined $old_index) {
1990 $ENV{GIT_INDEX_FILE} = $old_index;
1991 } else {
1992 delete $ENV{GIT_INDEX_FILE};
1993 }
1994 die $err if $err;
1995 wantarray ? @ret : $ret[0];
1996 }
1997
1998 sub assert_index_clean {
1999 my ($self, $treeish) = @_;
2000
2001 $self->tmp_index_do(sub {
2002 command_noisy('read-tree', $treeish) unless -e $self->{index};
2003 my $x = command_oneline('write-tree');
2004 my ($y) = (command(qw/cat-file commit/, $treeish) =~
2005 /^tree ($::sha1)/mo);
2006 return if $y eq $x;
2007
2008 warn "Index mismatch: $y != $x\nrereading $treeish\n";
2009 unlink $self->{index} or die "unlink $self->{index}: $!\n";
2010 command_noisy('read-tree', $treeish);
2011 $x = command_oneline('write-tree');
2012 if ($y ne $x) {
2013 ::fatal "trees ($treeish) $y != $x\n",
2014 "Something is seriously wrong...";
2015 }
2016 });
2017 }
2018
2019 sub get_commit_parents {
2020 my ($self, $log_entry) = @_;
2021 my (%seen, @ret, @tmp);
2022 # legacy support for 'set-tree'; this is only used by set_tree_cb:
2023 if (my $ip = $self->{inject_parents}) {
2024 if (my $commit = delete $ip->{$log_entry->{revision}}) {
2025 push @tmp, $commit;
2026 }
2027 }
2028 if (my $cur = ::verify_ref($self->refname.'^0')) {
2029 push @tmp, $cur;
2030 }
2031 if (my $ipd = $self->{inject_parents_dcommit}) {
2032 if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2033 push @tmp, @$commit;
2034 }
2035 }
2036 push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2037 while (my $p = shift @tmp) {
2038 next if $seen{$p};
2039 $seen{$p} = 1;
2040 push @ret, $p;
2041 # MAXPARENT is defined to 16 in commit-tree.c:
2042 last if @ret >= 16;
2043 }
2044 if (@tmp) {
2045 die "r$log_entry->{revision}: No room for parents:\n\t",
2046 join("\n\t", @tmp), "\n";
2047 }
2048 @ret;
2049 }
2050
2051 sub rewrite_root {
2052 my ($self) = @_;
2053 return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2054 my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2055 my $rwr = eval { command_oneline(qw/config --get/, $k) };
2056 if ($rwr) {
2057 $rwr =~ s#/+$##;
2058 if ($rwr !~ m#^[a-z\+]+://#) {
2059 die "$rwr is not a valid URL (key: $k)\n";
2060 }
2061 }
2062 $self->{-rewrite_root} = $rwr;
2063 }
2064
2065 sub metadata_url {
2066 my ($self) = @_;
2067 ($self->rewrite_root || $self->{url}) .
2068 (length $self->{path} ? '/' . $self->{path} : '');
2069 }
2070
2071 sub full_url {
2072 my ($self) = @_;
2073 $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2074 }
2075
2076
2077 sub set_commit_header_env {
2078 my ($log_entry) = @_;
2079 my %env;
2080 foreach my $ned (qw/NAME EMAIL DATE/) {
2081 foreach my $ac (qw/AUTHOR COMMITTER/) {
2082 $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2083 }
2084 }
2085
2086 $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2087 $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2088 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2089
2090 $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2091 ? $log_entry->{commit_name}
2092 : $log_entry->{name};
2093 $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2094 ? $log_entry->{commit_email}
2095 : $log_entry->{email};
2096 \%env;
2097 }
2098
2099 sub restore_commit_header_env {
2100 my ($env) = @_;
2101 foreach my $ned (qw/NAME EMAIL DATE/) {
2102 foreach my $ac (qw/AUTHOR COMMITTER/) {
2103 my $k = "GIT_${ac}_${ned}";
2104 if (defined $env->{$k}) {
2105 $ENV{$k} = $env->{$k};
2106 } else {
2107 delete $ENV{$k};
2108 }
2109 }
2110 }
2111 }
2112
2113 sub gc {
2114 command_noisy('gc', '--auto');
2115 };
2116
2117 sub do_git_commit {
2118 my ($self, $log_entry) = @_;
2119 my $lr = $self->last_rev;
2120 if (defined $lr && $lr >= $log_entry->{revision}) {
2121 die "Last fetched revision of ", $self->refname,
2122 " was r$lr, but we are about to fetch: ",
2123 "r$log_entry->{revision}!\n";
2124 }
2125 if (my $c = $self->rev_map_get($log_entry->{revision})) {
2126 croak "$log_entry->{revision} = $c already exists! ",
2127 "Why are we refetching it?\n";
2128 }
2129 my $old_env = set_commit_header_env($log_entry);
2130 my $tree = $log_entry->{tree};
2131 if (!defined $tree) {
2132 $tree = $self->tmp_index_do(sub {
2133 command_oneline('write-tree') });
2134 }
2135 die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2136
2137 my @exec = ('git-commit-tree', $tree);
2138 foreach ($self->get_commit_parents($log_entry)) {
2139 push @exec, '-p', $_;
2140 }
2141 defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2142 or croak $!;
2143 print $msg_fh $log_entry->{log} or croak $!;
2144 restore_commit_header_env($old_env);
2145 unless ($self->no_metadata) {
2146 print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2147 or croak $!;
2148 }
2149 $msg_fh->flush == 0 or croak $!;
2150 close $msg_fh or croak $!;
2151 chomp(my $commit = do { local $/; <$out_fh> });
2152 close $out_fh or croak $!;
2153 waitpid $pid, 0;
2154 croak $? if $?;
2155 if ($commit !~ /^$::sha1$/o) {
2156 die "Failed to commit, invalid sha1: $commit\n";
2157 }
2158
2159 $self->rev_map_set($log_entry->{revision}, $commit, 1);
2160
2161 $self->{last_rev} = $log_entry->{revision};
2162 $self->{last_commit} = $commit;
2163 print "r$log_entry->{revision}";
2164 if (defined $log_entry->{svm_revision}) {
2165 print " (\@$log_entry->{svm_revision})";
2166 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2167 0, $self->svm_uuid);
2168 }
2169 print " = $commit ($self->{ref_id})\n";
2170 if (--$_gc_nr == 0) {
2171 $_gc_nr = $_gc_period;
2172 gc();
2173 }
2174 return $commit;
2175 }
2176
2177 sub match_paths {
2178 my ($self, $paths, $r) = @_;
2179 return 1 if $self->{path} eq '';
2180 if (my $path = $paths->{"/$self->{path}"}) {
2181 return ($path->{action} eq 'D') ? 0 : 1;
2182 }
2183 $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
2184 if (grep /$self->{path_regex}/, keys %$paths) {
2185 return 1;
2186 }
2187 my $c = '';
2188 foreach (split m#/#, $self->{path}) {
2189 $c .= "/$_";
2190 next unless ($paths->{$c} &&
2191 ($paths->{$c}->{action} =~ /^[AR]$/));
2192 if ($self->ra->check_path($self->{path}, $r) ==
2193 $SVN::Node::dir) {
2194 return 1;
2195 }
2196 }
2197 return 0;
2198 }
2199
2200 sub find_parent_branch {
2201 my ($self, $paths, $rev) = @_;
2202 return undef unless $self->follow_parent;
2203 unless (defined $paths) {
2204 my $err_handler = $SVN::Error::handler;
2205 $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2206 $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1, sub {
2207 $paths =
2208 Git::SVN::Ra::dup_changed_paths($_[0]) });
2209 $SVN::Error::handler = $err_handler;
2210 }
2211 return undef unless defined $paths;
2212
2213 # look for a parent from another branch:
2214 my @b_path_components = split m#/#, $self->rel_path;
2215 my @a_path_components;
2216 my $i;
2217 while (@b_path_components) {
2218 $i = $paths->{'/'.join('/', @b_path_components)};
2219 last if $i && defined $i->{copyfrom_path};
2220 unshift(@a_path_components, pop(@b_path_components));
2221 }
2222 return undef unless defined $i && defined $i->{copyfrom_path};
2223 my $branch_from = $i->{copyfrom_path};
2224 if (@a_path_components) {
2225 print STDERR "branch_from: $branch_from => ";
2226 $branch_from .= '/'.join('/', @a_path_components);
2227 print STDERR $branch_from, "\n";
2228 }
2229 my $r = $i->{copyfrom_rev};
2230 my $repos_root = $self->ra->{repos_root};
2231 my $url = $self->ra->{url};
2232 my $new_url = $repos_root . $branch_from;
2233 print STDERR "Found possible branch point: ",
2234 "$new_url => ", $self->full_url, ", $r\n";
2235 $branch_from =~ s#^/##;
2236 my $gs = Git::SVN->find_by_url($new_url, $repos_root, $branch_from);
2237 unless ($gs) {
2238 my $ref_id = $self->{ref_id};
2239 $ref_id =~ s/\@\d+$//;
2240 $ref_id .= "\@$r";
2241 # just grow a tail if we're not unique enough :x
2242 $ref_id .= '-' while find_ref($ref_id);
2243 print STDERR "Initializing parent: $ref_id\n";
2244 my ($u, $p, $repo_id) = ($new_url, '', $ref_id);
2245 if ($u =~ s#^\Q$url\E(/|$)##) {
2246 $p = $u;
2247 $u = $url;
2248 $repo_id = $self->{repo_id};
2249 }
2250 $gs = Git::SVN->init($u, $p, $repo_id, $ref_id, 1);
2251 }
2252 my ($r0, $parent) = $gs->find_rev_before($r, 1);
2253 if (!defined $r0 || !defined $parent) {
2254 my ($base, $head) = parse_revision_argument(0, $r);
2255 if ($base <= $r) {
2256 $gs->fetch($base, $r);
2257 }
2258 ($r0, $parent) = $gs->last_rev_commit;
2259 }
2260 if (defined $r0 && defined $parent) {
2261 print STDERR "Found branch parent: ($self->{ref_id}) $parent\n";
2262 my $ed;
2263 if ($self->ra->can_do_switch) {
2264 $self->assert_index_clean($parent);
2265 print STDERR "Following parent with do_switch\n";
2266 # do_switch works with svn/trunk >= r22312, but that
2267 # is not included with SVN 1.4.3 (the latest version
2268 # at the moment), so we can't rely on it
2269 $self->{last_commit} = $parent;
2270 $ed = SVN::Git::Fetcher->new($self);
2271 $gs->ra->gs_do_switch($r0, $rev, $gs,
2272 $self->full_url, $ed)
2273 or die "SVN connection failed somewhere...\n";
2274 } elsif ($self->ra->trees_match($new_url, $r0,
2275 $self->full_url, $rev)) {
2276 print STDERR "Trees match:\n",
2277 " $new_url\@$r0\n",
2278 " ${\$self->full_url}\@$rev\n",
2279 "Following parent with no changes\n";
2280 $self->tmp_index_do(sub {
2281 command_noisy('read-tree', $parent);
2282 });
2283 $self->{last_commit} = $parent;
2284 } else {
2285 print STDERR "Following parent with do_update\n";
2286 $ed = SVN::Git::Fetcher->new($self);
2287 $self->ra->gs_do_update($rev, $rev, $self, $ed)
2288 or die "SVN connection failed somewhere...\n";
2289 }
2290 print STDERR "Successfully followed parent\n";
2291 return $self->make_log_entry($rev, [$parent], $ed);
2292 }
2293 return undef;
2294 }
2295
2296 sub do_fetch {
2297 my ($self, $paths, $rev) = @_;
2298 my $ed;
2299 my ($last_rev, @parents);
2300 if (my $lc = $self->last_commit) {
2301 # we can have a branch that was deleted, then re-added
2302 # under the same name but copied from another path, in
2303 # which case we'll have multiple parents (we don't
2304 # want to break the original ref, nor lose copypath info):
2305 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2306 push @{$log_entry->{parents}}, $lc;
2307 return $log_entry;
2308 }
2309 $ed = SVN::Git::Fetcher->new($self);
2310 $last_rev = $self->{last_rev};
2311 $ed->{c} = $lc;
2312 @parents = ($lc);
2313 } else {
2314 $last_rev = $rev;
2315 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2316 return $log_entry;
2317 }
2318 $ed = SVN::Git::Fetcher->new($self);
2319 }
2320 unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2321 die "SVN connection failed somewhere...\n";
2322 }
2323 $self->make_log_entry($rev, \@parents, $ed);
2324 }
2325
2326 sub get_untracked {
2327 my ($self, $ed) = @_;
2328 my @out;
2329 my $h = $ed->{empty};
2330 foreach (sort keys %$h) {
2331 my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2332 push @out, " $act: " . uri_encode($_);
2333 warn "W: $act: $_\n";
2334 }
2335 foreach my $t (qw/dir_prop file_prop/) {
2336 $h = $ed->{$t} or next;
2337 foreach my $path (sort keys %$h) {
2338 my $ppath = $path eq '' ? '.' : $path;
2339 foreach my $prop (sort keys %{$h->{$path}}) {
2340 next if $SKIP_PROP{$prop};
2341 my $v = $h->{$path}->{$prop};
2342 my $t_ppath_prop = "$t: " .
2343 uri_encode($ppath) . ' ' .
2344 uri_encode($prop);
2345 if (defined $v) {
2346 push @out, " +$t_ppath_prop " .
2347 uri_encode($v);
2348 } else {
2349 push @out, " -$t_ppath_prop";
2350 }
2351 }
2352 }
2353 }
2354 foreach my $t (qw/absent_file absent_directory/) {
2355 $h = $ed->{$t} or next;
2356 foreach my $parent (sort keys %$h) {
2357 foreach my $path (sort @{$h->{$parent}}) {
2358 push @out, " $t: " .
2359 uri_encode("$parent/$path");
2360 warn "W: $t: $parent/$path ",
2361 "Insufficient permissions?\n";
2362 }
2363 }
2364 }
2365 \@out;
2366 }
2367
2368 sub parse_svn_date {
2369 my $date = shift || return '+0000 1970-01-01 00:00:00';
2370 my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2371 (\d\d)\:(\d\d)\:(\d\d).\d+Z$/x) or
2372 croak "Unable to parse date: $date\n";
2373 "+0000 $Y-$m-$d $H:$M:$S";
2374 }
2375
2376 sub check_author {
2377 my ($author) = @_;
2378 if (!defined $author || length $author == 0) {
2379 $author = '(no author)';
2380 } elsif (defined $::_authors && ! defined $::users{$author}) {
2381 die "Author: $author not defined in $::_authors file\n";
2382 }
2383 $author;
2384 }
2385
2386 sub make_log_entry {
2387 my ($self, $rev, $parents, $ed) = @_;
2388 my $untracked = $self->get_untracked($ed);
2389
2390 open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
2391 print $un "r$rev\n" or croak $!;
2392 print $un $_, "\n" foreach @$untracked;
2393 my %log_entry = ( parents => $parents || [], revision => $rev,
2394 log => '');
2395
2396 my $headrev;
2397 my $logged = delete $self->{logged_rev_props};
2398 if (!$logged || $self->{-want_revprops}) {
2399 my $rp = $self->ra->rev_proplist($rev);
2400 foreach (sort keys %$rp) {
2401 my $v = $rp->{$_};
2402 if (/^svn:(author|date|log)$/) {
2403 $log_entry{$1} = $v;
2404 } elsif ($_ eq 'svm:headrev') {
2405 $headrev = $v;
2406 } else {
2407 print $un " rev_prop: ", uri_encode($_), ' ',
2408 uri_encode($v), "\n";
2409 }
2410 }
2411 } else {
2412 map { $log_entry{$_} = $logged->{$_} } keys %$logged;
2413 }
2414 close $un or croak $!;
2415
2416 $log_entry{date} = parse_svn_date($log_entry{date});
2417 $log_entry{log} .= "\n";
2418 my $author = $log_entry{author} = check_author($log_entry{author});
2419 my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
2420 : ($author, undef);
2421
2422 my ($commit_name, $commit_email) = ($name, $email);
2423 if ($_use_log_author) {
2424 my $name_field;
2425 if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
2426 $name_field = $1;
2427 } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
2428 $name_field = $1;
2429 }
2430 if (!defined $name_field) {
2431 #
2432 } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
2433 ($name, $email) = ($1, $2);
2434 } elsif ($name_field =~ /(.*)@/) {
2435 ($name, $email) = ($1, $name_field);
2436 } else {
2437 ($name, $email) = ($name_field, 'unknown');
2438 }
2439 }
2440 if (defined $headrev && $self->use_svm_props) {
2441 if ($self->rewrite_root) {
2442 die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
2443 "options set!\n";
2444 }
2445 my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$};
2446 # we don't want "SVM: initializing mirror for junk" ...
2447 return undef if $r == 0;
2448 my $svm = $self->svm;
2449 if ($uuid ne $svm->{uuid}) {
2450 die "UUID mismatch on SVM path:\n",
2451 "expected: $svm->{uuid}\n",
2452 " got: $uuid\n";
2453 }
2454 my $full_url = $self->full_url;
2455 $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
2456 die "Failed to replace '$svm->{replace}' with ",
2457 "'$svm->{source}' in $full_url\n";
2458 # throw away username for storing in records
2459 remove_username($full_url);
2460 $log_entry{metadata} = "$full_url\@$r $uuid";
2461 $log_entry{svm_revision} = $r;
2462 $email ||= "$author\@$uuid";
2463 $commit_email ||= "$author\@$uuid";
2464 } elsif ($self->use_svnsync_props) {
2465 my $full_url = $self->svnsync->{url};
2466 $full_url .= "/$self->{path}" if length $self->{path};
2467 remove_username($full_url);
2468 my $uuid = $self->svnsync->{uuid};
2469 $log_entry{metadata} = "$full_url\@$rev $uuid";
2470 $email ||= "$author\@$uuid";
2471 $commit_email ||= "$author\@$uuid";
2472 } else {
2473 my $url = $self->metadata_url;
2474 remove_username($url);
2475 $log_entry{metadata} = "$url\@$rev " .
2476 $self->ra->get_uuid;
2477 $email ||= "$author\@" . $self->ra->get_uuid;
2478 $commit_email ||= "$author\@" . $self->ra->get_uuid;
2479 }
2480 $log_entry{name} = $name;
2481 $log_entry{email} = $email;
2482 $log_entry{commit_name} = $commit_name;
2483 $log_entry{commit_email} = $commit_email;
2484 \%log_entry;
2485 }
2486
2487 sub fetch {
2488 my ($self, $min_rev, $max_rev, @parents) = @_;
2489 my ($last_rev, $last_commit) = $self->last_rev_commit;
2490 my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
2491 $self->ra->gs_fetch_loop_common($base, $head, [$self]);
2492 }
2493
2494 sub set_tree_cb {
2495 my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
2496 $self->{inject_parents} = { $rev => $tree };
2497 $self->fetch(undef, undef);
2498 }
2499
2500 sub set_tree {
2501 my ($self, $tree) = (shift, shift);
2502 my $log_entry = ::get_commit_entry($tree);
2503 unless ($self->{last_rev}) {
2504 fatal("Must have an existing revision to commit");
2505 }
2506 my %ed_opts = ( r => $self->{last_rev},
2507 log => $log_entry->{log},
2508 ra => $self->ra,
2509 tree_a => $self->{last_commit},
2510 tree_b => $tree,
2511 editor_cb => sub {
2512 $self->set_tree_cb($log_entry, $tree, @_) },
2513 svn_path => $self->{path} );
2514 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
2515 print "No changes\nr$self->{last_rev} = $tree\n";
2516 }
2517 }
2518
2519 sub rebuild_from_rev_db {
2520 my ($self, $path) = @_;
2521 my $r = -1;
2522 open my $fh, '<', $path or croak "open: $!";
2523 binmode $fh or croak "binmode: $!";
2524 while (<$fh>) {
2525 length($_) == 41 or croak "inconsistent size in ($_) != 41";
2526 chomp($_);
2527 ++$r;
2528 next if $_ eq ('0' x 40);
2529 $self->rev_map_set($r, $_);
2530 print "r$r = $_\n";
2531 }
2532 close $fh or croak "close: $!";
2533 unlink $path or croak "unlink: $!";
2534 }
2535
2536 sub rebuild {
2537 my ($self) = @_;
2538 my $map_path = $self->map_path;
2539 return if (-e $map_path && ! -z $map_path);
2540 return unless ::verify_ref($self->refname.'^0');
2541 if ($self->use_svm_props || $self->no_metadata) {
2542 my $rev_db = $self->rev_db_path;
2543 $self->rebuild_from_rev_db($rev_db);
2544 if ($self->use_svm_props) {
2545 my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
2546 $self->rebuild_from_rev_db($svm_rev_db);
2547 }
2548 $self->unlink_rev_db_symlink;
2549 return;
2550 }
2551 print "Rebuilding $map_path ...\n";
2552 my ($log, $ctx) =
2553 command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
2554 $self->refname, '--');
2555 my $full_url = $self->full_url;
2556 remove_username($full_url);
2557 my $svn_uuid = $self->ra_uuid;
2558 my $c;
2559 while (<$log>) {
2560 if ( m{^commit ($::sha1)$} ) {
2561 $c = $1;
2562 next;
2563 }
2564 next unless s{^\s*(git-svn-id:)}{$1};
2565 my ($url, $rev, $uuid) = ::extract_metadata($_);
2566 remove_username($url);
2567
2568 # ignore merges (from set-tree)
2569 next if (!defined $rev || !$uuid);
2570
2571 # if we merged or otherwise started elsewhere, this is
2572 # how we break out of it
2573 if (($uuid ne $svn_uuid) ||
2574 ($full_url && $url && ($url ne $full_url))) {
2575 next;
2576 }
2577
2578 $self->rev_map_set($rev, $c);
2579 print "r$rev = $c\n";
2580 }
2581 command_close_pipe($log, $ctx);
2582 print "Done rebuilding $map_path\n";
2583 my $rev_db_path = $self->rev_db_path;
2584 if (-f $self->rev_db_path) {
2585 unlink $self->rev_db_path or croak "unlink: $!";
2586 }
2587 $self->unlink_rev_db_symlink;
2588 }
2589
2590 # rev_map:
2591 # Tie::File seems to be prone to offset errors if revisions get sparse,
2592 # it's not that fast, either. Tie::File is also not in Perl 5.6. So
2593 # one of my favorite modules is out :< Next up would be one of the DBM
2594 # modules, but I'm not sure which is most portable...
2595 #
2596 # This is the replacement for the rev_db format, which was too big
2597 # and inefficient for large repositories with a lot of sparse history
2598 # (mainly tags)
2599 #
2600 # The format is this:
2601 # - 24 bytes for every record,
2602 # * 4 bytes for the integer representing an SVN revision number
2603 # * 20 bytes representing the sha1 of a git commit
2604 # - No empty padding records like the old format
2605 # (except the last record, which can be overwritten)
2606 # - new records are written append-only since SVN revision numbers
2607 # increase monotonically
2608 # - lookups on SVN revision number are done via a binary search
2609 # - Piping the file to xxd -c24 is a good way of dumping it for
2610 # viewing or editing (piped back through xxd -r), should the need
2611 # ever arise.
2612 # - The last record can be padding revision with an all-zero sha1
2613 # This is used to optimize fetch performance when using multiple
2614 # "fetch" directives in .git/config
2615 #
2616 # These files are disposable unless noMetadata or useSvmProps is set
2617
2618 sub _rev_map_set {
2619 my ($fh, $rev, $commit) = @_;
2620
2621 binmode $fh or croak "binmode: $!";
2622 my $size = (stat($fh))[7];
2623 ($size % 24) == 0 or croak "inconsistent size: $size";
2624
2625 my $wr_offset = 0;
2626 if ($size > 0) {
2627 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2628 my $read = sysread($fh, my $buf, 24) or croak "read: $!";
2629 $read == 24 or croak "read only $read bytes (!= 24)";
2630 my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
2631 if ($last_commit eq ('0' x40)) {
2632 if ($size >= 48) {
2633 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2634 $read = sysread($fh, $buf, 24) or
2635 croak "read: $!";
2636 $read == 24 or
2637 croak "read only $read bytes (!= 24)";
2638 ($last_rev, $last_commit) =
2639 unpack(rev_map_fmt, $buf);
2640 if ($last_commit eq ('0' x40)) {
2641 croak "inconsistent .rev_map\n";
2642 }
2643 }
2644 if ($last_rev >= $rev) {
2645 croak "last_rev is higher!: $last_rev >= $rev";
2646 }
2647 $wr_offset = -24;
2648 }
2649 }
2650 sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
2651 syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
2652 croak "write: $!";
2653 }
2654
2655 sub mkfile {
2656 my ($path) = @_;
2657 unless (-e $path) {
2658 my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
2659 mkpath([$dir]) unless -d $dir;
2660 open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
2661 close $fh or die "Couldn't close (create) $path: $!\n";
2662 }
2663 }
2664
2665 sub rev_map_set {
2666 my ($self, $rev, $commit, $update_ref, $uuid) = @_;
2667 length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
2668 my $db = $self->map_path($uuid);
2669 my $db_lock = "$db.lock";
2670 my $sig;
2671 if ($update_ref) {
2672 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2673 $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
2674 }
2675 mkfile($db);
2676
2677 $LOCKFILES{$db_lock} = 1;
2678 my $sync;
2679 # both of these options make our .rev_db file very, very important
2680 # and we can't afford to lose it because rebuild() won't work
2681 if ($self->use_svm_props || $self->no_metadata) {
2682 $sync = 1;
2683 copy($db, $db_lock) or die "rev_map_set(@_): ",
2684 "Failed to copy: ",
2685 "$db => $db_lock ($!)\n";
2686 } else {
2687 rename $db, $db_lock or die "rev_map_set(@_): ",
2688 "Failed to rename: ",
2689 "$db => $db_lock ($!)\n";
2690 }
2691
2692 sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
2693 or croak "Couldn't open $db_lock: $!\n";
2694 _rev_map_set($fh, $rev, $commit);
2695 if ($sync) {
2696 $fh->flush or die "Couldn't flush $db_lock: $!\n";
2697 $fh->sync or die "Couldn't sync $db_lock: $!\n";
2698 }
2699 close $fh or croak $!;
2700 if ($update_ref) {
2701 $_head = $self;
2702 command_noisy('update-ref', '-m', "r$rev",
2703 $self->refname, $commit);
2704 }
2705 rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
2706 "$db_lock => $db ($!)\n";
2707 delete $LOCKFILES{$db_lock};
2708 if ($update_ref) {
2709 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2710 $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
2711 kill $sig, $$ if defined $sig;
2712 }
2713 }
2714
2715 # If want_commit, this will return an array of (rev, commit) where
2716 # commit _must_ be a valid commit in the archive.
2717 # Otherwise, it'll return the max revision (whether or not the
2718 # commit is valid or just a 0x40 placeholder).
2719 sub rev_map_max {
2720 my ($self, $want_commit) = @_;
2721 $self->rebuild;
2722 my $map_path = $self->map_path;
2723 stat $map_path or return $want_commit ? (0, undef) : 0;
2724 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2725 binmode $fh or croak "binmode: $!";
2726 my $size = (stat($fh))[7];
2727 ($size % 24) == 0 or croak "inconsistent size: $size";
2728
2729 if ($size == 0) {
2730 close $fh or croak "close: $!";
2731 return $want_commit ? (0, undef) : 0;
2732 }
2733
2734 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2735 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2736 my ($r, $c) = unpack(rev_map_fmt, $buf);
2737 if ($want_commit && $c eq ('0' x40)) {
2738 if ($size < 48) {
2739 return $want_commit ? (0, undef) : 0;
2740 }
2741 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2742 sysread($fh, $buf, 24) == 24 or croak "read: $!";
2743 ($r, $c) = unpack(rev_map_fmt, $buf);
2744 if ($c eq ('0'x40)) {
2745 croak "Penultimate record is all-zeroes in $map_path";
2746 }
2747 }
2748 close $fh or croak "close: $!";
2749 $want_commit ? ($r, $c) : $r;
2750 }
2751
2752 sub rev_map_get {
2753 my ($self, $rev, $uuid) = @_;
2754 my $map_path = $self->map_path($uuid);
2755 return undef unless -e $map_path;
2756
2757 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2758 binmode $fh or croak "binmode: $!";
2759 my $size = (stat($fh))[7];
2760 ($size % 24) == 0 or croak "inconsistent size: $size";
2761
2762 if ($size == 0) {
2763 close $fh or croak "close: $fh";
2764 return undef;
2765 }
2766
2767 my ($l, $u) = (0, $size - 24);
2768 my ($r, $c, $buf);
2769
2770 while ($l <= $u) {
2771 my $i = int(($l/24 + $u/24) / 2) * 24;
2772 sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
2773 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2774 my ($r, $c) = unpack('NH40', $buf);
2775
2776 if ($r < $rev) {
2777 $l = $i + 24;
2778 } elsif ($r > $rev) {
2779 $u = $i - 24;
2780 } else { # $r == $rev
2781 close($fh) or croak "close: $!";
2782 return $c eq ('0' x 40) ? undef : $c;
2783 }
2784 }
2785 close($fh) or croak "close: $!";
2786 undef;
2787 }
2788
2789 # Finds the first svn revision that exists on (if $eq_ok is true) or
2790 # before $rev for the current branch. It will not search any lower
2791 # than $min_rev. Returns the git commit hash and svn revision number
2792 # if found, else (undef, undef).
2793 sub find_rev_before {
2794 my ($self, $rev, $eq_ok, $min_rev) = @_;
2795 --$rev unless $eq_ok;
2796 $min_rev ||= 1;
2797 while ($rev >= $min_rev) {
2798 if (my $c = $self->rev_map_get($rev)) {
2799 return ($rev, $c);
2800 }
2801 --$rev;
2802 }
2803 return (undef, undef);
2804 }
2805
2806 # Finds the first svn revision that exists on (if $eq_ok is true) or
2807 # after $rev for the current branch. It will not search any higher
2808 # than $max_rev. Returns the git commit hash and svn revision number
2809 # if found, else (undef, undef).
2810 sub find_rev_after {
2811 my ($self, $rev, $eq_ok, $max_rev) = @_;
2812 ++$rev unless $eq_ok;
2813 $max_rev ||= $self->rev_map_max;
2814 while ($rev <= $max_rev) {
2815 if (my $c = $self->rev_map_get($rev)) {
2816 return ($rev, $c);
2817 }
2818 ++$rev;
2819 }
2820 return (undef, undef);
2821 }
2822
2823 sub _new {
2824 my ($class, $repo_id, $ref_id, $path) = @_;
2825 unless (defined $repo_id && length $repo_id) {
2826 $repo_id = $Git::SVN::default_repo_id;
2827 }
2828 unless (defined $ref_id && length $ref_id) {
2829 $_[2] = $ref_id = $Git::SVN::default_ref_id;
2830 }
2831 $_[1] = $repo_id = sanitize_remote_name($repo_id);
2832 my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
2833 $_[3] = $path = '' unless (defined $path);
2834 mkpath(["$ENV{GIT_DIR}/svn"]);
2835 bless {
2836 ref_id => $ref_id, dir => $dir, index => "$dir/index",
2837 path => $path, config => "$ENV{GIT_DIR}/svn/config",
2838 map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
2839 }
2840
2841 # for read-only access of old .rev_db formats
2842 sub unlink_rev_db_symlink {
2843 my ($self) = @_;
2844 my $link = $self->rev_db_path;
2845 $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
2846 if (-l $link) {
2847 unlink $link or croak "unlink: $link failed!";
2848 }
2849 }
2850
2851 sub rev_db_path {
2852 my ($self, $uuid) = @_;
2853 my $db_path = $self->map_path($uuid);
2854 $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
2855 or croak "map_path: $db_path does not contain '/.rev_map.' !";
2856 $db_path;
2857 }
2858
2859 # the new replacement for .rev_db
2860 sub map_path {
2861 my ($self, $uuid) = @_;
2862 $uuid ||= $self->ra_uuid;
2863 "$self->{map_root}.$uuid";
2864 }
2865
2866 sub uri_encode {
2867 my ($f) = @_;
2868 $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
2869 $f
2870 }
2871
2872 sub remove_username {
2873 $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
2874 }
2875
2876 package Git::SVN::Prompt;
2877 use strict;
2878 use warnings;
2879 require SVN::Core;
2880 use vars qw/$_no_auth_cache $_username/;
2881
2882 sub simple {
2883 my ($cred, $realm, $default_username, $may_save, $pool) = @_;
2884 $may_save = undef if $_no_auth_cache;
2885 $default_username = $_username if defined $_username;
2886 if (defined $default_username && length $default_username) {
2887 if (defined $realm && length $realm) {
2888 print STDERR "Authentication realm: $realm\n";
2889 STDERR->flush;
2890 }
2891 $cred->username($default_username);
2892 } else {
2893 username($cred, $realm, $may_save, $pool);
2894 }
2895 $cred->password(_read_password("Password for '" .
2896 $cred->username . "': ", $realm));
2897 $cred->may_save($may_save);
2898 $SVN::_Core::SVN_NO_ERROR;
2899 }
2900
2901 sub ssl_server_trust {
2902 my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
2903 $may_save = undef if $_no_auth_cache;
2904 print STDERR "Error validating server certificate for '$realm':\n";
2905 {
2906 no warnings 'once';
2907 # All variables SVN::Auth::SSL::* are used only once,
2908 # so we're shutting up Perl warnings about this.
2909 if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
2910 print STDERR " - The certificate is not issued ",
2911 "by a trusted authority. Use the\n",
2912 " fingerprint to validate ",
2913 "the certificate manually!\n";
2914 }
2915 if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
2916 print STDERR " - The certificate hostname ",
2917 "does not match.\n";
2918 }
2919 if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
2920 print STDERR " - The certificate is not yet valid.\n";
2921 }
2922 if ($failures & $SVN::Auth::SSL::EXPIRED) {
2923 print STDERR " - The certificate has expired.\n";
2924 }
2925 if ($failures & $SVN::Auth::SSL::OTHER) {
2926 print STDERR " - The certificate has ",
2927 "an unknown error.\n";
2928 }
2929 } # no warnings 'once'
2930 printf STDERR
2931 "Certificate information:\n".
2932 " - Hostname: %s\n".
2933 " - Valid: from %s until %s\n".
2934 " - Issuer: %s\n".
2935 " - Fingerprint: %s\n",
2936 map $cert_info->$_, qw(hostname valid_from valid_until
2937 issuer_dname fingerprint);
2938 my $choice;
2939 prompt:
2940 print STDERR $may_save ?
2941 "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
2942 "(R)eject or accept (t)emporarily? ";
2943 STDERR->flush;
2944 $choice = lc(substr(<STDIN> || 'R', 0, 1));
2945 if ($choice =~ /^t$/i) {
2946 $cred->may_save(undef);
2947 } elsif ($choice =~ /^r$/i) {
2948 return -1;
2949 } elsif ($may_save && $choice =~ /^p$/i) {
2950 $cred->may_save($may_save);
2951 } else {
2952 goto prompt;
2953 }
2954 $cred->accepted_failures($failures);
2955 $SVN::_Core::SVN_NO_ERROR;
2956 }
2957
2958 sub ssl_client_cert {
2959 my ($cred, $realm, $may_save, $pool) = @_;
2960 $may_save = undef if $_no_auth_cache;
2961 print STDERR "Client certificate filename: ";
2962 STDERR->flush;
2963 chomp(my $filename = <STDIN>);
2964 $cred->cert_file($filename);
2965 $cred->may_save($may_save);
2966 $SVN::_Core::SVN_NO_ERROR;
2967 }
2968
2969 sub ssl_client_cert_pw {
2970 my ($cred, $realm, $may_save, $pool) = @_;
2971 $may_save = undef if $_no_auth_cache;
2972 $cred->password(_read_password("Password: ", $realm));
2973 $cred->may_save($may_save);
2974 $SVN::_Core::SVN_NO_ERROR;
2975 }
2976
2977 sub username {
2978 my ($cred, $realm, $may_save, $pool) = @_;
2979 $may_save = undef if $_no_auth_cache;
2980 if (defined $realm && length $realm) {
2981 print STDERR "Authentication realm: $realm\n";
2982 }
2983 my $username;
2984 if (defined $_username) {
2985 $username = $_username;
2986 } else {
2987 print STDERR "Username: ";
2988 STDERR->flush;
2989 chomp($username = <STDIN>);
2990 }
2991 $cred->username($username);
2992 $cred->may_save($may_save);
2993 $SVN::_Core::SVN_NO_ERROR;
2994 }
2995
2996 sub _read_password {
2997 my ($prompt, $realm) = @_;
2998 print STDERR $prompt;
2999 STDERR->flush;
3000 require Term::ReadKey;
3001 Term::ReadKey::ReadMode('noecho');
3002 my $password = '';
3003 while (defined(my $key = Term::ReadKey::ReadKey(0))) {
3004 last if $key =~ /[\012\015]/; # \n\r
3005 $password .= $key;
3006 }
3007 Term::ReadKey::ReadMode('restore');
3008 print STDERR "\n";
3009 STDERR->flush;
3010 $password;
3011 }
3012
3013 package SVN::Git::Fetcher;
3014 use vars qw/@ISA/;
3015 use strict;
3016 use warnings;
3017 use Carp qw/croak/;
3018 use IO::File qw//;
3019
3020 # file baton members: path, mode_a, mode_b, pool, fh, blob, base
3021 sub new {
3022 my ($class, $git_svn) = @_;
3023 my $self = SVN::Delta::Editor->new;
3024 bless $self, $class;
3025 $self->{c} = $git_svn->{last_commit} if exists $git_svn->{last_commit};
3026 $self->{empty} = {};
3027 $self->{dir_prop} = {};
3028 $self->{file_prop} = {};
3029 $self->{absent_dir} = {};
3030 $self->{absent_file} = {};
3031 $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
3032 $self;
3033 }
3034
3035 sub set_path_strip {
3036 my ($self, $path) = @_;
3037 $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
3038 }
3039
3040 sub open_root {
3041 { path => '' };
3042 }
3043
3044 sub open_directory {
3045 my ($self, $path, $pb, $rev) = @_;
3046 { path => $path };
3047 }
3048
3049 sub git_path {
3050 my ($self, $path) = @_;
3051 if ($self->{path_strip}) {
3052 $path =~ s!$self->{path_strip}!! or
3053 die "Failed to strip path '$path' ($self->{path_strip})\n";
3054 }
3055 $path;
3056 }
3057
3058 sub delete_entry {
3059 my ($self, $path, $rev, $pb) = @_;
3060
3061 my $gpath = $self->git_path($path);
3062 return undef if ($gpath eq '');
3063
3064 # remove entire directories.
3065 if (command('ls-tree', $self->{c}, '--', $gpath) =~ /^040000 tree/) {
3066 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3067 -r --name-only -z/,
3068 $self->{c}, '--', $gpath);
3069 local $/ = "\0";
3070 while (<$ls>) {
3071 chomp;
3072 $self->{gii}->remove($_);
3073 print "\tD\t$_\n" unless $::_q;
3074 }
3075 print "\tD\t$gpath/\n" unless $::_q;
3076 command_close_pipe($ls, $ctx);
3077 $self->{empty}->{$path} = 0
3078 } else {
3079 $self->{gii}->remove($gpath);
3080 print "\tD\t$gpath\n" unless $::_q;
3081 }
3082 undef;
3083 }
3084
3085 sub open_file {
3086 my ($self, $path, $pb, $rev) = @_;
3087 my $gpath = $self->git_path($path);
3088 my ($mode, $blob) = (command('ls-tree', $self->{c}, '--', $gpath)
3089 =~ /^(\d{6}) blob ([a-f\d]{40})\t/);
3090 unless (defined $mode && defined $blob) {
3091 die "$path was not found in commit $self->{c} (r$rev)\n";
3092 }
3093 { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3094 pool => SVN::Pool->new, action => 'M' };
3095 }
3096
3097 sub add_file {
3098 my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3099 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3100 delete $self->{empty}->{$dir};
3101 { path => $path, mode_a => 100644, mode_b => 100644,
3102 pool => SVN::Pool->new, action => 'A' };
3103 }
3104
3105 sub add_directory {
3106 my ($self, $path, $cp_path, $cp_rev) = @_;
3107 my $gpath = $self->git_path($path);
3108 if ($gpath eq '') {
3109 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3110 -r --name-only -z/,
3111 $self->{c});
3112 local $/ = "\0";
3113 while (<$ls>) {
3114 chomp;
3115 $self->{gii}->remove($_);
3116 print "\tD\t$_\n" unless $::_q;
3117 }
3118 command_close_pipe($ls, $ctx);
3119 $self->{empty}->{$path} = 0;
3120 }
3121 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3122 delete $self->{empty}->{$dir};
3123 $self->{empty}->{$path} = 1;
3124 { path => $path };
3125 }
3126
3127 sub change_dir_prop {
3128 my ($self, $db, $prop, $value) = @_;
3129 $self->{dir_prop}->{$db->{path}} ||= {};
3130 $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3131 undef;
3132 }
3133
3134 sub absent_directory {
3135 my ($self, $path, $pb) = @_;
3136 $self->{absent_dir}->{$pb->{path}} ||= [];
3137 push @{$self->{absent_dir}->{$pb->{path}}}, $path;
3138 undef;
3139 }
3140
3141 sub absent_file {
3142 my ($self, $path, $pb) = @_;
3143 $self->{absent_file}->{$pb->{path}} ||= [];
3144 push @{$self->{absent_file}->{$pb->{path}}}, $path;
3145 undef;
3146 }
3147
3148 sub change_file_prop {
3149 my ($self, $fb, $prop, $value) = @_;
3150 if ($prop eq 'svn:executable') {
3151 if ($fb->{mode_b} != 120000) {
3152 $fb->{mode_b} = defined $value ? 100755 : 100644;
3153 }
3154 } elsif ($prop eq 'svn:special') {
3155 $fb->{mode_b} = defined $value ? 120000 : 100644;
3156 } else {
3157 $self->{file_prop}->{$fb->{path}} ||= {};
3158 $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
3159 }
3160 undef;
3161 }
3162
3163 sub apply_textdelta {
3164 my ($self, $fb, $exp) = @_;
3165 my $fh = IO::File->new_tmpfile;
3166 $fh->autoflush(1);
3167 # $fh gets auto-closed() by SVN::TxDelta::apply(),
3168 # (but $base does not,) so dup() it for reading in close_file
3169 open my $dup, '<&', $fh or croak $!;
3170 my $base = IO::File->new_tmpfile;
3171 $base->autoflush(1);
3172 if ($fb->{blob}) {
3173 defined (my $pid = fork) or croak $!;
3174 if (!$pid) {
3175 open STDOUT, '>&', $base or croak $!;
3176 print STDOUT 'link ' if ($fb->{mode_a} == 120000);
3177 exec qw/git-cat-file blob/, $fb->{blob} or croak $!;
3178 }
3179 waitpid $pid, 0;
3180 croak $? if $?;
3181
3182 if (defined $exp) {
3183 seek $base, 0, 0 or croak $!;
3184 my $got = ::md5sum($base);
3185 die "Checksum mismatch: $fb->{path} $fb->{blob}\n",
3186 "expected: $exp\n",
3187 " got: $got\n" if ($got ne $exp);
3188 }
3189 }
3190 seek $base, 0, 0 or croak $!;
3191 $fb->{fh} = $dup;
3192 $fb->{base} = $base;
3193 [ SVN::TxDelta::apply($base, $fh, undef, $fb->{path}, $fb->{pool}) ];
3194 }
3195
3196 sub close_file {
3197 my ($self, $fb, $exp) = @_;
3198 my $hash;
3199 my $path = $self->git_path($fb->{path});
3200 if (my $fh = $fb->{fh}) {
3201 if (defined $exp) {
3202 seek($fh, 0, 0) or croak $!;
3203 my $got = ::md5sum($fh);
3204 if ($got ne $exp) {
3205 die "Checksum mismatch: $path\n",
3206 "expected: $exp\n got: $got\n";
3207 }
3208 }
3209 sysseek($fh, 0, 0) or croak $!;
3210 if ($fb->{mode_b} == 120000) {
3211 eval {
3212 sysread($fh, my $buf, 5) == 5 or croak $!;
3213 $buf eq 'link ' or die "$path has mode 120000",
3214 " but is not a link";
3215 };
3216 if ($@) {
3217 warn "$@\n";
3218 sysseek($fh, 0, 0) or croak $!;
3219 }
3220 }
3221 defined(my $pid = open my $out,'-|') or die "Can't fork: $!\n";
3222 if (!$pid) {
3223 open STDIN, '<&', $fh or croak $!;
3224 exec qw/git-hash-object -w --stdin/ or croak $!;
3225 }
3226 chomp($hash = do { local $/; <$out> });
3227 close $out or croak $!;
3228 close $fh or croak $!;
3229 $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
3230 close $fb->{base} or croak $!;
3231 } else {
3232 $hash = $fb->{blob} or die "no blob information\n";
3233 }
3234 $fb->{pool}->clear;
3235 $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
3236 print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
3237 undef;
3238 }
3239
3240 sub abort_edit {
3241 my $self = shift;
3242 $self->{nr} = $self->{gii}->{nr};
3243 delete $self->{gii};
3244 $self->SUPER::abort_edit(@_);
3245 }
3246
3247 sub close_edit {
3248 my $self = shift;
3249 $self->{git_commit_ok} = 1;
3250 $self->{nr} = $self->{gii}->{nr};
3251 delete $self->{gii};
3252 $self->SUPER::close_edit(@_);
3253 }
3254
3255 package SVN::Git::Editor;
3256 use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
3257 use strict;
3258 use warnings;
3259 use Carp qw/croak/;
3260 use IO::File;
3261
3262 sub new {
3263 my ($class, $opts) = @_;
3264 foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
3265 die "$_ required!\n" unless (defined $opts->{$_});
3266 }
3267
3268 my $pool = SVN::Pool->new;
3269 my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
3270 my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
3271 $opts->{r}, $mods);
3272
3273 # $opts->{ra} functions should not be used after this:
3274 my @ce = $opts->{ra}->get_commit_editor($opts->{log},
3275 $opts->{editor_cb}, $pool);
3276 my $self = SVN::Delta::Editor->new(@ce, $pool);
3277 bless $self, $class;
3278 foreach (qw/svn_path r tree_a tree_b/) {
3279 $self->{$_} = $opts->{$_};
3280 }
3281 $self->{url} = $opts->{ra}->{url};
3282 $self->{mods} = $mods;
3283 $self->{types} = $types;
3284 $self->{pool} = $pool;
3285 $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
3286 $self->{rm} = { };
3287 $self->{path_prefix} = length $self->{svn_path} ?
3288 "$self->{svn_path}/" : '';
3289 return $self;
3290 }
3291
3292 sub generate_diff {
3293 my ($tree_a, $tree_b) = @_;
3294 my @diff_tree = qw(diff-tree -z -r);
3295 if ($_cp_similarity) {
3296 push @diff_tree, "-C$_cp_similarity";
3297 } else {
3298 push @diff_tree, '-C';
3299 }
3300 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
3301 push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
3302 push @diff_tree, $tree_a, $tree_b;
3303 my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
3304 local $/ = "\0";
3305 my $state = 'meta';
3306 my @mods;
3307 while (<$diff_fh>) {
3308 chomp $_; # this gets rid of the trailing "\0"
3309 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
3310 $::sha1\s($::sha1)\s
3311 ([MTCRAD])\d*$/xo) {
3312 push @mods, { mode_a => $1, mode_b => $2,
3313 sha1_b => $3, chg => $4 };
3314 if ($4 =~ /^(?:C|R)$/) {
3315 $state = 'file_a';
3316 } else {
3317 $state = 'file_b';
3318 }
3319 } elsif ($state eq 'file_a') {
3320 my $x = $mods[$#mods] or croak "Empty array\n";
3321 if ($x->{chg} !~ /^(?:C|R)$/) {
3322 croak "Error parsing $_, $x->{chg}\n";
3323 }
3324 $x->{file_a} = $_;
3325 $state = 'file_b';
3326 } elsif ($state eq 'file_b') {
3327 my $x = $mods[$#mods] or croak "Empty array\n";
3328 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
3329 croak "Error parsing $_, $x->{chg}\n";
3330 }
3331 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
3332 croak "Error parsing $_, $x->{chg}\n";
3333 }
3334 $x->{file_b} = $_;
3335 $state = 'meta';
3336 } else {
3337 croak "Error parsing $_\n";
3338 }
3339 }
3340 command_close_pipe($diff_fh, $ctx);
3341 \@mods;
3342 }
3343
3344 sub check_diff_paths {
3345 my ($ra, $pfx, $rev, $mods) = @_;
3346 my %types;
3347 $pfx .= '/' if length $pfx;
3348
3349 sub type_diff_paths {
3350 my ($ra, $types, $path, $rev) = @_;
3351 my @p = split m#/+#, $path;
3352 my $c = shift @p;
3353 unless (defined $types->{$c}) {
3354 $types->{$c} = $ra->check_path($c, $rev);
3355 }
3356 while (@p) {
3357 $c .= '/' . shift @p;
3358 next if defined $types->{$c};
3359 $types->{$c} = $ra->check_path($c, $rev);
3360 }
3361 }
3362
3363 foreach my $m (@$mods) {
3364 foreach my $f (qw/file_a file_b/) {
3365 next unless defined $m->{$f};
3366 my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
3367 if (length $pfx.$dir && ! defined $types{$dir}) {
3368 type_diff_paths($ra, \%types, $pfx.$dir, $rev);
3369 }
3370 }
3371 }
3372 \%types;
3373 }
3374
3375 sub split_path {
3376 return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
3377 }
3378
3379 sub repo_path {
3380 my ($self, $path) = @_;
3381 $self->{path_prefix}.(defined $path ? $path : '');
3382 }
3383
3384 sub url_path {
3385 my ($self, $path) = @_;
3386 if ($self->{url} =~ m#^https?://#) {
3387 $path =~ s/([^a-zA-Z0-9_.-])/uc sprintf("%%%02x",ord($1))/eg;
3388 }
3389 $self->{url} . '/' . $self->repo_path($path);
3390 }
3391
3392 sub rmdirs {
3393 my ($self) = @_;
3394 my $rm = $self->{rm};
3395 delete $rm->{''}; # we never delete the url we're tracking
3396 return unless %$rm;
3397
3398 foreach (keys %$rm) {
3399 my @d = split m#/#, $_;
3400 my $c = shift @d;
3401 $rm->{$c} = 1;
3402 while (@d) {
3403 $c .= '/' . shift @d;
3404 $rm->{$c} = 1;
3405 }
3406 }
3407 delete $rm->{$self->{svn_path}};
3408 delete $rm->{''}; # we never delete the url we're tracking
3409 return unless %$rm;
3410
3411 my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
3412 $self->{tree_b});
3413 local $/ = "\0";
3414 while (<$fh>) {
3415 chomp;
3416 my @dn = split m#/#, $_;
3417 while (pop @dn) {
3418 delete $rm->{join '/', @dn};
3419 }
3420 unless (%$rm) {
3421 close $fh;
3422 return;
3423 }
3424 }
3425 command_close_pipe($fh, $ctx);
3426
3427 my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
3428 foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
3429 $self->close_directory($bat->{$d}, $p);
3430 my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
3431 print "\tD+\t$d/\n" unless $::_q;
3432 $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
3433 delete $bat->{$d};
3434 }
3435 }
3436
3437 sub open_or_add_dir {
3438 my ($self, $full_path, $baton) = @_;
3439 my $t = $self->{types}->{$full_path};
3440 if (!defined $t) {
3441 die "$full_path not known in r$self->{r} or we have a bug!\n";
3442 }
3443 {
3444 no warnings 'once';
3445 # SVN::Node::none and SVN::Node::file are used only once,
3446 # so we're shutting up Perl's warnings about them.
3447 if ($t == $SVN::Node::none) {
3448 return $self->add_directory($full_path, $baton,
3449 undef, -1, $self->{pool});
3450 } elsif ($t == $SVN::Node::dir) {
3451 return $self->open_directory($full_path, $baton,
3452 $self->{r}, $self->{pool});
3453 } # no warnings 'once'
3454 print STDERR "$full_path already exists in repository at ",
3455 "r$self->{r} and it is not a directory (",
3456 ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
3457 } # no warnings 'once'
3458 exit 1;
3459 }
3460
3461 sub ensure_path {
3462 my ($self, $path) = @_;
3463 my $bat = $self->{bat};
3464 my $repo_path = $self->repo_path($path);
3465 return $bat->{''} unless (length $repo_path);
3466 my @p = split m#/+#, $repo_path;
3467 my $c = shift @p;
3468 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
3469 while (@p) {
3470 my $c0 = $c;
3471 $c .= '/' . shift @p;
3472 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
3473 }
3474 return $bat->{$c};
3475 }
3476
3477 sub A {
3478 my ($self, $m) = @_;
3479 my ($dir, $file) = split_path($m->{file_b});
3480 my $pbat = $self->ensure_path($dir);
3481 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3482 undef, -1);
3483 print "\tA\t$m->{file_b}\n" unless $::_q;
3484 $self->chg_file($fbat, $m);
3485 $self->close_file($fbat,undef,$self->{pool});
3486 }
3487
3488 sub C {
3489 my ($self, $m) = @_;
3490 my ($dir, $file) = split_path($m->{file_b});
3491 my $pbat = $self->ensure_path($dir);
3492 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3493 $self->url_path($m->{file_a}), $self->{r});
3494 print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3495 $self->chg_file($fbat, $m);
3496 $self->close_file($fbat,undef,$self->{pool});
3497 }
3498
3499 sub delete_entry {
3500 my ($self, $path, $pbat) = @_;
3501 my $rpath = $self->repo_path($path);
3502 my ($dir, $file) = split_path($rpath);
3503 $self->{rm}->{$dir} = 1;
3504 $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
3505 }
3506
3507 sub R {
3508 my ($self, $m) = @_;
3509 my ($dir, $file) = split_path($m->{file_b});
3510 my $pbat = $self->ensure_path($dir);
3511 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3512 $self->url_path($m->{file_a}), $self->{r});
3513 print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3514 $self->chg_file($fbat, $m);
3515 $self->close_file($fbat,undef,$self->{pool});
3516
3517 ($dir, $file) = split_path($m->{file_a});
3518 $pbat = $self->ensure_path($dir);
3519 $self->delete_entry($m->{file_a}, $pbat);
3520 }
3521
3522 sub M {
3523 my ($self, $m) = @_;
3524 my ($dir, $file) = split_path($m->{file_b});
3525 my $pbat = $self->ensure_path($dir);
3526 my $fbat = $self->open_file($self->repo_path($m->{file_b}),
3527 $pbat,$self->{r},$self->{pool});
3528 print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
3529 $self->chg_file($fbat, $m);
3530 $self->close_file($fbat,undef,$self->{pool});
3531 }
3532
3533 sub T { shift->M(@_) }
3534
3535 sub change_file_prop {
3536 my ($self, $fbat, $pname, $pval) = @_;
3537 $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
3538 }
3539
3540 sub chg_file {
3541 my ($self, $fbat, $m) = @_;
3542 if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
3543 $self->change_file_prop($fbat,'svn:executable','*');
3544 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
3545 $self->change_file_prop($fbat,'svn:executable',undef);
3546 }
3547 my $fh = IO::File->new_tmpfile or croak $!;
3548 if ($m->{mode_b} =~ /^120/) {
3549 print $fh 'link ' or croak $!;
3550 $self->change_file_prop($fbat,'svn:special','*');
3551 } elsif ($m->{mode_a} =~ /^120/ && $m->{mode_b} !~ /^120/) {
3552 $self->change_file_prop($fbat,'svn:special',undef);
3553 }
3554 defined(my $pid = fork) or croak $!;
3555 if (!$pid) {
3556 open STDOUT, '>&', $fh or croak $!;
3557 exec qw/git-cat-file blob/, $m->{sha1_b} or croak $!;
3558 }
3559 waitpid $pid, 0;
3560 croak $? if $?;
3561 $fh->flush == 0 or croak $!;
3562 seek $fh, 0, 0 or croak $!;
3563
3564 my $exp = ::md5sum($fh);
3565 seek $fh, 0, 0 or croak $!;
3566
3567 my $pool = SVN::Pool->new;
3568 my $atd = $self->apply_textdelta($fbat, undef, $pool);
3569 my $got = SVN::TxDelta::send_stream($fh, @$atd, $pool);
3570 die "Checksum mismatch\nexpected: $exp\ngot: $got\n" if ($got ne $exp);
3571 $pool->clear;
3572
3573 close $fh or croak $!;
3574 }
3575
3576 sub D {
3577 my ($self, $m) = @_;
3578 my ($dir, $file) = split_path($m->{file_b});
3579 my $pbat = $self->ensure_path($dir);
3580 print "\tD\t$m->{file_b}\n" unless $::_q;
3581 $self->delete_entry($m->{file_b}, $pbat);
3582 }
3583
3584 sub close_edit {
3585 my ($self) = @_;
3586 my ($p,$bat) = ($self->{pool}, $self->{bat});
3587 foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
3588 next if $_ eq '';
3589 $self->close_directory($bat->{$_}, $p);
3590 }
3591 $self->close_directory($bat->{''}, $p);
3592 $self->SUPER::close_edit($p);
3593 $p->clear;
3594 }
3595
3596 sub abort_edit {
3597 my ($self) = @_;
3598 $self->SUPER::abort_edit($self->{pool});
3599 }
3600
3601 sub DESTROY {
3602 my $self = shift;
3603 $self->SUPER::DESTROY(@_);
3604 $self->{pool}->clear;
3605 }
3606
3607 # this drives the editor
3608 sub apply_diff {
3609 my ($self) = @_;
3610 my $mods = $self->{mods};
3611 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
3612 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
3613 my $f = $m->{chg};
3614 if (defined $o{$f}) {
3615 $self->$f($m);
3616 } else {
3617 fatal("Invalid change type: $f");
3618 }
3619 }
3620 $self->rmdirs if $_rmdir;
3621 if (@$mods == 0) {
3622 $self->abort_edit;
3623 } else {
3624 $self->close_edit;
3625 }
3626 return scalar @$mods;
3627 }
3628
3629 package Git::SVN::Ra;
3630 use vars qw/@ISA $config_dir $_log_window_size/;
3631 use strict;
3632 use warnings;
3633 my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
3634
3635 BEGIN {
3636 # enforce temporary pool usage for some simple functions
3637 no strict 'refs';
3638 for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root/) {
3639 my $SUPER = "SUPER::$f";
3640 *$f = sub {
3641 my $self = shift;
3642 my $pool = SVN::Pool->new;
3643 my @ret = $self->$SUPER(@_,$pool);
3644 $pool->clear;
3645 wantarray ? @ret : $ret[0];
3646 };
3647 }
3648 }
3649
3650 sub _auth_providers () {
3651 [
3652 SVN::Client::get_simple_provider(),
3653 SVN::Client::get_ssl_server_trust_file_provider(),
3654 SVN::Client::get_simple_prompt_provider(
3655 \&Git::SVN::Prompt::simple, 2),
3656 SVN::Client::get_ssl_client_cert_file_provider(),
3657 SVN::Client::get_ssl_client_cert_prompt_provider(
3658 \&Git::SVN::Prompt::ssl_client_cert, 2),
3659 SVN::Client::get_ssl_client_cert_pw_file_provider(),
3660 SVN::Client::get_ssl_client_cert_pw_prompt_provider(
3661 \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
3662 SVN::Client::get_username_provider(),
3663 SVN::Client::get_ssl_server_trust_prompt_provider(
3664 \&Git::SVN::Prompt::ssl_server_trust),
3665 SVN::Client::get_username_prompt_provider(
3666 \&Git::SVN::Prompt::username, 2)
3667 ]
3668 }
3669
3670 sub escape_uri_only {
3671 my ($uri) = @_;
3672 my @tmp;
3673 foreach (split m{/}, $uri) {
3674 s/([^\w.%-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
3675 push @tmp, $_;
3676 }
3677 join('/', @tmp);
3678 }
3679
3680 sub escape_url {
3681 my ($url) = @_;
3682 if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
3683 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
3684 $url = "$scheme://$domain$uri";
3685 }
3686 $url;
3687 }
3688
3689 sub new {
3690 my ($class, $url) = @_;
3691 $url =~ s!/+$!!;
3692 return $RA if ($RA && $RA->{url} eq $url);
3693
3694 SVN::_Core::svn_config_ensure($config_dir, undef);
3695 my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
3696 my $config = SVN::Core::config_get_config($config_dir);
3697 $RA = undef;
3698 my $dont_store_passwords = 1;
3699 my $conf_t = ${$config}{'config'};
3700 {
3701 no warnings 'once';
3702 # The usage of $SVN::_Core::SVN_CONFIG_* variables
3703 # produces warnings that variables are used only once.
3704 # I had not found the better way to shut them up, so
3705 # the warnings of type 'once' are disabled in this block.
3706 if (SVN::_Core::svn_config_get_bool($conf_t,
3707 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3708 $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
3709 1) == 0) {
3710 SVN::_Core::svn_auth_set_parameter($baton,
3711 $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
3712 bless (\$dont_store_passwords, "_p_void"));
3713 }
3714 if (SVN::_Core::svn_config_get_bool($conf_t,
3715 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3716 $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
3717 1) == 0) {
3718 $Git::SVN::Prompt::_no_auth_cache = 1;
3719 }
3720 } # no warnings 'once'
3721 my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
3722 config => $config,
3723 pool => SVN::Pool->new,
3724 auth_provider_callbacks => $callbacks);
3725 $self->{url} = $url;
3726 $self->{svn_path} = $url;
3727 $self->{repos_root} = $self->get_repos_root;
3728 $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
3729 $self->{cache} = { check_path => { r => 0, data => {} },
3730 get_dir => { r => 0, data => {} } };
3731 $RA = bless $self, $class;
3732 }
3733
3734 sub check_path {
3735 my ($self, $path, $r) = @_;
3736 my $cache = $self->{cache}->{check_path};
3737 if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
3738 return $cache->{data}->{$path};
3739 }
3740 my $pool = SVN::Pool->new;
3741 my $t = $self->SUPER::check_path($path, $r, $pool);
3742 $pool->clear;
3743 if ($r != $cache->{r}) {
3744 %{$cache->{data}} = ();
3745 $cache->{r} = $r;
3746 }
3747 $cache->{data}->{$path} = $t;
3748 }
3749
3750 sub get_dir {
3751 my ($self, $dir, $r) = @_;
3752 my $cache = $self->{cache}->{get_dir};
3753 if ($r == $cache->{r}) {
3754 if (my $x = $cache->{data}->{$dir}) {
3755 return wantarray ? @$x : $x->[0];
3756 }
3757 }
3758 my $pool = SVN::Pool->new;
3759 my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
3760 my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
3761 $pool->clear;
3762 if ($r != $cache->{r}) {
3763 %{$cache->{data}} = ();
3764 $cache->{r} = $r;
3765 }
3766 $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
3767 wantarray ? (\%dirents, $r, $props) : \%dirents;
3768 }
3769
3770 sub DESTROY {
3771 # do not call the real DESTROY since we store ourselves in $RA
3772 }
3773
3774 sub get_log {
3775 my ($self, @args) = @_;
3776 my $pool = SVN::Pool->new;
3777 splice(@args, 3, 1) if ($SVN::Core::VERSION le '1.2.0');
3778 my $ret = $self->SUPER::get_log(@args, $pool);
3779 $pool->clear;
3780 $ret;
3781 }
3782
3783 sub trees_match {
3784 my ($self, $url1, $rev1, $url2, $rev2) = @_;
3785 my $ctx = SVN::Client->new(auth => _auth_providers);
3786 my $out = IO::File->new_tmpfile;
3787
3788 # older SVN (1.1.x) doesn't take $pool as the last parameter for
3789 # $ctx->diff(), so we'll create a default one
3790 my $pool = SVN::Pool->new_default_sub;
3791
3792 $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
3793 $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
3794 $out->flush;
3795 my $ret = (($out->stat)[7] == 0);
3796 close $out or croak $!;
3797
3798 $ret;
3799 }
3800
3801 sub get_commit_editor {
3802 my ($self, $log, $cb, $pool) = @_;
3803 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
3804 $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
3805 }
3806
3807 sub gs_do_update {
3808 my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
3809 my $new = ($rev_a == $rev_b);
3810 my $path = $gs->{path};
3811
3812 if ($new && -e $gs->{index}) {
3813 unlink $gs->{index} or die
3814 "Couldn't unlink index: $gs->{index}: $!\n";
3815 }
3816 my $pool = SVN::Pool->new;
3817 $editor->set_path_strip($path);
3818 my (@pc) = split m#/#, $path;
3819 my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
3820 1, $editor, $pool);
3821 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3822
3823 # Since we can't rely on svn_ra_reparent being available, we'll
3824 # just have to do some magic with set_path to make it so
3825 # we only want a partial path.
3826 my $sp = '';
3827 my $final = join('/', @pc);
3828 while (@pc) {
3829 $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
3830 $sp .= '/' if length $sp;
3831 $sp .= shift @pc;
3832 }
3833 die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
3834
3835 $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
3836
3837 $reporter->finish_report($pool);
3838 $pool->clear;
3839 $editor->{git_commit_ok};
3840 }
3841
3842 # this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
3843 # svn_ra_reparent didn't work before 1.4)
3844 sub gs_do_switch {
3845 my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
3846 my $path = $gs->{path};
3847 my $pool = SVN::Pool->new;
3848
3849 my $full_url = $self->{url};
3850 my $old_url = $full_url;
3851 $full_url .= '/' . escape_uri_only($path) if length $path;
3852 my ($ra, $reparented);
3853 if ($old_url ne $full_url) {
3854 if ($old_url !~ m#^svn(\+ssh)?://#) {
3855 SVN::_Ra::svn_ra_reparent($self->{session}, $full_url,
3856 $pool);
3857 $self->{url} = $full_url;
3858 $reparented = 1;
3859 } else {
3860 $_[0] = undef;
3861 $self = undef;
3862 $RA = undef;
3863 $ra = Git::SVN::Ra->new($full_url);
3864 $ra_invalid = 1;
3865 }
3866 }
3867 $ra ||= $self;
3868 my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
3869 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3870 $reporter->set_path('', $rev_a, 0, @lock, $pool);
3871 $reporter->finish_report($pool);
3872
3873 if ($reparented) {
3874 SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
3875 $self->{url} = $old_url;
3876 }
3877
3878 $pool->clear;
3879 $editor->{git_commit_ok};
3880 }
3881
3882 sub longest_common_path {
3883 my ($gsv, $globs) = @_;
3884 my %common;
3885 my $common_max = scalar @$gsv;
3886
3887 foreach my $gs (@$gsv) {
3888 my @tmp = split m#/#, $gs->{path};
3889 my $p = '';
3890 foreach (@tmp) {
3891 $p .= length($p) ? "/$_" : $_;
3892 $common{$p} ||= 0;
3893 $common{$p}++;
3894 }
3895 }
3896 $globs ||= [];
3897 $common_max += scalar @$globs;
3898 foreach my $glob (@$globs) {
3899 my @tmp = split m#/#, $glob->{path}->{left};
3900 my $p = '';
3901 foreach (@tmp) {
3902 $p .= length($p) ? "/$_" : $_;
3903 $common{$p} ||= 0;
3904 $common{$p}++;
3905 }
3906 }
3907
3908 my $longest_path = '';
3909 foreach (sort {length $b <=> length $a} keys %common) {
3910 if ($common{$_} == $common_max) {
3911 $longest_path = $_;
3912 last;
3913 }
3914 }
3915 $longest_path;
3916 }
3917
3918 sub gs_fetch_loop_common {
3919 my ($self, $base, $head, $gsv, $globs) = @_;
3920 return if ($base > $head);
3921 my $inc = $_log_window_size;
3922 my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
3923 my $longest_path = longest_common_path($gsv, $globs);
3924 my $ra_url = $self->{url};
3925 while (1) {
3926 my %revs;
3927 my $err;
3928 my $err_handler = $SVN::Error::handler;
3929 $SVN::Error::handler = sub {
3930 ($err) = @_;
3931 skip_unknown_revs($err);
3932 };
3933 sub _cb {
3934 my ($paths, $r, $author, $date, $log) = @_;
3935 [ dup_changed_paths($paths),
3936 { author => $author, date => $date, log => $log } ];
3937 }
3938 $self->get_log([$longest_path], $min, $max, 0, 1, 1,
3939 sub { $revs{$_[1]} = _cb(@_) });
3940 if ($err && $max >= $head) {
3941 print STDERR "Path '$longest_path' ",
3942 "was probably deleted:\n",
3943 $err->expanded_message,
3944 "\nWill attempt to follow ",
3945 "revisions r$min .. r$max ",
3946 "committed before the deletion\n";
3947 my $hi = $max;
3948 while (--$hi >= $min) {
3949 my $ok;
3950 $self->get_log([$longest_path], $min, $hi,
3951 0, 1, 1, sub {
3952 $ok ||= $_[1];
3953 $revs{$_[1]} = _cb(@_) });
3954 if ($ok) {
3955 print STDERR "r$min .. r$ok OK\n";
3956 last;
3957 }
3958 }
3959 }
3960 $SVN::Error::handler = $err_handler;
3961
3962 my %exists = map { $_->{path} => $_ } @$gsv;
3963 foreach my $r (sort {$a <=> $b} keys %revs) {
3964 my ($paths, $logged) = @{$revs{$r}};
3965
3966 foreach my $gs ($self->match_globs(\%exists, $paths,
3967 $globs, $r)) {
3968 if ($gs->rev_map_max >= $r) {
3969 next;
3970 }
3971 next unless $gs->match_paths($paths, $r);
3972 $gs->{logged_rev_props} = $logged;
3973 if (my $last_commit = $gs->last_commit) {
3974 $gs->assert_index_clean($last_commit);
3975 }
3976 my $log_entry = $gs->do_fetch($paths, $r);
3977 if ($log_entry) {
3978 $gs->do_git_commit($log_entry);
3979 }
3980 $INDEX_FILES{$gs->{index}} = 1;
3981 }
3982 foreach my $g (@$globs) {
3983 my $k = "svn-remote.$g->{remote}." .
3984 "$g->{t}-maxRev";
3985 Git::SVN::tmp_config($k, $r);
3986 }
3987 if ($ra_invalid) {
3988 $_[0] = undef;
3989 $self = undef;
3990 $RA = undef;
3991 $self = Git::SVN::Ra->new($ra_url);
3992 $ra_invalid = undef;
3993 }
3994 }
3995 # pre-fill the .rev_db since it'll eventually get filled in
3996 # with '0' x40 if something new gets committed
3997 foreach my $gs (@$gsv) {
3998 next if $gs->rev_map_max >= $max;
3999 next if defined $gs->rev_map_get($max);
4000 $gs->rev_map_set($max, 0 x40);
4001 }
4002 foreach my $g (@$globs) {
4003 my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
4004 Git::SVN::tmp_config($k, $max);
4005 }
4006 last if $max >= $head;
4007 $min = $max + 1;
4008 $max += $inc;
4009 $max = $head if ($max > $head);
4010 }
4011 Git::SVN::gc();
4012 }
4013
4014 sub match_globs {
4015 my ($self, $exists, $paths, $globs, $r) = @_;
4016
4017 sub get_dir_check {
4018 my ($self, $exists, $g, $r) = @_;
4019 my @x = eval { $self->get_dir($g->{path}->{left}, $r) };
4020 return unless scalar @x == 3;
4021 my $dirents = $x[0];
4022 foreach my $de (keys %$dirents) {
4023 next if $dirents->{$de}->{kind} != $SVN::Node::dir;
4024 my $p = $g->{path}->full_path($de);
4025 next if $exists->{$p};
4026 next if (length $g->{path}->{right} &&
4027 ($self->check_path($p, $r) !=
4028 $SVN::Node::dir));
4029 $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
4030 $g->{ref}->full_path($de), 1);
4031 }
4032 }
4033 foreach my $g (@$globs) {
4034 if (my $path = $paths->{"/$g->{path}->{left}"}) {
4035 if ($path->{action} =~ /^[AR]$/) {
4036 get_dir_check($self, $exists, $g, $r);
4037 }
4038 }
4039 foreach (keys %$paths) {
4040 if (/$g->{path}->{left_regex}/ &&
4041 !/$g->{path}->{regex}/) {
4042 next if $paths->{$_}->{action} !~ /^[AR]$/;
4043 get_dir_check($self, $exists, $g, $r);
4044 }
4045 next unless /$g->{path}->{regex}/;
4046 my $p = $1;
4047 my $pathname = $g->{path}->full_path($p);
4048 next if $exists->{$pathname};
4049 next if ($self->check_path($pathname, $r) !=
4050 $SVN::Node::dir);
4051 $exists->{$pathname} = Git::SVN->init(
4052 $self->{url}, $pathname, undef,
4053 $g->{ref}->full_path($p), 1);
4054 }
4055 my $c = '';
4056 foreach (split m#/#, $g->{path}->{left}) {
4057 $c .= "/$_";
4058 next unless ($paths->{$c} &&
4059 ($paths->{$c}->{action} =~ /^[AR]$/));
4060 get_dir_check($self, $exists, $g, $r);
4061 }
4062 }
4063 values %$exists;
4064 }
4065
4066 sub minimize_url {
4067 my ($self) = @_;
4068 return $self->{url} if ($self->{url} eq $self->{repos_root});
4069 my $url = $self->{repos_root};
4070 my @components = split(m!/!, $self->{svn_path});
4071 my $c = '';
4072 do {
4073 $url .= "/$c" if length $c;
4074 eval { (ref $self)->new($url)->get_latest_revnum };
4075 } while ($@ && ($c = shift @components));
4076 $url;
4077 }
4078
4079 sub can_do_switch {
4080 my $self = shift;
4081 unless (defined $can_do_switch) {
4082 my $pool = SVN::Pool->new;
4083 my $rep = eval {
4084 $self->do_switch(1, '', 0, $self->{url},
4085 SVN::Delta::Editor->new, $pool);
4086 };
4087 if ($@) {
4088 $can_do_switch = 0;
4089 } else {
4090 $rep->abort_report($pool);
4091 $can_do_switch = 1;
4092 }
4093 $pool->clear;
4094 }
4095 $can_do_switch;
4096 }
4097
4098 sub skip_unknown_revs {
4099 my ($err) = @_;
4100 my $errno = $err->apr_err();
4101 # Maybe the branch we're tracking didn't
4102 # exist when the repo started, so it's
4103 # not an error if it doesn't, just continue
4104 #
4105 # Wonderfully consistent library, eh?
4106 # 160013 - svn:// and file://
4107 # 175002 - http(s)://
4108 # 175007 - http(s):// (this repo required authorization, too...)
4109 # More codes may be discovered later...
4110 if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
4111 my $err_key = $err->expanded_message;
4112 # revision numbers change every time, filter them out
4113 $err_key =~ s/\d+/\0/g;
4114 $err_key = "$errno\0$err_key";
4115 unless ($ignored_err{$err_key}) {
4116 warn "W: Ignoring error from SVN, path probably ",
4117 "does not exist: ($errno): ",
4118 $err->expanded_message,"\n";
4119 warn "W: Do not be alarmed at the above message ",
4120 "git-svn is just searching aggressively for ",
4121 "old history.\n",
4122 "This may take a while on large repositories\n";
4123 $ignored_err{$err_key} = 1;
4124 }
4125 return;
4126 }
4127 die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
4128 }
4129
4130 # svn_log_changed_path_t objects passed to get_log are likely to be
4131 # overwritten even if only the refs are copied to an external variable,
4132 # so we should dup the structures in their entirety. Using an externally
4133 # passed pool (instead of our temporary and quickly cleared pool in
4134 # Git::SVN::Ra) does not help matters at all...
4135 sub dup_changed_paths {
4136 my ($paths) = @_;
4137 return undef unless $paths;
4138 my %ret;
4139 foreach my $p (keys %$paths) {
4140 my $i = $paths->{$p};
4141 my %s = map { $_ => $i->$_ }
4142 qw/copyfrom_path copyfrom_rev action/;
4143 $ret{$p} = \%s;
4144 }
4145 \%ret;
4146 }
4147
4148 package Git::SVN::Log;
4149 use strict;
4150 use warnings;
4151 use POSIX qw/strftime/;
4152 use constant commit_log_separator => ('-' x 72) . "\n";
4153 use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
4154 %rusers $show_commit $incremental/;
4155 my $l_fmt;
4156
4157 sub cmt_showable {
4158 my ($c) = @_;
4159 return 1 if defined $c->{r};
4160
4161 # big commit message got truncated by the 16k pretty buffer in rev-list
4162 if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
4163 $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
4164 @{$c->{l}} = ();
4165 my @log = command(qw/cat-file commit/, $c->{c});
4166
4167 # shift off the headers
4168 shift @log while ($log[0] ne '');
4169 shift @log;
4170
4171 # TODO: make $c->{l} not have a trailing newline in the future
4172 @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
4173
4174 (undef, $c->{r}, undef) = ::extract_metadata(
4175 (grep(/^git-svn-id: /, @log))[-1]);
4176 }
4177 return defined $c->{r};
4178 }
4179
4180 sub log_use_color {
4181 return $color || Git->repository->get_colorbool('color.diff');
4182 }
4183
4184 sub git_svn_log_cmd {
4185 my ($r_min, $r_max, @args) = @_;
4186 my $head = 'HEAD';
4187 my (@files, @log_opts);
4188 foreach my $x (@args) {
4189 if ($x eq '--' || @files) {
4190 push @files, $x;
4191 } else {
4192 if (::verify_ref("$x^0")) {
4193 $head = $x;
4194 } else {
4195 push @log_opts, $x;
4196 }
4197 }
4198 }
4199
4200 my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
4201 $gs ||= Git::SVN->_new;
4202 my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
4203 $gs->refname);
4204 push @cmd, '-r' unless $non_recursive;
4205 push @cmd, qw/--raw --name-status/ if $verbose;
4206 push @cmd, '--color' if log_use_color();
4207 push @cmd, @log_opts;
4208 if (defined $r_max && $r_max == $r_min) {
4209 push @cmd, '--max-count=1';
4210 if (my $c = $gs->rev_map_get($r_max)) {
4211 push @cmd, $c;
4212 }
4213 } elsif (defined $r_max) {
4214 if ($r_max < $r_min) {
4215 ($r_min, $r_max) = ($r_max, $r_min);
4216 }
4217 my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
4218 my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
4219 # If there are no commits in the range, both $c_max and $c_min
4220 # will be undefined. If there is at least 1 commit in the
4221 # range, both will be defined.
4222 return () if !defined $c_min || !defined $c_max;
4223 if ($c_min eq $c_max) {
4224 push @cmd, '--max-count=1', $c_min;
4225 } else {
4226 push @cmd, '--boundary', "$c_min..$c_max";
4227 }
4228 }
4229 return (@cmd, @files);
4230 }
4231
4232 # adapted from pager.c
4233 sub config_pager {
4234 $pager ||= $ENV{GIT_PAGER} || $ENV{PAGER};
4235 if (!defined $pager) {
4236 $pager = 'less';
4237 } elsif (length $pager == 0 || $pager eq 'cat') {
4238 $pager = undef;
4239 }
4240 $ENV{GIT_PAGER_IN_USE} = defined($pager);
4241 }
4242
4243 sub run_pager {
4244 return unless -t *STDOUT && defined $pager;
4245 pipe my $rfd, my $wfd or return;
4246 defined(my $pid = fork) or ::fatal "Can't fork: $!";
4247 if (!$pid) {
4248 open STDOUT, '>&', $wfd or
4249 ::fatal "Can't redirect to stdout: $!";
4250 return;
4251 }
4252 open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
4253 $ENV{LESS} ||= 'FRSX';
4254 exec $pager or ::fatal "Can't run pager: $! ($pager)";
4255 }
4256
4257 sub format_svn_date {
4258 return strftime("%Y-%m-%d %H:%M:%S %z (%a, %d %b %Y)", localtime(shift));
4259 }
4260
4261 sub parse_git_date {
4262 my ($t, $tz) = @_;
4263 # Date::Parse isn't in the standard Perl distro :(
4264 if ($tz =~ s/^\+//) {
4265 $t += tz_to_s_offset($tz);
4266 } elsif ($tz =~ s/^\-//) {
4267 $t -= tz_to_s_offset($tz);
4268 }
4269 return $t;
4270 }
4271
4272 sub set_local_timezone {
4273 if (defined $TZ) {
4274 $ENV{TZ} = $TZ;
4275 } else {
4276 delete $ENV{TZ};
4277 }
4278 }
4279
4280 sub tz_to_s_offset {
4281 my ($tz) = @_;
4282 $tz =~ s/(\d\d)$//;
4283 return ($1 * 60) + ($tz * 3600);
4284 }
4285
4286 sub get_author_info {
4287 my ($dest, $author, $t, $tz) = @_;
4288 $author =~ s/(?:^\s*|\s*$)//g;
4289 $dest->{a_raw} = $author;
4290 my $au;
4291 if ($::_authors) {
4292 $au = $rusers{$author} || undef;
4293 }
4294 if (!$au) {
4295 ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
4296 }
4297 $dest->{t} = $t;
4298 $dest->{tz} = $tz;
4299 $dest->{a} = $au;
4300 $dest->{t_utc} = parse_git_date($t, $tz);
4301 }
4302
4303 sub process_commit {
4304 my ($c, $r_min, $r_max, $defer) = @_;
4305 if (defined $r_min && defined $r_max) {
4306 if ($r_min == $c->{r} && $r_min == $r_max) {
4307 show_commit($c);
4308 return 0;
4309 }
4310 return 1 if $r_min == $r_max;
4311 if ($r_min < $r_max) {
4312 # we need to reverse the print order
4313 return 0 if (defined $limit && --$limit < 0);
4314 push @$defer, $c;
4315 return 1;
4316 }
4317 if ($r_min != $r_max) {
4318 return 1 if ($r_min < $c->{r});
4319 return 1 if ($r_max > $c->{r});
4320 }
4321 }
4322 return 0 if (defined $limit && --$limit < 0);
4323 show_commit($c);
4324 return 1;
4325 }
4326
4327 sub show_commit {
4328 my $c = shift;
4329 if ($oneline) {
4330 my $x = "\n";
4331 if (my $l = $c->{l}) {
4332 while ($l->[0] =~ /^\s*$/) { shift @$l }
4333 $x = $l->[0];
4334 }
4335 $l_fmt ||= 'A' . length($c->{r});
4336 print 'r',pack($l_fmt, $c->{r}),' | ';
4337 print "$c->{c} | " if $show_commit;
4338 print $x;
4339 } else {
4340 show_commit_normal($c);
4341 }
4342 }
4343
4344 sub show_commit_changed_paths {
4345 my ($c) = @_;
4346 return unless $c->{changed};
4347 print "Changed paths:\n", @{$c->{changed}};
4348 }
4349
4350 sub show_commit_normal {
4351 my ($c) = @_;
4352 print commit_log_separator, "r$c->{r} | ";
4353 print "$c->{c} | " if $show_commit;
4354 print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
4355 my $nr_line = 0;
4356
4357 if (my $l = $c->{l}) {
4358 while ($l->[$#$l] eq "\n" && $#$l > 0
4359 && $l->[($#$l - 1)] eq "\n") {
4360 pop @$l;
4361 }
4362 $nr_line = scalar @$l;
4363 if (!$nr_line) {
4364 print "1 line\n\n\n";
4365 } else {
4366 if ($nr_line == 1) {
4367 $nr_line = '1 line';
4368 } else {
4369 $nr_line .= ' lines';
4370 }
4371 print $nr_line, "\n";
4372 show_commit_changed_paths($c);
4373 print "\n";
4374 print $_ foreach @$l;
4375 }
4376 } else {
4377 print "1 line\n";
4378 show_commit_changed_paths($c);
4379 print "\n";
4380
4381 }
4382 foreach my $x (qw/raw stat diff/) {
4383 if ($c->{$x}) {
4384 print "\n";
4385 print $_ foreach @{$c->{$x}}
4386 }
4387 }
4388 }
4389
4390 sub cmd_show_log {
4391 my (@args) = @_;
4392 my ($r_min, $r_max);
4393 my $r_last = -1; # prevent dupes
4394 set_local_timezone();
4395 if (defined $::_revision) {
4396 if ($::_revision =~ /^(\d+):(\d+)$/) {
4397 ($r_min, $r_max) = ($1, $2);
4398 } elsif ($::_revision =~ /^\d+$/) {
4399 $r_min = $r_max = $::_revision;
4400 } else {
4401 ::fatal "-r$::_revision is not supported, use ",
4402 "standard 'git log' arguments instead";
4403 }
4404 }
4405
4406 config_pager();
4407 @args = git_svn_log_cmd($r_min, $r_max, @args);
4408 if (!@args) {
4409 print commit_log_separator unless $incremental || $oneline;
4410 return;
4411 }
4412 my $log = command_output_pipe(@args);
4413 run_pager();
4414 my (@k, $c, $d, $stat);
4415 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
4416 while (<$log>) {
4417 if (/^${esc_color}commit -?($::sha1_short)/o) {
4418 my $cmt = $1;
4419 if ($c && cmt_showable($c) && $c->{r} != $r_last) {
4420 $r_last = $c->{r};
4421 process_commit($c, $r_min, $r_max, \@k) or
4422 goto out;
4423 }
4424 $d = undef;
4425 $c = { c => $cmt };
4426 } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
4427 get_author_info($c, $1, $2, $3);
4428 } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
4429 # ignore
4430 } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
4431 push @{$c->{raw}}, $_;
4432 } elsif (/^${esc_color}[ACRMDT]\t/) {
4433 # we could add $SVN->{svn_path} here, but that requires
4434 # remote access at the moment (repo_path_split)...
4435 s#^(${esc_color})([ACRMDT])\t#$1 $2 #o;
4436 push @{$c->{changed}}, $_;
4437 } elsif (/^${esc_color}diff /o) {
4438 $d = 1;
4439 push @{$c->{diff}}, $_;
4440 } elsif ($d) {
4441 push @{$c->{diff}}, $_;
4442 } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
4443 $esc_color*[\+\-]*$esc_color$/x) {
4444 $stat = 1;
4445 push @{$c->{stat}}, $_;
4446 } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
4447 push @{$c->{stat}}, $_;
4448 $stat = undef;
4449 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
4450 ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
4451 } elsif (s/^${esc_color} //o) {
4452 push @{$c->{l}}, $_;
4453 }
4454 }
4455 if ($c && defined $c->{r} && $c->{r} != $r_last) {
4456 $r_last = $c->{r};
4457 process_commit($c, $r_min, $r_max, \@k);
4458 }
4459 if (@k) {
4460 ($r_min, $r_max) = ($r_max, $r_min);
4461 process_commit($_, $r_min, $r_max) foreach reverse @k;
4462 }
4463 out:
4464 close $log;
4465 print commit_log_separator unless $incremental || $oneline;
4466 }
4467
4468 sub cmd_blame {
4469 my $path = shift;
4470
4471 config_pager();
4472 run_pager();
4473
4474 my ($fh, $ctx) = command_output_pipe('blame', @_, $path);
4475 while (my $line = <$fh>) {
4476 if ($line =~ /^\^?([[:xdigit:]]+)\s/) {
4477 my (undef, $rev, undef) = ::cmt_metadata($1);
4478 $rev = sprintf('%-10s', $rev);
4479 $line =~ s/^\^?[[:xdigit:]]+(\s)/$rev$1/;
4480 }
4481 print $line;
4482 }
4483 command_close_pipe($fh, $ctx);
4484 }
4485
4486 package Git::SVN::Migration;
4487 # these version numbers do NOT correspond to actual version numbers
4488 # of git nor git-svn. They are just relative.
4489 #
4490 # v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
4491 #
4492 # v1 layout: .git/$id/info/url, refs/remotes/$id
4493 #
4494 # v2 layout: .git/svn/$id/info/url, refs/remotes/$id
4495 #
4496 # v3 layout: .git/svn/$id, refs/remotes/$id
4497 # - info/url may remain for backwards compatibility
4498 # - this is what we migrate up to this layout automatically,
4499 # - this will be used by git svn init on single branches
4500 # v3.1 layout (auto migrated):
4501 # - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
4502 # for backwards compatibility
4503 #
4504 # v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
4505 # - this is only created for newly multi-init-ed
4506 # repositories. Similar in spirit to the
4507 # --use-separate-remotes option in git-clone (now default)
4508 # - we do not automatically migrate to this (following
4509 # the example set by core git)
4510 #
4511 # v5 layout: .rev_db.$UUID => .rev_map.$UUID
4512 # - newer, more-efficient format that uses 24-bytes per record
4513 # with no filler space.
4514 # - use xxd -c24 < .rev_map.$UUID to view and debug
4515 # - This is a one-way migration, repositories updated to the
4516 # new format will not be able to use old git-svn without
4517 # rebuilding the .rev_db. Rebuilding the rev_db is not
4518 # possible if noMetadata or useSvmProps are set; but should
4519 # be no problem for users that use the (sensible) defaults.
4520 use strict;
4521 use warnings;
4522 use Carp qw/croak/;
4523 use File::Path qw/mkpath/;
4524 use File::Basename qw/dirname basename/;
4525 use vars qw/$_minimize/;
4526
4527 sub migrate_from_v0 {
4528 my $git_dir = $ENV{GIT_DIR};
4529 return undef unless -d $git_dir;
4530 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4531 my $migrated = 0;
4532 while (<$fh>) {
4533 chomp;
4534 my ($id, $orig_ref) = ($_, $_);
4535 next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
4536 next unless -f "$git_dir/$id/info/url";
4537 my $new_ref = "refs/remotes/$id";
4538 if (::verify_ref("$new_ref^0")) {
4539 print STDERR "W: $orig_ref is probably an old ",
4540 "branch used by an ancient version of ",
4541 "git-svn.\n",
4542 "However, $new_ref also exists.\n",
4543 "We will not be able ",
4544 "to use this branch until this ",
4545 "ambiguity is resolved.\n";
4546 next;
4547 }
4548 print STDERR "Migrating from v0 layout...\n" if !$migrated;
4549 print STDERR "Renaming ref: $orig_ref => $new_ref\n";
4550 command_noisy('update-ref', $new_ref, $orig_ref);
4551 command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
4552 $migrated++;
4553 }
4554 command_close_pipe($fh, $ctx);
4555 print STDERR "Done migrating from v0 layout...\n" if $migrated;
4556 $migrated;
4557 }
4558
4559 sub migrate_from_v1 {
4560 my $git_dir = $ENV{GIT_DIR};
4561 my $migrated = 0;
4562 return $migrated unless -d $git_dir;
4563 my $svn_dir = "$git_dir/svn";
4564
4565 # just in case somebody used 'svn' as their $id at some point...
4566 return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
4567
4568 print STDERR "Migrating from a git-svn v1 layout...\n";
4569 mkpath([$svn_dir]);
4570 print STDERR "Data from a previous version of git-svn exists, but\n\t",
4571 "$svn_dir\n\t(required for this version ",
4572 "($::VERSION) of git-svn) does not. exist\n";
4573 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4574 while (<$fh>) {
4575 my $x = $_;
4576 next unless $x =~ s#^refs/remotes/##;
4577 chomp $x;
4578 next unless -f "$git_dir/$x/info/url";
4579 my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
4580 next unless $u;
4581 my $dn = dirname("$git_dir/svn/$x");
4582 mkpath([$dn]) unless -d $dn;
4583 if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
4584 mkpath(["$git_dir/svn/svn"]);
4585 print STDERR " - $git_dir/$x/info => ",
4586 "$git_dir/svn/$x/info\n";
4587 rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
4588 croak "$!: $x";
4589 # don't worry too much about these, they probably
4590 # don't exist with repos this old (save for index,
4591 # and we can easily regenerate that)
4592 foreach my $f (qw/unhandled.log index .rev_db/) {
4593 rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
4594 }
4595 } else {
4596 print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
4597 rename "$git_dir/$x", "$git_dir/svn/$x" or
4598 croak "$!: $x";
4599 }
4600 $migrated++;
4601 }
4602 command_close_pipe($fh, $ctx);
4603 print STDERR "Done migrating from a git-svn v1 layout\n";
4604 $migrated;
4605 }
4606
4607 sub read_old_urls {
4608 my ($l_map, $pfx, $path) = @_;
4609 my @dir;
4610 foreach (<$path/*>) {
4611 if (-r "$_/info/url") {
4612 $pfx .= '/' if $pfx && $pfx !~ m!/$!;
4613 my $ref_id = $pfx . basename $_;
4614 my $url = ::file_to_s("$_/info/url");
4615 $l_map->{$ref_id} = $url;
4616 } elsif (-d $_) {
4617 push @dir, $_;
4618 }
4619 }
4620 foreach (@dir) {
4621 my $x = $_;
4622 $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
4623 read_old_urls($l_map, $x, $_);
4624 }
4625 }
4626
4627 sub migrate_from_v2 {
4628 my @cfg = command(qw/config -l/);
4629 return if grep /^svn-remote\..+\.url=/, @cfg;
4630 my %l_map;
4631 read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
4632 my $migrated = 0;
4633
4634 foreach my $ref_id (sort keys %l_map) {
4635 eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
4636 if ($@) {
4637 Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
4638 }
4639 $migrated++;
4640 }
4641 $migrated;
4642 }
4643
4644 sub minimize_connections {
4645 my $r = Git::SVN::read_all_remotes();
4646 my $new_urls = {};
4647 my $root_repos = {};
4648 foreach my $repo_id (keys %$r) {
4649 my $url = $r->{$repo_id}->{url} or next;
4650 my $fetch = $r->{$repo_id}->{fetch} or next;
4651 my $ra = Git::SVN::Ra->new($url);
4652
4653 # skip existing cases where we already connect to the root
4654 if (($ra->{url} eq $ra->{repos_root}) ||
4655 (Git::SVN::sanitize_remote_name($ra->{repos_root}) eq
4656 $repo_id)) {
4657 $root_repos->{$ra->{url}} = $repo_id;
4658 next;
4659 }
4660
4661 my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
4662 my $root_path = $ra->{url};
4663 $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
4664 foreach my $path (keys %$fetch) {
4665 my $ref_id = $fetch->{$path};
4666 my $gs = Git::SVN->new($ref_id, $repo_id, $path);
4667
4668 # make sure we can read when connecting to
4669 # a higher level of a repository
4670 my ($last_rev, undef) = $gs->last_rev_commit;
4671 if (!defined $last_rev) {
4672 $last_rev = eval {
4673 $root_ra->get_latest_revnum;
4674 };
4675 next if $@;
4676 }
4677 my $new = $root_path;
4678 $new .= length $path ? "/$path" : '';
4679 eval {
4680 $root_ra->get_log([$new], $last_rev, $last_rev,
4681 0, 0, 1, sub { });
4682 };
4683 next if $@;
4684 $new_urls->{$ra->{repos_root}}->{$new} =
4685 { ref_id => $ref_id,
4686 old_repo_id => $repo_id,
4687 old_path => $path };
4688 }
4689 }
4690
4691 my @emptied;
4692 foreach my $url (keys %$new_urls) {
4693 # see if we can re-use an existing [svn-remote "repo_id"]
4694 # instead of creating a(n ugly) new section:
4695 my $repo_id = $root_repos->{$url} ||
4696 Git::SVN::sanitize_remote_name($url);
4697
4698 my $fetch = $new_urls->{$url};
4699 foreach my $path (keys %$fetch) {
4700 my $x = $fetch->{$path};
4701 Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
4702 my $pfx = "svn-remote.$x->{old_repo_id}";
4703
4704 my $old_fetch = quotemeta("$x->{old_path}:".
4705 "refs/remotes/$x->{ref_id}");
4706 command_noisy(qw/config --unset/,
4707 "$pfx.fetch", '^'. $old_fetch . '$');
4708 delete $r->{$x->{old_repo_id}}->
4709 {fetch}->{$x->{old_path}};
4710 if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
4711 command_noisy(qw/config --unset/,
4712 "$pfx.url");
4713 push @emptied, $x->{old_repo_id}
4714 }
4715 }
4716 }
4717 if (@emptied) {
4718 my $file = $ENV{GIT_CONFIG} || $ENV{GIT_CONFIG_LOCAL} ||
4719 "$ENV{GIT_DIR}/config";
4720 print STDERR <<EOF;
4721 The following [svn-remote] sections in your config file ($file) are empty
4722 and can be safely removed:
4723 EOF
4724 print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
4725 }
4726 }
4727
4728 sub migration_check {
4729 migrate_from_v0();
4730 migrate_from_v1();
4731 migrate_from_v2();
4732 minimize_connections() if $_minimize;
4733 }
4734
4735 package Git::IndexInfo;
4736 use strict;
4737 use warnings;
4738 use Git qw/command_input_pipe command_close_pipe/;
4739
4740 sub new {
4741 my ($class) = @_;
4742 my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
4743 bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
4744 }
4745
4746 sub remove {
4747 my ($self, $path) = @_;
4748 if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
4749 return ++$self->{nr};
4750 }
4751 undef;
4752 }
4753
4754 sub update {
4755 my ($self, $mode, $hash, $path) = @_;
4756 if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
4757 return ++$self->{nr};
4758 }
4759 undef;
4760 }
4761
4762 sub DESTROY {
4763 my ($self) = @_;
4764 command_close_pipe($self->{gui}, $self->{ctx});
4765 }
4766
4767 package Git::SVN::GlobSpec;
4768 use strict;
4769 use warnings;
4770
4771 sub new {
4772 my ($class, $glob) = @_;
4773 my $re = $glob;
4774 $re =~ s!/+$!!g; # no need for trailing slashes
4775 my $nr = ($re =~ s!^(.*)\*(.*)$!\(\[^/\]+\)!g);
4776 my ($left, $right) = ($1, $2);
4777 if ($nr > 1) {
4778 die "Only one '*' wildcard expansion ",
4779 "is supported (got $nr): '$glob'\n";
4780 } elsif ($nr == 0) {
4781 die "One '*' is needed for glob: '$glob'\n";
4782 }
4783 $re = quotemeta($left) . $re . quotemeta($right);
4784 if (length $left && !($left =~ s!/+$!!g)) {
4785 die "Missing trailing '/' on left side of: '$glob' ($left)\n";
4786 }
4787 if (length $right && !($right =~ s!^/+!!g)) {
4788 die "Missing leading '/' on right side of: '$glob' ($right)\n";
4789 }
4790 my $left_re = qr/^\/\Q$left\E(\/|$)/;
4791 bless { left => $left, right => $right, left_regex => $left_re,
4792 regex => qr/$re/, glob => $glob }, $class;
4793 }
4794
4795 sub full_path {
4796 my ($self, $path) = @_;
4797 return (length $self->{left} ? "$self->{left}/" : '') .
4798 $path . (length $self->{right} ? "/$self->{right}" : '');
4799 }
4800
4801 __END__
4802
4803 Data structures:
4804
4805
4806 $remotes = { # returned by read_all_remotes()
4807 'svn' => {
4808 # svn-remote.svn.url=https://svn.musicpd.org
4809 url => 'https://svn.musicpd.org',
4810 # svn-remote.svn.fetch=mpd/trunk:trunk
4811 fetch => {
4812 'mpd/trunk' => 'trunk',
4813 },
4814 # svn-remote.svn.tags=mpd/tags/*:tags/*
4815 tags => {
4816 path => {
4817 left => 'mpd/tags',
4818 right => '',
4819 regex => qr!mpd/tags/([^/]+)$!,
4820 glob => 'tags/*',
4821 },
4822 ref => {
4823 left => 'tags',
4824 right => '',
4825 regex => qr!tags/([^/]+)$!,
4826 glob => 'tags/*',
4827 },
4828 }
4829 }
4830 };
4831
4832 $log_entry hashref as returned by libsvn_log_entry()
4833 {
4834 log => 'whitespace-formatted log entry
4835 ', # trailing newline is preserved
4836 revision => '8', # integer
4837 date => '2004-02-24T17:01:44.108345Z', # commit date
4838 author => 'committer name'
4839 };
4840
4841
4842 # this is generated by generate_diff();
4843 @mods = array of diff-index line hashes, each element represents one line
4844 of diff-index output
4845
4846 diff-index line ($m hash)
4847 {
4848 mode_a => first column of diff-index output, no leading ':',
4849 mode_b => second column of diff-index output,
4850 sha1_b => sha1sum of the final blob,
4851 chg => change type [MCRADT],
4852 file_a => original file name of a file (iff chg is 'C' or 'R')
4853 file_b => new/current file name of a file (any chg)
4854 }
4855 ;
4856
4857 # retval of read_url_paths{,_all}();
4858 $l_map = {
4859 # repository root url
4860 'https://svn.musicpd.org' => {
4861 # repository path # GIT_SVN_ID
4862 'mpd/trunk' => 'trunk',
4863 'mpd/tags/0.11.5' => 'tags/0.11.5',
4864 },
4865 }
4866
4867 Notes:
4868 I don't trust the each() function on unless I created %hash myself
4869 because the internal iterator may not have started at base.
Jonathan Nieder’s miscellaneous git branches