?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Merge branch 'ew/svn-rev-db'
[git.git] / git-svn.perl
blob54d784469af7b787190e80d277abf2a4782c7d3f
1 #!/usr/bin/env perl
2 # Copyright (C) 2006, Eric Wong <normalperson@yhbt.net>
3 # License: GPL v2 or later
4 use warnings;
5 use strict;
6 use vars qw/ $AUTHOR $VERSION
7 $sha1 $sha1_short $_revision
8 $_q $_authors %users/;
9 $AUTHOR = 'Eric Wong <normalperson@yhbt.net>';
10 $VERSION = '@@GIT_VERSION@@';
11
12 # From which subdir have we been invoked?
13 my $cmd_dir_prefix = eval {
14 command_oneline([qw/rev-parse --show-prefix/], STDERR => 0)
15 } || '';
16
17 my $git_dir_user_set = 1 if defined $ENV{GIT_DIR};
18 $ENV{GIT_DIR} ||= '.git';
19 $Git::SVN::default_repo_id = 'svn';
20 $Git::SVN::default_ref_id = $ENV{GIT_SVN_ID} || 'git-svn';
21 $Git::SVN::Ra::_log_window_size = 100;
22
23 $Git::SVN::Log::TZ = $ENV{TZ};
24 $ENV{TZ} = 'UTC';
25 $| = 1; # unbuffer STDOUT
26
27 sub fatal (@) { print STDERR "@_\n"; exit 1 }
28 require SVN::Core; # use()-ing this causes segfaults for me... *shrug*
29 require SVN::Ra;
30 require SVN::Delta;
31 if ($SVN::Core::VERSION lt '1.1.0') {
32 fatal "Need SVN::Core 1.1.0 or better (got $SVN::Core::VERSION)";
33 }
34 push @Git::SVN::Ra::ISA, 'SVN::Ra';
35 push @SVN::Git::Editor::ISA, 'SVN::Delta::Editor';
36 push @SVN::Git::Fetcher::ISA, 'SVN::Delta::Editor';
37 use Carp qw/croak/;
38 use Digest::MD5;
39 use IO::File qw//;
40 use File::Basename qw/dirname basename/;
41 use File::Path qw/mkpath/;
42 use Getopt::Long qw/:config gnu_getopt no_ignore_case auto_abbrev/;
43 use IPC::Open3;
44 use Git;
45
46 BEGIN {
47 # import functions from Git into our packages, en masse
48 no strict 'refs';
49 foreach (qw/command command_oneline command_noisy command_output_pipe
50 command_input_pipe command_close_pipe/) {
51 for my $package ( qw(SVN::Git::Editor SVN::Git::Fetcher
52 Git::SVN::Migration Git::SVN::Log Git::SVN),
53 __PACKAGE__) {
54 *{"${package}::$_"} = \&{"Git::$_"};
55 }
56 }
57 }
58
59 my ($SVN);
60
61 $sha1 = qr/[a-f\d]{40}/;
62 $sha1_short = qr/[a-f\d]{4,40}/;
63 my ($_stdin, $_help, $_edit,
64 $_message, $_file,
65 $_template, $_shared,
66 $_version, $_fetch_all, $_no_rebase,
67 $_merge, $_strategy, $_dry_run, $_local,
68 $_prefix, $_no_checkout, $_url, $_verbose);
69 $Git::SVN::_follow_parent = 1;
70 my %remote_opts = ( 'username=s' => \$Git::SVN::Prompt::_username,
71 'config-dir=s' => \$Git::SVN::Ra::config_dir,
72 'no-auth-cache' => \$Git::SVN::Prompt::_no_auth_cache );
73 my %fc_opts = ( 'follow-parent|follow!' => \$Git::SVN::_follow_parent,
74 'authors-file|A=s' => \$_authors,
75 'repack:i' => \$Git::SVN::_repack,
76 'noMetadata' => \$Git::SVN::_no_metadata,
77 'useSvmProps' => \$Git::SVN::_use_svm_props,
78 'useSvnsyncProps' => \$Git::SVN::_use_svnsync_props,
79 'log-window-size=i' => \$Git::SVN::Ra::_log_window_size,
80 'no-checkout' => \$_no_checkout,
81 'quiet|q' => \$_q,
82 'repack-flags|repack-args|repack-opts=s' =>
83 \$Git::SVN::_repack_flags,
84 'use-log-author' => \$Git::SVN::_use_log_author,
85 %remote_opts );
86
87 my ($_trunk, $_tags, $_branches, $_stdlayout);
88 my %icv;
89 my %init_opts = ( 'template=s' => \$_template, 'shared:s' => \$_shared,
90 'trunk|T=s' => \$_trunk, 'tags|t=s' => \$_tags,
91 'branches|b=s' => \$_branches, 'prefix=s' => \$_prefix,
92 'stdlayout|s' => \$_stdlayout,
93 'minimize-url|m' => \$Git::SVN::_minimize_url,
94 'no-metadata' => sub { $icv{noMetadata} = 1 },
95 'use-svm-props' => sub { $icv{useSvmProps} = 1 },
96 'use-svnsync-props' => sub { $icv{useSvnsyncProps} = 1 },
97 'rewrite-root=s' => sub { $icv{rewriteRoot} = $_[1] },
98 %remote_opts );
99 my %cmt_opts = ( 'edit|e' => \$_edit,
100 'rmdir' => \$SVN::Git::Editor::_rmdir,
101 'find-copies-harder' => \$SVN::Git::Editor::_find_copies_harder,
102 'l=i' => \$SVN::Git::Editor::_rename_limit,
103 'copy-similarity|C=i'=> \$SVN::Git::Editor::_cp_similarity
104 );
105
106 my %cmd = (
107 fetch => [ \&cmd_fetch, "Download new revisions from SVN",
108 { 'revision|r=s' => \$_revision,
109 'fetch-all|all' => \$_fetch_all,
110 %fc_opts } ],
111 clone => [ \&cmd_clone, "Initialize and fetch revisions",
112 { 'revision|r=s' => \$_revision,
113 %fc_opts, %init_opts } ],
114 init => [ \&cmd_init, "Initialize a repo for tracking" .
115 " (requires URL argument)",
116 \%init_opts ],
117 'multi-init' => [ \&cmd_multi_init,
118 "Deprecated alias for ".
119 "'$0 init -T<trunk> -b<branches> -t<tags>'",
120 \%init_opts ],
121 dcommit => [ \&cmd_dcommit,
122 'Commit several diffs to merge with upstream',
123 { 'merge|m|M' => \$_merge,
124 'strategy|s=s' => \$_strategy,
125 'verbose|v' => \$_verbose,
126 'dry-run|n' => \$_dry_run,
127 'fetch-all|all' => \$_fetch_all,
128 'no-rebase' => \$_no_rebase,
129 %cmt_opts, %fc_opts } ],
130 'set-tree' => [ \&cmd_set_tree,
131 "Set an SVN repository to a git tree-ish",
132 { 'stdin|' => \$_stdin, %cmt_opts, %fc_opts, } ],
133 'create-ignore' => [ \&cmd_create_ignore,
134 'Create a .gitignore per svn:ignore',
135 { 'revision|r=i' => \$_revision
136 } ],
137 'propget' => [ \&cmd_propget,
138 'Print the value of a property on a file or directory',
139 { 'revision|r=i' => \$_revision } ],
140 'proplist' => [ \&cmd_proplist,
141 'List all properties of a file or directory',
142 { 'revision|r=i' => \$_revision } ],
143 'show-ignore' => [ \&cmd_show_ignore, "Show svn:ignore listings",
144 { 'revision|r=i' => \$_revision
145 } ],
146 'show-externals' => [ \&cmd_show_externals, "Show svn:externals listings",
147 { 'revision|r=i' => \$_revision
148 } ],
149 'multi-fetch' => [ \&cmd_multi_fetch,
150 "Deprecated alias for $0 fetch --all",
151 { 'revision|r=s' => \$_revision, %fc_opts } ],
152 'migrate' => [ sub { },
153 # no-op, we automatically run this anyways,
154 'Migrate configuration/metadata/layout from
155 previous versions of git-svn',
156 { 'minimize' => \$Git::SVN::Migration::_minimize,
157 %remote_opts } ],
158 'log' => [ \&Git::SVN::Log::cmd_show_log, 'Show commit logs',
159 { 'limit=i' => \$Git::SVN::Log::limit,
160 'revision|r=s' => \$_revision,
161 'verbose|v' => \$Git::SVN::Log::verbose,
162 'incremental' => \$Git::SVN::Log::incremental,
163 'oneline' => \$Git::SVN::Log::oneline,
164 'show-commit' => \$Git::SVN::Log::show_commit,
165 'non-recursive' => \$Git::SVN::Log::non_recursive,
166 'authors-file|A=s' => \$_authors,
167 'color' => \$Git::SVN::Log::color,
168 'pager=s' => \$Git::SVN::Log::pager
169 } ],
170 'find-rev' => [ \&cmd_find_rev, "Translate between SVN revision numbers and tree-ish",
171 {} ],
172 'rebase' => [ \&cmd_rebase, "Fetch and rebase your working directory",
173 { 'merge|m|M' => \$_merge,
174 'verbose|v' => \$_verbose,
175 'strategy|s=s' => \$_strategy,
176 'local|l' => \$_local,
177 'fetch-all|all' => \$_fetch_all,
178 %fc_opts } ],
179 'commit-diff' => [ \&cmd_commit_diff,
180 'Commit a diff between two trees',
181 { 'message|m=s' => \$_message,
182 'file|F=s' => \$_file,
183 'revision|r=s' => \$_revision,
184 %cmt_opts } ],
185 'info' => [ \&cmd_info,
186 "Show info about the latest SVN revision
187 on the current branch",
188 { 'url' => \$_url, } ],
189 );
190
191 my $cmd;
192 for (my $i = 0; $i < @ARGV; $i++) {
193 if (defined $cmd{$ARGV[$i]}) {
194 $cmd = $ARGV[$i];
195 splice @ARGV, $i, 1;
196 last;
197 }
198 };
199
200 # make sure we're always running
201 unless ($cmd =~ /(?:clone|init|multi-init)$/) {
202 unless (-d $ENV{GIT_DIR}) {
203 if ($git_dir_user_set) {
204 die "GIT_DIR=$ENV{GIT_DIR} explicitly set, ",
205 "but it is not a directory\n";
206 }
207 my $git_dir = delete $ENV{GIT_DIR};
208 chomp(my $cdup = command_oneline(qw/rev-parse --show-cdup/));
209 unless (length $cdup) {
210 die "Already at toplevel, but $git_dir ",
211 "not found '$cdup'\n";
212 }
213 chdir $cdup or die "Unable to chdir up to '$cdup'\n";
214 unless (-d $git_dir) {
215 die "$git_dir still not found after going to ",
216 "'$cdup'\n";
217 }
218 $ENV{GIT_DIR} = $git_dir;
219 }
220 }
221
222 my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
223
224 read_repo_config(\%opts);
225 Getopt::Long::Configure('pass_through') if ($cmd && $cmd eq 'log');
226 my $rv = GetOptions(%opts, 'help|H|h' => \$_help, 'version|V' => \$_version,
227 'minimize-connections' => \$Git::SVN::Migration::_minimize,
228 'id|i=s' => \$Git::SVN::default_ref_id,
229 'svn-remote|remote|R=s' => sub {
230 $Git::SVN::no_reuse_existing = 1;
231 $Git::SVN::default_repo_id = $_[1] });
232 exit 1 if (!$rv && $cmd && $cmd ne 'log');
233
234 usage(0) if $_help;
235 version() if $_version;
236 usage(1) unless defined $cmd;
237 load_authors() if $_authors;
238
239 unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
240 Git::SVN::Migration::migration_check();
241 }
242 Git::SVN::init_vars();
243 eval {
244 Git::SVN::verify_remotes_sanity();
245 $cmd{$cmd}->[0]->(@ARGV);
246 };
247 fatal $@ if $@;
248 post_fetch_checkout();
249 exit 0;
250
251 ####################### primary functions ######################
252 sub usage {
253 my $exit = shift || 0;
254 my $fd = $exit ? \*STDERR : \*STDOUT;
255 print $fd <<"";
256 git-svn - bidirectional operations between a single Subversion tree and git
257 Usage: $0 <command> [options] [arguments]\n
258
259 print $fd "Available commands:\n" unless $cmd;
260
261 foreach (sort keys %cmd) {
262 next if $cmd && $cmd ne $_;
263 next if /^multi-/; # don't show deprecated commands
264 print $fd ' ',pack('A17',$_),$cmd{$_}->[1],"\n";
265 foreach (sort keys %{$cmd{$_}->[2]}) {
266 # mixed-case options are for .git/config only
267 next if /[A-Z]/ && /^[a-z]+$/i;
268 # prints out arguments as they should be passed:
269 my $x = s#[:=]s$## ? '<arg>' : s#[:=]i$## ? '<num>' : '';
270 print $fd ' ' x 21, join(', ', map { length $_ > 1 ?
271 "--$_" : "-$_" }
272 split /\|/,$_)," $x\n";
273 }
274 }
275 print $fd <<"";
276 \nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
277 arbitrary identifier if you're tracking multiple SVN branches/repositories in
278 one git repository and want to keep them separate. See git-svn(1) for more
279 information.
280
281 exit $exit;
282 }
283
284 sub version {
285 print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
286 exit 0;
287 }
288
289 sub do_git_init_db {
290 unless (-d $ENV{GIT_DIR}) {
291 my @init_db = ('init');
292 push @init_db, "--template=$_template" if defined $_template;
293 if (defined $_shared) {
294 if ($_shared =~ /[a-z]/) {
295 push @init_db, "--shared=$_shared";
296 } else {
297 push @init_db, "--shared";
298 }
299 }
300 command_noisy(@init_db);
301 }
302 my $set;
303 my $pfx = "svn-remote.$Git::SVN::default_repo_id";
304 foreach my $i (keys %icv) {
305 die "'$set' and '$i' cannot both be set\n" if $set;
306 next unless defined $icv{$i};
307 command_noisy('config', "$pfx.$i", $icv{$i});
308 $set = $i;
309 }
310 }
311
312 sub init_subdir {
313 my $repo_path = shift or return;
314 mkpath([$repo_path]) unless -d $repo_path;
315 chdir $repo_path or die "Couldn't chdir to $repo_path: $!\n";
316 $ENV{GIT_DIR} = '.git';
317 }
318
319 sub cmd_clone {
320 my ($url, $path) = @_;
321 if (!defined $path &&
322 (defined $_trunk || defined $_branches || defined $_tags ||
323 defined $_stdlayout) &&
324 $url !~ m#^[a-z\+]+://#) {
325 $path = $url;
326 }
327 $path = basename($url) if !defined $path || !length $path;
328 cmd_init($url, $path);
329 Git::SVN::fetch_all($Git::SVN::default_repo_id);
330 }
331
332 sub cmd_init {
333 if (defined $_stdlayout) {
334 $_trunk = 'trunk' if (!defined $_trunk);
335 $_tags = 'tags' if (!defined $_tags);
336 $_branches = 'branches' if (!defined $_branches);
337 }
338 if (defined $_trunk || defined $_branches || defined $_tags) {
339 return cmd_multi_init(@_);
340 }
341 my $url = shift or die "SVN repository location required ",
342 "as a command-line argument\n";
343 init_subdir(@_);
344 do_git_init_db();
345
346 Git::SVN->init($url);
347 }
348
349 sub cmd_fetch {
350 if (grep /^\d+=./, @_) {
351 die "'<rev>=<commit>' fetch arguments are ",
352 "no longer supported.\n";
353 }
354 my ($remote) = @_;
355 if (@_ > 1) {
356 die "Usage: $0 fetch [--all] [svn-remote]\n";
357 }
358 $remote ||= $Git::SVN::default_repo_id;
359 if ($_fetch_all) {
360 cmd_multi_fetch();
361 } else {
362 Git::SVN::fetch_all($remote, Git::SVN::read_all_remotes());
363 }
364 }
365
366 sub cmd_set_tree {
367 my (@commits) = @_;
368 if ($_stdin || !@commits) {
369 print "Reading from stdin...\n";
370 @commits = ();
371 while (<STDIN>) {
372 if (/\b($sha1_short)\b/o) {
373 unshift @commits, $1;
374 }
375 }
376 }
377 my @revs;
378 foreach my $c (@commits) {
379 my @tmp = command('rev-parse',$c);
380 if (scalar @tmp == 1) {
381 push @revs, $tmp[0];
382 } elsif (scalar @tmp > 1) {
383 push @revs, reverse(command('rev-list',@tmp));
384 } else {
385 fatal "Failed to rev-parse $c";
386 }
387 }
388 my $gs = Git::SVN->new;
389 my ($r_last, $cmt_last) = $gs->last_rev_commit;
390 $gs->fetch;
391 if (defined $gs->{last_rev} && $r_last != $gs->{last_rev}) {
392 fatal "There are new revisions that were fetched ",
393 "and need to be merged (or acknowledged) ",
394 "before committing.\nlast rev: $r_last\n",
395 " current: $gs->{last_rev}";
396 }
397 $gs->set_tree($_) foreach @revs;
398 print "Done committing ",scalar @revs," revisions to SVN\n";
399 }
400
401 sub cmd_dcommit {
402 my $head = shift;
403 git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
404 'Cannot dcommit with a dirty index. Commit your changes first, '
405 . "or stash them with `git stash'.\n";
406 $head ||= 'HEAD';
407 my @refs;
408 my ($url, $rev, $uuid, $gs) = working_head_info($head, \@refs);
409 print "Committing to $url ...\n";
410 unless ($gs) {
411 die "Unable to determine upstream SVN information from ",
412 "$head history\n";
413 }
414 my $last_rev;
415 my ($linear_refs, $parents) = linearize_history($gs, \@refs);
416 if ($_no_rebase && scalar(@$linear_refs) > 1) {
417 warn "Attempting to commit more than one change while ",
418 "--no-rebase is enabled.\n",
419 "If these changes depend on each other, re-running ",
420 "without --no-rebase will be required."
421 }
422 while (1) {
423 my $d = shift @$linear_refs or last;
424 unless (defined $last_rev) {
425 (undef, $last_rev, undef) = cmt_metadata("$d~1");
426 unless (defined $last_rev) {
427 fatal "Unable to extract revision information ",
428 "from commit $d~1";
429 }
430 }
431 if ($_dry_run) {
432 print "diff-tree $d~1 $d\n";
433 } else {
434 my $cmt_rev;
435 my %ed_opts = ( r => $last_rev,
436 log => get_commit_entry($d)->{log},
437 ra => Git::SVN::Ra->new($gs->full_url),
438 config => SVN::Core::config_get_config(
439 $Git::SVN::Ra::config_dir
440 ),
441 tree_a => "$d~1",
442 tree_b => $d,
443 editor_cb => sub {
444 print "Committed r$_[0]\n";
445 $cmt_rev = $_[0];
446 },
447 svn_path => '');
448 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
449 print "No changes\n$d~1 == $d\n";
450 } elsif ($parents->{$d} && @{$parents->{$d}}) {
451 $gs->{inject_parents_dcommit}->{$cmt_rev} =
452 $parents->{$d};
453 }
454 $_fetch_all ? $gs->fetch_all : $gs->fetch;
455 next if $_no_rebase;
456
457 # we always want to rebase against the current HEAD,
458 # not any head that was passed to us
459 my @diff = command('diff-tree', $d,
460 $gs->refname, '--');
461 my @finish;
462 if (@diff) {
463 @finish = rebase_cmd();
464 print STDERR "W: $d and ", $gs->refname,
465 " differ, using @finish:\n",
466 join("\n", @diff), "\n";
467 } else {
468 print "No changes between current HEAD and ",
469 $gs->refname,
470 "\nResetting to the latest ",
471 $gs->refname, "\n";
472 @finish = qw/reset --mixed/;
473 }
474 command_noisy(@finish, $gs->refname);
475 if (@diff) {
476 @refs = ();
477 my ($url_, $rev_, $uuid_, $gs_) =
478 working_head_info($head, \@refs);
479 my ($linear_refs_, $parents_) =
480 linearize_history($gs_, \@refs);
481 if (scalar(@$linear_refs) !=
482 scalar(@$linear_refs_)) {
483 fatal "# of revisions changed ",
484 "\nbefore:\n",
485 join("\n", @$linear_refs),
486 "\n\nafter:\n",
487 join("\n", @$linear_refs_), "\n",
488 'If you are attempting to commit ',
489 "merges, try running:\n\t",
490 'git rebase --interactive',
491 '--preserve-merges ',
492 $gs->refname,
493 "\nBefore dcommitting";
494 }
495 if ($url_ ne $url) {
496 fatal "URL mismatch after rebase: ",
497 "$url_ != $url";
498 }
499 if ($uuid_ ne $uuid) {
500 fatal "uuid mismatch after rebase: ",
501 "$uuid_ != $uuid";
502 }
503 # remap parents
504 my (%p, @l, $i);
505 for ($i = 0; $i < scalar @$linear_refs; $i++) {
506 my $new = $linear_refs_->[$i] or next;
507 $p{$new} =
508 $parents->{$linear_refs->[$i]};
509 push @l, $new;
510 }
511 $parents = \%p;
512 $linear_refs = \@l;
513 }
514 $last_rev = $cmt_rev;
515 }
516 }
517 }
518
519 sub cmd_find_rev {
520 my $revision_or_hash = shift;
521 my $result;
522 if ($revision_or_hash =~ /^r\d+$/) {
523 my $head = shift;
524 $head ||= 'HEAD';
525 my @refs;
526 my (undef, undef, undef, $gs) = working_head_info($head, \@refs);
527 unless ($gs) {
528 die "Unable to determine upstream SVN information from ",
529 "$head history\n";
530 }
531 my $desired_revision = substr($revision_or_hash, 1);
532 $result = $gs->rev_map_get($desired_revision);
533 } else {
534 my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
535 $result = $rev;
536 }
537 print "$result\n" if $result;
538 }
539
540 sub cmd_rebase {
541 command_noisy(qw/update-index --refresh/);
542 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
543 unless ($gs) {
544 die "Unable to determine upstream SVN information from ",
545 "working tree history\n";
546 }
547 if (command(qw/diff-index HEAD --/)) {
548 print STDERR "Cannot rebase with uncommited changes:\n";
549 command_noisy('status');
550 exit 1;
551 }
552 unless ($_local) {
553 # rebase will checkout for us, so no need to do it explicitly
554 $_no_checkout = 'true';
555 $_fetch_all ? $gs->fetch_all : $gs->fetch;
556 }
557 command_noisy(rebase_cmd(), $gs->refname);
558 }
559
560 sub cmd_show_ignore {
561 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
562 $gs ||= Git::SVN->new;
563 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
564 $gs->prop_walk($gs->{path}, $r, sub {
565 my ($gs, $path, $props) = @_;
566 print STDOUT "\n# $path\n";
567 my $s = $props->{'svn:ignore'} or return;
568 $s =~ s/[\r\n]+/\n/g;
569 chomp $s;
570 $s =~ s#^#$path#gm;
571 print STDOUT "$s\n";
572 });
573 }
574
575 sub cmd_show_externals {
576 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
577 $gs ||= Git::SVN->new;
578 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
579 $gs->prop_walk($gs->{path}, $r, sub {
580 my ($gs, $path, $props) = @_;
581 print STDOUT "\n# $path\n";
582 my $s = $props->{'svn:externals'} or return;
583 $s =~ s/[\r\n]+/\n/g;
584 chomp $s;
585 $s =~ s#^#$path#gm;
586 print STDOUT "$s\n";
587 });
588 }
589
590 sub cmd_create_ignore {
591 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
592 $gs ||= Git::SVN->new;
593 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
594 $gs->prop_walk($gs->{path}, $r, sub {
595 my ($gs, $path, $props) = @_;
596 # $path is of the form /path/to/dir/
597 my $ignore = '.' . $path . '.gitignore';
598 my $s = $props->{'svn:ignore'} or return;
599 open(GITIGNORE, '>', $ignore)
600 or fatal("Failed to open `$ignore' for writing: $!");
601 $s =~ s/[\r\n]+/\n/g;
602 chomp $s;
603 # Prefix all patterns so that the ignore doesn't apply
604 # to sub-directories.
605 $s =~ s#^#/#gm;
606 print GITIGNORE "$s\n";
607 close(GITIGNORE)
608 or fatal("Failed to close `$ignore': $!");
609 command_noisy('add', $ignore);
610 });
611 }
612
613 sub canonicalize_path {
614 my ($path) = @_;
615 my $dot_slash_added = 0;
616 if (substr($path, 0, 1) ne "/") {
617 $path = "./" . $path;
618 $dot_slash_added = 1;
619 }
620 # File::Spec->canonpath doesn't collapse x/../y into y (for a
621 # good reason), so let's do this manually.
622 $path =~ s#/+#/#g;
623 $path =~ s#/\.(?:/|$)#/#g;
624 $path =~ s#/[^/]+/\.\.##g;
625 $path =~ s#/$##g;
626 $path =~ s#^\./## if $dot_slash_added;
627 return $path;
628 }
629
630 # get_svnprops(PATH)
631 # ------------------
632 # Helper for cmd_propget and cmd_proplist below.
633 sub get_svnprops {
634 my $path = shift;
635 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
636 $gs ||= Git::SVN->new;
637
638 # prefix THE PATH by the sub-directory from which the user
639 # invoked us.
640 $path = $cmd_dir_prefix . $path;
641 fatal("No such file or directory: $path") unless -e $path;
642 my $is_dir = -d $path ? 1 : 0;
643 $path = $gs->{path} . '/' . $path;
644
645 # canonicalize the path (otherwise libsvn will abort or fail to
646 # find the file)
647 $path = canonicalize_path($path);
648
649 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
650 my $props;
651 if ($is_dir) {
652 (undef, undef, $props) = $gs->ra->get_dir($path, $r);
653 }
654 else {
655 (undef, $props) = $gs->ra->get_file($path, $r, undef);
656 }
657 return $props;
658 }
659
660 # cmd_propget (PROP, PATH)
661 # ------------------------
662 # Print the SVN property PROP for PATH.
663 sub cmd_propget {
664 my ($prop, $path) = @_;
665 $path = '.' if not defined $path;
666 usage(1) if not defined $prop;
667 my $props = get_svnprops($path);
668 if (not defined $props->{$prop}) {
669 fatal("`$path' does not have a `$prop' SVN property.");
670 }
671 print $props->{$prop} . "\n";
672 }
673
674 # cmd_proplist (PATH)
675 # -------------------
676 # Print the list of SVN properties for PATH.
677 sub cmd_proplist {
678 my $path = shift;
679 $path = '.' if not defined $path;
680 my $props = get_svnprops($path);
681 print "Properties on '$path':\n";
682 foreach (sort keys %{$props}) {
683 print " $_\n";
684 }
685 }
686
687 sub cmd_multi_init {
688 my $url = shift;
689 unless (defined $_trunk || defined $_branches || defined $_tags) {
690 usage(1);
691 }
692
693 # there are currently some bugs that prevent multi-init/multi-fetch
694 # setups from working well without this.
695 $Git::SVN::_minimize_url = 1;
696
697 $_prefix = '' unless defined $_prefix;
698 if (defined $url) {
699 $url =~ s#/+$##;
700 init_subdir(@_);
701 }
702 do_git_init_db();
703 if (defined $_trunk) {
704 my $trunk_ref = $_prefix . 'trunk';
705 # try both old-style and new-style lookups:
706 my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
707 unless ($gs_trunk) {
708 my ($trunk_url, $trunk_path) =
709 complete_svn_url($url, $_trunk);
710 $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
711 undef, $trunk_ref);
712 }
713 }
714 return unless defined $_branches || defined $_tags;
715 my $ra = $url ? Git::SVN::Ra->new($url) : undef;
716 complete_url_ls_init($ra, $_branches, '--branches/-b', $_prefix);
717 complete_url_ls_init($ra, $_tags, '--tags/-t', $_prefix . 'tags/');
718 }
719
720 sub cmd_multi_fetch {
721 my $remotes = Git::SVN::read_all_remotes();
722 foreach my $repo_id (sort keys %$remotes) {
723 if ($remotes->{$repo_id}->{url}) {
724 Git::SVN::fetch_all($repo_id, $remotes);
725 }
726 }
727 }
728
729 # this command is special because it requires no metadata
730 sub cmd_commit_diff {
731 my ($ta, $tb, $url) = @_;
732 my $usage = "Usage: $0 commit-diff -r<revision> ".
733 "<tree-ish> <tree-ish> [<URL>]";
734 fatal($usage) if (!defined $ta || !defined $tb);
735 my $svn_path;
736 if (!defined $url) {
737 my $gs = eval { Git::SVN->new };
738 if (!$gs) {
739 fatal("Needed URL or usable git-svn --id in ",
740 "the command-line\n", $usage);
741 }
742 $url = $gs->{url};
743 $svn_path = $gs->{path};
744 }
745 unless (defined $_revision) {
746 fatal("-r|--revision is a required argument\n", $usage);
747 }
748 if (defined $_message && defined $_file) {
749 fatal("Both --message/-m and --file/-F specified ",
750 "for the commit message.\n",
751 "I have no idea what you mean");
752 }
753 if (defined $_file) {
754 $_message = file_to_s($_file);
755 } else {
756 $_message ||= get_commit_entry($tb)->{log};
757 }
758 my $ra ||= Git::SVN::Ra->new($url);
759 $svn_path ||= $ra->{svn_path};
760 my $r = $_revision;
761 if ($r eq 'HEAD') {
762 $r = $ra->get_latest_revnum;
763 } elsif ($r !~ /^\d+$/) {
764 die "revision argument: $r not understood by git-svn\n";
765 }
766 my %ed_opts = ( r => $r,
767 log => $_message,
768 ra => $ra,
769 tree_a => $ta,
770 tree_b => $tb,
771 editor_cb => sub { print "Committed r$_[0]\n" },
772 svn_path => $svn_path );
773 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
774 print "No changes\n$ta == $tb\n";
775 }
776 }
777
778 sub cmd_info {
779 my $path = canonicalize_path(shift or ".");
780 unless (scalar(@_) == 0) {
781 die "Too many arguments specified\n";
782 }
783
784 my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
785
786 if (!$file_type && !$diff_status) {
787 print STDERR "$path: (Not a versioned resource)\n\n";
788 return;
789 }
790
791 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
792 unless ($gs) {
793 die "Unable to determine upstream SVN information from ",
794 "working tree history\n";
795 }
796 my $full_url = $url . ($path eq "." ? "" : "/$path");
797
798 if ($_url) {
799 print $full_url, "\n";
800 return;
801 }
802
803 my $result = "Path: $path\n";
804 $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
805 $result .= "URL: " . $full_url . "\n";
806
807 eval {
808 my $repos_root = $gs->repos_root;
809 Git::SVN::remove_username($repos_root);
810 $result .= "Repository Root: $repos_root\n";
811 };
812 if ($@) {
813 $result .= "Repository Root: (offline)\n";
814 }
815 $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A";
816 $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
817
818 $result .= "Node Kind: " .
819 ($file_type eq "dir" ? "directory" : "file") . "\n";
820
821 my $schedule = $diff_status eq "A"
822 ? "add"
823 : ($diff_status eq "D" ? "delete" : "normal");
824 $result .= "Schedule: $schedule\n";
825
826 if ($diff_status eq "A") {
827 print $result, "\n";
828 return;
829 }
830
831 my ($lc_author, $lc_rev, $lc_date_utc);
832 my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $path);
833 my $log = command_output_pipe(@args);
834 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
835 while (<$log>) {
836 if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
837 $lc_author = $1;
838 $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
839 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
840 (undef, $lc_rev, undef) = ::extract_metadata($1);
841 }
842 }
843 close $log;
844
845 Git::SVN::Log::set_local_timezone();
846
847 $result .= "Last Changed Author: $lc_author\n";
848 $result .= "Last Changed Rev: $lc_rev\n";
849 $result .= "Last Changed Date: " .
850 Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
851
852 if ($file_type ne "dir") {
853 my $text_last_updated_date =
854 ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
855 $result .=
856 "Text Last Updated: " .
857 Git::SVN::Log::format_svn_date($text_last_updated_date) .
858 "\n";
859 my $checksum;
860 if ($diff_status eq "D") {
861 my ($fh, $ctx) =
862 command_output_pipe(qw(cat-file blob), "HEAD:$path");
863 if ($file_type eq "link") {
864 my $file_name = <$fh>;
865 $checksum = md5sum("link $file_name");
866 } else {
867 $checksum = md5sum($fh);
868 }
869 command_close_pipe($fh, $ctx);
870 } elsif ($file_type eq "link") {
871 my $file_name =
872 command(qw(cat-file blob), "HEAD:$path");
873 $checksum =
874 md5sum("link " . $file_name);
875 } else {
876 open FILE, "<", $path or die $!;
877 $checksum = md5sum(\*FILE);
878 close FILE or die $!;
879 }
880 $result .= "Checksum: " . $checksum . "\n";
881 }
882
883 print $result, "\n";
884 }
885
886 ########################### utility functions #########################
887
888 sub rebase_cmd {
889 my @cmd = qw/rebase/;
890 push @cmd, '-v' if $_verbose;
891 push @cmd, qw/--merge/ if $_merge;
892 push @cmd, "--strategy=$_strategy" if $_strategy;
893 @cmd;
894 }
895
896 sub post_fetch_checkout {
897 return if $_no_checkout;
898 my $gs = $Git::SVN::_head or return;
899 return if verify_ref('refs/heads/master^0');
900
901 my $valid_head = verify_ref('HEAD^0');
902 command_noisy(qw(update-ref refs/heads/master), $gs->refname);
903 return if ($valid_head || !verify_ref('HEAD^0'));
904
905 return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
906 my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
907 return if -f $index;
908
909 return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
910 return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
911 command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
912 print STDERR "Checked out HEAD:\n ",
913 $gs->full_url, " r", $gs->last_rev, "\n";
914 }
915
916 sub complete_svn_url {
917 my ($url, $path) = @_;
918 $path =~ s#/+$##;
919 if ($path !~ m#^[a-z\+]+://#) {
920 if (!defined $url || $url !~ m#^[a-z\+]+://#) {
921 fatal("E: '$path' is not a complete URL ",
922 "and a separate URL is not specified");
923 }
924 return ($url, $path);
925 }
926 return ($path, '');
927 }
928
929 sub complete_url_ls_init {
930 my ($ra, $repo_path, $switch, $pfx) = @_;
931 unless ($repo_path) {
932 print STDERR "W: $switch not specified\n";
933 return;
934 }
935 $repo_path =~ s#/+$##;
936 if ($repo_path =~ m#^[a-z\+]+://#) {
937 $ra = Git::SVN::Ra->new($repo_path);
938 $repo_path = '';
939 } else {
940 $repo_path =~ s#^/+##;
941 unless ($ra) {
942 fatal("E: '$repo_path' is not a complete URL ",
943 "and a separate URL is not specified");
944 }
945 }
946 my $url = $ra->{url};
947 my $gs = Git::SVN->init($url, undef, undef, undef, 1);
948 my $k = "svn-remote.$gs->{repo_id}.url";
949 my $orig_url = eval { command_oneline(qw/config --get/, $k) };
950 if ($orig_url && ($orig_url ne $gs->{url})) {
951 die "$k already set: $orig_url\n",
952 "wanted to set to: $gs->{url}\n";
953 }
954 command_oneline('config', $k, $gs->{url}) unless $orig_url;
955 my $remote_path = "$ra->{svn_path}/$repo_path/*";
956 $remote_path =~ s#/+#/#g;
957 $remote_path =~ s#^/##g;
958 my ($n) = ($switch =~ /^--(\w+)/);
959 if (length $pfx && $pfx !~ m#/$#) {
960 die "--prefix='$pfx' must have a trailing slash '/'\n";
961 }
962 command_noisy('config', "svn-remote.$gs->{repo_id}.$n",
963 "$remote_path:refs/remotes/$pfx*");
964 }
965
966 sub verify_ref {
967 my ($ref) = @_;
968 eval { command_oneline([ 'rev-parse', '--verify', $ref ],
969 { STDERR => 0 }); };
970 }
971
972 sub get_tree_from_treeish {
973 my ($treeish) = @_;
974 # $treeish can be a symbolic ref, too:
975 my $type = command_oneline(qw/cat-file -t/, $treeish);
976 my $expected;
977 while ($type eq 'tag') {
978 ($treeish, $type) = command(qw/cat-file tag/, $treeish);
979 }
980 if ($type eq 'commit') {
981 $expected = (grep /^tree /, command(qw/cat-file commit/,
982 $treeish))[0];
983 ($expected) = ($expected =~ /^tree ($sha1)$/o);
984 die "Unable to get tree from $treeish\n" unless $expected;
985 } elsif ($type eq 'tree') {
986 $expected = $treeish;
987 } else {
988 die "$treeish is a $type, expected tree, tag or commit\n";
989 }
990 return $expected;
991 }
992
993 sub get_commit_entry {
994 my ($treeish) = shift;
995 my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
996 my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
997 my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
998 open my $log_fh, '>', $commit_editmsg or croak $!;
999
1000 my $type = command_oneline(qw/cat-file -t/, $treeish);
1001 if ($type eq 'commit' || $type eq 'tag') {
1002 my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1003 $type, $treeish);
1004 my $in_msg = 0;
1005 while (<$msg_fh>) {
1006 if (!$in_msg) {
1007 $in_msg = 1 if (/^\s*$/);
1008 } elsif (/^git-svn-id: /) {
1009 # skip this for now, we regenerate the
1010 # correct one on re-fetch anyways
1011 # TODO: set *:merge properties or like...
1012 } else {
1013 print $log_fh $_ or croak $!;
1014 }
1015 }
1016 command_close_pipe($msg_fh, $ctx);
1017 }
1018 close $log_fh or croak $!;
1019
1020 if ($_edit || ($type eq 'tree')) {
1021 my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
1022 # TODO: strip out spaces, comments, like git-commit.sh
1023 system($editor, $commit_editmsg);
1024 }
1025 rename $commit_editmsg, $commit_msg or croak $!;
1026 open $log_fh, '<', $commit_msg or croak $!;
1027 { local $/; chomp($log_entry{log} = <$log_fh>); }
1028 close $log_fh or croak $!;
1029 unlink $commit_msg;
1030 \%log_entry;
1031 }
1032
1033 sub s_to_file {
1034 my ($str, $file, $mode) = @_;
1035 open my $fd,'>',$file or croak $!;
1036 print $fd $str,"\n" or croak $!;
1037 close $fd or croak $!;
1038 chmod ($mode &~ umask, $file) if (defined $mode);
1039 }
1040
1041 sub file_to_s {
1042 my $file = shift;
1043 open my $fd,'<',$file or croak "$!: file: $file\n";
1044 local $/;
1045 my $ret = <$fd>;
1046 close $fd or croak $!;
1047 $ret =~ s/\s*$//s;
1048 return $ret;
1049 }
1050
1051 # '<svn username> = real-name <email address>' mapping based on git-svnimport:
1052 sub load_authors {
1053 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1054 my $log = $cmd eq 'log';
1055 while (<$authors>) {
1056 chomp;
1057 next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1058 my ($user, $name, $email) = ($1, $2, $3);
1059 if ($log) {
1060 $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1061 } else {
1062 $users{$user} = [$name, $email];
1063 }
1064 }
1065 close $authors or croak $!;
1066 }
1067
1068 # convert GetOpt::Long specs for use by git-config
1069 sub read_repo_config {
1070 return unless -d $ENV{GIT_DIR};
1071 my $opts = shift;
1072 my @config_only;
1073 foreach my $o (keys %$opts) {
1074 # if we have mixedCase and a long option-only, then
1075 # it's a config-only variable that we don't need for
1076 # the command-line.
1077 push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1078 my $v = $opts->{$o};
1079 my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1080 $key =~ s/-//g;
1081 my $arg = 'git-config';
1082 $arg .= ' --int' if ($o =~ /[:=]i$/);
1083 $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1084 if (ref $v eq 'ARRAY') {
1085 chomp(my @tmp = `$arg --get-all svn.$key`);
1086 @$v = @tmp if @tmp;
1087 } else {
1088 chomp(my $tmp = `$arg --get svn.$key`);
1089 if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1090 $$v = $tmp;
1091 }
1092 }
1093 }
1094 delete @$opts{@config_only} if @config_only;
1095 }
1096
1097 sub extract_metadata {
1098 my $id = shift or return (undef, undef, undef);
1099 my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1100 \s([a-f\d\-]+)$/x);
1101 if (!defined $rev || !$uuid || !$url) {
1102 # some of the original repositories I made had
1103 # identifiers like this:
1104 ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/);
1105 }
1106 return ($url, $rev, $uuid);
1107 }
1108
1109 sub cmt_metadata {
1110 return extract_metadata((grep(/^git-svn-id: /,
1111 command(qw/cat-file commit/, shift)))[-1]);
1112 }
1113
1114 sub working_head_info {
1115 my ($head, $refs) = @_;
1116 my @args = ('log', '--no-color', '--first-parent');
1117 my ($fh, $ctx) = command_output_pipe(@args, $head);
1118 my $hash;
1119 my %max;
1120 while (<$fh>) {
1121 if ( m{^commit ($::sha1)$} ) {
1122 unshift @$refs, $hash if $hash and $refs;
1123 $hash = $1;
1124 next;
1125 }
1126 next unless s{^\s*(git-svn-id:)}{$1};
1127 my ($url, $rev, $uuid) = extract_metadata($_);
1128 if (defined $url && defined $rev) {
1129 next if $max{$url} and $max{$url} < $rev;
1130 if (my $gs = Git::SVN->find_by_url($url)) {
1131 my $c = $gs->rev_map_get($rev);
1132 if ($c && $c eq $hash) {
1133 close $fh; # break the pipe
1134 return ($url, $rev, $uuid, $gs);
1135 } else {
1136 $max{$url} ||= $gs->rev_map_max;
1137 }
1138 }
1139 }
1140 }
1141 command_close_pipe($fh, $ctx);
1142 (undef, undef, undef, undef);
1143 }
1144
1145 sub read_commit_parents {
1146 my ($parents, $c) = @_;
1147 chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1148 $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1149 @{$parents->{$c}} = split(/ /, $p);
1150 }
1151
1152 sub linearize_history {
1153 my ($gs, $refs) = @_;
1154 my %parents;
1155 foreach my $c (@$refs) {
1156 read_commit_parents(\%parents, $c);
1157 }
1158
1159 my @linear_refs;
1160 my %skip = ();
1161 my $last_svn_commit = $gs->last_commit;
1162 foreach my $c (reverse @$refs) {
1163 next if $c eq $last_svn_commit;
1164 last if $skip{$c};
1165
1166 unshift @linear_refs, $c;
1167 $skip{$c} = 1;
1168
1169 # we only want the first parent to diff against for linear
1170 # history, we save the rest to inject when we finalize the
1171 # svn commit
1172 my $fp_a = verify_ref("$c~1");
1173 my $fp_b = shift @{$parents{$c}} if $parents{$c};
1174 if (!$fp_a || !$fp_b) {
1175 die "Commit $c\n",
1176 "has no parent commit, and therefore ",
1177 "nothing to diff against.\n",
1178 "You should be working from a repository ",
1179 "originally created by git-svn\n";
1180 }
1181 if ($fp_a ne $fp_b) {
1182 die "$c~1 = $fp_a, however parsing commit $c ",
1183 "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1184 }
1185
1186 foreach my $p (@{$parents{$c}}) {
1187 $skip{$p} = 1;
1188 }
1189 }
1190 (\@linear_refs, \%parents);
1191 }
1192
1193 sub find_file_type_and_diff_status {
1194 my ($path) = @_;
1195 return ('dir', '') if $path eq '.';
1196
1197 my $diff_output =
1198 command_oneline(qw(diff --cached --name-status --), $path) || "";
1199 my $diff_status = (split(' ', $diff_output))[0] || "";
1200
1201 my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1202
1203 return (undef, undef) if !$diff_status && !$ls_tree;
1204
1205 if ($diff_status eq "A") {
1206 return ("link", $diff_status) if -l $path;
1207 return ("dir", $diff_status) if -d $path;
1208 return ("file", $diff_status);
1209 }
1210
1211 my $mode = (split(' ', $ls_tree))[0] || "";
1212
1213 return ("link", $diff_status) if $mode eq "120000";
1214 return ("dir", $diff_status) if $mode eq "040000";
1215 return ("file", $diff_status);
1216 }
1217
1218 sub md5sum {
1219 my $arg = shift;
1220 my $ref = ref $arg;
1221 my $md5 = Digest::MD5->new();
1222 if ($ref eq 'GLOB' || $ref eq 'IO::File') {
1223 $md5->addfile($arg) or croak $!;
1224 } elsif ($ref eq 'SCALAR') {
1225 $md5->add($$arg) or croak $!;
1226 } elsif (!$ref) {
1227 $md5->add($arg) or croak $!;
1228 } else {
1229 ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
1230 }
1231 return $md5->hexdigest();
1232 }
1233
1234 package Git::SVN;
1235 use strict;
1236 use warnings;
1237 use Fcntl qw/:DEFAULT :seek/;
1238 use constant rev_map_fmt => 'NH40';
1239 use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1240 $_repack $_repack_flags $_use_svm_props $_head
1241 $_use_svnsync_props $no_reuse_existing $_minimize_url
1242 $_use_log_author/;
1243 use Carp qw/croak/;
1244 use File::Path qw/mkpath/;
1245 use File::Copy qw/copy/;
1246 use IPC::Open3;
1247
1248 my $_repack_nr;
1249 # properties that we do not log:
1250 my %SKIP_PROP;
1251 BEGIN {
1252 %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1253 svn:special svn:executable
1254 svn:entry:committed-rev
1255 svn:entry:last-author
1256 svn:entry:uuid
1257 svn:entry:committed-date/;
1258
1259 # some options are read globally, but can be overridden locally
1260 # per [svn-remote "..."] section. Command-line options will *NOT*
1261 # override options set in an [svn-remote "..."] section
1262 no strict 'refs';
1263 for my $option (qw/follow_parent no_metadata use_svm_props
1264 use_svnsync_props/) {
1265 my $key = $option;
1266 $key =~ tr/_//d;
1267 my $prop = "-$option";
1268 *$option = sub {
1269 my ($self) = @_;
1270 return $self->{$prop} if exists $self->{$prop};
1271 my $k = "svn-remote.$self->{repo_id}.$key";
1272 eval { command_oneline(qw/config --get/, $k) };
1273 if ($@) {
1274 $self->{$prop} = ${"Git::SVN::_$option"};
1275 } else {
1276 my $v = command_oneline(qw/config --bool/,$k);
1277 $self->{$prop} = $v eq 'false' ? 0 : 1;
1278 }
1279 return $self->{$prop};
1280 }
1281 }
1282 }
1283
1284 my %LOCKFILES;
1285 END { unlink keys %LOCKFILES if %LOCKFILES }
1286
1287 sub resolve_local_globs {
1288 my ($url, $fetch, $glob_spec) = @_;
1289 return unless defined $glob_spec;
1290 my $ref = $glob_spec->{ref};
1291 my $path = $glob_spec->{path};
1292 foreach (command(qw#for-each-ref --format=%(refname) refs/remotes#)) {
1293 next unless m#^refs/remotes/$ref->{regex}$#;
1294 my $p = $1;
1295 my $pathname = desanitize_refname($path->full_path($p));
1296 my $refname = desanitize_refname($ref->full_path($p));
1297 if (my $existing = $fetch->{$pathname}) {
1298 if ($existing ne $refname) {
1299 die "Refspec conflict:\n",
1300 "existing: refs/remotes/$existing\n",
1301 " globbed: refs/remotes/$refname\n";
1302 }
1303 my $u = (::cmt_metadata("refs/remotes/$refname"))[0];
1304 $u =~ s!^\Q$url\E(/|$)!! or die
1305 "refs/remotes/$refname: '$url' not found in '$u'\n";
1306 if ($pathname ne $u) {
1307 warn "W: Refspec glob conflict ",
1308 "(ref: refs/remotes/$refname):\n",
1309 "expected path: $pathname\n",
1310 " real path: $u\n",
1311 "Continuing ahead with $u\n";
1312 next;
1313 }
1314 } else {
1315 $fetch->{$pathname} = $refname;
1316 }
1317 }
1318 }
1319
1320 sub parse_revision_argument {
1321 my ($base, $head) = @_;
1322 if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1323 return ($base, $head);
1324 }
1325 return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1326 return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1327 return ($head, $head) if ($::_revision eq 'HEAD');
1328 return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1329 return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1330 die "revision argument: $::_revision not understood by git-svn\n";
1331 }
1332
1333 sub fetch_all {
1334 my ($repo_id, $remotes) = @_;
1335 if (ref $repo_id) {
1336 my $gs = $repo_id;
1337 $repo_id = undef;
1338 $repo_id = $gs->{repo_id};
1339 }
1340 $remotes ||= read_all_remotes();
1341 my $remote = $remotes->{$repo_id} or
1342 die "[svn-remote \"$repo_id\"] unknown\n";
1343 my $fetch = $remote->{fetch};
1344 my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1345 my (@gs, @globs);
1346 my $ra = Git::SVN::Ra->new($url);
1347 my $uuid = $ra->get_uuid;
1348 my $head = $ra->get_latest_revnum;
1349 my $base = defined $fetch ? $head : 0;
1350
1351 # read the max revs for wildcard expansion (branches/*, tags/*)
1352 foreach my $t (qw/branches tags/) {
1353 defined $remote->{$t} or next;
1354 push @globs, $remote->{$t};
1355 my $max_rev = eval { tmp_config(qw/--int --get/,
1356 "svn-remote.$repo_id.${t}-maxRev") };
1357 if (defined $max_rev && ($max_rev < $base)) {
1358 $base = $max_rev;
1359 } elsif (!defined $max_rev) {
1360 $base = 0;
1361 }
1362 }
1363
1364 if ($fetch) {
1365 foreach my $p (sort keys %$fetch) {
1366 my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1367 my $lr = $gs->rev_map_max;
1368 if (defined $lr) {
1369 $base = $lr if ($lr < $base);
1370 }
1371 push @gs, $gs;
1372 }
1373 }
1374
1375 ($base, $head) = parse_revision_argument($base, $head);
1376 $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1377 }
1378
1379 sub read_all_remotes {
1380 my $r = {};
1381 foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1382 if (m!^(.+)\.fetch=\s*(.*)\s*:\s*refs/remotes/(.+)\s*$!) {
1383 my ($remote, $local_ref, $remote_ref) = ($1, $2, $3);
1384 $local_ref =~ s{^/}{};
1385 $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1386 } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1387 $r->{$1}->{url} = $2;
1388 } elsif (m!^(.+)\.(branches|tags)=
1389 (.*):refs/remotes/(.+)\s*$/!x) {
1390 my ($p, $g) = ($3, $4);
1391 my $rs = $r->{$1}->{$2} = {
1392 t => $2,
1393 remote => $1,
1394 path => Git::SVN::GlobSpec->new($p),
1395 ref => Git::SVN::GlobSpec->new($g) };
1396 if (length($rs->{ref}->{right}) != 0) {
1397 die "The '*' glob character must be the last ",
1398 "character of '$g'\n";
1399 }
1400 }
1401 }
1402 $r;
1403 }
1404
1405 sub init_vars {
1406 if (defined $_repack) {
1407 $_repack = 1000 if ($_repack <= 0);
1408 $_repack_nr = $_repack;
1409 $_repack_flags ||= '-d';
1410 }
1411 }
1412
1413 sub verify_remotes_sanity {
1414 return unless -d $ENV{GIT_DIR};
1415 my %seen;
1416 foreach (command(qw/config -l/)) {
1417 if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1418 if ($seen{$1}) {
1419 die "Remote ref refs/remote/$1 is tracked by",
1420 "\n \"$_\"\nand\n \"$seen{$1}\"\n",
1421 "Please resolve this ambiguity in ",
1422 "your git configuration file before ",
1423 "continuing\n";
1424 }
1425 $seen{$1} = $_;
1426 }
1427 }
1428 }
1429
1430 # we allow more chars than remotes2config.sh...
1431 sub sanitize_remote_name {
1432 my ($name) = @_;
1433 $name =~ tr{A-Za-z0-9:,/+-}{.}c;
1434 $name;
1435 }
1436
1437 sub find_existing_remote {
1438 my ($url, $remotes) = @_;
1439 return undef if $no_reuse_existing;
1440 my $existing;
1441 foreach my $repo_id (keys %$remotes) {
1442 my $u = $remotes->{$repo_id}->{url} or next;
1443 next if $u ne $url;
1444 $existing = $repo_id;
1445 last;
1446 }
1447 $existing;
1448 }
1449
1450 sub init_remote_config {
1451 my ($self, $url, $no_write) = @_;
1452 $url =~ s!/+$!!; # strip trailing slash
1453 my $r = read_all_remotes();
1454 my $existing = find_existing_remote($url, $r);
1455 if ($existing) {
1456 unless ($no_write) {
1457 print STDERR "Using existing ",
1458 "[svn-remote \"$existing\"]\n";
1459 }
1460 $self->{repo_id} = $existing;
1461 } elsif ($_minimize_url) {
1462 my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1463 $existing = find_existing_remote($min_url, $r);
1464 if ($existing) {
1465 unless ($no_write) {
1466 print STDERR "Using existing ",
1467 "[svn-remote \"$existing\"]\n";
1468 }
1469 $self->{repo_id} = $existing;
1470 }
1471 if ($min_url ne $url) {
1472 unless ($no_write) {
1473 print STDERR "Using higher level of URL: ",
1474 "$url => $min_url\n";
1475 }
1476 my $old_path = $self->{path};
1477 $self->{path} = $url;
1478 $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1479 if (length $old_path) {
1480 $self->{path} .= "/$old_path";
1481 }
1482 $url = $min_url;
1483 }
1484 }
1485 my $orig_url;
1486 if (!$existing) {
1487 # verify that we aren't overwriting anything:
1488 $orig_url = eval {
1489 command_oneline('config', '--get',
1490 "svn-remote.$self->{repo_id}.url")
1491 };
1492 if ($orig_url && ($orig_url ne $url)) {
1493 die "svn-remote.$self->{repo_id}.url already set: ",
1494 "$orig_url\nwanted to set to: $url\n";
1495 }
1496 }
1497 my ($xrepo_id, $xpath) = find_ref($self->refname);
1498 if (defined $xpath) {
1499 die "svn-remote.$xrepo_id.fetch already set to track ",
1500 "$xpath:refs/remotes/", $self->refname, "\n";
1501 }
1502 unless ($no_write) {
1503 command_noisy('config',
1504 "svn-remote.$self->{repo_id}.url", $url);
1505 $self->{path} =~ s{^/}{};
1506 command_noisy('config', '--add',
1507 "svn-remote.$self->{repo_id}.fetch",
1508 "$self->{path}:".$self->refname);
1509 }
1510 $self->{url} = $url;
1511 }
1512
1513 sub find_by_url { # repos_root and, path are optional
1514 my ($class, $full_url, $repos_root, $path) = @_;
1515
1516 return undef unless defined $full_url;
1517 remove_username($full_url);
1518 remove_username($repos_root) if defined $repos_root;
1519 my $remotes = read_all_remotes();
1520 if (defined $full_url && defined $repos_root && !defined $path) {
1521 $path = $full_url;
1522 $path =~ s#^\Q$repos_root\E(?:/|$)##;
1523 }
1524 foreach my $repo_id (keys %$remotes) {
1525 my $u = $remotes->{$repo_id}->{url} or next;
1526 remove_username($u);
1527 next if defined $repos_root && $repos_root ne $u;
1528
1529 my $fetch = $remotes->{$repo_id}->{fetch} || {};
1530 foreach (qw/branches tags/) {
1531 resolve_local_globs($u, $fetch,
1532 $remotes->{$repo_id}->{$_});
1533 }
1534 my $p = $path;
1535 unless (defined $p) {
1536 $p = $full_url;
1537 $p =~ s#^\Q$u\E(?:/|$)## or next;
1538 }
1539 foreach my $f (keys %$fetch) {
1540 next if $f ne $p;
1541 return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1542 }
1543 }
1544 undef;
1545 }
1546
1547 sub init {
1548 my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1549 my $self = _new($class, $repo_id, $ref_id, $path);
1550 if (defined $url) {
1551 $self->init_remote_config($url, $no_write);
1552 }
1553 $self;
1554 }
1555
1556 sub find_ref {
1557 my ($ref_id) = @_;
1558 foreach (command(qw/config -l/)) {
1559 next unless m!^svn-remote\.(.+)\.fetch=
1560 \s*(.*)\s*:\s*refs/remotes/(.+)\s*$!x;
1561 my ($repo_id, $path, $ref) = ($1, $2, $3);
1562 if ($ref eq $ref_id) {
1563 $path = '' if ($path =~ m#^\./?#);
1564 return ($repo_id, $path);
1565 }
1566 }
1567 (undef, undef, undef);
1568 }
1569
1570 sub new {
1571 my ($class, $ref_id, $repo_id, $path) = @_;
1572 if (defined $ref_id && !defined $repo_id && !defined $path) {
1573 ($repo_id, $path) = find_ref($ref_id);
1574 if (!defined $repo_id) {
1575 die "Could not find a \"svn-remote.*.fetch\" key ",
1576 "in the repository configuration matching: ",
1577 "refs/remotes/$ref_id\n";
1578 }
1579 }
1580 my $self = _new($class, $repo_id, $ref_id, $path);
1581 if (!defined $self->{path} || !length $self->{path}) {
1582 my $fetch = command_oneline('config', '--get',
1583 "svn-remote.$repo_id.fetch",
1584 ":refs/remotes/$ref_id\$") or
1585 die "Failed to read \"svn-remote.$repo_id.fetch\" ",
1586 "\":refs/remotes/$ref_id\$\" in config\n";
1587 ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
1588 }
1589 $self->{url} = command_oneline('config', '--get',
1590 "svn-remote.$repo_id.url") or
1591 die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
1592 $self->rebuild;
1593 $self;
1594 }
1595
1596 sub refname {
1597 my ($refname) = "refs/remotes/$_[0]->{ref_id}" ;
1598
1599 # It cannot end with a slash /, we'll throw up on this because
1600 # SVN can't have directories with a slash in their name, either:
1601 if ($refname =~ m{/$}) {
1602 die "ref: '$refname' ends with a trailing slash, this is ",
1603 "not permitted by git nor Subversion\n";
1604 }
1605
1606 # It cannot have ASCII control character space, tilde ~, caret ^,
1607 # colon :, question-mark ?, asterisk *, space, or open bracket [
1608 # anywhere.
1609 #
1610 # Additionally, % must be escaped because it is used for escaping
1611 # and we want our escaped refname to be reversible
1612 $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
1613
1614 # no slash-separated component can begin with a dot .
1615 # /.* becomes /%2E*
1616 $refname =~ s{/\.}{/%2E}g;
1617
1618 # It cannot have two consecutive dots .. anywhere
1619 # .. becomes %2E%2E
1620 $refname =~ s{\.\.}{%2E%2E}g;
1621
1622 return $refname;
1623 }
1624
1625 sub desanitize_refname {
1626 my ($refname) = @_;
1627 $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
1628 return $refname;
1629 }
1630
1631 sub svm_uuid {
1632 my ($self) = @_;
1633 return $self->{svm}->{uuid} if $self->svm;
1634 $self->ra;
1635 unless ($self->{svm}) {
1636 die "SVM UUID not cached, and reading remotely failed\n";
1637 }
1638 $self->{svm}->{uuid};
1639 }
1640
1641 sub svm {
1642 my ($self) = @_;
1643 return $self->{svm} if $self->{svm};
1644 my $svm;
1645 # see if we have it in our config, first:
1646 eval {
1647 my $section = "svn-remote.$self->{repo_id}";
1648 $svm = {
1649 source => tmp_config('--get', "$section.svm-source"),
1650 uuid => tmp_config('--get', "$section.svm-uuid"),
1651 replace => tmp_config('--get', "$section.svm-replace"),
1652 }
1653 };
1654 if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
1655 $self->{svm} = $svm;
1656 }
1657 $self->{svm};
1658 }
1659
1660 sub _set_svm_vars {
1661 my ($self, $ra) = @_;
1662 return $ra if $self->svm;
1663
1664 my @err = ( "useSvmProps set, but failed to read SVM properties\n",
1665 "(svm:source, svm:uuid) ",
1666 "from the following URLs:\n" );
1667 sub read_svm_props {
1668 my ($self, $ra, $path, $r) = @_;
1669 my $props = ($ra->get_dir($path, $r))[2];
1670 my $src = $props->{'svm:source'};
1671 my $uuid = $props->{'svm:uuid'};
1672 return undef if (!$src || !$uuid);
1673
1674 chomp($src, $uuid);
1675
1676 $uuid =~ m{^[0-9a-f\-]{30,}$}
1677 or die "doesn't look right - svm:uuid is '$uuid'\n";
1678
1679 # the '!' is used to mark the repos_root!/relative/path
1680 $src =~ s{/?!/?}{/};
1681 $src =~ s{/+$}{}; # no trailing slashes please
1682 # username is of no interest
1683 $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
1684
1685 my $replace = $ra->{url};
1686 $replace .= "/$path" if length $path;
1687
1688 my $section = "svn-remote.$self->{repo_id}";
1689 tmp_config("$section.svm-source", $src);
1690 tmp_config("$section.svm-replace", $replace);
1691 tmp_config("$section.svm-uuid", $uuid);
1692 $self->{svm} = {
1693 source => $src,
1694 uuid => $uuid,
1695 replace => $replace
1696 };
1697 }
1698
1699 my $r = $ra->get_latest_revnum;
1700 my $path = $self->{path};
1701 my %tried;
1702 while (length $path) {
1703 unless ($tried{"$self->{url}/$path"}) {
1704 return $ra if $self->read_svm_props($ra, $path, $r);
1705 $tried{"$self->{url}/$path"} = 1;
1706 }
1707 $path =~ s#/?[^/]+$##;
1708 }
1709 die "Path: '$path' should be ''\n" if $path ne '';
1710 return $ra if $self->read_svm_props($ra, $path, $r);
1711 $tried{"$self->{url}/$path"} = 1;
1712
1713 if ($ra->{repos_root} eq $self->{url}) {
1714 die @err, (map { " $_\n" } keys %tried), "\n";
1715 }
1716
1717 # nope, make sure we're connected to the repository root:
1718 my $ok;
1719 my @tried_b;
1720 $path = $ra->{svn_path};
1721 $ra = Git::SVN::Ra->new($ra->{repos_root});
1722 while (length $path) {
1723 unless ($tried{"$ra->{url}/$path"}) {
1724 $ok = $self->read_svm_props($ra, $path, $r);
1725 last if $ok;
1726 $tried{"$ra->{url}/$path"} = 1;
1727 }
1728 $path =~ s#/?[^/]+$##;
1729 }
1730 die "Path: '$path' should be ''\n" if $path ne '';
1731 $ok ||= $self->read_svm_props($ra, $path, $r);
1732 $tried{"$ra->{url}/$path"} = 1;
1733 if (!$ok) {
1734 die @err, (map { " $_\n" } keys %tried), "\n";
1735 }
1736 Git::SVN::Ra->new($self->{url});
1737 }
1738
1739 sub svnsync {
1740 my ($self) = @_;
1741 return $self->{svnsync} if $self->{svnsync};
1742
1743 if ($self->no_metadata) {
1744 die "Can't have both 'noMetadata' and ",
1745 "'useSvnsyncProps' options set!\n";
1746 }
1747 if ($self->rewrite_root) {
1748 die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
1749 "options set!\n";
1750 }
1751
1752 my $svnsync;
1753 # see if we have it in our config, first:
1754 eval {
1755 my $section = "svn-remote.$self->{repo_id}";
1756 $svnsync = {
1757 url => tmp_config('--get', "$section.svnsync-url"),
1758 uuid => tmp_config('--get', "$section.svnsync-uuid"),
1759 }
1760 };
1761 if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
1762 return $self->{svnsync} = $svnsync;
1763 }
1764
1765 my $err = "useSvnsyncProps set, but failed to read " .
1766 "svnsync property: svn:sync-from-";
1767 my $rp = $self->ra->rev_proplist(0);
1768
1769 my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
1770 $url =~ m{^[a-z\+]+://} or
1771 die "doesn't look right - svn:sync-from-url is '$url'\n";
1772
1773 my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
1774 $uuid =~ m{^[0-9a-f\-]{30,}$} or
1775 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1776
1777 my $section = "svn-remote.$self->{repo_id}";
1778 tmp_config('--add', "$section.svnsync-uuid", $uuid);
1779 tmp_config('--add', "$section.svnsync-url", $url);
1780 return $self->{svnsync} = { url => $url, uuid => $uuid };
1781 }
1782
1783 # this allows us to memoize our SVN::Ra UUID locally and avoid a
1784 # remote lookup (useful for 'git svn log').
1785 sub ra_uuid {
1786 my ($self) = @_;
1787 unless ($self->{ra_uuid}) {
1788 my $key = "svn-remote.$self->{repo_id}.uuid";
1789 my $uuid = eval { tmp_config('--get', $key) };
1790 if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/) {
1791 $self->{ra_uuid} = $uuid;
1792 } else {
1793 die "ra_uuid called without URL\n" unless $self->{url};
1794 $self->{ra_uuid} = $self->ra->get_uuid;
1795 tmp_config('--add', $key, $self->{ra_uuid});
1796 }
1797 }
1798 $self->{ra_uuid};
1799 }
1800
1801 sub _set_repos_root {
1802 my ($self, $repos_root) = @_;
1803 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1804 $repos_root ||= $self->ra->{repos_root};
1805 tmp_config($k, $repos_root);
1806 $repos_root;
1807 }
1808
1809 sub repos_root {
1810 my ($self) = @_;
1811 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1812 eval { tmp_config('--get', $k) } || $self->_set_repos_root;
1813 }
1814
1815 sub ra {
1816 my ($self) = shift;
1817 my $ra = Git::SVN::Ra->new($self->{url});
1818 $self->_set_repos_root($ra->{repos_root});
1819 if ($self->use_svm_props && !$self->{svm}) {
1820 if ($self->no_metadata) {
1821 die "Can't have both 'noMetadata' and ",
1822 "'useSvmProps' options set!\n";
1823 } elsif ($self->use_svnsync_props) {
1824 die "Can't have both 'useSvnsyncProps' and ",
1825 "'useSvmProps' options set!\n";
1826 }
1827 $ra = $self->_set_svm_vars($ra);
1828 $self->{-want_revprops} = 1;
1829 }
1830 $ra;
1831 }
1832
1833 sub rel_path {
1834 my ($self) = @_;
1835 my $repos_root = $self->ra->{repos_root};
1836 return $self->{path} if ($self->{url} eq $repos_root);
1837 my $url = $self->{url} .
1838 (length $self->{path} ? "/$self->{path}" : $self->{path});
1839 $url =~ s!^\Q$repos_root\E(?:/+|$)!!g;
1840 $url;
1841 }
1842
1843 # prop_walk(PATH, REV, SUB)
1844 # -------------------------
1845 # Recursively traverse PATH at revision REV and invoke SUB for each
1846 # directory that contains a SVN property. SUB will be invoked as
1847 # follows: &SUB(gs, path, props); where `gs' is this instance of
1848 # Git::SVN, `path' the path to the directory where the properties
1849 # `props' were found. The `path' will be relative to point of checkout,
1850 # that is, if url://repo/trunk is the current Git branch, and that
1851 # directory contains a sub-directory `d', SUB will be invoked with `/d/'
1852 # as `path' (note the trailing `/').
1853 sub prop_walk {
1854 my ($self, $path, $rev, $sub) = @_;
1855
1856 my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
1857 $path =~ s#^/*#/#g;
1858 my $p = $path;
1859 # Strip the irrelevant part of the path.
1860 $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
1861 # Ensure the path is terminated by a `/'.
1862 $p =~ s#/*$#/#;
1863
1864 # The properties contain all the internal SVN stuff nobody
1865 # (usually) cares about.
1866 my $interesting_props = 0;
1867 foreach (keys %{$props}) {
1868 # If it doesn't start with `svn:', it must be a
1869 # user-defined property.
1870 ++$interesting_props and next if $_ !~ /^svn:/;
1871 # FIXME: Fragile, if SVN adds new public properties,
1872 # this needs to be updated.
1873 ++$interesting_props if /^svn:(?:ignore|keywords|executable
1874 |eol-style|mime-type
1875 |externals|needs-lock)$/x;
1876 }
1877 &$sub($self, $p, $props) if $interesting_props;
1878
1879 foreach (sort keys %$dirent) {
1880 next if $dirent->{$_}->{kind} != $SVN::Node::dir;
1881 $self->prop_walk($path . '/' . $_, $rev, $sub);
1882 }
1883 }
1884
1885 sub last_rev { ($_[0]->last_rev_commit)[0] }
1886 sub last_commit { ($_[0]->last_rev_commit)[1] }
1887
1888 # returns the newest SVN revision number and newest commit SHA1
1889 sub last_rev_commit {
1890 my ($self) = @_;
1891 if (defined $self->{last_rev} && defined $self->{last_commit}) {
1892 return ($self->{last_rev}, $self->{last_commit});
1893 }
1894 my $c = ::verify_ref($self->refname.'^0');
1895 if ($c && !$self->use_svm_props && !$self->no_metadata) {
1896 my $rev = (::cmt_metadata($c))[1];
1897 if (defined $rev) {
1898 ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
1899 return ($rev, $c);
1900 }
1901 }
1902 my $map_path = $self->map_path;
1903 unless (-e $map_path) {
1904 ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
1905 return (undef, undef);
1906 }
1907 my ($rev, $commit) = $self->rev_map_max(1);
1908 ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
1909 return ($rev, $commit);
1910 }
1911
1912 sub get_fetch_range {
1913 my ($self, $min, $max) = @_;
1914 $max ||= $self->ra->get_latest_revnum;
1915 $min ||= $self->rev_map_max;
1916 (++$min, $max);
1917 }
1918
1919 sub tmp_config {
1920 my (@args) = @_;
1921 my $old_def_config = "$ENV{GIT_DIR}/svn/config";
1922 my $config = "$ENV{GIT_DIR}/svn/.metadata";
1923 if (! -f $config && -f $old_def_config) {
1924 rename $old_def_config, $config or
1925 die "Failed rename $old_def_config => $config: $!\n";
1926 }
1927 my $old_config = $ENV{GIT_CONFIG};
1928 $ENV{GIT_CONFIG} = $config;
1929 $@ = undef;
1930 my @ret = eval {
1931 unless (-f $config) {
1932 mkfile($config);
1933 open my $fh, '>', $config or
1934 die "Can't open $config: $!\n";
1935 print $fh "; This file is used internally by ",
1936 "git-svn\n" or die
1937 "Couldn't write to $config: $!\n";
1938 print $fh "; You should not have to edit it\n" or
1939 die "Couldn't write to $config: $!\n";
1940 close $fh or die "Couldn't close $config: $!\n";
1941 }
1942 command('config', @args);
1943 };
1944 my $err = $@;
1945 if (defined $old_config) {
1946 $ENV{GIT_CONFIG} = $old_config;
1947 } else {
1948 delete $ENV{GIT_CONFIG};
1949 }
1950 die $err if $err;
1951 wantarray ? @ret : $ret[0];
1952 }
1953
1954 sub tmp_index_do {
1955 my ($self, $sub) = @_;
1956 my $old_index = $ENV{GIT_INDEX_FILE};
1957 $ENV{GIT_INDEX_FILE} = $self->{index};
1958 $@ = undef;
1959 my @ret = eval {
1960 my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
1961 mkpath([$dir]) unless -d $dir;
1962 &$sub;
1963 };
1964 my $err = $@;
1965 if (defined $old_index) {
1966 $ENV{GIT_INDEX_FILE} = $old_index;
1967 } else {
1968 delete $ENV{GIT_INDEX_FILE};
1969 }
1970 die $err if $err;
1971 wantarray ? @ret : $ret[0];
1972 }
1973
1974 sub assert_index_clean {
1975 my ($self, $treeish) = @_;
1976
1977 $self->tmp_index_do(sub {
1978 command_noisy('read-tree', $treeish) unless -e $self->{index};
1979 my $x = command_oneline('write-tree');
1980 my ($y) = (command(qw/cat-file commit/, $treeish) =~
1981 /^tree ($::sha1)/mo);
1982 return if $y eq $x;
1983
1984 warn "Index mismatch: $y != $x\nrereading $treeish\n";
1985 unlink $self->{index} or die "unlink $self->{index}: $!\n";
1986 command_noisy('read-tree', $treeish);
1987 $x = command_oneline('write-tree');
1988 if ($y ne $x) {
1989 ::fatal "trees ($treeish) $y != $x\n",
1990 "Something is seriously wrong...";
1991 }
1992 });
1993 }
1994
1995 sub get_commit_parents {
1996 my ($self, $log_entry) = @_;
1997 my (%seen, @ret, @tmp);
1998 # legacy support for 'set-tree'; this is only used by set_tree_cb:
1999 if (my $ip = $self->{inject_parents}) {
2000 if (my $commit = delete $ip->{$log_entry->{revision}}) {
2001 push @tmp, $commit;
2002 }
2003 }
2004 if (my $cur = ::verify_ref($self->refname.'^0')) {
2005 push @tmp, $cur;
2006 }
2007 if (my $ipd = $self->{inject_parents_dcommit}) {
2008 if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2009 push @tmp, @$commit;
2010 }
2011 }
2012 push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2013 while (my $p = shift @tmp) {
2014 next if $seen{$p};
2015 $seen{$p} = 1;
2016 push @ret, $p;
2017 # MAXPARENT is defined to 16 in commit-tree.c:
2018 last if @ret >= 16;
2019 }
2020 if (@tmp) {
2021 die "r$log_entry->{revision}: No room for parents:\n\t",
2022 join("\n\t", @tmp), "\n";
2023 }
2024 @ret;
2025 }
2026
2027 sub rewrite_root {
2028 my ($self) = @_;
2029 return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2030 my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2031 my $rwr = eval { command_oneline(qw/config --get/, $k) };
2032 if ($rwr) {
2033 $rwr =~ s#/+$##;
2034 if ($rwr !~ m#^[a-z\+]+://#) {
2035 die "$rwr is not a valid URL (key: $k)\n";
2036 }
2037 }
2038 $self->{-rewrite_root} = $rwr;
2039 }
2040
2041 sub metadata_url {
2042 my ($self) = @_;
2043 ($self->rewrite_root || $self->{url}) .
2044 (length $self->{path} ? '/' . $self->{path} : '');
2045 }
2046
2047 sub full_url {
2048 my ($self) = @_;
2049 $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2050 }
2051
2052 sub do_git_commit {
2053 my ($self, $log_entry) = @_;
2054 my $lr = $self->last_rev;
2055 if (defined $lr && $lr >= $log_entry->{revision}) {
2056 die "Last fetched revision of ", $self->refname,
2057 " was r$lr, but we are about to fetch: ",
2058 "r$log_entry->{revision}!\n";
2059 }
2060 if (my $c = $self->rev_map_get($log_entry->{revision})) {
2061 croak "$log_entry->{revision} = $c already exists! ",
2062 "Why are we refetching it?\n";
2063 }
2064 $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2065 $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2066 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2067
2068 $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2069 ? $log_entry->{commit_name}
2070 : $log_entry->{name};
2071 $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2072 ? $log_entry->{commit_email}
2073 : $log_entry->{email};
2074
2075 my $tree = $log_entry->{tree};
2076 if (!defined $tree) {
2077 $tree = $self->tmp_index_do(sub {
2078 command_oneline('write-tree') });
2079 }
2080 die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2081
2082 my @exec = ('git-commit-tree', $tree);
2083 foreach ($self->get_commit_parents($log_entry)) {
2084 push @exec, '-p', $_;
2085 }
2086 defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2087 or croak $!;
2088 print $msg_fh $log_entry->{log} or croak $!;
2089 unless ($self->no_metadata) {
2090 print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2091 or croak $!;
2092 }
2093 $msg_fh->flush == 0 or croak $!;
2094 close $msg_fh or croak $!;
2095 chomp(my $commit = do { local $/; <$out_fh> });
2096 close $out_fh or croak $!;
2097 waitpid $pid, 0;
2098 croak $? if $?;
2099 if ($commit !~ /^$::sha1$/o) {
2100 die "Failed to commit, invalid sha1: $commit\n";
2101 }
2102
2103 $self->rev_map_set($log_entry->{revision}, $commit, 1);
2104
2105 $self->{last_rev} = $log_entry->{revision};
2106 $self->{last_commit} = $commit;
2107 print "r$log_entry->{revision}";
2108 if (defined $log_entry->{svm_revision}) {
2109 print " (\@$log_entry->{svm_revision})";
2110 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2111 0, $self->svm_uuid);
2112 }
2113 print " = $commit ($self->{ref_id})\n";
2114 if (defined $_repack && (--$_repack_nr == 0)) {
2115 $_repack_nr = $_repack;
2116 # repack doesn't use any arguments with spaces in them, does it?
2117 print "Running git repack $_repack_flags ...\n";
2118 command_noisy('repack', split(/\s+/, $_repack_flags));
2119 print "Done repacking\n";
2120 }
2121 return $commit;
2122 }
2123
2124 sub match_paths {
2125 my ($self, $paths, $r) = @_;
2126 return 1 if $self->{path} eq '';
2127 if (my $path = $paths->{"/$self->{path}"}) {
2128 return ($path->{action} eq 'D') ? 0 : 1;
2129 }
2130 $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
2131 if (grep /$self->{path_regex}/, keys %$paths) {
2132 return 1;
2133 }
2134 my $c = '';
2135 foreach (split m#/#, $self->{path}) {
2136 $c .= "/$_";
2137 next unless ($paths->{$c} &&
2138 ($paths->{$c}->{action} =~ /^[AR]$/));
2139 if ($self->ra->check_path($self->{path}, $r) ==
2140 $SVN::Node::dir) {
2141 return 1;
2142 }
2143 }
2144 return 0;
2145 }
2146
2147 sub find_parent_branch {
2148 my ($self, $paths, $rev) = @_;
2149 return undef unless $self->follow_parent;
2150 unless (defined $paths) {
2151 my $err_handler = $SVN::Error::handler;
2152 $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2153 $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1, sub {
2154 $paths =
2155 Git::SVN::Ra::dup_changed_paths($_[0]) });
2156 $SVN::Error::handler = $err_handler;
2157 }
2158 return undef unless defined $paths;
2159
2160 # look for a parent from another branch:
2161 my @b_path_components = split m#/#, $self->rel_path;
2162 my @a_path_components;
2163 my $i;
2164 while (@b_path_components) {
2165 $i = $paths->{'/'.join('/', @b_path_components)};
2166 last if $i && defined $i->{copyfrom_path};
2167 unshift(@a_path_components, pop(@b_path_components));
2168 }
2169 return undef unless defined $i && defined $i->{copyfrom_path};
2170 my $branch_from = $i->{copyfrom_path};
2171 if (@a_path_components) {
2172 print STDERR "branch_from: $branch_from => ";
2173 $branch_from .= '/'.join('/', @a_path_components);
2174 print STDERR $branch_from, "\n";
2175 }
2176 my $r = $i->{copyfrom_rev};
2177 my $repos_root = $self->ra->{repos_root};
2178 my $url = $self->ra->{url};
2179 my $new_url = $repos_root . $branch_from;
2180 print STDERR "Found possible branch point: ",
2181 "$new_url => ", $self->full_url, ", $r\n";
2182 $branch_from =~ s#^/##;
2183 my $gs = Git::SVN->find_by_url($new_url, $repos_root, $branch_from);
2184 unless ($gs) {
2185 my $ref_id = $self->{ref_id};
2186 $ref_id =~ s/\@\d+$//;
2187 $ref_id .= "\@$r";
2188 # just grow a tail if we're not unique enough :x
2189 $ref_id .= '-' while find_ref($ref_id);
2190 print STDERR "Initializing parent: $ref_id\n";
2191 $gs = Git::SVN->init($new_url, '', $ref_id, $ref_id, 1);
2192 }
2193 my ($r0, $parent) = $gs->find_rev_before($r, 1);
2194 if (!defined $r0 || !defined $parent) {
2195 my ($base, $head) = parse_revision_argument(0, $r);
2196 if ($base <= $r) {
2197 $gs->fetch($base, $r);
2198 }
2199 ($r0, $parent) = $gs->last_rev_commit;
2200 }
2201 if (defined $r0 && defined $parent) {
2202 print STDERR "Found branch parent: ($self->{ref_id}) $parent\n";
2203 my $ed;
2204 if ($self->ra->can_do_switch) {
2205 $self->assert_index_clean($parent);
2206 print STDERR "Following parent with do_switch\n";
2207 # do_switch works with svn/trunk >= r22312, but that
2208 # is not included with SVN 1.4.3 (the latest version
2209 # at the moment), so we can't rely on it
2210 $self->{last_commit} = $parent;
2211 $ed = SVN::Git::Fetcher->new($self);
2212 $gs->ra->gs_do_switch($r0, $rev, $gs,
2213 $self->full_url, $ed)
2214 or die "SVN connection failed somewhere...\n";
2215 } elsif ($self->ra->trees_match($new_url, $r0,
2216 $self->full_url, $rev)) {
2217 print STDERR "Trees match:\n",
2218 " $new_url\@$r0\n",
2219 " ${\$self->full_url}\@$rev\n",
2220 "Following parent with no changes\n";
2221 $self->tmp_index_do(sub {
2222 command_noisy('read-tree', $parent);
2223 });
2224 $self->{last_commit} = $parent;
2225 } else {
2226 print STDERR "Following parent with do_update\n";
2227 $ed = SVN::Git::Fetcher->new($self);
2228 $self->ra->gs_do_update($rev, $rev, $self, $ed)
2229 or die "SVN connection failed somewhere...\n";
2230 }
2231 print STDERR "Successfully followed parent\n";
2232 return $self->make_log_entry($rev, [$parent], $ed);
2233 }
2234 return undef;
2235 }
2236
2237 sub do_fetch {
2238 my ($self, $paths, $rev) = @_;
2239 my $ed;
2240 my ($last_rev, @parents);
2241 if (my $lc = $self->last_commit) {
2242 # we can have a branch that was deleted, then re-added
2243 # under the same name but copied from another path, in
2244 # which case we'll have multiple parents (we don't
2245 # want to break the original ref, nor lose copypath info):
2246 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2247 push @{$log_entry->{parents}}, $lc;
2248 return $log_entry;
2249 }
2250 $ed = SVN::Git::Fetcher->new($self);
2251 $last_rev = $self->{last_rev};
2252 $ed->{c} = $lc;
2253 @parents = ($lc);
2254 } else {
2255 $last_rev = $rev;
2256 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2257 return $log_entry;
2258 }
2259 $ed = SVN::Git::Fetcher->new($self);
2260 }
2261 unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2262 die "SVN connection failed somewhere...\n";
2263 }
2264 $self->make_log_entry($rev, \@parents, $ed);
2265 }
2266
2267 sub get_untracked {
2268 my ($self, $ed) = @_;
2269 my @out;
2270 my $h = $ed->{empty};
2271 foreach (sort keys %$h) {
2272 my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2273 push @out, " $act: " . uri_encode($_);
2274 warn "W: $act: $_\n";
2275 }
2276 foreach my $t (qw/dir_prop file_prop/) {
2277 $h = $ed->{$t} or next;
2278 foreach my $path (sort keys %$h) {
2279 my $ppath = $path eq '' ? '.' : $path;
2280 foreach my $prop (sort keys %{$h->{$path}}) {
2281 next if $SKIP_PROP{$prop};
2282 my $v = $h->{$path}->{$prop};
2283 my $t_ppath_prop = "$t: " .
2284 uri_encode($ppath) . ' ' .
2285 uri_encode($prop);
2286 if (defined $v) {
2287 push @out, " +$t_ppath_prop " .
2288 uri_encode($v);
2289 } else {
2290 push @out, " -$t_ppath_prop";
2291 }
2292 }
2293 }
2294 }
2295 foreach my $t (qw/absent_file absent_directory/) {
2296 $h = $ed->{$t} or next;
2297 foreach my $parent (sort keys %$h) {
2298 foreach my $path (sort @{$h->{$parent}}) {
2299 push @out, " $t: " .
2300 uri_encode("$parent/$path");
2301 warn "W: $t: $parent/$path ",
2302 "Insufficient permissions?\n";
2303 }
2304 }
2305 }
2306 \@out;
2307 }
2308
2309 sub parse_svn_date {
2310 my $date = shift || return '+0000 1970-01-01 00:00:00';
2311 my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2312 (\d\d)\:(\d\d)\:(\d\d).\d+Z$/x) or
2313 croak "Unable to parse date: $date\n";
2314 "+0000 $Y-$m-$d $H:$M:$S";
2315 }
2316
2317 sub check_author {
2318 my ($author) = @_;
2319 if (!defined $author || length $author == 0) {
2320 $author = '(no author)';
2321 }
2322 if (defined $::_authors && ! defined $::users{$author}) {
2323 die "Author: $author not defined in $::_authors file\n";
2324 }
2325 $author;
2326 }
2327
2328 sub make_log_entry {
2329 my ($self, $rev, $parents, $ed) = @_;
2330 my $untracked = $self->get_untracked($ed);
2331
2332 open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
2333 print $un "r$rev\n" or croak $!;
2334 print $un $_, "\n" foreach @$untracked;
2335 my %log_entry = ( parents => $parents || [], revision => $rev,
2336 log => '');
2337
2338 my $headrev;
2339 my $logged = delete $self->{logged_rev_props};
2340 if (!$logged || $self->{-want_revprops}) {
2341 my $rp = $self->ra->rev_proplist($rev);
2342 foreach (sort keys %$rp) {
2343 my $v = $rp->{$_};
2344 if (/^svn:(author|date|log)$/) {
2345 $log_entry{$1} = $v;
2346 } elsif ($_ eq 'svm:headrev') {
2347 $headrev = $v;
2348 } else {
2349 print $un " rev_prop: ", uri_encode($_), ' ',
2350 uri_encode($v), "\n";
2351 }
2352 }
2353 } else {
2354 map { $log_entry{$_} = $logged->{$_} } keys %$logged;
2355 }
2356 close $un or croak $!;
2357
2358 $log_entry{date} = parse_svn_date($log_entry{date});
2359 $log_entry{log} .= "\n";
2360 my $author = $log_entry{author} = check_author($log_entry{author});
2361 my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
2362 : ($author, undef);
2363
2364 my ($commit_name, $commit_email) = ($name, $email);
2365 if ($_use_log_author) {
2366 if ($log_entry{log} =~ /From:\s+(.*?)\s+<(.*)>\s*\n/) {
2367 ($name, $email) = ($1, $2);
2368 } elsif ($log_entry{log} =~
2369 /Signed-off-by:\s+(.*?)\s+<(.*)>\s*\n/) {
2370 ($name, $email) = ($1, $2);
2371 }
2372 }
2373 if (defined $headrev && $self->use_svm_props) {
2374 if ($self->rewrite_root) {
2375 die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
2376 "options set!\n";
2377 }
2378 my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$};
2379 # we don't want "SVM: initializing mirror for junk" ...
2380 return undef if $r == 0;
2381 my $svm = $self->svm;
2382 if ($uuid ne $svm->{uuid}) {
2383 die "UUID mismatch on SVM path:\n",
2384 "expected: $svm->{uuid}\n",
2385 " got: $uuid\n";
2386 }
2387 my $full_url = $self->full_url;
2388 $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
2389 die "Failed to replace '$svm->{replace}' with ",
2390 "'$svm->{source}' in $full_url\n";
2391 # throw away username for storing in records
2392 remove_username($full_url);
2393 $log_entry{metadata} = "$full_url\@$r $uuid";
2394 $log_entry{svm_revision} = $r;
2395 $email ||= "$author\@$uuid";
2396 $commit_email ||= "$author\@$uuid";
2397 } elsif ($self->use_svnsync_props) {
2398 my $full_url = $self->svnsync->{url};
2399 $full_url .= "/$self->{path}" if length $self->{path};
2400 remove_username($full_url);
2401 my $uuid = $self->svnsync->{uuid};
2402 $log_entry{metadata} = "$full_url\@$rev $uuid";
2403 $email ||= "$author\@$uuid";
2404 $commit_email ||= "$author\@$uuid";
2405 } else {
2406 my $url = $self->metadata_url;
2407 remove_username($url);
2408 $log_entry{metadata} = "$url\@$rev " .
2409 $self->ra->get_uuid;
2410 $email ||= "$author\@" . $self->ra->get_uuid;
2411 $commit_email ||= "$author\@" . $self->ra->get_uuid;
2412 }
2413 $log_entry{name} = $name;
2414 $log_entry{email} = $email;
2415 $log_entry{commit_name} = $commit_name;
2416 $log_entry{commit_email} = $commit_email;
2417 \%log_entry;
2418 }
2419
2420 sub fetch {
2421 my ($self, $min_rev, $max_rev, @parents) = @_;
2422 my ($last_rev, $last_commit) = $self->last_rev_commit;
2423 my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
2424 $self->ra->gs_fetch_loop_common($base, $head, [$self]);
2425 }
2426
2427 sub set_tree_cb {
2428 my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
2429 $self->{inject_parents} = { $rev => $tree };
2430 $self->fetch(undef, undef);
2431 }
2432
2433 sub set_tree {
2434 my ($self, $tree) = (shift, shift);
2435 my $log_entry = ::get_commit_entry($tree);
2436 unless ($self->{last_rev}) {
2437 fatal("Must have an existing revision to commit");
2438 }
2439 my %ed_opts = ( r => $self->{last_rev},
2440 log => $log_entry->{log},
2441 ra => $self->ra,
2442 tree_a => $self->{last_commit},
2443 tree_b => $tree,
2444 editor_cb => sub {
2445 $self->set_tree_cb($log_entry, $tree, @_) },
2446 svn_path => $self->{path} );
2447 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
2448 print "No changes\nr$self->{last_rev} = $tree\n";
2449 }
2450 }
2451
2452 sub rebuild_from_rev_db {
2453 my ($self, $path) = @_;
2454 my $r = -1;
2455 open my $fh, '<', $path or croak "open: $!";
2456 while (<$fh>) {
2457 length($_) == 41 or croak "inconsistent size in ($_) != 41";
2458 chomp($_);
2459 ++$r;
2460 next if $_ eq ('0' x 40);
2461 $self->rev_map_set($r, $_);
2462 print "r$r = $_\n";
2463 }
2464 close $fh or croak "close: $!";
2465 unlink $path or croak "unlink: $!";
2466 }
2467
2468 sub rebuild {
2469 my ($self) = @_;
2470 my $map_path = $self->map_path;
2471 return if (-e $map_path && ! -z $map_path);
2472 return unless ::verify_ref($self->refname.'^0');
2473 if ($self->use_svm_props || $self->no_metadata) {
2474 my $rev_db = $self->rev_db_path;
2475 $self->rebuild_from_rev_db($rev_db);
2476 if ($self->use_svm_props) {
2477 my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
2478 $self->rebuild_from_rev_db($svm_rev_db);
2479 }
2480 $self->unlink_rev_db_symlink;
2481 return;
2482 }
2483 print "Rebuilding $map_path ...\n";
2484 my ($log, $ctx) =
2485 command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
2486 $self->refname, '--');
2487 my $full_url = $self->full_url;
2488 remove_username($full_url);
2489 my $svn_uuid = $self->ra_uuid;
2490 my $c;
2491 while (<$log>) {
2492 if ( m{^commit ($::sha1)$} ) {
2493 $c = $1;
2494 next;
2495 }
2496 next unless s{^\s*(git-svn-id:)}{$1};
2497 my ($url, $rev, $uuid) = ::extract_metadata($_);
2498 remove_username($url);
2499
2500 # ignore merges (from set-tree)
2501 next if (!defined $rev || !$uuid);
2502
2503 # if we merged or otherwise started elsewhere, this is
2504 # how we break out of it
2505 if (($uuid ne $svn_uuid) ||
2506 ($full_url && $url && ($url ne $full_url))) {
2507 next;
2508 }
2509
2510 $self->rev_map_set($rev, $c);
2511 print "r$rev = $c\n";
2512 }
2513 command_close_pipe($log, $ctx);
2514 print "Done rebuilding $map_path\n";
2515 my $rev_db_path = $self->rev_db_path;
2516 if (-f $self->rev_db_path) {
2517 unlink $self->rev_db_path or croak "unlink: $!";
2518 }
2519 $self->unlink_rev_db_symlink;
2520 }
2521
2522 # rev_map:
2523 # Tie::File seems to be prone to offset errors if revisions get sparse,
2524 # it's not that fast, either. Tie::File is also not in Perl 5.6. So
2525 # one of my favorite modules is out :< Next up would be one of the DBM
2526 # modules, but I'm not sure which is most portable...
2527 #
2528 # This is the replacement for the rev_db format, which was too big
2529 # and inefficient for large repositories with a lot of sparse history
2530 # (mainly tags)
2531 #
2532 # The format is this:
2533 # - 24 bytes for every record,
2534 # * 4 bytes for the integer representing an SVN revision number
2535 # * 20 bytes representing the sha1 of a git commit
2536 # - No empty padding records like the old format
2537 # (except the last record, which can be overwritten)
2538 # - new records are written append-only since SVN revision numbers
2539 # increase monotonically
2540 # - lookups on SVN revision number are done via a binary search
2541 # - Piping the file to xxd -c24 is a good way of dumping it for
2542 # viewing or editing (piped back through xxd -r), should the need
2543 # ever arise.
2544 # - The last record can be padding revision with an all-zero sha1
2545 # This is used to optimize fetch performance when using multiple
2546 # "fetch" directives in .git/config
2547 #
2548 # These files are disposable unless noMetadata or useSvmProps is set
2549
2550 sub _rev_map_set {
2551 my ($fh, $rev, $commit) = @_;
2552
2553 my $size = (stat($fh))[7];
2554 ($size % 24) == 0 or croak "inconsistent size: $size";
2555
2556 my $wr_offset = 0;
2557 if ($size > 0) {
2558 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2559 my $read = sysread($fh, my $buf, 24) or croak "read: $!";
2560 $read == 24 or croak "read only $read bytes (!= 24)";
2561 my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
2562 if ($last_commit eq ('0' x40)) {
2563 if ($size >= 48) {
2564 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2565 $read = sysread($fh, $buf, 24) or
2566 croak "read: $!";
2567 $read == 24 or
2568 croak "read only $read bytes (!= 24)";
2569 ($last_rev, $last_commit) =
2570 unpack(rev_map_fmt, $buf);
2571 if ($last_commit eq ('0' x40)) {
2572 croak "inconsistent .rev_map\n";
2573 }
2574 }
2575 if ($last_rev >= $rev) {
2576 croak "last_rev is higher!: $last_rev >= $rev";
2577 }
2578 $wr_offset = -24;
2579 }
2580 }
2581 sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
2582 syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
2583 croak "write: $!";
2584 }
2585
2586 sub mkfile {
2587 my ($path) = @_;
2588 unless (-e $path) {
2589 my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
2590 mkpath([$dir]) unless -d $dir;
2591 open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
2592 close $fh or die "Couldn't close (create) $path: $!\n";
2593 }
2594 }
2595
2596 sub rev_map_set {
2597 my ($self, $rev, $commit, $update_ref, $uuid) = @_;
2598 length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
2599 my $db = $self->map_path($uuid);
2600 my $db_lock = "$db.lock";
2601 my $sig;
2602 if ($update_ref) {
2603 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2604 $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
2605 }
2606 mkfile($db);
2607
2608 $LOCKFILES{$db_lock} = 1;
2609 my $sync;
2610 # both of these options make our .rev_db file very, very important
2611 # and we can't afford to lose it because rebuild() won't work
2612 if ($self->use_svm_props || $self->no_metadata) {
2613 $sync = 1;
2614 copy($db, $db_lock) or die "rev_map_set(@_): ",
2615 "Failed to copy: ",
2616 "$db => $db_lock ($!)\n";
2617 } else {
2618 rename $db, $db_lock or die "rev_map_set(@_): ",
2619 "Failed to rename: ",
2620 "$db => $db_lock ($!)\n";
2621 }
2622
2623 sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
2624 or croak "Couldn't open $db_lock: $!\n";
2625 _rev_map_set($fh, $rev, $commit);
2626 if ($sync) {
2627 $fh->flush or die "Couldn't flush $db_lock: $!\n";
2628 $fh->sync or die "Couldn't sync $db_lock: $!\n";
2629 }
2630 close $fh or croak $!;
2631 if ($update_ref) {
2632 $_head = $self;
2633 command_noisy('update-ref', '-m', "r$rev",
2634 $self->refname, $commit);
2635 }
2636 rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
2637 "$db_lock => $db ($!)\n";
2638 delete $LOCKFILES{$db_lock};
2639 if ($update_ref) {
2640 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2641 $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
2642 kill $sig, $$ if defined $sig;
2643 }
2644 }
2645
2646 # If want_commit, this will return an array of (rev, commit) where
2647 # commit _must_ be a valid commit in the archive.
2648 # Otherwise, it'll return the max revision (whether or not the
2649 # commit is valid or just a 0x40 placeholder).
2650 sub rev_map_max {
2651 my ($self, $want_commit) = @_;
2652 $self->rebuild;
2653 my $map_path = $self->map_path;
2654 stat $map_path or return $want_commit ? (0, undef) : 0;
2655 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2656 my $size = (stat($fh))[7];
2657 ($size % 24) == 0 or croak "inconsistent size: $size";
2658
2659 if ($size == 0) {
2660 close $fh or croak "close: $!";
2661 return $want_commit ? (0, undef) : 0;
2662 }
2663
2664 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2665 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2666 my ($r, $c) = unpack(rev_map_fmt, $buf);
2667 if ($want_commit && $c eq ('0' x40)) {
2668 if ($size < 48) {
2669 return $want_commit ? (0, undef) : 0;
2670 }
2671 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2672 sysread($fh, $buf, 24) == 24 or croak "read: $!";
2673 ($r, $c) = unpack(rev_map_fmt, $buf);
2674 if ($c eq ('0'x40)) {
2675 croak "Penultimate record is all-zeroes in $map_path";
2676 }
2677 }
2678 close $fh or croak "close: $!";
2679 $want_commit ? ($r, $c) : $r;
2680 }
2681
2682 sub rev_map_get {
2683 my ($self, $rev, $uuid) = @_;
2684 my $map_path = $self->map_path($uuid);
2685 return undef unless -e $map_path;
2686
2687 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2688 my $size = (stat($fh))[7];
2689 ($size % 24) == 0 or croak "inconsistent size: $size";
2690
2691 if ($size == 0) {
2692 close $fh or croak "close: $fh";
2693 return undef;
2694 }
2695
2696 my ($l, $u) = (0, $size - 24);
2697 my ($r, $c, $buf);
2698
2699 while ($l <= $u) {
2700 my $i = int(($l/24 + $u/24) / 2) * 24;
2701 sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
2702 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2703 my ($r, $c) = unpack('NH40', $buf);
2704
2705 if ($r < $rev) {
2706 $l = $i + 24;
2707 } elsif ($r > $rev) {
2708 $u = $i - 24;
2709 } else { # $r == $rev
2710 close($fh) or croak "close: $!";
2711 return $c eq ('0' x 40) ? undef : $c;
2712 }
2713 }
2714 close($fh) or croak "close: $!";
2715 undef;
2716 }
2717
2718 # Finds the first svn revision that exists on (if $eq_ok is true) or
2719 # before $rev for the current branch. It will not search any lower
2720 # than $min_rev. Returns the git commit hash and svn revision number
2721 # if found, else (undef, undef).
2722 sub find_rev_before {
2723 my ($self, $rev, $eq_ok, $min_rev) = @_;
2724 --$rev unless $eq_ok;
2725 $min_rev ||= 1;
2726 while ($rev >= $min_rev) {
2727 if (my $c = $self->rev_map_get($rev)) {
2728 return ($rev, $c);
2729 }
2730 --$rev;
2731 }
2732 return (undef, undef);
2733 }
2734
2735 # Finds the first svn revision that exists on (if $eq_ok is true) or
2736 # after $rev for the current branch. It will not search any higher
2737 # than $max_rev. Returns the git commit hash and svn revision number
2738 # if found, else (undef, undef).
2739 sub find_rev_after {
2740 my ($self, $rev, $eq_ok, $max_rev) = @_;
2741 ++$rev unless $eq_ok;
2742 $max_rev ||= $self->rev_map_max;
2743 while ($rev <= $max_rev) {
2744 if (my $c = $self->rev_map_get($rev)) {
2745 return ($rev, $c);
2746 }
2747 ++$rev;
2748 }
2749 return (undef, undef);
2750 }
2751
2752 sub _new {
2753 my ($class, $repo_id, $ref_id, $path) = @_;
2754 unless (defined $repo_id && length $repo_id) {
2755 $repo_id = $Git::SVN::default_repo_id;
2756 }
2757 unless (defined $ref_id && length $ref_id) {
2758 $_[2] = $ref_id = $Git::SVN::default_ref_id;
2759 }
2760 $_[1] = $repo_id = sanitize_remote_name($repo_id);
2761 my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
2762 $_[3] = $path = '' unless (defined $path);
2763 mkpath(["$ENV{GIT_DIR}/svn"]);
2764 bless {
2765 ref_id => $ref_id, dir => $dir, index => "$dir/index",
2766 path => $path, config => "$ENV{GIT_DIR}/svn/config",
2767 map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
2768 }
2769
2770 # for read-only access of old .rev_db formats
2771 sub unlink_rev_db_symlink {
2772 my ($self) = @_;
2773 my $link = $self->rev_db_path;
2774 $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
2775 if (-l $link) {
2776 unlink $link or croak "unlink: $link failed!";
2777 }
2778 }
2779
2780 sub rev_db_path {
2781 my ($self, $uuid) = @_;
2782 my $db_path = $self->map_path($uuid);
2783 $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
2784 or croak "map_path: $db_path does not contain '/.rev_map.' !";
2785 $db_path;
2786 }
2787
2788 # the new replacement for .rev_db
2789 sub map_path {
2790 my ($self, $uuid) = @_;
2791 $uuid ||= $self->ra_uuid;
2792 "$self->{map_root}.$uuid";
2793 }
2794
2795 sub uri_encode {
2796 my ($f) = @_;
2797 $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
2798 $f
2799 }
2800
2801 sub remove_username {
2802 $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
2803 }
2804
2805 package Git::SVN::Prompt;
2806 use strict;
2807 use warnings;
2808 require SVN::Core;
2809 use vars qw/$_no_auth_cache $_username/;
2810
2811 sub simple {
2812 my ($cred, $realm, $default_username, $may_save, $pool) = @_;
2813 $may_save = undef if $_no_auth_cache;
2814 $default_username = $_username if defined $_username;
2815 if (defined $default_username && length $default_username) {
2816 if (defined $realm && length $realm) {
2817 print STDERR "Authentication realm: $realm\n";
2818 STDERR->flush;
2819 }
2820 $cred->username($default_username);
2821 } else {
2822 username($cred, $realm, $may_save, $pool);
2823 }
2824 $cred->password(_read_password("Password for '" .
2825 $cred->username . "': ", $realm));
2826 $cred->may_save($may_save);
2827 $SVN::_Core::SVN_NO_ERROR;
2828 }
2829
2830 sub ssl_server_trust {
2831 my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
2832 $may_save = undef if $_no_auth_cache;
2833 print STDERR "Error validating server certificate for '$realm':\n";
2834 {
2835 no warnings 'once';
2836 # All variables SVN::Auth::SSL::* are used only once,
2837 # so we're shutting up Perl warnings about this.
2838 if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
2839 print STDERR " - The certificate is not issued ",
2840 "by a trusted authority. Use the\n",
2841 " fingerprint to validate ",
2842 "the certificate manually!\n";
2843 }
2844 if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
2845 print STDERR " - The certificate hostname ",
2846 "does not match.\n";
2847 }
2848 if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
2849 print STDERR " - The certificate is not yet valid.\n";
2850 }
2851 if ($failures & $SVN::Auth::SSL::EXPIRED) {
2852 print STDERR " - The certificate has expired.\n";
2853 }
2854 if ($failures & $SVN::Auth::SSL::OTHER) {
2855 print STDERR " - The certificate has ",
2856 "an unknown error.\n";
2857 }
2858 } # no warnings 'once'
2859 printf STDERR
2860 "Certificate information:\n".
2861 " - Hostname: %s\n".
2862 " - Valid: from %s until %s\n".
2863 " - Issuer: %s\n".
2864 " - Fingerprint: %s\n",
2865 map $cert_info->$_, qw(hostname valid_from valid_until
2866 issuer_dname fingerprint);
2867 my $choice;
2868 prompt:
2869 print STDERR $may_save ?
2870 "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
2871 "(R)eject or accept (t)emporarily? ";
2872 STDERR->flush;
2873 $choice = lc(substr(<STDIN> || 'R', 0, 1));
2874 if ($choice =~ /^t$/i) {
2875 $cred->may_save(undef);
2876 } elsif ($choice =~ /^r$/i) {
2877 return -1;
2878 } elsif ($may_save && $choice =~ /^p$/i) {
2879 $cred->may_save($may_save);
2880 } else {
2881 goto prompt;
2882 }
2883 $cred->accepted_failures($failures);
2884 $SVN::_Core::SVN_NO_ERROR;
2885 }
2886
2887 sub ssl_client_cert {
2888 my ($cred, $realm, $may_save, $pool) = @_;
2889 $may_save = undef if $_no_auth_cache;
2890 print STDERR "Client certificate filename: ";
2891 STDERR->flush;
2892 chomp(my $filename = <STDIN>);
2893 $cred->cert_file($filename);
2894 $cred->may_save($may_save);
2895 $SVN::_Core::SVN_NO_ERROR;
2896 }
2897
2898 sub ssl_client_cert_pw {
2899 my ($cred, $realm, $may_save, $pool) = @_;
2900 $may_save = undef if $_no_auth_cache;
2901 $cred->password(_read_password("Password: ", $realm));
2902 $cred->may_save($may_save);
2903 $SVN::_Core::SVN_NO_ERROR;
2904 }
2905
2906 sub username {
2907 my ($cred, $realm, $may_save, $pool) = @_;
2908 $may_save = undef if $_no_auth_cache;
2909 if (defined $realm && length $realm) {
2910 print STDERR "Authentication realm: $realm\n";
2911 }
2912 my $username;
2913 if (defined $_username) {
2914 $username = $_username;
2915 } else {
2916 print STDERR "Username: ";
2917 STDERR->flush;
2918 chomp($username = <STDIN>);
2919 }
2920 $cred->username($username);
2921 $cred->may_save($may_save);
2922 $SVN::_Core::SVN_NO_ERROR;
2923 }
2924
2925 sub _read_password {
2926 my ($prompt, $realm) = @_;
2927 print STDERR $prompt;
2928 STDERR->flush;
2929 require Term::ReadKey;
2930 Term::ReadKey::ReadMode('noecho');
2931 my $password = '';
2932 while (defined(my $key = Term::ReadKey::ReadKey(0))) {
2933 last if $key =~ /[\012\015]/; # \n\r
2934 $password .= $key;
2935 }
2936 Term::ReadKey::ReadMode('restore');
2937 print STDERR "\n";
2938 STDERR->flush;
2939 $password;
2940 }
2941
2942 package SVN::Git::Fetcher;
2943 use vars qw/@ISA/;
2944 use strict;
2945 use warnings;
2946 use Carp qw/croak/;
2947 use IO::File qw//;
2948
2949 # file baton members: path, mode_a, mode_b, pool, fh, blob, base
2950 sub new {
2951 my ($class, $git_svn) = @_;
2952 my $self = SVN::Delta::Editor->new;
2953 bless $self, $class;
2954 $self->{c} = $git_svn->{last_commit} if exists $git_svn->{last_commit};
2955 $self->{empty} = {};
2956 $self->{dir_prop} = {};
2957 $self->{file_prop} = {};
2958 $self->{absent_dir} = {};
2959 $self->{absent_file} = {};
2960 $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
2961 $self;
2962 }
2963
2964 sub set_path_strip {
2965 my ($self, $path) = @_;
2966 $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
2967 }
2968
2969 sub open_root {
2970 { path => '' };
2971 }
2972
2973 sub open_directory {
2974 my ($self, $path, $pb, $rev) = @_;
2975 { path => $path };
2976 }
2977
2978 sub git_path {
2979 my ($self, $path) = @_;
2980 if ($self->{path_strip}) {
2981 $path =~ s!$self->{path_strip}!! or
2982 die "Failed to strip path '$path' ($self->{path_strip})\n";
2983 }
2984 $path;
2985 }
2986
2987 sub delete_entry {
2988 my ($self, $path, $rev, $pb) = @_;
2989
2990 my $gpath = $self->git_path($path);
2991 return undef if ($gpath eq '');
2992
2993 # remove entire directories.
2994 if (command('ls-tree', $self->{c}, '--', $gpath) =~ /^040000 tree/) {
2995 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
2996 -r --name-only -z/,
2997 $self->{c}, '--', $gpath);
2998 local $/ = "\0";
2999 while (<$ls>) {
3000 chomp;
3001 $self->{gii}->remove($_);
3002 print "\tD\t$_\n" unless $::_q;
3003 }
3004 print "\tD\t$gpath/\n" unless $::_q;
3005 command_close_pipe($ls, $ctx);
3006 $self->{empty}->{$path} = 0
3007 } else {
3008 $self->{gii}->remove($gpath);
3009 print "\tD\t$gpath\n" unless $::_q;
3010 }
3011 undef;
3012 }
3013
3014 sub open_file {
3015 my ($self, $path, $pb, $rev) = @_;
3016 my $gpath = $self->git_path($path);
3017 my ($mode, $blob) = (command('ls-tree', $self->{c}, '--', $gpath)
3018 =~ /^(\d{6}) blob ([a-f\d]{40})\t/);
3019 unless (defined $mode && defined $blob) {
3020 die "$path was not found in commit $self->{c} (r$rev)\n";
3021 }
3022 { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3023 pool => SVN::Pool->new, action => 'M' };
3024 }
3025
3026 sub add_file {
3027 my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3028 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3029 delete $self->{empty}->{$dir};
3030 { path => $path, mode_a => 100644, mode_b => 100644,
3031 pool => SVN::Pool->new, action => 'A' };
3032 }
3033
3034 sub add_directory {
3035 my ($self, $path, $cp_path, $cp_rev) = @_;
3036 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3037 delete $self->{empty}->{$dir};
3038 $self->{empty}->{$path} = 1;
3039 { path => $path };
3040 }
3041
3042 sub change_dir_prop {
3043 my ($self, $db, $prop, $value) = @_;
3044 $self->{dir_prop}->{$db->{path}} ||= {};
3045 $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3046 undef;
3047 }
3048
3049 sub absent_directory {
3050 my ($self, $path, $pb) = @_;
3051 $self->{absent_dir}->{$pb->{path}} ||= [];
3052 push @{$self->{absent_dir}->{$pb->{path}}}, $path;
3053 undef;
3054 }
3055
3056 sub absent_file {
3057 my ($self, $path, $pb) = @_;
3058 $self->{absent_file}->{$pb->{path}} ||= [];
3059 push @{$self->{absent_file}->{$pb->{path}}}, $path;
3060 undef;
3061 }
3062
3063 sub change_file_prop {
3064 my ($self, $fb, $prop, $value) = @_;
3065 if ($prop eq 'svn:executable') {
3066 if ($fb->{mode_b} != 120000) {
3067 $fb->{mode_b} = defined $value ? 100755 : 100644;
3068 }
3069 } elsif ($prop eq 'svn:special') {
3070 $fb->{mode_b} = defined $value ? 120000 : 100644;
3071 } else {
3072 $self->{file_prop}->{$fb->{path}} ||= {};
3073 $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
3074 }
3075 undef;
3076 }
3077
3078 sub apply_textdelta {
3079 my ($self, $fb, $exp) = @_;
3080 my $fh = IO::File->new_tmpfile;
3081 $fh->autoflush(1);
3082 # $fh gets auto-closed() by SVN::TxDelta::apply(),
3083 # (but $base does not,) so dup() it for reading in close_file
3084 open my $dup, '<&', $fh or croak $!;
3085 my $base = IO::File->new_tmpfile;
3086 $base->autoflush(1);
3087 if ($fb->{blob}) {
3088 defined (my $pid = fork) or croak $!;
3089 if (!$pid) {
3090 open STDOUT, '>&', $base or croak $!;
3091 print STDOUT 'link ' if ($fb->{mode_a} == 120000);
3092 exec qw/git-cat-file blob/, $fb->{blob} or croak $!;
3093 }
3094 waitpid $pid, 0;
3095 croak $? if $?;
3096
3097 if (defined $exp) {
3098 seek $base, 0, 0 or croak $!;
3099 my $got = ::md5sum($base);
3100 die "Checksum mismatch: $fb->{path} $fb->{blob}\n",
3101 "expected: $exp\n",
3102 " got: $got\n" if ($got ne $exp);
3103 }
3104 }
3105 seek $base, 0, 0 or croak $!;
3106 $fb->{fh} = $dup;
3107 $fb->{base} = $base;
3108 [ SVN::TxDelta::apply($base, $fh, undef, $fb->{path}, $fb->{pool}) ];
3109 }
3110
3111 sub close_file {
3112 my ($self, $fb, $exp) = @_;
3113 my $hash;
3114 my $path = $self->git_path($fb->{path});
3115 if (my $fh = $fb->{fh}) {
3116 if (defined $exp) {
3117 seek($fh, 0, 0) or croak $!;
3118 my $got = ::md5sum($fh);
3119 if ($got ne $exp) {
3120 die "Checksum mismatch: $path\n",
3121 "expected: $exp\n got: $got\n";
3122 }
3123 }
3124 sysseek($fh, 0, 0) or croak $!;
3125 if ($fb->{mode_b} == 120000) {
3126 sysread($fh, my $buf, 5) == 5 or croak $!;
3127 $buf eq 'link ' or die "$path has mode 120000",
3128 "but is not a link\n";
3129 }
3130 defined(my $pid = open my $out,'-|') or die "Can't fork: $!\n";
3131 if (!$pid) {
3132 open STDIN, '<&', $fh or croak $!;
3133 exec qw/git-hash-object -w --stdin/ or croak $!;
3134 }
3135 chomp($hash = do { local $/; <$out> });
3136 close $out or croak $!;
3137 close $fh or croak $!;
3138 $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
3139 close $fb->{base} or croak $!;
3140 } else {
3141 $hash = $fb->{blob} or die "no blob information\n";
3142 }
3143 $fb->{pool}->clear;
3144 $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
3145 print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
3146 undef;
3147 }
3148
3149 sub abort_edit {
3150 my $self = shift;
3151 $self->{nr} = $self->{gii}->{nr};
3152 delete $self->{gii};
3153 $self->SUPER::abort_edit(@_);
3154 }
3155
3156 sub close_edit {
3157 my $self = shift;
3158 $self->{git_commit_ok} = 1;
3159 $self->{nr} = $self->{gii}->{nr};
3160 delete $self->{gii};
3161 $self->SUPER::close_edit(@_);
3162 }
3163
3164 package SVN::Git::Editor;
3165 use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
3166 use strict;
3167 use warnings;
3168 use Carp qw/croak/;
3169 use IO::File;
3170
3171 sub new {
3172 my ($class, $opts) = @_;
3173 foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
3174 die "$_ required!\n" unless (defined $opts->{$_});
3175 }
3176
3177 my $pool = SVN::Pool->new;
3178 my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
3179 my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
3180 $opts->{r}, $mods);
3181
3182 # $opts->{ra} functions should not be used after this:
3183 my @ce = $opts->{ra}->get_commit_editor($opts->{log},
3184 $opts->{editor_cb}, $pool);
3185 my $self = SVN::Delta::Editor->new(@ce, $pool);
3186 bless $self, $class;
3187 foreach (qw/svn_path r tree_a tree_b/) {
3188 $self->{$_} = $opts->{$_};
3189 }
3190 $self->{url} = $opts->{ra}->{url};
3191 $self->{mods} = $mods;
3192 $self->{types} = $types;
3193 $self->{pool} = $pool;
3194 $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
3195 $self->{rm} = { };
3196 $self->{path_prefix} = length $self->{svn_path} ?
3197 "$self->{svn_path}/" : '';
3198 return $self;
3199 }
3200
3201 sub generate_diff {
3202 my ($tree_a, $tree_b) = @_;
3203 my @diff_tree = qw(diff-tree -z -r);
3204 if ($_cp_similarity) {
3205 push @diff_tree, "-C$_cp_similarity";
3206 } else {
3207 push @diff_tree, '-C';
3208 }
3209 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
3210 push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
3211 push @diff_tree, $tree_a, $tree_b;
3212 my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
3213 local $/ = "\0";
3214 my $state = 'meta';
3215 my @mods;
3216 while (<$diff_fh>) {
3217 chomp $_; # this gets rid of the trailing "\0"
3218 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
3219 $::sha1\s($::sha1)\s
3220 ([MTCRAD])\d*$/xo) {
3221 push @mods, { mode_a => $1, mode_b => $2,
3222 sha1_b => $3, chg => $4 };
3223 if ($4 =~ /^(?:C|R)$/) {
3224 $state = 'file_a';
3225 } else {
3226 $state = 'file_b';
3227 }
3228 } elsif ($state eq 'file_a') {
3229 my $x = $mods[$#mods] or croak "Empty array\n";
3230 if ($x->{chg} !~ /^(?:C|R)$/) {
3231 croak "Error parsing $_, $x->{chg}\n";
3232 }
3233 $x->{file_a} = $_;
3234 $state = 'file_b';
3235 } elsif ($state eq 'file_b') {
3236 my $x = $mods[$#mods] or croak "Empty array\n";
3237 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
3238 croak "Error parsing $_, $x->{chg}\n";
3239 }
3240 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
3241 croak "Error parsing $_, $x->{chg}\n";
3242 }
3243 $x->{file_b} = $_;
3244 $state = 'meta';
3245 } else {
3246 croak "Error parsing $_\n";
3247 }
3248 }
3249 command_close_pipe($diff_fh, $ctx);
3250 \@mods;
3251 }
3252
3253 sub check_diff_paths {
3254 my ($ra, $pfx, $rev, $mods) = @_;
3255 my %types;
3256 $pfx .= '/' if length $pfx;
3257
3258 sub type_diff_paths {
3259 my ($ra, $types, $path, $rev) = @_;
3260 my @p = split m#/+#, $path;
3261 my $c = shift @p;
3262 unless (defined $types->{$c}) {
3263 $types->{$c} = $ra->check_path($c, $rev);
3264 }
3265 while (@p) {
3266 $c .= '/' . shift @p;
3267 next if defined $types->{$c};
3268 $types->{$c} = $ra->check_path($c, $rev);
3269 }
3270 }
3271
3272 foreach my $m (@$mods) {
3273 foreach my $f (qw/file_a file_b/) {
3274 next unless defined $m->{$f};
3275 my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
3276 if (length $pfx.$dir && ! defined $types{$dir}) {
3277 type_diff_paths($ra, \%types, $pfx.$dir, $rev);
3278 }
3279 }
3280 }
3281 \%types;
3282 }
3283
3284 sub split_path {
3285 return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
3286 }
3287
3288 sub repo_path {
3289 my ($self, $path) = @_;
3290 $self->{path_prefix}.(defined $path ? $path : '');
3291 }
3292
3293 sub url_path {
3294 my ($self, $path) = @_;
3295 if ($self->{url} =~ m#^https?://#) {
3296 $path =~ s/([^a-zA-Z0-9_.-])/uc sprintf("%%%02x",ord($1))/eg;
3297 }
3298 $self->{url} . '/' . $self->repo_path($path);
3299 }
3300
3301 sub rmdirs {
3302 my ($self) = @_;
3303 my $rm = $self->{rm};
3304 delete $rm->{''}; # we never delete the url we're tracking
3305 return unless %$rm;
3306
3307 foreach (keys %$rm) {
3308 my @d = split m#/#, $_;
3309 my $c = shift @d;
3310 $rm->{$c} = 1;
3311 while (@d) {
3312 $c .= '/' . shift @d;
3313 $rm->{$c} = 1;
3314 }
3315 }
3316 delete $rm->{$self->{svn_path}};
3317 delete $rm->{''}; # we never delete the url we're tracking
3318 return unless %$rm;
3319
3320 my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
3321 $self->{tree_b});
3322 local $/ = "\0";
3323 while (<$fh>) {
3324 chomp;
3325 my @dn = split m#/#, $_;
3326 while (pop @dn) {
3327 delete $rm->{join '/', @dn};
3328 }
3329 unless (%$rm) {
3330 close $fh;
3331 return;
3332 }
3333 }
3334 command_close_pipe($fh, $ctx);
3335
3336 my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
3337 foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
3338 $self->close_directory($bat->{$d}, $p);
3339 my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
3340 print "\tD+\t$d/\n" unless $::_q;
3341 $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
3342 delete $bat->{$d};
3343 }
3344 }
3345
3346 sub open_or_add_dir {
3347 my ($self, $full_path, $baton) = @_;
3348 my $t = $self->{types}->{$full_path};
3349 if (!defined $t) {
3350 die "$full_path not known in r$self->{r} or we have a bug!\n";
3351 }
3352 {
3353 no warnings 'once';
3354 # SVN::Node::none and SVN::Node::file are used only once,
3355 # so we're shutting up Perl's warnings about them.
3356 if ($t == $SVN::Node::none) {
3357 return $self->add_directory($full_path, $baton,
3358 undef, -1, $self->{pool});
3359 } elsif ($t == $SVN::Node::dir) {
3360 return $self->open_directory($full_path, $baton,
3361 $self->{r}, $self->{pool});
3362 } # no warnings 'once'
3363 print STDERR "$full_path already exists in repository at ",
3364 "r$self->{r} and it is not a directory (",
3365 ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
3366 } # no warnings 'once'
3367 exit 1;
3368 }
3369
3370 sub ensure_path {
3371 my ($self, $path) = @_;
3372 my $bat = $self->{bat};
3373 my $repo_path = $self->repo_path($path);
3374 return $bat->{''} unless (length $repo_path);
3375 my @p = split m#/+#, $repo_path;
3376 my $c = shift @p;
3377 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
3378 while (@p) {
3379 my $c0 = $c;
3380 $c .= '/' . shift @p;
3381 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
3382 }
3383 return $bat->{$c};
3384 }
3385
3386 sub A {
3387 my ($self, $m) = @_;
3388 my ($dir, $file) = split_path($m->{file_b});
3389 my $pbat = $self->ensure_path($dir);
3390 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3391 undef, -1);
3392 print "\tA\t$m->{file_b}\n" unless $::_q;
3393 $self->chg_file($fbat, $m);
3394 $self->close_file($fbat,undef,$self->{pool});
3395 }
3396
3397 sub C {
3398 my ($self, $m) = @_;
3399 my ($dir, $file) = split_path($m->{file_b});
3400 my $pbat = $self->ensure_path($dir);
3401 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3402 $self->url_path($m->{file_a}), $self->{r});
3403 print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3404 $self->chg_file($fbat, $m);
3405 $self->close_file($fbat,undef,$self->{pool});
3406 }
3407
3408 sub delete_entry {
3409 my ($self, $path, $pbat) = @_;
3410 my $rpath = $self->repo_path($path);
3411 my ($dir, $file) = split_path($rpath);
3412 $self->{rm}->{$dir} = 1;
3413 $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
3414 }
3415
3416 sub R {
3417 my ($self, $m) = @_;
3418 my ($dir, $file) = split_path($m->{file_b});
3419 my $pbat = $self->ensure_path($dir);
3420 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3421 $self->url_path($m->{file_a}), $self->{r});
3422 print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3423 $self->chg_file($fbat, $m);
3424 $self->close_file($fbat,undef,$self->{pool});
3425
3426 ($dir, $file) = split_path($m->{file_a});
3427 $pbat = $self->ensure_path($dir);
3428 $self->delete_entry($m->{file_a}, $pbat);
3429 }
3430
3431 sub M {
3432 my ($self, $m) = @_;
3433 my ($dir, $file) = split_path($m->{file_b});
3434 my $pbat = $self->ensure_path($dir);
3435 my $fbat = $self->open_file($self->repo_path($m->{file_b}),
3436 $pbat,$self->{r},$self->{pool});
3437 print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
3438 $self->chg_file($fbat, $m);
3439 $self->close_file($fbat,undef,$self->{pool});
3440 }
3441
3442 sub T { shift->M(@_) }
3443
3444 sub change_file_prop {
3445 my ($self, $fbat, $pname, $pval) = @_;
3446 $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
3447 }
3448
3449 sub chg_file {
3450 my ($self, $fbat, $m) = @_;
3451 if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
3452 $self->change_file_prop($fbat,'svn:executable','*');
3453 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
3454 $self->change_file_prop($fbat,'svn:executable',undef);
3455 }
3456 my $fh = IO::File->new_tmpfile or croak $!;
3457 if ($m->{mode_b} =~ /^120/) {
3458 print $fh 'link ' or croak $!;
3459 $self->change_file_prop($fbat,'svn:special','*');
3460 } elsif ($m->{mode_a} =~ /^120/ && $m->{mode_b} !~ /^120/) {
3461 $self->change_file_prop($fbat,'svn:special',undef);
3462 }
3463 defined(my $pid = fork) or croak $!;
3464 if (!$pid) {
3465 open STDOUT, '>&', $fh or croak $!;
3466 exec qw/git-cat-file blob/, $m->{sha1_b} or croak $!;
3467 }
3468 waitpid $pid, 0;
3469 croak $? if $?;
3470 $fh->flush == 0 or croak $!;
3471 seek $fh, 0, 0 or croak $!;
3472
3473 my $exp = ::md5sum($fh);
3474 seek $fh, 0, 0 or croak $!;
3475
3476 my $pool = SVN::Pool->new;
3477 my $atd = $self->apply_textdelta($fbat, undef, $pool);
3478 my $got = SVN::TxDelta::send_stream($fh, @$atd, $pool);
3479 die "Checksum mismatch\nexpected: $exp\ngot: $got\n" if ($got ne $exp);
3480 $pool->clear;
3481
3482 close $fh or croak $!;
3483 }
3484
3485 sub D {
3486 my ($self, $m) = @_;
3487 my ($dir, $file) = split_path($m->{file_b});
3488 my $pbat = $self->ensure_path($dir);
3489 print "\tD\t$m->{file_b}\n" unless $::_q;
3490 $self->delete_entry($m->{file_b}, $pbat);
3491 }
3492
3493 sub close_edit {
3494 my ($self) = @_;
3495 my ($p,$bat) = ($self->{pool}, $self->{bat});
3496 foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
3497 next if $_ eq '';
3498 $self->close_directory($bat->{$_}, $p);
3499 }
3500 $self->close_directory($bat->{''}, $p);
3501 $self->SUPER::close_edit($p);
3502 $p->clear;
3503 }
3504
3505 sub abort_edit {
3506 my ($self) = @_;
3507 $self->SUPER::abort_edit($self->{pool});
3508 }
3509
3510 sub DESTROY {
3511 my $self = shift;
3512 $self->SUPER::DESTROY(@_);
3513 $self->{pool}->clear;
3514 }
3515
3516 # this drives the editor
3517 sub apply_diff {
3518 my ($self) = @_;
3519 my $mods = $self->{mods};
3520 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
3521 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
3522 my $f = $m->{chg};
3523 if (defined $o{$f}) {
3524 $self->$f($m);
3525 } else {
3526 fatal("Invalid change type: $f");
3527 }
3528 }
3529 $self->rmdirs if $_rmdir;
3530 if (@$mods == 0) {
3531 $self->abort_edit;
3532 } else {
3533 $self->close_edit;
3534 }
3535 return scalar @$mods;
3536 }
3537
3538 package Git::SVN::Ra;
3539 use vars qw/@ISA $config_dir $_log_window_size/;
3540 use strict;
3541 use warnings;
3542 my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
3543
3544 BEGIN {
3545 # enforce temporary pool usage for some simple functions
3546 no strict 'refs';
3547 for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root/) {
3548 my $SUPER = "SUPER::$f";
3549 *$f = sub {
3550 my $self = shift;
3551 my $pool = SVN::Pool->new;
3552 my @ret = $self->$SUPER(@_,$pool);
3553 $pool->clear;
3554 wantarray ? @ret : $ret[0];
3555 };
3556 }
3557 }
3558
3559 sub _auth_providers () {
3560 [
3561 SVN::Client::get_simple_provider(),
3562 SVN::Client::get_ssl_server_trust_file_provider(),
3563 SVN::Client::get_simple_prompt_provider(
3564 \&Git::SVN::Prompt::simple, 2),
3565 SVN::Client::get_ssl_client_cert_file_provider(),
3566 SVN::Client::get_ssl_client_cert_prompt_provider(
3567 \&Git::SVN::Prompt::ssl_client_cert, 2),
3568 SVN::Client::get_ssl_client_cert_pw_prompt_provider(
3569 \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
3570 SVN::Client::get_username_provider(),
3571 SVN::Client::get_ssl_server_trust_prompt_provider(
3572 \&Git::SVN::Prompt::ssl_server_trust),
3573 SVN::Client::get_username_prompt_provider(
3574 \&Git::SVN::Prompt::username, 2)
3575 ]
3576 }
3577
3578 sub escape_uri_only {
3579 my ($uri) = @_;
3580 my @tmp;
3581 foreach (split m{/}, $uri) {
3582 s/([^\w.-])/sprintf("%%%02X",ord($1))/eg;
3583 push @tmp, $_;
3584 }
3585 join('/', @tmp);
3586 }
3587
3588 sub escape_url {
3589 my ($url) = @_;
3590 if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
3591 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
3592 $url = "$scheme://$domain$uri";
3593 }
3594 $url;
3595 }
3596
3597 sub new {
3598 my ($class, $url) = @_;
3599 $url =~ s!/+$!!;
3600 return $RA if ($RA && $RA->{url} eq $url);
3601
3602 SVN::_Core::svn_config_ensure($config_dir, undef);
3603 my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
3604 my $config = SVN::Core::config_get_config($config_dir);
3605 $RA = undef;
3606 my $dont_store_passwords = 1;
3607 my $conf_t = ${$config}{'config'};
3608 {
3609 no warnings 'once';
3610 # The usage of $SVN::_Core::SVN_CONFIG_* variables
3611 # produces warnings that variables are used only once.
3612 # I had not found the better way to shut them up, so
3613 # the warnings of type 'once' are disabled in this block.
3614 if (SVN::_Core::svn_config_get_bool($conf_t,
3615 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3616 $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
3617 1) == 0) {
3618 SVN::_Core::svn_auth_set_parameter($baton,
3619 $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
3620 bless (\$dont_store_passwords, "_p_void"));
3621 }
3622 if (SVN::_Core::svn_config_get_bool($conf_t,
3623 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3624 $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
3625 1) == 0) {
3626 $Git::SVN::Prompt::_no_auth_cache = 1;
3627 }
3628 } # no warnings 'once'
3629 my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
3630 config => $config,
3631 pool => SVN::Pool->new,
3632 auth_provider_callbacks => $callbacks);
3633 $self->{url} = $url;
3634 $self->{svn_path} = $url;
3635 $self->{repos_root} = $self->get_repos_root;
3636 $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
3637 $self->{cache} = { check_path => { r => 0, data => {} },
3638 get_dir => { r => 0, data => {} } };
3639 $RA = bless $self, $class;
3640 }
3641
3642 sub check_path {
3643 my ($self, $path, $r) = @_;
3644 my $cache = $self->{cache}->{check_path};
3645 if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
3646 return $cache->{data}->{$path};
3647 }
3648 my $pool = SVN::Pool->new;
3649 my $t = $self->SUPER::check_path($path, $r, $pool);
3650 $pool->clear;
3651 if ($r != $cache->{r}) {
3652 %{$cache->{data}} = ();
3653 $cache->{r} = $r;
3654 }
3655 $cache->{data}->{$path} = $t;
3656 }
3657
3658 sub get_dir {
3659 my ($self, $dir, $r) = @_;
3660 my $cache = $self->{cache}->{get_dir};
3661 if ($r == $cache->{r}) {
3662 if (my $x = $cache->{data}->{$dir}) {
3663 return wantarray ? @$x : $x->[0];
3664 }
3665 }
3666 my $pool = SVN::Pool->new;
3667 my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
3668 my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
3669 $pool->clear;
3670 if ($r != $cache->{r}) {
3671 %{$cache->{data}} = ();
3672 $cache->{r} = $r;
3673 }
3674 $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
3675 wantarray ? (\%dirents, $r, $props) : \%dirents;
3676 }
3677
3678 sub DESTROY {
3679 # do not call the real DESTROY since we store ourselves in $RA
3680 }
3681
3682 sub get_log {
3683 my ($self, @args) = @_;
3684 my $pool = SVN::Pool->new;
3685 splice(@args, 3, 1) if ($SVN::Core::VERSION le '1.2.0');
3686 my $ret = $self->SUPER::get_log(@args, $pool);
3687 $pool->clear;
3688 $ret;
3689 }
3690
3691 sub trees_match {
3692 my ($self, $url1, $rev1, $url2, $rev2) = @_;
3693 my $ctx = SVN::Client->new(auth => _auth_providers);
3694 my $out = IO::File->new_tmpfile;
3695
3696 # older SVN (1.1.x) doesn't take $pool as the last parameter for
3697 # $ctx->diff(), so we'll create a default one
3698 my $pool = SVN::Pool->new_default_sub;
3699
3700 $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
3701 $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
3702 $out->flush;
3703 my $ret = (($out->stat)[7] == 0);
3704 close $out or croak $!;
3705
3706 $ret;
3707 }
3708
3709 sub get_commit_editor {
3710 my ($self, $log, $cb, $pool) = @_;
3711 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
3712 $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
3713 }
3714
3715 sub gs_do_update {
3716 my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
3717 my $new = ($rev_a == $rev_b);
3718 my $path = $gs->{path};
3719
3720 if ($new && -e $gs->{index}) {
3721 unlink $gs->{index} or die
3722 "Couldn't unlink index: $gs->{index}: $!\n";
3723 }
3724 my $pool = SVN::Pool->new;
3725 $editor->set_path_strip($path);
3726 my (@pc) = split m#/#, $path;
3727 my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
3728 1, $editor, $pool);
3729 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3730
3731 # Since we can't rely on svn_ra_reparent being available, we'll
3732 # just have to do some magic with set_path to make it so
3733 # we only want a partial path.
3734 my $sp = '';
3735 my $final = join('/', @pc);
3736 while (@pc) {
3737 $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
3738 $sp .= '/' if length $sp;
3739 $sp .= shift @pc;
3740 }
3741 die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
3742
3743 $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
3744
3745 $reporter->finish_report($pool);
3746 $pool->clear;
3747 $editor->{git_commit_ok};
3748 }
3749
3750 # this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
3751 # svn_ra_reparent didn't work before 1.4)
3752 sub gs_do_switch {
3753 my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
3754 my $path = $gs->{path};
3755 my $pool = SVN::Pool->new;
3756
3757 my $full_url = $self->{url};
3758 my $old_url = $full_url;
3759 $full_url .= '/' . escape_uri_only($path) if length $path;
3760 my ($ra, $reparented);
3761 if ($old_url ne $full_url) {
3762 if ($old_url !~ m#^svn(\+ssh)?://#) {
3763 SVN::_Ra::svn_ra_reparent($self->{session}, $full_url,
3764 $pool);
3765 $self->{url} = $full_url;
3766 $reparented = 1;
3767 } else {
3768 $_[0] = undef;
3769 $self = undef;
3770 $RA = undef;
3771 $ra = Git::SVN::Ra->new($full_url);
3772 $ra_invalid = 1;
3773 }
3774 }
3775 $ra ||= $self;
3776 my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
3777 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3778 $reporter->set_path('', $rev_a, 0, @lock, $pool);
3779 $reporter->finish_report($pool);
3780
3781 if ($reparented) {
3782 SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
3783 $self->{url} = $old_url;
3784 }
3785
3786 $pool->clear;
3787 $editor->{git_commit_ok};
3788 }
3789
3790 sub longest_common_path {
3791 my ($gsv, $globs) = @_;
3792 my %common;
3793 my $common_max = scalar @$gsv;
3794
3795 foreach my $gs (@$gsv) {
3796 my @tmp = split m#/#, $gs->{path};
3797 my $p = '';
3798 foreach (@tmp) {
3799 $p .= length($p) ? "/$_" : $_;
3800 $common{$p} ||= 0;
3801 $common{$p}++;
3802 }
3803 }
3804 $globs ||= [];
3805 $common_max += scalar @$globs;
3806 foreach my $glob (@$globs) {
3807 my @tmp = split m#/#, $glob->{path}->{left};
3808 my $p = '';
3809 foreach (@tmp) {
3810 $p .= length($p) ? "/$_" : $_;
3811 $common{$p} ||= 0;
3812 $common{$p}++;
3813 }
3814 }
3815
3816 my $longest_path = '';
3817 foreach (sort {length $b <=> length $a} keys %common) {
3818 if ($common{$_} == $common_max) {
3819 $longest_path = $_;
3820 last;
3821 }
3822 }
3823 $longest_path;
3824 }
3825
3826 sub gs_fetch_loop_common {
3827 my ($self, $base, $head, $gsv, $globs) = @_;
3828 return if ($base > $head);
3829 my $inc = $_log_window_size;
3830 my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
3831 my $longest_path = longest_common_path($gsv, $globs);
3832 my $ra_url = $self->{url};
3833 while (1) {
3834 my %revs;
3835 my $err;
3836 my $err_handler = $SVN::Error::handler;
3837 $SVN::Error::handler = sub {
3838 ($err) = @_;
3839 skip_unknown_revs($err);
3840 };
3841 sub _cb {
3842 my ($paths, $r, $author, $date, $log) = @_;
3843 [ dup_changed_paths($paths),
3844 { author => $author, date => $date, log => $log } ];
3845 }
3846 $self->get_log([$longest_path], $min, $max, 0, 1, 1,
3847 sub { $revs{$_[1]} = _cb(@_) });
3848 if ($err && $max >= $head) {
3849 print STDERR "Path '$longest_path' ",
3850 "was probably deleted:\n",
3851 $err->expanded_message,
3852 "\nWill attempt to follow ",
3853 "revisions r$min .. r$max ",
3854 "committed before the deletion\n";
3855 my $hi = $max;
3856 while (--$hi >= $min) {
3857 my $ok;
3858 $self->get_log([$longest_path], $min, $hi,
3859 0, 1, 1, sub {
3860 $ok ||= $_[1];
3861 $revs{$_[1]} = _cb(@_) });
3862 if ($ok) {
3863 print STDERR "r$min .. r$ok OK\n";
3864 last;
3865 }
3866 }
3867 }
3868 $SVN::Error::handler = $err_handler;
3869
3870 my %exists = map { $_->{path} => $_ } @$gsv;
3871 foreach my $r (sort {$a <=> $b} keys %revs) {
3872 my ($paths, $logged) = @{$revs{$r}};
3873
3874 foreach my $gs ($self->match_globs(\%exists, $paths,
3875 $globs, $r)) {
3876 if ($gs->rev_map_max >= $r) {
3877 next;
3878 }
3879 next unless $gs->match_paths($paths, $r);
3880 $gs->{logged_rev_props} = $logged;
3881 if (my $last_commit = $gs->last_commit) {
3882 $gs->assert_index_clean($last_commit);
3883 }
3884 my $log_entry = $gs->do_fetch($paths, $r);
3885 if ($log_entry) {
3886 $gs->do_git_commit($log_entry);
3887 }
3888 }
3889 foreach my $g (@$globs) {
3890 my $k = "svn-remote.$g->{remote}." .
3891 "$g->{t}-maxRev";
3892 Git::SVN::tmp_config($k, $r);
3893 }
3894 if ($ra_invalid) {
3895 $_[0] = undef;
3896 $self = undef;
3897 $RA = undef;
3898 $self = Git::SVN::Ra->new($ra_url);
3899 $ra_invalid = undef;
3900 }
3901 }
3902 # pre-fill the .rev_db since it'll eventually get filled in
3903 # with '0' x40 if something new gets committed
3904 foreach my $gs (@$gsv) {
3905 next if $gs->rev_map_max >= $max;
3906 next if defined $gs->rev_map_get($max);
3907 $gs->rev_map_set($max, 0 x40);
3908 }
3909 foreach my $g (@$globs) {
3910 my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
3911 Git::SVN::tmp_config($k, $max);
3912 }
3913 last if $max >= $head;
3914 $min = $max + 1;
3915 $max += $inc;
3916 $max = $head if ($max > $head);
3917 }
3918 }
3919
3920 sub match_globs {
3921 my ($self, $exists, $paths, $globs, $r) = @_;
3922
3923 sub get_dir_check {
3924 my ($self, $exists, $g, $r) = @_;
3925 my @x = eval { $self->get_dir($g->{path}->{left}, $r) };
3926 return unless scalar @x == 3;
3927 my $dirents = $x[0];
3928 foreach my $de (keys %$dirents) {
3929 next if $dirents->{$de}->{kind} != $SVN::Node::dir;
3930 my $p = $g->{path}->full_path($de);
3931 next if $exists->{$p};
3932 next if (length $g->{path}->{right} &&
3933 ($self->check_path($p, $r) !=
3934 $SVN::Node::dir));
3935 $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
3936 $g->{ref}->full_path($de), 1);
3937 }
3938 }
3939 foreach my $g (@$globs) {
3940 if (my $path = $paths->{"/$g->{path}->{left}"}) {
3941 if ($path->{action} =~ /^[AR]$/) {
3942 get_dir_check($self, $exists, $g, $r);
3943 }
3944 }
3945 foreach (keys %$paths) {
3946 if (/$g->{path}->{left_regex}/ &&
3947 !/$g->{path}->{regex}/) {
3948 next if $paths->{$_}->{action} !~ /^[AR]$/;
3949 get_dir_check($self, $exists, $g, $r);
3950 }
3951 next unless /$g->{path}->{regex}/;
3952 my $p = $1;
3953 my $pathname = $g->{path}->full_path($p);
3954 next if $exists->{$pathname};
3955 next if ($self->check_path($pathname, $r) !=
3956 $SVN::Node::dir);
3957 $exists->{$pathname} = Git::SVN->init(
3958 $self->{url}, $pathname, undef,
3959 $g->{ref}->full_path($p), 1);
3960 }
3961 my $c = '';
3962 foreach (split m#/#, $g->{path}->{left}) {
3963 $c .= "/$_";
3964 next unless ($paths->{$c} &&
3965 ($paths->{$c}->{action} =~ /^[AR]$/));
3966 get_dir_check($self, $exists, $g, $r);
3967 }
3968 }
3969 values %$exists;
3970 }
3971
3972 sub minimize_url {
3973 my ($self) = @_;
3974 return $self->{url} if ($self->{url} eq $self->{repos_root});
3975 my $url = $self->{repos_root};
3976 my @components = split(m!/!, $self->{svn_path});
3977 my $c = '';
3978 do {
3979 $url .= "/$c" if length $c;
3980 eval { (ref $self)->new($url)->get_latest_revnum };
3981 } while ($@ && ($c = shift @components));
3982 $url;
3983 }
3984
3985 sub can_do_switch {
3986 my $self = shift;
3987 unless (defined $can_do_switch) {
3988 my $pool = SVN::Pool->new;
3989 my $rep = eval {
3990 $self->do_switch(1, '', 0, $self->{url},
3991 SVN::Delta::Editor->new, $pool);
3992 };
3993 if ($@) {
3994 $can_do_switch = 0;
3995 } else {
3996 $rep->abort_report($pool);
3997 $can_do_switch = 1;
3998 }
3999 $pool->clear;
4000 }
4001 $can_do_switch;
4002 }
4003
4004 sub skip_unknown_revs {
4005 my ($err) = @_;
4006 my $errno = $err->apr_err();
4007 # Maybe the branch we're tracking didn't
4008 # exist when the repo started, so it's
4009 # not an error if it doesn't, just continue
4010 #
4011 # Wonderfully consistent library, eh?
4012 # 160013 - svn:// and file://
4013 # 175002 - http(s)://
4014 # 175007 - http(s):// (this repo required authorization, too...)
4015 # More codes may be discovered later...
4016 if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
4017 my $err_key = $err->expanded_message;
4018 # revision numbers change every time, filter them out
4019 $err_key =~ s/\d+/\0/g;
4020 $err_key = "$errno\0$err_key";
4021 unless ($ignored_err{$err_key}) {
4022 warn "W: Ignoring error from SVN, path probably ",
4023 "does not exist: ($errno): ",
4024 $err->expanded_message,"\n";
4025 $ignored_err{$err_key} = 1;
4026 }
4027 return;
4028 }
4029 die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
4030 }
4031
4032 # svn_log_changed_path_t objects passed to get_log are likely to be
4033 # overwritten even if only the refs are copied to an external variable,
4034 # so we should dup the structures in their entirety. Using an externally
4035 # passed pool (instead of our temporary and quickly cleared pool in
4036 # Git::SVN::Ra) does not help matters at all...
4037 sub dup_changed_paths {
4038 my ($paths) = @_;
4039 return undef unless $paths;
4040 my %ret;
4041 foreach my $p (keys %$paths) {
4042 my $i = $paths->{$p};
4043 my %s = map { $_ => $i->$_ }
4044 qw/copyfrom_path copyfrom_rev action/;
4045 $ret{$p} = \%s;
4046 }
4047 \%ret;
4048 }
4049
4050 package Git::SVN::Log;
4051 use strict;
4052 use warnings;
4053 use POSIX qw/strftime/;
4054 use constant commit_log_separator => ('-' x 72) . "\n";
4055 use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
4056 %rusers $show_commit $incremental/;
4057 my $l_fmt;
4058
4059 sub cmt_showable {
4060 my ($c) = @_;
4061 return 1 if defined $c->{r};
4062
4063 # big commit message got truncated by the 16k pretty buffer in rev-list
4064 if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
4065 $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
4066 @{$c->{l}} = ();
4067 my @log = command(qw/cat-file commit/, $c->{c});
4068
4069 # shift off the headers
4070 shift @log while ($log[0] ne '');
4071 shift @log;
4072
4073 # TODO: make $c->{l} not have a trailing newline in the future
4074 @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
4075
4076 (undef, $c->{r}, undef) = ::extract_metadata(
4077 (grep(/^git-svn-id: /, @log))[-1]);
4078 }
4079 return defined $c->{r};
4080 }
4081
4082 sub log_use_color {
4083 return $color || Git->repository->get_colorbool('color.diff');
4084 }
4085
4086 sub git_svn_log_cmd {
4087 my ($r_min, $r_max, @args) = @_;
4088 my $head = 'HEAD';
4089 my (@files, @log_opts);
4090 foreach my $x (@args) {
4091 if ($x eq '--' || @files) {
4092 push @files, $x;
4093 } else {
4094 if (::verify_ref("$x^0")) {
4095 $head = $x;
4096 } else {
4097 push @log_opts, $x;
4098 }
4099 }
4100 }
4101
4102 my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
4103 $gs ||= Git::SVN->_new;
4104 my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
4105 $gs->refname);
4106 push @cmd, '-r' unless $non_recursive;
4107 push @cmd, qw/--raw --name-status/ if $verbose;
4108 push @cmd, '--color' if log_use_color();
4109 push @cmd, @log_opts;
4110 if (defined $r_max && $r_max == $r_min) {
4111 push @cmd, '--max-count=1';
4112 if (my $c = $gs->rev_map_get($r_max)) {
4113 push @cmd, $c;
4114 }
4115 } elsif (defined $r_max) {
4116 if ($r_max < $r_min) {
4117 ($r_min, $r_max) = ($r_max, $r_min);
4118 }
4119 my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
4120 my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
4121 # If there are no commits in the range, both $c_max and $c_min
4122 # will be undefined. If there is at least 1 commit in the
4123 # range, both will be defined.
4124 return () if !defined $c_min || !defined $c_max;
4125 if ($c_min eq $c_max) {
4126 push @cmd, '--max-count=1', $c_min;
4127 } else {
4128 push @cmd, '--boundary', "$c_min..$c_max";
4129 }
4130 }
4131 return (@cmd, @files);
4132 }
4133
4134 # adapted from pager.c
4135 sub config_pager {
4136 $pager ||= $ENV{GIT_PAGER} || $ENV{PAGER};
4137 if (!defined $pager) {
4138 $pager = 'less';
4139 } elsif (length $pager == 0 || $pager eq 'cat') {
4140 $pager = undef;
4141 }
4142 $ENV{GIT_PAGER_IN_USE} = defined($pager);
4143 }
4144
4145 sub run_pager {
4146 return unless -t *STDOUT && defined $pager;
4147 pipe my $rfd, my $wfd or return;
4148 defined(my $pid = fork) or ::fatal "Can't fork: $!";
4149 if (!$pid) {
4150 open STDOUT, '>&', $wfd or
4151 ::fatal "Can't redirect to stdout: $!";
4152 return;
4153 }
4154 open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
4155 $ENV{LESS} ||= 'FRSX';
4156 exec $pager or ::fatal "Can't run pager: $! ($pager)";
4157 }
4158
4159 sub format_svn_date {
4160 return strftime("%Y-%m-%d %H:%M:%S %z (%a, %d %b %Y)", localtime(shift));
4161 }
4162
4163 sub parse_git_date {
4164 my ($t, $tz) = @_;
4165 # Date::Parse isn't in the standard Perl distro :(
4166 if ($tz =~ s/^\+//) {
4167 $t += tz_to_s_offset($tz);
4168 } elsif ($tz =~ s/^\-//) {
4169 $t -= tz_to_s_offset($tz);
4170 }
4171 return $t;
4172 }
4173
4174 sub set_local_timezone {
4175 if (defined $TZ) {
4176 $ENV{TZ} = $TZ;
4177 } else {
4178 delete $ENV{TZ};
4179 }
4180 }
4181
4182 sub tz_to_s_offset {
4183 my ($tz) = @_;
4184 $tz =~ s/(\d\d)$//;
4185 return ($1 * 60) + ($tz * 3600);
4186 }
4187
4188 sub get_author_info {
4189 my ($dest, $author, $t, $tz) = @_;
4190 $author =~ s/(?:^\s*|\s*$)//g;
4191 $dest->{a_raw} = $author;
4192 my $au;
4193 if ($::_authors) {
4194 $au = $rusers{$author} || undef;
4195 }
4196 if (!$au) {
4197 ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
4198 }
4199 $dest->{t} = $t;
4200 $dest->{tz} = $tz;
4201 $dest->{a} = $au;
4202 $dest->{t_utc} = parse_git_date($t, $tz);
4203 }
4204
4205 sub process_commit {
4206 my ($c, $r_min, $r_max, $defer) = @_;
4207 if (defined $r_min && defined $r_max) {
4208 if ($r_min == $c->{r} && $r_min == $r_max) {
4209 show_commit($c);
4210 return 0;
4211 }
4212 return 1 if $r_min == $r_max;
4213 if ($r_min < $r_max) {
4214 # we need to reverse the print order
4215 return 0 if (defined $limit && --$limit < 0);
4216 push @$defer, $c;
4217 return 1;
4218 }
4219 if ($r_min != $r_max) {
4220 return 1 if ($r_min < $c->{r});
4221 return 1 if ($r_max > $c->{r});
4222 }
4223 }
4224 return 0 if (defined $limit && --$limit < 0);
4225 show_commit($c);
4226 return 1;
4227 }
4228
4229 sub show_commit {
4230 my $c = shift;
4231 if ($oneline) {
4232 my $x = "\n";
4233 if (my $l = $c->{l}) {
4234 while ($l->[0] =~ /^\s*$/) { shift @$l }
4235 $x = $l->[0];
4236 }
4237 $l_fmt ||= 'A' . length($c->{r});
4238 print 'r',pack($l_fmt, $c->{r}),' | ';
4239 print "$c->{c} | " if $show_commit;
4240 print $x;
4241 } else {
4242 show_commit_normal($c);
4243 }
4244 }
4245
4246 sub show_commit_changed_paths {
4247 my ($c) = @_;
4248 return unless $c->{changed};
4249 print "Changed paths:\n", @{$c->{changed}};
4250 }
4251
4252 sub show_commit_normal {
4253 my ($c) = @_;
4254 print commit_log_separator, "r$c->{r} | ";
4255 print "$c->{c} | " if $show_commit;
4256 print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
4257 my $nr_line = 0;
4258
4259 if (my $l = $c->{l}) {
4260 while ($l->[$#$l] eq "\n" && $#$l > 0
4261 && $l->[($#$l - 1)] eq "\n") {
4262 pop @$l;
4263 }
4264 $nr_line = scalar @$l;
4265 if (!$nr_line) {
4266 print "1 line\n\n\n";
4267 } else {
4268 if ($nr_line == 1) {
4269 $nr_line = '1 line';
4270 } else {
4271 $nr_line .= ' lines';
4272 }
4273 print $nr_line, "\n";
4274 show_commit_changed_paths($c);
4275 print "\n";
4276 print $_ foreach @$l;
4277 }
4278 } else {
4279 print "1 line\n";
4280 show_commit_changed_paths($c);
4281 print "\n";
4282
4283 }
4284 foreach my $x (qw/raw stat diff/) {
4285 if ($c->{$x}) {
4286 print "\n";
4287 print $_ foreach @{$c->{$x}}
4288 }
4289 }
4290 }
4291
4292 sub cmd_show_log {
4293 my (@args) = @_;
4294 my ($r_min, $r_max);
4295 my $r_last = -1; # prevent dupes
4296 set_local_timezone();
4297 if (defined $::_revision) {
4298 if ($::_revision =~ /^(\d+):(\d+)$/) {
4299 ($r_min, $r_max) = ($1, $2);
4300 } elsif ($::_revision =~ /^\d+$/) {
4301 $r_min = $r_max = $::_revision;
4302 } else {
4303 ::fatal "-r$::_revision is not supported, use ",
4304 "standard 'git log' arguments instead";
4305 }
4306 }
4307
4308 config_pager();
4309 @args = git_svn_log_cmd($r_min, $r_max, @args);
4310 if (!@args) {
4311 print commit_log_separator unless $incremental || $oneline;
4312 return;
4313 }
4314 my $log = command_output_pipe(@args);
4315 run_pager();
4316 my (@k, $c, $d, $stat);
4317 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
4318 while (<$log>) {
4319 if (/^${esc_color}commit -?($::sha1_short)/o) {
4320 my $cmt = $1;
4321 if ($c && cmt_showable($c) && $c->{r} != $r_last) {
4322 $r_last = $c->{r};
4323 process_commit($c, $r_min, $r_max, \@k) or
4324 goto out;
4325 }
4326 $d = undef;
4327 $c = { c => $cmt };
4328 } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
4329 get_author_info($c, $1, $2, $3);
4330 } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
4331 # ignore
4332 } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
4333 push @{$c->{raw}}, $_;
4334 } elsif (/^${esc_color}[ACRMDT]\t/) {
4335 # we could add $SVN->{svn_path} here, but that requires
4336 # remote access at the moment (repo_path_split)...
4337 s#^(${esc_color})([ACRMDT])\t#$1 $2 #o;
4338 push @{$c->{changed}}, $_;
4339 } elsif (/^${esc_color}diff /o) {
4340 $d = 1;
4341 push @{$c->{diff}}, $_;
4342 } elsif ($d) {
4343 push @{$c->{diff}}, $_;
4344 } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
4345 $esc_color*[\+\-]*$esc_color$/x) {
4346 $stat = 1;
4347 push @{$c->{stat}}, $_;
4348 } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
4349 push @{$c->{stat}}, $_;
4350 $stat = undef;
4351 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
4352 ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
4353 } elsif (s/^${esc_color} //o) {
4354 push @{$c->{l}}, $_;
4355 }
4356 }
4357 if ($c && defined $c->{r} && $c->{r} != $r_last) {
4358 $r_last = $c->{r};
4359 process_commit($c, $r_min, $r_max, \@k);
4360 }
4361 if (@k) {
4362 ($r_min, $r_max) = ($r_max, $r_min);
4363 process_commit($_, $r_min, $r_max) foreach reverse @k;
4364 }
4365 out:
4366 close $log;
4367 print commit_log_separator unless $incremental || $oneline;
4368 }
4369
4370 package Git::SVN::Migration;
4371 # these version numbers do NOT correspond to actual version numbers
4372 # of git nor git-svn. They are just relative.
4373 #
4374 # v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
4375 #
4376 # v1 layout: .git/$id/info/url, refs/remotes/$id
4377 #
4378 # v2 layout: .git/svn/$id/info/url, refs/remotes/$id
4379 #
4380 # v3 layout: .git/svn/$id, refs/remotes/$id
4381 # - info/url may remain for backwards compatibility
4382 # - this is what we migrate up to this layout automatically,
4383 # - this will be used by git svn init on single branches
4384 # v3.1 layout (auto migrated):
4385 # - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
4386 # for backwards compatibility
4387 #
4388 # v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
4389 # - this is only created for newly multi-init-ed
4390 # repositories. Similar in spirit to the
4391 # --use-separate-remotes option in git-clone (now default)
4392 # - we do not automatically migrate to this (following
4393 # the example set by core git)
4394 #
4395 # v5 layout: .rev_db.$UUID => .rev_map.$UUID
4396 # - newer, more-efficient format that uses 24-bytes per record
4397 # with no filler space.
4398 # - use xxd -c24 < .rev_map.$UUID to view and debug
4399 # - This is a one-way migration, repositories updated to the
4400 # new format will not be able to use old git-svn without
4401 # rebuilding the .rev_db. Rebuilding the rev_db is not
4402 # possible if noMetadata or useSvmProps are set; but should
4403 # be no problem for users that use the (sensible) defaults.
4404 use strict;
4405 use warnings;
4406 use Carp qw/croak/;
4407 use File::Path qw/mkpath/;
4408 use File::Basename qw/dirname basename/;
4409 use vars qw/$_minimize/;
4410
4411 sub migrate_from_v0 {
4412 my $git_dir = $ENV{GIT_DIR};
4413 return undef unless -d $git_dir;
4414 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4415 my $migrated = 0;
4416 while (<$fh>) {
4417 chomp;
4418 my ($id, $orig_ref) = ($_, $_);
4419 next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
4420 next unless -f "$git_dir/$id/info/url";
4421 my $new_ref = "refs/remotes/$id";
4422 if (::verify_ref("$new_ref^0")) {
4423 print STDERR "W: $orig_ref is probably an old ",
4424 "branch used by an ancient version of ",
4425 "git-svn.\n",
4426 "However, $new_ref also exists.\n",
4427 "We will not be able ",
4428 "to use this branch until this ",
4429 "ambiguity is resolved.\n";
4430 next;
4431 }
4432 print STDERR "Migrating from v0 layout...\n" if !$migrated;
4433 print STDERR "Renaming ref: $orig_ref => $new_ref\n";
4434 command_noisy('update-ref', $new_ref, $orig_ref);
4435 command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
4436 $migrated++;
4437 }
4438 command_close_pipe($fh, $ctx);
4439 print STDERR "Done migrating from v0 layout...\n" if $migrated;
4440 $migrated;
4441 }
4442
4443 sub migrate_from_v1 {
4444 my $git_dir = $ENV{GIT_DIR};
4445 my $migrated = 0;
4446 return $migrated unless -d $git_dir;
4447 my $svn_dir = "$git_dir/svn";
4448
4449 # just in case somebody used 'svn' as their $id at some point...
4450 return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
4451
4452 print STDERR "Migrating from a git-svn v1 layout...\n";
4453 mkpath([$svn_dir]);
4454 print STDERR "Data from a previous version of git-svn exists, but\n\t",
4455 "$svn_dir\n\t(required for this version ",
4456 "($::VERSION) of git-svn) does not. exist\n";
4457 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4458 while (<$fh>) {
4459 my $x = $_;
4460 next unless $x =~ s#^refs/remotes/##;
4461 chomp $x;
4462 next unless -f "$git_dir/$x/info/url";
4463 my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
4464 next unless $u;
4465 my $dn = dirname("$git_dir/svn/$x");
4466 mkpath([$dn]) unless -d $dn;
4467 if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
4468 mkpath(["$git_dir/svn/svn"]);
4469 print STDERR " - $git_dir/$x/info => ",
4470 "$git_dir/svn/$x/info\n";
4471 rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
4472 croak "$!: $x";
4473 # don't worry too much about these, they probably
4474 # don't exist with repos this old (save for index,
4475 # and we can easily regenerate that)
4476 foreach my $f (qw/unhandled.log index .rev_db/) {
4477 rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
4478 }
4479 } else {
4480 print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
4481 rename "$git_dir/$x", "$git_dir/svn/$x" or
4482 croak "$!: $x";
4483 }
4484 $migrated++;
4485 }
4486 command_close_pipe($fh, $ctx);
4487 print STDERR "Done migrating from a git-svn v1 layout\n";
4488 $migrated;
4489 }
4490
4491 sub read_old_urls {
4492 my ($l_map, $pfx, $path) = @_;
4493 my @dir;
4494 foreach (<$path/*>) {
4495 if (-r "$_/info/url") {
4496 $pfx .= '/' if $pfx && $pfx !~ m!/$!;
4497 my $ref_id = $pfx . basename $_;
4498 my $url = ::file_to_s("$_/info/url");
4499 $l_map->{$ref_id} = $url;
4500 } elsif (-d $_) {
4501 push @dir, $_;
4502 }
4503 }
4504 foreach (@dir) {
4505 my $x = $_;
4506 $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
4507 read_old_urls($l_map, $x, $_);
4508 }
4509 }
4510
4511 sub migrate_from_v2 {
4512 my @cfg = command(qw/config -l/);
4513 return if grep /^svn-remote\..+\.url=/, @cfg;
4514 my %l_map;
4515 read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
4516 my $migrated = 0;
4517
4518 foreach my $ref_id (sort keys %l_map) {
4519 eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
4520 if ($@) {
4521 Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
4522 }
4523 $migrated++;
4524 }
4525 $migrated;
4526 }
4527
4528 sub minimize_connections {
4529 my $r = Git::SVN::read_all_remotes();
4530 my $new_urls = {};
4531 my $root_repos = {};
4532 foreach my $repo_id (keys %$r) {
4533 my $url = $r->{$repo_id}->{url} or next;
4534 my $fetch = $r->{$repo_id}->{fetch} or next;
4535 my $ra = Git::SVN::Ra->new($url);
4536
4537 # skip existing cases where we already connect to the root
4538 if (($ra->{url} eq $ra->{repos_root}) ||
4539 (Git::SVN::sanitize_remote_name($ra->{repos_root}) eq
4540 $repo_id)) {
4541 $root_repos->{$ra->{url}} = $repo_id;
4542 next;
4543 }
4544
4545 my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
4546 my $root_path = $ra->{url};
4547 $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
4548 foreach my $path (keys %$fetch) {
4549 my $ref_id = $fetch->{$path};
4550 my $gs = Git::SVN->new($ref_id, $repo_id, $path);
4551
4552 # make sure we can read when connecting to
4553 # a higher level of a repository
4554 my ($last_rev, undef) = $gs->last_rev_commit;
4555 if (!defined $last_rev) {
4556 $last_rev = eval {
4557 $root_ra->get_latest_revnum;
4558 };
4559 next if $@;
4560 }
4561 my $new = $root_path;
4562 $new .= length $path ? "/$path" : '';
4563 eval {
4564 $root_ra->get_log([$new], $last_rev, $last_rev,
4565 0, 0, 1, sub { });
4566 };
4567 next if $@;
4568 $new_urls->{$ra->{repos_root}}->{$new} =
4569 { ref_id => $ref_id,
4570 old_repo_id => $repo_id,
4571 old_path => $path };
4572 }
4573 }
4574
4575 my @emptied;
4576 foreach my $url (keys %$new_urls) {
4577 # see if we can re-use an existing [svn-remote "repo_id"]
4578 # instead of creating a(n ugly) new section:
4579 my $repo_id = $root_repos->{$url} ||
4580 Git::SVN::sanitize_remote_name($url);
4581
4582 my $fetch = $new_urls->{$url};
4583 foreach my $path (keys %$fetch) {
4584 my $x = $fetch->{$path};
4585 Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
4586 my $pfx = "svn-remote.$x->{old_repo_id}";
4587
4588 my $old_fetch = quotemeta("$x->{old_path}:".
4589 "refs/remotes/$x->{ref_id}");
4590 command_noisy(qw/config --unset/,
4591 "$pfx.fetch", '^'. $old_fetch . '$');
4592 delete $r->{$x->{old_repo_id}}->
4593 {fetch}->{$x->{old_path}};
4594 if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
4595 command_noisy(qw/config --unset/,
4596 "$pfx.url");
4597 push @emptied, $x->{old_repo_id}
4598 }
4599 }
4600 }
4601 if (@emptied) {
4602 my $file = $ENV{GIT_CONFIG} || $ENV{GIT_CONFIG_LOCAL} ||
4603 "$ENV{GIT_DIR}/config";
4604 print STDERR <<EOF;
4605 The following [svn-remote] sections in your config file ($file) are empty
4606 and can be safely removed:
4607 EOF
4608 print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
4609 }
4610 }
4611
4612 sub migration_check {
4613 migrate_from_v0();
4614 migrate_from_v1();
4615 migrate_from_v2();
4616 minimize_connections() if $_minimize;
4617 }
4618
4619 package Git::IndexInfo;
4620 use strict;
4621 use warnings;
4622 use Git qw/command_input_pipe command_close_pipe/;
4623
4624 sub new {
4625 my ($class) = @_;
4626 my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
4627 bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
4628 }
4629
4630 sub remove {
4631 my ($self, $path) = @_;
4632 if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
4633 return ++$self->{nr};
4634 }
4635 undef;
4636 }
4637
4638 sub update {
4639 my ($self, $mode, $hash, $path) = @_;
4640 if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
4641 return ++$self->{nr};
4642 }
4643 undef;
4644 }
4645
4646 sub DESTROY {
4647 my ($self) = @_;
4648 command_close_pipe($self->{gui}, $self->{ctx});
4649 }
4650
4651 package Git::SVN::GlobSpec;
4652 use strict;
4653 use warnings;
4654
4655 sub new {
4656 my ($class, $glob) = @_;
4657 my $re = $glob;
4658 $re =~ s!/+$!!g; # no need for trailing slashes
4659 my $nr = ($re =~ s!^(.*)\*(.*)$!\(\[^/\]+\)!g);
4660 my ($left, $right) = ($1, $2);
4661 if ($nr > 1) {
4662 die "Only one '*' wildcard expansion ",
4663 "is supported (got $nr): '$glob'\n";
4664 } elsif ($nr == 0) {
4665 die "One '*' is needed for glob: '$glob'\n";
4666 }
4667 $re = quotemeta($left) . $re . quotemeta($right);
4668 if (length $left && !($left =~ s!/+$!!g)) {
4669 die "Missing trailing '/' on left side of: '$glob' ($left)\n";
4670 }
4671 if (length $right && !($right =~ s!^/+!!g)) {
4672 die "Missing leading '/' on right side of: '$glob' ($right)\n";
4673 }
4674 my $left_re = qr/^\/\Q$left\E(\/|$)/;
4675 bless { left => $left, right => $right, left_regex => $left_re,
4676 regex => qr/$re/, glob => $glob }, $class;
4677 }
4678
4679 sub full_path {
4680 my ($self, $path) = @_;
4681 return (length $self->{left} ? "$self->{left}/" : '') .
4682 $path . (length $self->{right} ? "/$self->{right}" : '');
4683 }
4684
4685 __END__
4686
4687 Data structures:
4688
4689
4690 $remotes = { # returned by read_all_remotes()
4691 'svn' => {
4692 # svn-remote.svn.url=https://svn.musicpd.org
4693 url => 'https://svn.musicpd.org',
4694 # svn-remote.svn.fetch=mpd/trunk:trunk
4695 fetch => {
4696 'mpd/trunk' => 'trunk',
4697 },
4698 # svn-remote.svn.tags=mpd/tags/*:tags/*
4699 tags => {
4700 path => {
4701 left => 'mpd/tags',
4702 right => '',
4703 regex => qr!mpd/tags/([^/]+)$!,
4704 glob => 'tags/*',
4705 },
4706 ref => {
4707 left => 'tags',
4708 right => '',
4709 regex => qr!tags/([^/]+)$!,
4710 glob => 'tags/*',
4711 },
4712 }
4713 }
4714 };
4715
4716 $log_entry hashref as returned by libsvn_log_entry()
4717 {
4718 log => 'whitespace-formatted log entry
4719 ', # trailing newline is preserved
4720 revision => '8', # integer
4721 date => '2004-02-24T17:01:44.108345Z', # commit date
4722 author => 'committer name'
4723 };
4724
4725
4726 # this is generated by generate_diff();
4727 @mods = array of diff-index line hashes, each element represents one line
4728 of diff-index output
4729
4730 diff-index line ($m hash)
4731 {
4732 mode_a => first column of diff-index output, no leading ':',
4733 mode_b => second column of diff-index output,
4734 sha1_b => sha1sum of the final blob,
4735 chg => change type [MCRADT],
4736 file_a => original file name of a file (iff chg is 'C' or 'R')
4737 file_b => new/current file name of a file (any chg)
4738 }
4739 ;
4740
4741 # retval of read_url_paths{,_all}();
4742 $l_map = {
4743 # repository root url
4744 'https://svn.musicpd.org' => {
4745 # repository path # GIT_SVN_ID
4746 'mpd/trunk' => 'trunk',
4747 'mpd/tags/0.11.5' => 'tags/0.11.5',
4748 },
4749 }
4750
4751 Notes:
4752 I don't trust the each() function on unless I created %hash myself
4753 because the internal iterator may not have started at base.
The Best VCS Around