?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
git-svn: avoid leaving leftover committer/author info in rebase
[git/kirr.git] / git-svn.perl
blob7cd62fc5c48acc5bdeeb261b46b9c0df955db678
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 unlink $gs->{index};
400 }
401
402 sub cmd_dcommit {
403 my $head = shift;
404 git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
405 'Cannot dcommit with a dirty index. Commit your changes first, '
406 . "or stash them with `git stash'.\n";
407 $head ||= 'HEAD';
408 my @refs;
409 my ($url, $rev, $uuid, $gs) = working_head_info($head, \@refs);
410 print "Committing to $url ...\n";
411 unless ($gs) {
412 die "Unable to determine upstream SVN information from ",
413 "$head history\n";
414 }
415 my $last_rev;
416 my ($linear_refs, $parents) = linearize_history($gs, \@refs);
417 if ($_no_rebase && scalar(@$linear_refs) > 1) {
418 warn "Attempting to commit more than one change while ",
419 "--no-rebase is enabled.\n",
420 "If these changes depend on each other, re-running ",
421 "without --no-rebase will be required."
422 }
423 while (1) {
424 my $d = shift @$linear_refs or last;
425 unless (defined $last_rev) {
426 (undef, $last_rev, undef) = cmt_metadata("$d~1");
427 unless (defined $last_rev) {
428 fatal "Unable to extract revision information ",
429 "from commit $d~1";
430 }
431 }
432 if ($_dry_run) {
433 print "diff-tree $d~1 $d\n";
434 } else {
435 my $cmt_rev;
436 my %ed_opts = ( r => $last_rev,
437 log => get_commit_entry($d)->{log},
438 ra => Git::SVN::Ra->new($gs->full_url),
439 config => SVN::Core::config_get_config(
440 $Git::SVN::Ra::config_dir
441 ),
442 tree_a => "$d~1",
443 tree_b => $d,
444 editor_cb => sub {
445 print "Committed r$_[0]\n";
446 $cmt_rev = $_[0];
447 },
448 svn_path => '');
449 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
450 print "No changes\n$d~1 == $d\n";
451 } elsif ($parents->{$d} && @{$parents->{$d}}) {
452 $gs->{inject_parents_dcommit}->{$cmt_rev} =
453 $parents->{$d};
454 }
455 $_fetch_all ? $gs->fetch_all : $gs->fetch;
456 next if $_no_rebase;
457
458 # we always want to rebase against the current HEAD,
459 # not any head that was passed to us
460 my @diff = command('diff-tree', $d,
461 $gs->refname, '--');
462 my @finish;
463 if (@diff) {
464 @finish = rebase_cmd();
465 print STDERR "W: $d and ", $gs->refname,
466 " differ, using @finish:\n",
467 join("\n", @diff), "\n";
468 } else {
469 print "No changes between current HEAD and ",
470 $gs->refname,
471 "\nResetting to the latest ",
472 $gs->refname, "\n";
473 @finish = qw/reset --mixed/;
474 }
475 command_noisy(@finish, $gs->refname);
476 if (@diff) {
477 @refs = ();
478 my ($url_, $rev_, $uuid_, $gs_) =
479 working_head_info($head, \@refs);
480 my ($linear_refs_, $parents_) =
481 linearize_history($gs_, \@refs);
482 if (scalar(@$linear_refs) !=
483 scalar(@$linear_refs_)) {
484 fatal "# of revisions changed ",
485 "\nbefore:\n",
486 join("\n", @$linear_refs),
487 "\n\nafter:\n",
488 join("\n", @$linear_refs_), "\n",
489 'If you are attempting to commit ',
490 "merges, try running:\n\t",
491 'git rebase --interactive',
492 '--preserve-merges ',
493 $gs->refname,
494 "\nBefore dcommitting";
495 }
496 if ($url_ ne $url) {
497 fatal "URL mismatch after rebase: ",
498 "$url_ != $url";
499 }
500 if ($uuid_ ne $uuid) {
501 fatal "uuid mismatch after rebase: ",
502 "$uuid_ != $uuid";
503 }
504 # remap parents
505 my (%p, @l, $i);
506 for ($i = 0; $i < scalar @$linear_refs; $i++) {
507 my $new = $linear_refs_->[$i] or next;
508 $p{$new} =
509 $parents->{$linear_refs->[$i]};
510 push @l, $new;
511 }
512 $parents = \%p;
513 $linear_refs = \@l;
514 }
515 $last_rev = $cmt_rev;
516 }
517 }
518 unlink $gs->{index};
519 }
520
521 sub cmd_find_rev {
522 my $revision_or_hash = shift;
523 my $result;
524 if ($revision_or_hash =~ /^r\d+$/) {
525 my $head = shift;
526 $head ||= 'HEAD';
527 my @refs;
528 my (undef, undef, undef, $gs) = working_head_info($head, \@refs);
529 unless ($gs) {
530 die "Unable to determine upstream SVN information from ",
531 "$head history\n";
532 }
533 my $desired_revision = substr($revision_or_hash, 1);
534 $result = $gs->rev_map_get($desired_revision);
535 } else {
536 my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
537 $result = $rev;
538 }
539 print "$result\n" if $result;
540 }
541
542 sub cmd_rebase {
543 command_noisy(qw/update-index --refresh/);
544 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
545 unless ($gs) {
546 die "Unable to determine upstream SVN information from ",
547 "working tree history\n";
548 }
549 if (command(qw/diff-index HEAD --/)) {
550 print STDERR "Cannot rebase with uncommited changes:\n";
551 command_noisy('status');
552 exit 1;
553 }
554 unless ($_local) {
555 # rebase will checkout for us, so no need to do it explicitly
556 $_no_checkout = 'true';
557 $_fetch_all ? $gs->fetch_all : $gs->fetch;
558 }
559 command_noisy(rebase_cmd(), $gs->refname);
560 }
561
562 sub cmd_show_ignore {
563 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
564 $gs ||= Git::SVN->new;
565 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
566 $gs->prop_walk($gs->{path}, $r, sub {
567 my ($gs, $path, $props) = @_;
568 print STDOUT "\n# $path\n";
569 my $s = $props->{'svn:ignore'} or return;
570 $s =~ s/[\r\n]+/\n/g;
571 chomp $s;
572 $s =~ s#^#$path#gm;
573 print STDOUT "$s\n";
574 });
575 }
576
577 sub cmd_show_externals {
578 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
579 $gs ||= Git::SVN->new;
580 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
581 $gs->prop_walk($gs->{path}, $r, sub {
582 my ($gs, $path, $props) = @_;
583 print STDOUT "\n# $path\n";
584 my $s = $props->{'svn:externals'} or return;
585 $s =~ s/[\r\n]+/\n/g;
586 chomp $s;
587 $s =~ s#^#$path#gm;
588 print STDOUT "$s\n";
589 });
590 }
591
592 sub cmd_create_ignore {
593 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
594 $gs ||= Git::SVN->new;
595 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
596 $gs->prop_walk($gs->{path}, $r, sub {
597 my ($gs, $path, $props) = @_;
598 # $path is of the form /path/to/dir/
599 my $ignore = '.' . $path . '.gitignore';
600 my $s = $props->{'svn:ignore'} or return;
601 open(GITIGNORE, '>', $ignore)
602 or fatal("Failed to open `$ignore' for writing: $!");
603 $s =~ s/[\r\n]+/\n/g;
604 chomp $s;
605 # Prefix all patterns so that the ignore doesn't apply
606 # to sub-directories.
607 $s =~ s#^#/#gm;
608 print GITIGNORE "$s\n";
609 close(GITIGNORE)
610 or fatal("Failed to close `$ignore': $!");
611 command_noisy('add', $ignore);
612 });
613 }
614
615 sub canonicalize_path {
616 my ($path) = @_;
617 my $dot_slash_added = 0;
618 if (substr($path, 0, 1) ne "/") {
619 $path = "./" . $path;
620 $dot_slash_added = 1;
621 }
622 # File::Spec->canonpath doesn't collapse x/../y into y (for a
623 # good reason), so let's do this manually.
624 $path =~ s#/+#/#g;
625 $path =~ s#/\.(?:/|$)#/#g;
626 $path =~ s#/[^/]+/\.\.##g;
627 $path =~ s#/$##g;
628 $path =~ s#^\./## if $dot_slash_added;
629 return $path;
630 }
631
632 # get_svnprops(PATH)
633 # ------------------
634 # Helper for cmd_propget and cmd_proplist below.
635 sub get_svnprops {
636 my $path = shift;
637 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
638 $gs ||= Git::SVN->new;
639
640 # prefix THE PATH by the sub-directory from which the user
641 # invoked us.
642 $path = $cmd_dir_prefix . $path;
643 fatal("No such file or directory: $path") unless -e $path;
644 my $is_dir = -d $path ? 1 : 0;
645 $path = $gs->{path} . '/' . $path;
646
647 # canonicalize the path (otherwise libsvn will abort or fail to
648 # find the file)
649 $path = canonicalize_path($path);
650
651 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
652 my $props;
653 if ($is_dir) {
654 (undef, undef, $props) = $gs->ra->get_dir($path, $r);
655 }
656 else {
657 (undef, $props) = $gs->ra->get_file($path, $r, undef);
658 }
659 return $props;
660 }
661
662 # cmd_propget (PROP, PATH)
663 # ------------------------
664 # Print the SVN property PROP for PATH.
665 sub cmd_propget {
666 my ($prop, $path) = @_;
667 $path = '.' if not defined $path;
668 usage(1) if not defined $prop;
669 my $props = get_svnprops($path);
670 if (not defined $props->{$prop}) {
671 fatal("`$path' does not have a `$prop' SVN property.");
672 }
673 print $props->{$prop} . "\n";
674 }
675
676 # cmd_proplist (PATH)
677 # -------------------
678 # Print the list of SVN properties for PATH.
679 sub cmd_proplist {
680 my $path = shift;
681 $path = '.' if not defined $path;
682 my $props = get_svnprops($path);
683 print "Properties on '$path':\n";
684 foreach (sort keys %{$props}) {
685 print " $_\n";
686 }
687 }
688
689 sub cmd_multi_init {
690 my $url = shift;
691 unless (defined $_trunk || defined $_branches || defined $_tags) {
692 usage(1);
693 }
694
695 # there are currently some bugs that prevent multi-init/multi-fetch
696 # setups from working well without this.
697 $Git::SVN::_minimize_url = 1;
698
699 $_prefix = '' unless defined $_prefix;
700 if (defined $url) {
701 $url =~ s#/+$##;
702 init_subdir(@_);
703 }
704 do_git_init_db();
705 if (defined $_trunk) {
706 my $trunk_ref = $_prefix . 'trunk';
707 # try both old-style and new-style lookups:
708 my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
709 unless ($gs_trunk) {
710 my ($trunk_url, $trunk_path) =
711 complete_svn_url($url, $_trunk);
712 $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
713 undef, $trunk_ref);
714 }
715 }
716 return unless defined $_branches || defined $_tags;
717 my $ra = $url ? Git::SVN::Ra->new($url) : undef;
718 complete_url_ls_init($ra, $_branches, '--branches/-b', $_prefix);
719 complete_url_ls_init($ra, $_tags, '--tags/-t', $_prefix . 'tags/');
720 }
721
722 sub cmd_multi_fetch {
723 my $remotes = Git::SVN::read_all_remotes();
724 foreach my $repo_id (sort keys %$remotes) {
725 if ($remotes->{$repo_id}->{url}) {
726 Git::SVN::fetch_all($repo_id, $remotes);
727 }
728 }
729 }
730
731 # this command is special because it requires no metadata
732 sub cmd_commit_diff {
733 my ($ta, $tb, $url) = @_;
734 my $usage = "Usage: $0 commit-diff -r<revision> ".
735 "<tree-ish> <tree-ish> [<URL>]";
736 fatal($usage) if (!defined $ta || !defined $tb);
737 my $svn_path;
738 if (!defined $url) {
739 my $gs = eval { Git::SVN->new };
740 if (!$gs) {
741 fatal("Needed URL or usable git-svn --id in ",
742 "the command-line\n", $usage);
743 }
744 $url = $gs->{url};
745 $svn_path = $gs->{path};
746 }
747 unless (defined $_revision) {
748 fatal("-r|--revision is a required argument\n", $usage);
749 }
750 if (defined $_message && defined $_file) {
751 fatal("Both --message/-m and --file/-F specified ",
752 "for the commit message.\n",
753 "I have no idea what you mean");
754 }
755 if (defined $_file) {
756 $_message = file_to_s($_file);
757 } else {
758 $_message ||= get_commit_entry($tb)->{log};
759 }
760 my $ra ||= Git::SVN::Ra->new($url);
761 $svn_path ||= $ra->{svn_path};
762 my $r = $_revision;
763 if ($r eq 'HEAD') {
764 $r = $ra->get_latest_revnum;
765 } elsif ($r !~ /^\d+$/) {
766 die "revision argument: $r not understood by git-svn\n";
767 }
768 my %ed_opts = ( r => $r,
769 log => $_message,
770 ra => $ra,
771 tree_a => $ta,
772 tree_b => $tb,
773 editor_cb => sub { print "Committed r$_[0]\n" },
774 svn_path => $svn_path );
775 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
776 print "No changes\n$ta == $tb\n";
777 }
778 }
779
780 sub cmd_info {
781 my $path = canonicalize_path(shift or ".");
782 unless (scalar(@_) == 0) {
783 die "Too many arguments specified\n";
784 }
785
786 my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
787
788 if (!$file_type && !$diff_status) {
789 print STDERR "$path: (Not a versioned resource)\n\n";
790 return;
791 }
792
793 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
794 unless ($gs) {
795 die "Unable to determine upstream SVN information from ",
796 "working tree history\n";
797 }
798 my $full_url = $url . ($path eq "." ? "" : "/$path");
799
800 if ($_url) {
801 print $full_url, "\n";
802 return;
803 }
804
805 my $result = "Path: $path\n";
806 $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
807 $result .= "URL: " . $full_url . "\n";
808
809 eval {
810 my $repos_root = $gs->repos_root;
811 Git::SVN::remove_username($repos_root);
812 $result .= "Repository Root: $repos_root\n";
813 };
814 if ($@) {
815 $result .= "Repository Root: (offline)\n";
816 }
817 $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A";
818 $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
819
820 $result .= "Node Kind: " .
821 ($file_type eq "dir" ? "directory" : "file") . "\n";
822
823 my $schedule = $diff_status eq "A"
824 ? "add"
825 : ($diff_status eq "D" ? "delete" : "normal");
826 $result .= "Schedule: $schedule\n";
827
828 if ($diff_status eq "A") {
829 print $result, "\n";
830 return;
831 }
832
833 my ($lc_author, $lc_rev, $lc_date_utc);
834 my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $path);
835 my $log = command_output_pipe(@args);
836 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
837 while (<$log>) {
838 if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
839 $lc_author = $1;
840 $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
841 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
842 (undef, $lc_rev, undef) = ::extract_metadata($1);
843 }
844 }
845 close $log;
846
847 Git::SVN::Log::set_local_timezone();
848
849 $result .= "Last Changed Author: $lc_author\n";
850 $result .= "Last Changed Rev: $lc_rev\n";
851 $result .= "Last Changed Date: " .
852 Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
853
854 if ($file_type ne "dir") {
855 my $text_last_updated_date =
856 ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
857 $result .=
858 "Text Last Updated: " .
859 Git::SVN::Log::format_svn_date($text_last_updated_date) .
860 "\n";
861 my $checksum;
862 if ($diff_status eq "D") {
863 my ($fh, $ctx) =
864 command_output_pipe(qw(cat-file blob), "HEAD:$path");
865 if ($file_type eq "link") {
866 my $file_name = <$fh>;
867 $checksum = md5sum("link $file_name");
868 } else {
869 $checksum = md5sum($fh);
870 }
871 command_close_pipe($fh, $ctx);
872 } elsif ($file_type eq "link") {
873 my $file_name =
874 command(qw(cat-file blob), "HEAD:$path");
875 $checksum =
876 md5sum("link " . $file_name);
877 } else {
878 open FILE, "<", $path or die $!;
879 $checksum = md5sum(\*FILE);
880 close FILE or die $!;
881 }
882 $result .= "Checksum: " . $checksum . "\n";
883 }
884
885 print $result, "\n";
886 }
887
888 ########################### utility functions #########################
889
890 sub rebase_cmd {
891 my @cmd = qw/rebase/;
892 push @cmd, '-v' if $_verbose;
893 push @cmd, qw/--merge/ if $_merge;
894 push @cmd, "--strategy=$_strategy" if $_strategy;
895 @cmd;
896 }
897
898 sub post_fetch_checkout {
899 return if $_no_checkout;
900 my $gs = $Git::SVN::_head or return;
901 return if verify_ref('refs/heads/master^0');
902
903 my $valid_head = verify_ref('HEAD^0');
904 command_noisy(qw(update-ref refs/heads/master), $gs->refname);
905 return if ($valid_head || !verify_ref('HEAD^0'));
906
907 return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
908 my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
909 return if -f $index;
910
911 return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
912 return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
913 command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
914 print STDERR "Checked out HEAD:\n ",
915 $gs->full_url, " r", $gs->last_rev, "\n";
916 }
917
918 sub complete_svn_url {
919 my ($url, $path) = @_;
920 $path =~ s#/+$##;
921 if ($path !~ m#^[a-z\+]+://#) {
922 if (!defined $url || $url !~ m#^[a-z\+]+://#) {
923 fatal("E: '$path' is not a complete URL ",
924 "and a separate URL is not specified");
925 }
926 return ($url, $path);
927 }
928 return ($path, '');
929 }
930
931 sub complete_url_ls_init {
932 my ($ra, $repo_path, $switch, $pfx) = @_;
933 unless ($repo_path) {
934 print STDERR "W: $switch not specified\n";
935 return;
936 }
937 $repo_path =~ s#/+$##;
938 if ($repo_path =~ m#^[a-z\+]+://#) {
939 $ra = Git::SVN::Ra->new($repo_path);
940 $repo_path = '';
941 } else {
942 $repo_path =~ s#^/+##;
943 unless ($ra) {
944 fatal("E: '$repo_path' is not a complete URL ",
945 "and a separate URL is not specified");
946 }
947 }
948 my $url = $ra->{url};
949 my $gs = Git::SVN->init($url, undef, undef, undef, 1);
950 my $k = "svn-remote.$gs->{repo_id}.url";
951 my $orig_url = eval { command_oneline(qw/config --get/, $k) };
952 if ($orig_url && ($orig_url ne $gs->{url})) {
953 die "$k already set: $orig_url\n",
954 "wanted to set to: $gs->{url}\n";
955 }
956 command_oneline('config', $k, $gs->{url}) unless $orig_url;
957 my $remote_path = "$ra->{svn_path}/$repo_path/*";
958 $remote_path =~ s#/+#/#g;
959 $remote_path =~ s#^/##g;
960 my ($n) = ($switch =~ /^--(\w+)/);
961 if (length $pfx && $pfx !~ m#/$#) {
962 die "--prefix='$pfx' must have a trailing slash '/'\n";
963 }
964 command_noisy('config', "svn-remote.$gs->{repo_id}.$n",
965 "$remote_path:refs/remotes/$pfx*");
966 }
967
968 sub verify_ref {
969 my ($ref) = @_;
970 eval { command_oneline([ 'rev-parse', '--verify', $ref ],
971 { STDERR => 0 }); };
972 }
973
974 sub get_tree_from_treeish {
975 my ($treeish) = @_;
976 # $treeish can be a symbolic ref, too:
977 my $type = command_oneline(qw/cat-file -t/, $treeish);
978 my $expected;
979 while ($type eq 'tag') {
980 ($treeish, $type) = command(qw/cat-file tag/, $treeish);
981 }
982 if ($type eq 'commit') {
983 $expected = (grep /^tree /, command(qw/cat-file commit/,
984 $treeish))[0];
985 ($expected) = ($expected =~ /^tree ($sha1)$/o);
986 die "Unable to get tree from $treeish\n" unless $expected;
987 } elsif ($type eq 'tree') {
988 $expected = $treeish;
989 } else {
990 die "$treeish is a $type, expected tree, tag or commit\n";
991 }
992 return $expected;
993 }
994
995 sub get_commit_entry {
996 my ($treeish) = shift;
997 my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
998 my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
999 my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
1000 open my $log_fh, '>', $commit_editmsg or croak $!;
1001
1002 my $type = command_oneline(qw/cat-file -t/, $treeish);
1003 if ($type eq 'commit' || $type eq 'tag') {
1004 my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1005 $type, $treeish);
1006 my $in_msg = 0;
1007 while (<$msg_fh>) {
1008 if (!$in_msg) {
1009 $in_msg = 1 if (/^\s*$/);
1010 } elsif (/^git-svn-id: /) {
1011 # skip this for now, we regenerate the
1012 # correct one on re-fetch anyways
1013 # TODO: set *:merge properties or like...
1014 } else {
1015 print $log_fh $_ or croak $!;
1016 }
1017 }
1018 command_close_pipe($msg_fh, $ctx);
1019 }
1020 close $log_fh or croak $!;
1021
1022 if ($_edit || ($type eq 'tree')) {
1023 my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
1024 # TODO: strip out spaces, comments, like git-commit.sh
1025 system($editor, $commit_editmsg);
1026 }
1027 rename $commit_editmsg, $commit_msg or croak $!;
1028 open $log_fh, '<', $commit_msg or croak $!;
1029 { local $/; chomp($log_entry{log} = <$log_fh>); }
1030 close $log_fh or croak $!;
1031 unlink $commit_msg;
1032 \%log_entry;
1033 }
1034
1035 sub s_to_file {
1036 my ($str, $file, $mode) = @_;
1037 open my $fd,'>',$file or croak $!;
1038 print $fd $str,"\n" or croak $!;
1039 close $fd or croak $!;
1040 chmod ($mode &~ umask, $file) if (defined $mode);
1041 }
1042
1043 sub file_to_s {
1044 my $file = shift;
1045 open my $fd,'<',$file or croak "$!: file: $file\n";
1046 local $/;
1047 my $ret = <$fd>;
1048 close $fd or croak $!;
1049 $ret =~ s/\s*$//s;
1050 return $ret;
1051 }
1052
1053 # '<svn username> = real-name <email address>' mapping based on git-svnimport:
1054 sub load_authors {
1055 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1056 my $log = $cmd eq 'log';
1057 while (<$authors>) {
1058 chomp;
1059 next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1060 my ($user, $name, $email) = ($1, $2, $3);
1061 if ($log) {
1062 $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1063 } else {
1064 $users{$user} = [$name, $email];
1065 }
1066 }
1067 close $authors or croak $!;
1068 }
1069
1070 # convert GetOpt::Long specs for use by git-config
1071 sub read_repo_config {
1072 return unless -d $ENV{GIT_DIR};
1073 my $opts = shift;
1074 my @config_only;
1075 foreach my $o (keys %$opts) {
1076 # if we have mixedCase and a long option-only, then
1077 # it's a config-only variable that we don't need for
1078 # the command-line.
1079 push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1080 my $v = $opts->{$o};
1081 my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1082 $key =~ s/-//g;
1083 my $arg = 'git-config';
1084 $arg .= ' --int' if ($o =~ /[:=]i$/);
1085 $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1086 if (ref $v eq 'ARRAY') {
1087 chomp(my @tmp = `$arg --get-all svn.$key`);
1088 @$v = @tmp if @tmp;
1089 } else {
1090 chomp(my $tmp = `$arg --get svn.$key`);
1091 if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1092 $$v = $tmp;
1093 }
1094 }
1095 }
1096 delete @$opts{@config_only} if @config_only;
1097 }
1098
1099 sub extract_metadata {
1100 my $id = shift or return (undef, undef, undef);
1101 my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1102 \s([a-f\d\-]+)$/x);
1103 if (!defined $rev || !$uuid || !$url) {
1104 # some of the original repositories I made had
1105 # identifiers like this:
1106 ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/);
1107 }
1108 return ($url, $rev, $uuid);
1109 }
1110
1111 sub cmt_metadata {
1112 return extract_metadata((grep(/^git-svn-id: /,
1113 command(qw/cat-file commit/, shift)))[-1]);
1114 }
1115
1116 sub working_head_info {
1117 my ($head, $refs) = @_;
1118 my @args = ('log', '--no-color', '--first-parent');
1119 my ($fh, $ctx) = command_output_pipe(@args, $head);
1120 my $hash;
1121 my %max;
1122 while (<$fh>) {
1123 if ( m{^commit ($::sha1)$} ) {
1124 unshift @$refs, $hash if $hash and $refs;
1125 $hash = $1;
1126 next;
1127 }
1128 next unless s{^\s*(git-svn-id:)}{$1};
1129 my ($url, $rev, $uuid) = extract_metadata($_);
1130 if (defined $url && defined $rev) {
1131 next if $max{$url} and $max{$url} < $rev;
1132 if (my $gs = Git::SVN->find_by_url($url)) {
1133 my $c = $gs->rev_map_get($rev);
1134 if ($c && $c eq $hash) {
1135 close $fh; # break the pipe
1136 return ($url, $rev, $uuid, $gs);
1137 } else {
1138 $max{$url} ||= $gs->rev_map_max;
1139 }
1140 }
1141 }
1142 }
1143 command_close_pipe($fh, $ctx);
1144 (undef, undef, undef, undef);
1145 }
1146
1147 sub read_commit_parents {
1148 my ($parents, $c) = @_;
1149 chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1150 $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1151 @{$parents->{$c}} = split(/ /, $p);
1152 }
1153
1154 sub linearize_history {
1155 my ($gs, $refs) = @_;
1156 my %parents;
1157 foreach my $c (@$refs) {
1158 read_commit_parents(\%parents, $c);
1159 }
1160
1161 my @linear_refs;
1162 my %skip = ();
1163 my $last_svn_commit = $gs->last_commit;
1164 foreach my $c (reverse @$refs) {
1165 next if $c eq $last_svn_commit;
1166 last if $skip{$c};
1167
1168 unshift @linear_refs, $c;
1169 $skip{$c} = 1;
1170
1171 # we only want the first parent to diff against for linear
1172 # history, we save the rest to inject when we finalize the
1173 # svn commit
1174 my $fp_a = verify_ref("$c~1");
1175 my $fp_b = shift @{$parents{$c}} if $parents{$c};
1176 if (!$fp_a || !$fp_b) {
1177 die "Commit $c\n",
1178 "has no parent commit, and therefore ",
1179 "nothing to diff against.\n",
1180 "You should be working from a repository ",
1181 "originally created by git-svn\n";
1182 }
1183 if ($fp_a ne $fp_b) {
1184 die "$c~1 = $fp_a, however parsing commit $c ",
1185 "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1186 }
1187
1188 foreach my $p (@{$parents{$c}}) {
1189 $skip{$p} = 1;
1190 }
1191 }
1192 (\@linear_refs, \%parents);
1193 }
1194
1195 sub find_file_type_and_diff_status {
1196 my ($path) = @_;
1197 return ('dir', '') if $path eq '.';
1198
1199 my $diff_output =
1200 command_oneline(qw(diff --cached --name-status --), $path) || "";
1201 my $diff_status = (split(' ', $diff_output))[0] || "";
1202
1203 my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1204
1205 return (undef, undef) if !$diff_status && !$ls_tree;
1206
1207 if ($diff_status eq "A") {
1208 return ("link", $diff_status) if -l $path;
1209 return ("dir", $diff_status) if -d $path;
1210 return ("file", $diff_status);
1211 }
1212
1213 my $mode = (split(' ', $ls_tree))[0] || "";
1214
1215 return ("link", $diff_status) if $mode eq "120000";
1216 return ("dir", $diff_status) if $mode eq "040000";
1217 return ("file", $diff_status);
1218 }
1219
1220 sub md5sum {
1221 my $arg = shift;
1222 my $ref = ref $arg;
1223 my $md5 = Digest::MD5->new();
1224 if ($ref eq 'GLOB' || $ref eq 'IO::File') {
1225 $md5->addfile($arg) or croak $!;
1226 } elsif ($ref eq 'SCALAR') {
1227 $md5->add($$arg) or croak $!;
1228 } elsif (!$ref) {
1229 $md5->add($arg) or croak $!;
1230 } else {
1231 ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
1232 }
1233 return $md5->hexdigest();
1234 }
1235
1236 package Git::SVN;
1237 use strict;
1238 use warnings;
1239 use Fcntl qw/:DEFAULT :seek/;
1240 use constant rev_map_fmt => 'NH40';
1241 use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1242 $_repack $_repack_flags $_use_svm_props $_head
1243 $_use_svnsync_props $no_reuse_existing $_minimize_url
1244 $_use_log_author/;
1245 use Carp qw/croak/;
1246 use File::Path qw/mkpath/;
1247 use File::Copy qw/copy/;
1248 use IPC::Open3;
1249
1250 my $_repack_nr;
1251 # properties that we do not log:
1252 my %SKIP_PROP;
1253 BEGIN {
1254 %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1255 svn:special svn:executable
1256 svn:entry:committed-rev
1257 svn:entry:last-author
1258 svn:entry:uuid
1259 svn:entry:committed-date/;
1260
1261 # some options are read globally, but can be overridden locally
1262 # per [svn-remote "..."] section. Command-line options will *NOT*
1263 # override options set in an [svn-remote "..."] section
1264 no strict 'refs';
1265 for my $option (qw/follow_parent no_metadata use_svm_props
1266 use_svnsync_props/) {
1267 my $key = $option;
1268 $key =~ tr/_//d;
1269 my $prop = "-$option";
1270 *$option = sub {
1271 my ($self) = @_;
1272 return $self->{$prop} if exists $self->{$prop};
1273 my $k = "svn-remote.$self->{repo_id}.$key";
1274 eval { command_oneline(qw/config --get/, $k) };
1275 if ($@) {
1276 $self->{$prop} = ${"Git::SVN::_$option"};
1277 } else {
1278 my $v = command_oneline(qw/config --bool/,$k);
1279 $self->{$prop} = $v eq 'false' ? 0 : 1;
1280 }
1281 return $self->{$prop};
1282 }
1283 }
1284 }
1285
1286 my %LOCKFILES;
1287 END { unlink keys %LOCKFILES if %LOCKFILES }
1288
1289 sub resolve_local_globs {
1290 my ($url, $fetch, $glob_spec) = @_;
1291 return unless defined $glob_spec;
1292 my $ref = $glob_spec->{ref};
1293 my $path = $glob_spec->{path};
1294 foreach (command(qw#for-each-ref --format=%(refname) refs/remotes#)) {
1295 next unless m#^refs/remotes/$ref->{regex}$#;
1296 my $p = $1;
1297 my $pathname = desanitize_refname($path->full_path($p));
1298 my $refname = desanitize_refname($ref->full_path($p));
1299 if (my $existing = $fetch->{$pathname}) {
1300 if ($existing ne $refname) {
1301 die "Refspec conflict:\n",
1302 "existing: refs/remotes/$existing\n",
1303 " globbed: refs/remotes/$refname\n";
1304 }
1305 my $u = (::cmt_metadata("refs/remotes/$refname"))[0];
1306 $u =~ s!^\Q$url\E(/|$)!! or die
1307 "refs/remotes/$refname: '$url' not found in '$u'\n";
1308 if ($pathname ne $u) {
1309 warn "W: Refspec glob conflict ",
1310 "(ref: refs/remotes/$refname):\n",
1311 "expected path: $pathname\n",
1312 " real path: $u\n",
1313 "Continuing ahead with $u\n";
1314 next;
1315 }
1316 } else {
1317 $fetch->{$pathname} = $refname;
1318 }
1319 }
1320 }
1321
1322 sub parse_revision_argument {
1323 my ($base, $head) = @_;
1324 if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1325 return ($base, $head);
1326 }
1327 return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1328 return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1329 return ($head, $head) if ($::_revision eq 'HEAD');
1330 return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1331 return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1332 die "revision argument: $::_revision not understood by git-svn\n";
1333 }
1334
1335 sub fetch_all {
1336 my ($repo_id, $remotes) = @_;
1337 if (ref $repo_id) {
1338 my $gs = $repo_id;
1339 $repo_id = undef;
1340 $repo_id = $gs->{repo_id};
1341 }
1342 $remotes ||= read_all_remotes();
1343 my $remote = $remotes->{$repo_id} or
1344 die "[svn-remote \"$repo_id\"] unknown\n";
1345 my $fetch = $remote->{fetch};
1346 my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1347 my (@gs, @globs);
1348 my $ra = Git::SVN::Ra->new($url);
1349 my $uuid = $ra->get_uuid;
1350 my $head = $ra->get_latest_revnum;
1351 my $base = defined $fetch ? $head : 0;
1352
1353 # read the max revs for wildcard expansion (branches/*, tags/*)
1354 foreach my $t (qw/branches tags/) {
1355 defined $remote->{$t} or next;
1356 push @globs, $remote->{$t};
1357 my $max_rev = eval { tmp_config(qw/--int --get/,
1358 "svn-remote.$repo_id.${t}-maxRev") };
1359 if (defined $max_rev && ($max_rev < $base)) {
1360 $base = $max_rev;
1361 } elsif (!defined $max_rev) {
1362 $base = 0;
1363 }
1364 }
1365
1366 if ($fetch) {
1367 foreach my $p (sort keys %$fetch) {
1368 my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1369 my $lr = $gs->rev_map_max;
1370 if (defined $lr) {
1371 $base = $lr if ($lr < $base);
1372 }
1373 push @gs, $gs;
1374 }
1375 }
1376
1377 ($base, $head) = parse_revision_argument($base, $head);
1378 $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1379 unlink $_->{index} foreach @gs;
1380 }
1381
1382 sub read_all_remotes {
1383 my $r = {};
1384 foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1385 if (m!^(.+)\.fetch=\s*(.*)\s*:\s*refs/remotes/(.+)\s*$!) {
1386 my ($remote, $local_ref, $remote_ref) = ($1, $2, $3);
1387 $local_ref =~ s{^/}{};
1388 $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1389 } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1390 $r->{$1}->{url} = $2;
1391 } elsif (m!^(.+)\.(branches|tags)=
1392 (.*):refs/remotes/(.+)\s*$/!x) {
1393 my ($p, $g) = ($3, $4);
1394 my $rs = $r->{$1}->{$2} = {
1395 t => $2,
1396 remote => $1,
1397 path => Git::SVN::GlobSpec->new($p),
1398 ref => Git::SVN::GlobSpec->new($g) };
1399 if (length($rs->{ref}->{right}) != 0) {
1400 die "The '*' glob character must be the last ",
1401 "character of '$g'\n";
1402 }
1403 }
1404 }
1405 $r;
1406 }
1407
1408 sub init_vars {
1409 if (defined $_repack) {
1410 $_repack = 1000 if ($_repack <= 0);
1411 $_repack_nr = $_repack;
1412 $_repack_flags ||= '-d';
1413 }
1414 }
1415
1416 sub verify_remotes_sanity {
1417 return unless -d $ENV{GIT_DIR};
1418 my %seen;
1419 foreach (command(qw/config -l/)) {
1420 if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1421 if ($seen{$1}) {
1422 die "Remote ref refs/remote/$1 is tracked by",
1423 "\n \"$_\"\nand\n \"$seen{$1}\"\n",
1424 "Please resolve this ambiguity in ",
1425 "your git configuration file before ",
1426 "continuing\n";
1427 }
1428 $seen{$1} = $_;
1429 }
1430 }
1431 }
1432
1433 # we allow more chars than remotes2config.sh...
1434 sub sanitize_remote_name {
1435 my ($name) = @_;
1436 $name =~ tr{A-Za-z0-9:,/+-}{.}c;
1437 $name;
1438 }
1439
1440 sub find_existing_remote {
1441 my ($url, $remotes) = @_;
1442 return undef if $no_reuse_existing;
1443 my $existing;
1444 foreach my $repo_id (keys %$remotes) {
1445 my $u = $remotes->{$repo_id}->{url} or next;
1446 next if $u ne $url;
1447 $existing = $repo_id;
1448 last;
1449 }
1450 $existing;
1451 }
1452
1453 sub init_remote_config {
1454 my ($self, $url, $no_write) = @_;
1455 $url =~ s!/+$!!; # strip trailing slash
1456 my $r = read_all_remotes();
1457 my $existing = find_existing_remote($url, $r);
1458 if ($existing) {
1459 unless ($no_write) {
1460 print STDERR "Using existing ",
1461 "[svn-remote \"$existing\"]\n";
1462 }
1463 $self->{repo_id} = $existing;
1464 } elsif ($_minimize_url) {
1465 my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1466 $existing = find_existing_remote($min_url, $r);
1467 if ($existing) {
1468 unless ($no_write) {
1469 print STDERR "Using existing ",
1470 "[svn-remote \"$existing\"]\n";
1471 }
1472 $self->{repo_id} = $existing;
1473 }
1474 if ($min_url ne $url) {
1475 unless ($no_write) {
1476 print STDERR "Using higher level of URL: ",
1477 "$url => $min_url\n";
1478 }
1479 my $old_path = $self->{path};
1480 $self->{path} = $url;
1481 $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1482 if (length $old_path) {
1483 $self->{path} .= "/$old_path";
1484 }
1485 $url = $min_url;
1486 }
1487 }
1488 my $orig_url;
1489 if (!$existing) {
1490 # verify that we aren't overwriting anything:
1491 $orig_url = eval {
1492 command_oneline('config', '--get',
1493 "svn-remote.$self->{repo_id}.url")
1494 };
1495 if ($orig_url && ($orig_url ne $url)) {
1496 die "svn-remote.$self->{repo_id}.url already set: ",
1497 "$orig_url\nwanted to set to: $url\n";
1498 }
1499 }
1500 my ($xrepo_id, $xpath) = find_ref($self->refname);
1501 if (defined $xpath) {
1502 die "svn-remote.$xrepo_id.fetch already set to track ",
1503 "$xpath:refs/remotes/", $self->refname, "\n";
1504 }
1505 unless ($no_write) {
1506 command_noisy('config',
1507 "svn-remote.$self->{repo_id}.url", $url);
1508 $self->{path} =~ s{^/}{};
1509 command_noisy('config', '--add',
1510 "svn-remote.$self->{repo_id}.fetch",
1511 "$self->{path}:".$self->refname);
1512 }
1513 $self->{url} = $url;
1514 }
1515
1516 sub find_by_url { # repos_root and, path are optional
1517 my ($class, $full_url, $repos_root, $path) = @_;
1518
1519 return undef unless defined $full_url;
1520 remove_username($full_url);
1521 remove_username($repos_root) if defined $repos_root;
1522 my $remotes = read_all_remotes();
1523 if (defined $full_url && defined $repos_root && !defined $path) {
1524 $path = $full_url;
1525 $path =~ s#^\Q$repos_root\E(?:/|$)##;
1526 }
1527 foreach my $repo_id (keys %$remotes) {
1528 my $u = $remotes->{$repo_id}->{url} or next;
1529 remove_username($u);
1530 next if defined $repos_root && $repos_root ne $u;
1531
1532 my $fetch = $remotes->{$repo_id}->{fetch} || {};
1533 foreach (qw/branches tags/) {
1534 resolve_local_globs($u, $fetch,
1535 $remotes->{$repo_id}->{$_});
1536 }
1537 my $p = $path;
1538 unless (defined $p) {
1539 $p = $full_url;
1540 $p =~ s#^\Q$u\E(?:/|$)## or next;
1541 }
1542 foreach my $f (keys %$fetch) {
1543 next if $f ne $p;
1544 return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1545 }
1546 }
1547 undef;
1548 }
1549
1550 sub init {
1551 my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1552 my $self = _new($class, $repo_id, $ref_id, $path);
1553 if (defined $url) {
1554 $self->init_remote_config($url, $no_write);
1555 }
1556 $self;
1557 }
1558
1559 sub find_ref {
1560 my ($ref_id) = @_;
1561 foreach (command(qw/config -l/)) {
1562 next unless m!^svn-remote\.(.+)\.fetch=
1563 \s*(.*)\s*:\s*refs/remotes/(.+)\s*$!x;
1564 my ($repo_id, $path, $ref) = ($1, $2, $3);
1565 if ($ref eq $ref_id) {
1566 $path = '' if ($path =~ m#^\./?#);
1567 return ($repo_id, $path);
1568 }
1569 }
1570 (undef, undef, undef);
1571 }
1572
1573 sub new {
1574 my ($class, $ref_id, $repo_id, $path) = @_;
1575 if (defined $ref_id && !defined $repo_id && !defined $path) {
1576 ($repo_id, $path) = find_ref($ref_id);
1577 if (!defined $repo_id) {
1578 die "Could not find a \"svn-remote.*.fetch\" key ",
1579 "in the repository configuration matching: ",
1580 "refs/remotes/$ref_id\n";
1581 }
1582 }
1583 my $self = _new($class, $repo_id, $ref_id, $path);
1584 if (!defined $self->{path} || !length $self->{path}) {
1585 my $fetch = command_oneline('config', '--get',
1586 "svn-remote.$repo_id.fetch",
1587 ":refs/remotes/$ref_id\$") or
1588 die "Failed to read \"svn-remote.$repo_id.fetch\" ",
1589 "\":refs/remotes/$ref_id\$\" in config\n";
1590 ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
1591 }
1592 $self->{url} = command_oneline('config', '--get',
1593 "svn-remote.$repo_id.url") or
1594 die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
1595 $self->rebuild;
1596 $self;
1597 }
1598
1599 sub refname {
1600 my ($refname) = "refs/remotes/$_[0]->{ref_id}" ;
1601
1602 # It cannot end with a slash /, we'll throw up on this because
1603 # SVN can't have directories with a slash in their name, either:
1604 if ($refname =~ m{/$}) {
1605 die "ref: '$refname' ends with a trailing slash, this is ",
1606 "not permitted by git nor Subversion\n";
1607 }
1608
1609 # It cannot have ASCII control character space, tilde ~, caret ^,
1610 # colon :, question-mark ?, asterisk *, space, or open bracket [
1611 # anywhere.
1612 #
1613 # Additionally, % must be escaped because it is used for escaping
1614 # and we want our escaped refname to be reversible
1615 $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
1616
1617 # no slash-separated component can begin with a dot .
1618 # /.* becomes /%2E*
1619 $refname =~ s{/\.}{/%2E}g;
1620
1621 # It cannot have two consecutive dots .. anywhere
1622 # .. becomes %2E%2E
1623 $refname =~ s{\.\.}{%2E%2E}g;
1624
1625 return $refname;
1626 }
1627
1628 sub desanitize_refname {
1629 my ($refname) = @_;
1630 $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
1631 return $refname;
1632 }
1633
1634 sub svm_uuid {
1635 my ($self) = @_;
1636 return $self->{svm}->{uuid} if $self->svm;
1637 $self->ra;
1638 unless ($self->{svm}) {
1639 die "SVM UUID not cached, and reading remotely failed\n";
1640 }
1641 $self->{svm}->{uuid};
1642 }
1643
1644 sub svm {
1645 my ($self) = @_;
1646 return $self->{svm} if $self->{svm};
1647 my $svm;
1648 # see if we have it in our config, first:
1649 eval {
1650 my $section = "svn-remote.$self->{repo_id}";
1651 $svm = {
1652 source => tmp_config('--get', "$section.svm-source"),
1653 uuid => tmp_config('--get', "$section.svm-uuid"),
1654 replace => tmp_config('--get', "$section.svm-replace"),
1655 }
1656 };
1657 if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
1658 $self->{svm} = $svm;
1659 }
1660 $self->{svm};
1661 }
1662
1663 sub _set_svm_vars {
1664 my ($self, $ra) = @_;
1665 return $ra if $self->svm;
1666
1667 my @err = ( "useSvmProps set, but failed to read SVM properties\n",
1668 "(svm:source, svm:uuid) ",
1669 "from the following URLs:\n" );
1670 sub read_svm_props {
1671 my ($self, $ra, $path, $r) = @_;
1672 my $props = ($ra->get_dir($path, $r))[2];
1673 my $src = $props->{'svm:source'};
1674 my $uuid = $props->{'svm:uuid'};
1675 return undef if (!$src || !$uuid);
1676
1677 chomp($src, $uuid);
1678
1679 $uuid =~ m{^[0-9a-f\-]{30,}$}
1680 or die "doesn't look right - svm:uuid is '$uuid'\n";
1681
1682 # the '!' is used to mark the repos_root!/relative/path
1683 $src =~ s{/?!/?}{/};
1684 $src =~ s{/+$}{}; # no trailing slashes please
1685 # username is of no interest
1686 $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
1687
1688 my $replace = $ra->{url};
1689 $replace .= "/$path" if length $path;
1690
1691 my $section = "svn-remote.$self->{repo_id}";
1692 tmp_config("$section.svm-source", $src);
1693 tmp_config("$section.svm-replace", $replace);
1694 tmp_config("$section.svm-uuid", $uuid);
1695 $self->{svm} = {
1696 source => $src,
1697 uuid => $uuid,
1698 replace => $replace
1699 };
1700 }
1701
1702 my $r = $ra->get_latest_revnum;
1703 my $path = $self->{path};
1704 my %tried;
1705 while (length $path) {
1706 unless ($tried{"$self->{url}/$path"}) {
1707 return $ra if $self->read_svm_props($ra, $path, $r);
1708 $tried{"$self->{url}/$path"} = 1;
1709 }
1710 $path =~ s#/?[^/]+$##;
1711 }
1712 die "Path: '$path' should be ''\n" if $path ne '';
1713 return $ra if $self->read_svm_props($ra, $path, $r);
1714 $tried{"$self->{url}/$path"} = 1;
1715
1716 if ($ra->{repos_root} eq $self->{url}) {
1717 die @err, (map { " $_\n" } keys %tried), "\n";
1718 }
1719
1720 # nope, make sure we're connected to the repository root:
1721 my $ok;
1722 my @tried_b;
1723 $path = $ra->{svn_path};
1724 $ra = Git::SVN::Ra->new($ra->{repos_root});
1725 while (length $path) {
1726 unless ($tried{"$ra->{url}/$path"}) {
1727 $ok = $self->read_svm_props($ra, $path, $r);
1728 last if $ok;
1729 $tried{"$ra->{url}/$path"} = 1;
1730 }
1731 $path =~ s#/?[^/]+$##;
1732 }
1733 die "Path: '$path' should be ''\n" if $path ne '';
1734 $ok ||= $self->read_svm_props($ra, $path, $r);
1735 $tried{"$ra->{url}/$path"} = 1;
1736 if (!$ok) {
1737 die @err, (map { " $_\n" } keys %tried), "\n";
1738 }
1739 Git::SVN::Ra->new($self->{url});
1740 }
1741
1742 sub svnsync {
1743 my ($self) = @_;
1744 return $self->{svnsync} if $self->{svnsync};
1745
1746 if ($self->no_metadata) {
1747 die "Can't have both 'noMetadata' and ",
1748 "'useSvnsyncProps' options set!\n";
1749 }
1750 if ($self->rewrite_root) {
1751 die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
1752 "options set!\n";
1753 }
1754
1755 my $svnsync;
1756 # see if we have it in our config, first:
1757 eval {
1758 my $section = "svn-remote.$self->{repo_id}";
1759 $svnsync = {
1760 url => tmp_config('--get', "$section.svnsync-url"),
1761 uuid => tmp_config('--get', "$section.svnsync-uuid"),
1762 }
1763 };
1764 if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
1765 return $self->{svnsync} = $svnsync;
1766 }
1767
1768 my $err = "useSvnsyncProps set, but failed to read " .
1769 "svnsync property: svn:sync-from-";
1770 my $rp = $self->ra->rev_proplist(0);
1771
1772 my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
1773 $url =~ m{^[a-z\+]+://} or
1774 die "doesn't look right - svn:sync-from-url is '$url'\n";
1775
1776 my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
1777 $uuid =~ m{^[0-9a-f\-]{30,}$} or
1778 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1779
1780 my $section = "svn-remote.$self->{repo_id}";
1781 tmp_config('--add', "$section.svnsync-uuid", $uuid);
1782 tmp_config('--add', "$section.svnsync-url", $url);
1783 return $self->{svnsync} = { url => $url, uuid => $uuid };
1784 }
1785
1786 # this allows us to memoize our SVN::Ra UUID locally and avoid a
1787 # remote lookup (useful for 'git svn log').
1788 sub ra_uuid {
1789 my ($self) = @_;
1790 unless ($self->{ra_uuid}) {
1791 my $key = "svn-remote.$self->{repo_id}.uuid";
1792 my $uuid = eval { tmp_config('--get', $key) };
1793 if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/) {
1794 $self->{ra_uuid} = $uuid;
1795 } else {
1796 die "ra_uuid called without URL\n" unless $self->{url};
1797 $self->{ra_uuid} = $self->ra->get_uuid;
1798 tmp_config('--add', $key, $self->{ra_uuid});
1799 }
1800 }
1801 $self->{ra_uuid};
1802 }
1803
1804 sub _set_repos_root {
1805 my ($self, $repos_root) = @_;
1806 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1807 $repos_root ||= $self->ra->{repos_root};
1808 tmp_config($k, $repos_root);
1809 $repos_root;
1810 }
1811
1812 sub repos_root {
1813 my ($self) = @_;
1814 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1815 eval { tmp_config('--get', $k) } || $self->_set_repos_root;
1816 }
1817
1818 sub ra {
1819 my ($self) = shift;
1820 my $ra = Git::SVN::Ra->new($self->{url});
1821 $self->_set_repos_root($ra->{repos_root});
1822 if ($self->use_svm_props && !$self->{svm}) {
1823 if ($self->no_metadata) {
1824 die "Can't have both 'noMetadata' and ",
1825 "'useSvmProps' options set!\n";
1826 } elsif ($self->use_svnsync_props) {
1827 die "Can't have both 'useSvnsyncProps' and ",
1828 "'useSvmProps' options set!\n";
1829 }
1830 $ra = $self->_set_svm_vars($ra);
1831 $self->{-want_revprops} = 1;
1832 }
1833 $ra;
1834 }
1835
1836 sub rel_path {
1837 my ($self) = @_;
1838 my $repos_root = $self->ra->{repos_root};
1839 return $self->{path} if ($self->{url} eq $repos_root);
1840 my $url = $self->{url} .
1841 (length $self->{path} ? "/$self->{path}" : $self->{path});
1842 $url =~ s!^\Q$repos_root\E(?:/+|$)!!g;
1843 $url;
1844 }
1845
1846 # prop_walk(PATH, REV, SUB)
1847 # -------------------------
1848 # Recursively traverse PATH at revision REV and invoke SUB for each
1849 # directory that contains a SVN property. SUB will be invoked as
1850 # follows: &SUB(gs, path, props); where `gs' is this instance of
1851 # Git::SVN, `path' the path to the directory where the properties
1852 # `props' were found. The `path' will be relative to point of checkout,
1853 # that is, if url://repo/trunk is the current Git branch, and that
1854 # directory contains a sub-directory `d', SUB will be invoked with `/d/'
1855 # as `path' (note the trailing `/').
1856 sub prop_walk {
1857 my ($self, $path, $rev, $sub) = @_;
1858
1859 my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
1860 $path =~ s#^/*#/#g;
1861 my $p = $path;
1862 # Strip the irrelevant part of the path.
1863 $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
1864 # Ensure the path is terminated by a `/'.
1865 $p =~ s#/*$#/#;
1866
1867 # The properties contain all the internal SVN stuff nobody
1868 # (usually) cares about.
1869 my $interesting_props = 0;
1870 foreach (keys %{$props}) {
1871 # If it doesn't start with `svn:', it must be a
1872 # user-defined property.
1873 ++$interesting_props and next if $_ !~ /^svn:/;
1874 # FIXME: Fragile, if SVN adds new public properties,
1875 # this needs to be updated.
1876 ++$interesting_props if /^svn:(?:ignore|keywords|executable
1877 |eol-style|mime-type
1878 |externals|needs-lock)$/x;
1879 }
1880 &$sub($self, $p, $props) if $interesting_props;
1881
1882 foreach (sort keys %$dirent) {
1883 next if $dirent->{$_}->{kind} != $SVN::Node::dir;
1884 $self->prop_walk($path . '/' . $_, $rev, $sub);
1885 }
1886 }
1887
1888 sub last_rev { ($_[0]->last_rev_commit)[0] }
1889 sub last_commit { ($_[0]->last_rev_commit)[1] }
1890
1891 # returns the newest SVN revision number and newest commit SHA1
1892 sub last_rev_commit {
1893 my ($self) = @_;
1894 if (defined $self->{last_rev} && defined $self->{last_commit}) {
1895 return ($self->{last_rev}, $self->{last_commit});
1896 }
1897 my $c = ::verify_ref($self->refname.'^0');
1898 if ($c && !$self->use_svm_props && !$self->no_metadata) {
1899 my $rev = (::cmt_metadata($c))[1];
1900 if (defined $rev) {
1901 ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
1902 return ($rev, $c);
1903 }
1904 }
1905 my $map_path = $self->map_path;
1906 unless (-e $map_path) {
1907 ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
1908 return (undef, undef);
1909 }
1910 my ($rev, $commit) = $self->rev_map_max(1);
1911 ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
1912 return ($rev, $commit);
1913 }
1914
1915 sub get_fetch_range {
1916 my ($self, $min, $max) = @_;
1917 $max ||= $self->ra->get_latest_revnum;
1918 $min ||= $self->rev_map_max;
1919 (++$min, $max);
1920 }
1921
1922 sub tmp_config {
1923 my (@args) = @_;
1924 my $old_def_config = "$ENV{GIT_DIR}/svn/config";
1925 my $config = "$ENV{GIT_DIR}/svn/.metadata";
1926 if (! -f $config && -f $old_def_config) {
1927 rename $old_def_config, $config or
1928 die "Failed rename $old_def_config => $config: $!\n";
1929 }
1930 my $old_config = $ENV{GIT_CONFIG};
1931 $ENV{GIT_CONFIG} = $config;
1932 $@ = undef;
1933 my @ret = eval {
1934 unless (-f $config) {
1935 mkfile($config);
1936 open my $fh, '>', $config or
1937 die "Can't open $config: $!\n";
1938 print $fh "; This file is used internally by ",
1939 "git-svn\n" or die
1940 "Couldn't write to $config: $!\n";
1941 print $fh "; You should not have to edit it\n" or
1942 die "Couldn't write to $config: $!\n";
1943 close $fh or die "Couldn't close $config: $!\n";
1944 }
1945 command('config', @args);
1946 };
1947 my $err = $@;
1948 if (defined $old_config) {
1949 $ENV{GIT_CONFIG} = $old_config;
1950 } else {
1951 delete $ENV{GIT_CONFIG};
1952 }
1953 die $err if $err;
1954 wantarray ? @ret : $ret[0];
1955 }
1956
1957 sub tmp_index_do {
1958 my ($self, $sub) = @_;
1959 my $old_index = $ENV{GIT_INDEX_FILE};
1960 $ENV{GIT_INDEX_FILE} = $self->{index};
1961 $@ = undef;
1962 my @ret = eval {
1963 my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
1964 mkpath([$dir]) unless -d $dir;
1965 &$sub;
1966 };
1967 my $err = $@;
1968 if (defined $old_index) {
1969 $ENV{GIT_INDEX_FILE} = $old_index;
1970 } else {
1971 delete $ENV{GIT_INDEX_FILE};
1972 }
1973 die $err if $err;
1974 wantarray ? @ret : $ret[0];
1975 }
1976
1977 sub assert_index_clean {
1978 my ($self, $treeish) = @_;
1979
1980 $self->tmp_index_do(sub {
1981 command_noisy('read-tree', $treeish) unless -e $self->{index};
1982 my $x = command_oneline('write-tree');
1983 my ($y) = (command(qw/cat-file commit/, $treeish) =~
1984 /^tree ($::sha1)/mo);
1985 return if $y eq $x;
1986
1987 warn "Index mismatch: $y != $x\nrereading $treeish\n";
1988 unlink $self->{index} or die "unlink $self->{index}: $!\n";
1989 command_noisy('read-tree', $treeish);
1990 $x = command_oneline('write-tree');
1991 if ($y ne $x) {
1992 ::fatal "trees ($treeish) $y != $x\n",
1993 "Something is seriously wrong...";
1994 }
1995 });
1996 }
1997
1998 sub get_commit_parents {
1999 my ($self, $log_entry) = @_;
2000 my (%seen, @ret, @tmp);
2001 # legacy support for 'set-tree'; this is only used by set_tree_cb:
2002 if (my $ip = $self->{inject_parents}) {
2003 if (my $commit = delete $ip->{$log_entry->{revision}}) {
2004 push @tmp, $commit;
2005 }
2006 }
2007 if (my $cur = ::verify_ref($self->refname.'^0')) {
2008 push @tmp, $cur;
2009 }
2010 if (my $ipd = $self->{inject_parents_dcommit}) {
2011 if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2012 push @tmp, @$commit;
2013 }
2014 }
2015 push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2016 while (my $p = shift @tmp) {
2017 next if $seen{$p};
2018 $seen{$p} = 1;
2019 push @ret, $p;
2020 # MAXPARENT is defined to 16 in commit-tree.c:
2021 last if @ret >= 16;
2022 }
2023 if (@tmp) {
2024 die "r$log_entry->{revision}: No room for parents:\n\t",
2025 join("\n\t", @tmp), "\n";
2026 }
2027 @ret;
2028 }
2029
2030 sub rewrite_root {
2031 my ($self) = @_;
2032 return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2033 my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2034 my $rwr = eval { command_oneline(qw/config --get/, $k) };
2035 if ($rwr) {
2036 $rwr =~ s#/+$##;
2037 if ($rwr !~ m#^[a-z\+]+://#) {
2038 die "$rwr is not a valid URL (key: $k)\n";
2039 }
2040 }
2041 $self->{-rewrite_root} = $rwr;
2042 }
2043
2044 sub metadata_url {
2045 my ($self) = @_;
2046 ($self->rewrite_root || $self->{url}) .
2047 (length $self->{path} ? '/' . $self->{path} : '');
2048 }
2049
2050 sub full_url {
2051 my ($self) = @_;
2052 $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2053 }
2054
2055
2056 sub set_commit_header_env {
2057 my ($log_entry) = @_;
2058 my %env;
2059 foreach my $ned (qw/NAME EMAIL DATE/) {
2060 foreach my $ac (qw/AUTHOR COMMITTER/) {
2061 $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2062 }
2063 }
2064
2065 $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2066 $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2067 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2068
2069 $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2070 ? $log_entry->{commit_name}
2071 : $log_entry->{name};
2072 $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2073 ? $log_entry->{commit_email}
2074 : $log_entry->{email};
2075 \%env;
2076 }
2077
2078 sub restore_commit_header_env {
2079 my ($env) = @_;
2080 foreach my $ned (qw/NAME EMAIL DATE/) {
2081 foreach my $ac (qw/AUTHOR COMMITTER/) {
2082 my $k = "GIT_${ac}_${ned}";
2083 if (defined $env->{$k}) {
2084 $ENV{$k} = $env->{$k};
2085 } else {
2086 delete $ENV{$k};
2087 }
2088 }
2089 }
2090 }
2091
2092 sub do_git_commit {
2093 my ($self, $log_entry) = @_;
2094 my $lr = $self->last_rev;
2095 if (defined $lr && $lr >= $log_entry->{revision}) {
2096 die "Last fetched revision of ", $self->refname,
2097 " was r$lr, but we are about to fetch: ",
2098 "r$log_entry->{revision}!\n";
2099 }
2100 if (my $c = $self->rev_map_get($log_entry->{revision})) {
2101 croak "$log_entry->{revision} = $c already exists! ",
2102 "Why are we refetching it?\n";
2103 }
2104 my $old_env = set_commit_header_env($log_entry);
2105 my $tree = $log_entry->{tree};
2106 if (!defined $tree) {
2107 $tree = $self->tmp_index_do(sub {
2108 command_oneline('write-tree') });
2109 }
2110 die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2111
2112 my @exec = ('git-commit-tree', $tree);
2113 foreach ($self->get_commit_parents($log_entry)) {
2114 push @exec, '-p', $_;
2115 }
2116 defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2117 or croak $!;
2118 print $msg_fh $log_entry->{log} or croak $!;
2119 restore_commit_header_env($old_env);
2120 unless ($self->no_metadata) {
2121 print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2122 or croak $!;
2123 }
2124 $msg_fh->flush == 0 or croak $!;
2125 close $msg_fh or croak $!;
2126 chomp(my $commit = do { local $/; <$out_fh> });
2127 close $out_fh or croak $!;
2128 waitpid $pid, 0;
2129 croak $? if $?;
2130 if ($commit !~ /^$::sha1$/o) {
2131 die "Failed to commit, invalid sha1: $commit\n";
2132 }
2133
2134 $self->rev_map_set($log_entry->{revision}, $commit, 1);
2135
2136 $self->{last_rev} = $log_entry->{revision};
2137 $self->{last_commit} = $commit;
2138 print "r$log_entry->{revision}";
2139 if (defined $log_entry->{svm_revision}) {
2140 print " (\@$log_entry->{svm_revision})";
2141 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2142 0, $self->svm_uuid);
2143 }
2144 print " = $commit ($self->{ref_id})\n";
2145 if (defined $_repack && (--$_repack_nr == 0)) {
2146 $_repack_nr = $_repack;
2147 # repack doesn't use any arguments with spaces in them, does it?
2148 print "Running git repack $_repack_flags ...\n";
2149 command_noisy('repack', split(/\s+/, $_repack_flags));
2150 print "Done repacking\n";
2151 }
2152 return $commit;
2153 }
2154
2155 sub match_paths {
2156 my ($self, $paths, $r) = @_;
2157 return 1 if $self->{path} eq '';
2158 if (my $path = $paths->{"/$self->{path}"}) {
2159 return ($path->{action} eq 'D') ? 0 : 1;
2160 }
2161 $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
2162 if (grep /$self->{path_regex}/, keys %$paths) {
2163 return 1;
2164 }
2165 my $c = '';
2166 foreach (split m#/#, $self->{path}) {
2167 $c .= "/$_";
2168 next unless ($paths->{$c} &&
2169 ($paths->{$c}->{action} =~ /^[AR]$/));
2170 if ($self->ra->check_path($self->{path}, $r) ==
2171 $SVN::Node::dir) {
2172 return 1;
2173 }
2174 }
2175 return 0;
2176 }
2177
2178 sub find_parent_branch {
2179 my ($self, $paths, $rev) = @_;
2180 return undef unless $self->follow_parent;
2181 unless (defined $paths) {
2182 my $err_handler = $SVN::Error::handler;
2183 $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2184 $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1, sub {
2185 $paths =
2186 Git::SVN::Ra::dup_changed_paths($_[0]) });
2187 $SVN::Error::handler = $err_handler;
2188 }
2189 return undef unless defined $paths;
2190
2191 # look for a parent from another branch:
2192 my @b_path_components = split m#/#, $self->rel_path;
2193 my @a_path_components;
2194 my $i;
2195 while (@b_path_components) {
2196 $i = $paths->{'/'.join('/', @b_path_components)};
2197 last if $i && defined $i->{copyfrom_path};
2198 unshift(@a_path_components, pop(@b_path_components));
2199 }
2200 return undef unless defined $i && defined $i->{copyfrom_path};
2201 my $branch_from = $i->{copyfrom_path};
2202 if (@a_path_components) {
2203 print STDERR "branch_from: $branch_from => ";
2204 $branch_from .= '/'.join('/', @a_path_components);
2205 print STDERR $branch_from, "\n";
2206 }
2207 my $r = $i->{copyfrom_rev};
2208 my $repos_root = $self->ra->{repos_root};
2209 my $url = $self->ra->{url};
2210 my $new_url = $repos_root . $branch_from;
2211 print STDERR "Found possible branch point: ",
2212 "$new_url => ", $self->full_url, ", $r\n";
2213 $branch_from =~ s#^/##;
2214 my $gs = Git::SVN->find_by_url($new_url, $repos_root, $branch_from);
2215 unless ($gs) {
2216 my $ref_id = $self->{ref_id};
2217 $ref_id =~ s/\@\d+$//;
2218 $ref_id .= "\@$r";
2219 # just grow a tail if we're not unique enough :x
2220 $ref_id .= '-' while find_ref($ref_id);
2221 print STDERR "Initializing parent: $ref_id\n";
2222 $gs = Git::SVN->init($new_url, '', $ref_id, $ref_id, 1);
2223 }
2224 my ($r0, $parent) = $gs->find_rev_before($r, 1);
2225 if (!defined $r0 || !defined $parent) {
2226 my ($base, $head) = parse_revision_argument(0, $r);
2227 if ($base <= $r) {
2228 $gs->fetch($base, $r);
2229 }
2230 ($r0, $parent) = $gs->last_rev_commit;
2231 }
2232 if (defined $r0 && defined $parent) {
2233 print STDERR "Found branch parent: ($self->{ref_id}) $parent\n";
2234 my $ed;
2235 if ($self->ra->can_do_switch) {
2236 $self->assert_index_clean($parent);
2237 print STDERR "Following parent with do_switch\n";
2238 # do_switch works with svn/trunk >= r22312, but that
2239 # is not included with SVN 1.4.3 (the latest version
2240 # at the moment), so we can't rely on it
2241 $self->{last_commit} = $parent;
2242 $ed = SVN::Git::Fetcher->new($self);
2243 $gs->ra->gs_do_switch($r0, $rev, $gs,
2244 $self->full_url, $ed)
2245 or die "SVN connection failed somewhere...\n";
2246 } elsif ($self->ra->trees_match($new_url, $r0,
2247 $self->full_url, $rev)) {
2248 print STDERR "Trees match:\n",
2249 " $new_url\@$r0\n",
2250 " ${\$self->full_url}\@$rev\n",
2251 "Following parent with no changes\n";
2252 $self->tmp_index_do(sub {
2253 command_noisy('read-tree', $parent);
2254 });
2255 $self->{last_commit} = $parent;
2256 } else {
2257 print STDERR "Following parent with do_update\n";
2258 $ed = SVN::Git::Fetcher->new($self);
2259 $self->ra->gs_do_update($rev, $rev, $self, $ed)
2260 or die "SVN connection failed somewhere...\n";
2261 }
2262 print STDERR "Successfully followed parent\n";
2263 return $self->make_log_entry($rev, [$parent], $ed);
2264 }
2265 return undef;
2266 }
2267
2268 sub do_fetch {
2269 my ($self, $paths, $rev) = @_;
2270 my $ed;
2271 my ($last_rev, @parents);
2272 if (my $lc = $self->last_commit) {
2273 # we can have a branch that was deleted, then re-added
2274 # under the same name but copied from another path, in
2275 # which case we'll have multiple parents (we don't
2276 # want to break the original ref, nor lose copypath info):
2277 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2278 push @{$log_entry->{parents}}, $lc;
2279 return $log_entry;
2280 }
2281 $ed = SVN::Git::Fetcher->new($self);
2282 $last_rev = $self->{last_rev};
2283 $ed->{c} = $lc;
2284 @parents = ($lc);
2285 } else {
2286 $last_rev = $rev;
2287 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2288 return $log_entry;
2289 }
2290 $ed = SVN::Git::Fetcher->new($self);
2291 }
2292 unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2293 die "SVN connection failed somewhere...\n";
2294 }
2295 $self->make_log_entry($rev, \@parents, $ed);
2296 }
2297
2298 sub get_untracked {
2299 my ($self, $ed) = @_;
2300 my @out;
2301 my $h = $ed->{empty};
2302 foreach (sort keys %$h) {
2303 my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2304 push @out, " $act: " . uri_encode($_);
2305 warn "W: $act: $_\n";
2306 }
2307 foreach my $t (qw/dir_prop file_prop/) {
2308 $h = $ed->{$t} or next;
2309 foreach my $path (sort keys %$h) {
2310 my $ppath = $path eq '' ? '.' : $path;
2311 foreach my $prop (sort keys %{$h->{$path}}) {
2312 next if $SKIP_PROP{$prop};
2313 my $v = $h->{$path}->{$prop};
2314 my $t_ppath_prop = "$t: " .
2315 uri_encode($ppath) . ' ' .
2316 uri_encode($prop);
2317 if (defined $v) {
2318 push @out, " +$t_ppath_prop " .
2319 uri_encode($v);
2320 } else {
2321 push @out, " -$t_ppath_prop";
2322 }
2323 }
2324 }
2325 }
2326 foreach my $t (qw/absent_file absent_directory/) {
2327 $h = $ed->{$t} or next;
2328 foreach my $parent (sort keys %$h) {
2329 foreach my $path (sort @{$h->{$parent}}) {
2330 push @out, " $t: " .
2331 uri_encode("$parent/$path");
2332 warn "W: $t: $parent/$path ",
2333 "Insufficient permissions?\n";
2334 }
2335 }
2336 }
2337 \@out;
2338 }
2339
2340 sub parse_svn_date {
2341 my $date = shift || return '+0000 1970-01-01 00:00:00';
2342 my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2343 (\d\d)\:(\d\d)\:(\d\d).\d+Z$/x) or
2344 croak "Unable to parse date: $date\n";
2345 "+0000 $Y-$m-$d $H:$M:$S";
2346 }
2347
2348 sub check_author {
2349 my ($author) = @_;
2350 if (!defined $author || length $author == 0) {
2351 $author = '(no author)';
2352 }
2353 if (defined $::_authors && ! defined $::users{$author}) {
2354 die "Author: $author not defined in $::_authors file\n";
2355 }
2356 $author;
2357 }
2358
2359 sub make_log_entry {
2360 my ($self, $rev, $parents, $ed) = @_;
2361 my $untracked = $self->get_untracked($ed);
2362
2363 open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
2364 print $un "r$rev\n" or croak $!;
2365 print $un $_, "\n" foreach @$untracked;
2366 my %log_entry = ( parents => $parents || [], revision => $rev,
2367 log => '');
2368
2369 my $headrev;
2370 my $logged = delete $self->{logged_rev_props};
2371 if (!$logged || $self->{-want_revprops}) {
2372 my $rp = $self->ra->rev_proplist($rev);
2373 foreach (sort keys %$rp) {
2374 my $v = $rp->{$_};
2375 if (/^svn:(author|date|log)$/) {
2376 $log_entry{$1} = $v;
2377 } elsif ($_ eq 'svm:headrev') {
2378 $headrev = $v;
2379 } else {
2380 print $un " rev_prop: ", uri_encode($_), ' ',
2381 uri_encode($v), "\n";
2382 }
2383 }
2384 } else {
2385 map { $log_entry{$_} = $logged->{$_} } keys %$logged;
2386 }
2387 close $un or croak $!;
2388
2389 $log_entry{date} = parse_svn_date($log_entry{date});
2390 $log_entry{log} .= "\n";
2391 my $author = $log_entry{author} = check_author($log_entry{author});
2392 my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
2393 : ($author, undef);
2394
2395 my ($commit_name, $commit_email) = ($name, $email);
2396 if ($_use_log_author) {
2397 my $name_field;
2398 if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
2399 $name_field = $1;
2400 } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
2401 $name_field = $1;
2402 }
2403 if (!defined $name_field) {
2404 #
2405 } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
2406 ($name, $email) = ($1, $2);
2407 } elsif ($name_field =~ /(.*)@/) {
2408 ($name, $email) = ($1, $name_field);
2409 } else {
2410 ($name, $email) = ($name_field, 'unknown');
2411 }
2412 }
2413 if (defined $headrev && $self->use_svm_props) {
2414 if ($self->rewrite_root) {
2415 die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
2416 "options set!\n";
2417 }
2418 my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$};
2419 # we don't want "SVM: initializing mirror for junk" ...
2420 return undef if $r == 0;
2421 my $svm = $self->svm;
2422 if ($uuid ne $svm->{uuid}) {
2423 die "UUID mismatch on SVM path:\n",
2424 "expected: $svm->{uuid}\n",
2425 " got: $uuid\n";
2426 }
2427 my $full_url = $self->full_url;
2428 $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
2429 die "Failed to replace '$svm->{replace}' with ",
2430 "'$svm->{source}' in $full_url\n";
2431 # throw away username for storing in records
2432 remove_username($full_url);
2433 $log_entry{metadata} = "$full_url\@$r $uuid";
2434 $log_entry{svm_revision} = $r;
2435 $email ||= "$author\@$uuid";
2436 $commit_email ||= "$author\@$uuid";
2437 } elsif ($self->use_svnsync_props) {
2438 my $full_url = $self->svnsync->{url};
2439 $full_url .= "/$self->{path}" if length $self->{path};
2440 remove_username($full_url);
2441 my $uuid = $self->svnsync->{uuid};
2442 $log_entry{metadata} = "$full_url\@$rev $uuid";
2443 $email ||= "$author\@$uuid";
2444 $commit_email ||= "$author\@$uuid";
2445 } else {
2446 my $url = $self->metadata_url;
2447 remove_username($url);
2448 $log_entry{metadata} = "$url\@$rev " .
2449 $self->ra->get_uuid;
2450 $email ||= "$author\@" . $self->ra->get_uuid;
2451 $commit_email ||= "$author\@" . $self->ra->get_uuid;
2452 }
2453 $log_entry{name} = $name;
2454 $log_entry{email} = $email;
2455 $log_entry{commit_name} = $commit_name;
2456 $log_entry{commit_email} = $commit_email;
2457 \%log_entry;
2458 }
2459
2460 sub fetch {
2461 my ($self, $min_rev, $max_rev, @parents) = @_;
2462 my ($last_rev, $last_commit) = $self->last_rev_commit;
2463 my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
2464 $self->ra->gs_fetch_loop_common($base, $head, [$self]);
2465 }
2466
2467 sub set_tree_cb {
2468 my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
2469 $self->{inject_parents} = { $rev => $tree };
2470 $self->fetch(undef, undef);
2471 }
2472
2473 sub set_tree {
2474 my ($self, $tree) = (shift, shift);
2475 my $log_entry = ::get_commit_entry($tree);
2476 unless ($self->{last_rev}) {
2477 fatal("Must have an existing revision to commit");
2478 }
2479 my %ed_opts = ( r => $self->{last_rev},
2480 log => $log_entry->{log},
2481 ra => $self->ra,
2482 tree_a => $self->{last_commit},
2483 tree_b => $tree,
2484 editor_cb => sub {
2485 $self->set_tree_cb($log_entry, $tree, @_) },
2486 svn_path => $self->{path} );
2487 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
2488 print "No changes\nr$self->{last_rev} = $tree\n";
2489 }
2490 }
2491
2492 sub rebuild_from_rev_db {
2493 my ($self, $path) = @_;
2494 my $r = -1;
2495 open my $fh, '<', $path or croak "open: $!";
2496 while (<$fh>) {
2497 length($_) == 41 or croak "inconsistent size in ($_) != 41";
2498 chomp($_);
2499 ++$r;
2500 next if $_ eq ('0' x 40);
2501 $self->rev_map_set($r, $_);
2502 print "r$r = $_\n";
2503 }
2504 close $fh or croak "close: $!";
2505 unlink $path or croak "unlink: $!";
2506 }
2507
2508 sub rebuild {
2509 my ($self) = @_;
2510 my $map_path = $self->map_path;
2511 return if (-e $map_path && ! -z $map_path);
2512 return unless ::verify_ref($self->refname.'^0');
2513 if ($self->use_svm_props || $self->no_metadata) {
2514 my $rev_db = $self->rev_db_path;
2515 $self->rebuild_from_rev_db($rev_db);
2516 if ($self->use_svm_props) {
2517 my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
2518 $self->rebuild_from_rev_db($svm_rev_db);
2519 }
2520 $self->unlink_rev_db_symlink;
2521 return;
2522 }
2523 print "Rebuilding $map_path ...\n";
2524 my ($log, $ctx) =
2525 command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
2526 $self->refname, '--');
2527 my $full_url = $self->full_url;
2528 remove_username($full_url);
2529 my $svn_uuid = $self->ra_uuid;
2530 my $c;
2531 while (<$log>) {
2532 if ( m{^commit ($::sha1)$} ) {
2533 $c = $1;
2534 next;
2535 }
2536 next unless s{^\s*(git-svn-id:)}{$1};
2537 my ($url, $rev, $uuid) = ::extract_metadata($_);
2538 remove_username($url);
2539
2540 # ignore merges (from set-tree)
2541 next if (!defined $rev || !$uuid);
2542
2543 # if we merged or otherwise started elsewhere, this is
2544 # how we break out of it
2545 if (($uuid ne $svn_uuid) ||
2546 ($full_url && $url && ($url ne $full_url))) {
2547 next;
2548 }
2549
2550 $self->rev_map_set($rev, $c);
2551 print "r$rev = $c\n";
2552 }
2553 command_close_pipe($log, $ctx);
2554 print "Done rebuilding $map_path\n";
2555 my $rev_db_path = $self->rev_db_path;
2556 if (-f $self->rev_db_path) {
2557 unlink $self->rev_db_path or croak "unlink: $!";
2558 }
2559 $self->unlink_rev_db_symlink;
2560 }
2561
2562 # rev_map:
2563 # Tie::File seems to be prone to offset errors if revisions get sparse,
2564 # it's not that fast, either. Tie::File is also not in Perl 5.6. So
2565 # one of my favorite modules is out :< Next up would be one of the DBM
2566 # modules, but I'm not sure which is most portable...
2567 #
2568 # This is the replacement for the rev_db format, which was too big
2569 # and inefficient for large repositories with a lot of sparse history
2570 # (mainly tags)
2571 #
2572 # The format is this:
2573 # - 24 bytes for every record,
2574 # * 4 bytes for the integer representing an SVN revision number
2575 # * 20 bytes representing the sha1 of a git commit
2576 # - No empty padding records like the old format
2577 # (except the last record, which can be overwritten)
2578 # - new records are written append-only since SVN revision numbers
2579 # increase monotonically
2580 # - lookups on SVN revision number are done via a binary search
2581 # - Piping the file to xxd -c24 is a good way of dumping it for
2582 # viewing or editing (piped back through xxd -r), should the need
2583 # ever arise.
2584 # - The last record can be padding revision with an all-zero sha1
2585 # This is used to optimize fetch performance when using multiple
2586 # "fetch" directives in .git/config
2587 #
2588 # These files are disposable unless noMetadata or useSvmProps is set
2589
2590 sub _rev_map_set {
2591 my ($fh, $rev, $commit) = @_;
2592
2593 my $size = (stat($fh))[7];
2594 ($size % 24) == 0 or croak "inconsistent size: $size";
2595
2596 my $wr_offset = 0;
2597 if ($size > 0) {
2598 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2599 my $read = sysread($fh, my $buf, 24) or croak "read: $!";
2600 $read == 24 or croak "read only $read bytes (!= 24)";
2601 my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
2602 if ($last_commit eq ('0' x40)) {
2603 if ($size >= 48) {
2604 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2605 $read = sysread($fh, $buf, 24) or
2606 croak "read: $!";
2607 $read == 24 or
2608 croak "read only $read bytes (!= 24)";
2609 ($last_rev, $last_commit) =
2610 unpack(rev_map_fmt, $buf);
2611 if ($last_commit eq ('0' x40)) {
2612 croak "inconsistent .rev_map\n";
2613 }
2614 }
2615 if ($last_rev >= $rev) {
2616 croak "last_rev is higher!: $last_rev >= $rev";
2617 }
2618 $wr_offset = -24;
2619 }
2620 }
2621 sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
2622 syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
2623 croak "write: $!";
2624 }
2625
2626 sub mkfile {
2627 my ($path) = @_;
2628 unless (-e $path) {
2629 my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
2630 mkpath([$dir]) unless -d $dir;
2631 open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
2632 close $fh or die "Couldn't close (create) $path: $!\n";
2633 }
2634 }
2635
2636 sub rev_map_set {
2637 my ($self, $rev, $commit, $update_ref, $uuid) = @_;
2638 length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
2639 my $db = $self->map_path($uuid);
2640 my $db_lock = "$db.lock";
2641 my $sig;
2642 if ($update_ref) {
2643 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2644 $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
2645 }
2646 mkfile($db);
2647
2648 $LOCKFILES{$db_lock} = 1;
2649 my $sync;
2650 # both of these options make our .rev_db file very, very important
2651 # and we can't afford to lose it because rebuild() won't work
2652 if ($self->use_svm_props || $self->no_metadata) {
2653 $sync = 1;
2654 copy($db, $db_lock) or die "rev_map_set(@_): ",
2655 "Failed to copy: ",
2656 "$db => $db_lock ($!)\n";
2657 } else {
2658 rename $db, $db_lock or die "rev_map_set(@_): ",
2659 "Failed to rename: ",
2660 "$db => $db_lock ($!)\n";
2661 }
2662
2663 sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
2664 or croak "Couldn't open $db_lock: $!\n";
2665 _rev_map_set($fh, $rev, $commit);
2666 if ($sync) {
2667 $fh->flush or die "Couldn't flush $db_lock: $!\n";
2668 $fh->sync or die "Couldn't sync $db_lock: $!\n";
2669 }
2670 close $fh or croak $!;
2671 if ($update_ref) {
2672 $_head = $self;
2673 command_noisy('update-ref', '-m', "r$rev",
2674 $self->refname, $commit);
2675 }
2676 rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
2677 "$db_lock => $db ($!)\n";
2678 delete $LOCKFILES{$db_lock};
2679 if ($update_ref) {
2680 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2681 $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
2682 kill $sig, $$ if defined $sig;
2683 }
2684 }
2685
2686 # If want_commit, this will return an array of (rev, commit) where
2687 # commit _must_ be a valid commit in the archive.
2688 # Otherwise, it'll return the max revision (whether or not the
2689 # commit is valid or just a 0x40 placeholder).
2690 sub rev_map_max {
2691 my ($self, $want_commit) = @_;
2692 $self->rebuild;
2693 my $map_path = $self->map_path;
2694 stat $map_path or return $want_commit ? (0, undef) : 0;
2695 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2696 my $size = (stat($fh))[7];
2697 ($size % 24) == 0 or croak "inconsistent size: $size";
2698
2699 if ($size == 0) {
2700 close $fh or croak "close: $!";
2701 return $want_commit ? (0, undef) : 0;
2702 }
2703
2704 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2705 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2706 my ($r, $c) = unpack(rev_map_fmt, $buf);
2707 if ($want_commit && $c eq ('0' x40)) {
2708 if ($size < 48) {
2709 return $want_commit ? (0, undef) : 0;
2710 }
2711 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2712 sysread($fh, $buf, 24) == 24 or croak "read: $!";
2713 ($r, $c) = unpack(rev_map_fmt, $buf);
2714 if ($c eq ('0'x40)) {
2715 croak "Penultimate record is all-zeroes in $map_path";
2716 }
2717 }
2718 close $fh or croak "close: $!";
2719 $want_commit ? ($r, $c) : $r;
2720 }
2721
2722 sub rev_map_get {
2723 my ($self, $rev, $uuid) = @_;
2724 my $map_path = $self->map_path($uuid);
2725 return undef unless -e $map_path;
2726
2727 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2728 my $size = (stat($fh))[7];
2729 ($size % 24) == 0 or croak "inconsistent size: $size";
2730
2731 if ($size == 0) {
2732 close $fh or croak "close: $fh";
2733 return undef;
2734 }
2735
2736 my ($l, $u) = (0, $size - 24);
2737 my ($r, $c, $buf);
2738
2739 while ($l <= $u) {
2740 my $i = int(($l/24 + $u/24) / 2) * 24;
2741 sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
2742 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2743 my ($r, $c) = unpack('NH40', $buf);
2744
2745 if ($r < $rev) {
2746 $l = $i + 24;
2747 } elsif ($r > $rev) {
2748 $u = $i - 24;
2749 } else { # $r == $rev
2750 close($fh) or croak "close: $!";
2751 return $c eq ('0' x 40) ? undef : $c;
2752 }
2753 }
2754 close($fh) or croak "close: $!";
2755 undef;
2756 }
2757
2758 # Finds the first svn revision that exists on (if $eq_ok is true) or
2759 # before $rev for the current branch. It will not search any lower
2760 # than $min_rev. Returns the git commit hash and svn revision number
2761 # if found, else (undef, undef).
2762 sub find_rev_before {
2763 my ($self, $rev, $eq_ok, $min_rev) = @_;
2764 --$rev unless $eq_ok;
2765 $min_rev ||= 1;
2766 while ($rev >= $min_rev) {
2767 if (my $c = $self->rev_map_get($rev)) {
2768 return ($rev, $c);
2769 }
2770 --$rev;
2771 }
2772 return (undef, undef);
2773 }
2774
2775 # Finds the first svn revision that exists on (if $eq_ok is true) or
2776 # after $rev for the current branch. It will not search any higher
2777 # than $max_rev. Returns the git commit hash and svn revision number
2778 # if found, else (undef, undef).
2779 sub find_rev_after {
2780 my ($self, $rev, $eq_ok, $max_rev) = @_;
2781 ++$rev unless $eq_ok;
2782 $max_rev ||= $self->rev_map_max;
2783 while ($rev <= $max_rev) {
2784 if (my $c = $self->rev_map_get($rev)) {
2785 return ($rev, $c);
2786 }
2787 ++$rev;
2788 }
2789 return (undef, undef);
2790 }
2791
2792 sub _new {
2793 my ($class, $repo_id, $ref_id, $path) = @_;
2794 unless (defined $repo_id && length $repo_id) {
2795 $repo_id = $Git::SVN::default_repo_id;
2796 }
2797 unless (defined $ref_id && length $ref_id) {
2798 $_[2] = $ref_id = $Git::SVN::default_ref_id;
2799 }
2800 $_[1] = $repo_id = sanitize_remote_name($repo_id);
2801 my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
2802 $_[3] = $path = '' unless (defined $path);
2803 mkpath(["$ENV{GIT_DIR}/svn"]);
2804 bless {
2805 ref_id => $ref_id, dir => $dir, index => "$dir/index",
2806 path => $path, config => "$ENV{GIT_DIR}/svn/config",
2807 map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
2808 }
2809
2810 # for read-only access of old .rev_db formats
2811 sub unlink_rev_db_symlink {
2812 my ($self) = @_;
2813 my $link = $self->rev_db_path;
2814 $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
2815 if (-l $link) {
2816 unlink $link or croak "unlink: $link failed!";
2817 }
2818 }
2819
2820 sub rev_db_path {
2821 my ($self, $uuid) = @_;
2822 my $db_path = $self->map_path($uuid);
2823 $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
2824 or croak "map_path: $db_path does not contain '/.rev_map.' !";
2825 $db_path;
2826 }
2827
2828 # the new replacement for .rev_db
2829 sub map_path {
2830 my ($self, $uuid) = @_;
2831 $uuid ||= $self->ra_uuid;
2832 "$self->{map_root}.$uuid";
2833 }
2834
2835 sub uri_encode {
2836 my ($f) = @_;
2837 $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
2838 $f
2839 }
2840
2841 sub remove_username {
2842 $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
2843 }
2844
2845 package Git::SVN::Prompt;
2846 use strict;
2847 use warnings;
2848 require SVN::Core;
2849 use vars qw/$_no_auth_cache $_username/;
2850
2851 sub simple {
2852 my ($cred, $realm, $default_username, $may_save, $pool) = @_;
2853 $may_save = undef if $_no_auth_cache;
2854 $default_username = $_username if defined $_username;
2855 if (defined $default_username && length $default_username) {
2856 if (defined $realm && length $realm) {
2857 print STDERR "Authentication realm: $realm\n";
2858 STDERR->flush;
2859 }
2860 $cred->username($default_username);
2861 } else {
2862 username($cred, $realm, $may_save, $pool);
2863 }
2864 $cred->password(_read_password("Password for '" .
2865 $cred->username . "': ", $realm));
2866 $cred->may_save($may_save);
2867 $SVN::_Core::SVN_NO_ERROR;
2868 }
2869
2870 sub ssl_server_trust {
2871 my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
2872 $may_save = undef if $_no_auth_cache;
2873 print STDERR "Error validating server certificate for '$realm':\n";
2874 {
2875 no warnings 'once';
2876 # All variables SVN::Auth::SSL::* are used only once,
2877 # so we're shutting up Perl warnings about this.
2878 if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
2879 print STDERR " - The certificate is not issued ",
2880 "by a trusted authority. Use the\n",
2881 " fingerprint to validate ",
2882 "the certificate manually!\n";
2883 }
2884 if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
2885 print STDERR " - The certificate hostname ",
2886 "does not match.\n";
2887 }
2888 if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
2889 print STDERR " - The certificate is not yet valid.\n";
2890 }
2891 if ($failures & $SVN::Auth::SSL::EXPIRED) {
2892 print STDERR " - The certificate has expired.\n";
2893 }
2894 if ($failures & $SVN::Auth::SSL::OTHER) {
2895 print STDERR " - The certificate has ",
2896 "an unknown error.\n";
2897 }
2898 } # no warnings 'once'
2899 printf STDERR
2900 "Certificate information:\n".
2901 " - Hostname: %s\n".
2902 " - Valid: from %s until %s\n".
2903 " - Issuer: %s\n".
2904 " - Fingerprint: %s\n",
2905 map $cert_info->$_, qw(hostname valid_from valid_until
2906 issuer_dname fingerprint);
2907 my $choice;
2908 prompt:
2909 print STDERR $may_save ?
2910 "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
2911 "(R)eject or accept (t)emporarily? ";
2912 STDERR->flush;
2913 $choice = lc(substr(<STDIN> || 'R', 0, 1));
2914 if ($choice =~ /^t$/i) {
2915 $cred->may_save(undef);
2916 } elsif ($choice =~ /^r$/i) {
2917 return -1;
2918 } elsif ($may_save && $choice =~ /^p$/i) {
2919 $cred->may_save($may_save);
2920 } else {
2921 goto prompt;
2922 }
2923 $cred->accepted_failures($failures);
2924 $SVN::_Core::SVN_NO_ERROR;
2925 }
2926
2927 sub ssl_client_cert {
2928 my ($cred, $realm, $may_save, $pool) = @_;
2929 $may_save = undef if $_no_auth_cache;
2930 print STDERR "Client certificate filename: ";
2931 STDERR->flush;
2932 chomp(my $filename = <STDIN>);
2933 $cred->cert_file($filename);
2934 $cred->may_save($may_save);
2935 $SVN::_Core::SVN_NO_ERROR;
2936 }
2937
2938 sub ssl_client_cert_pw {
2939 my ($cred, $realm, $may_save, $pool) = @_;
2940 $may_save = undef if $_no_auth_cache;
2941 $cred->password(_read_password("Password: ", $realm));
2942 $cred->may_save($may_save);
2943 $SVN::_Core::SVN_NO_ERROR;
2944 }
2945
2946 sub username {
2947 my ($cred, $realm, $may_save, $pool) = @_;
2948 $may_save = undef if $_no_auth_cache;
2949 if (defined $realm && length $realm) {
2950 print STDERR "Authentication realm: $realm\n";
2951 }
2952 my $username;
2953 if (defined $_username) {
2954 $username = $_username;
2955 } else {
2956 print STDERR "Username: ";
2957 STDERR->flush;
2958 chomp($username = <STDIN>);
2959 }
2960 $cred->username($username);
2961 $cred->may_save($may_save);
2962 $SVN::_Core::SVN_NO_ERROR;
2963 }
2964
2965 sub _read_password {
2966 my ($prompt, $realm) = @_;
2967 print STDERR $prompt;
2968 STDERR->flush;
2969 require Term::ReadKey;
2970 Term::ReadKey::ReadMode('noecho');
2971 my $password = '';
2972 while (defined(my $key = Term::ReadKey::ReadKey(0))) {
2973 last if $key =~ /[\012\015]/; # \n\r
2974 $password .= $key;
2975 }
2976 Term::ReadKey::ReadMode('restore');
2977 print STDERR "\n";
2978 STDERR->flush;
2979 $password;
2980 }
2981
2982 package SVN::Git::Fetcher;
2983 use vars qw/@ISA/;
2984 use strict;
2985 use warnings;
2986 use Carp qw/croak/;
2987 use IO::File qw//;
2988
2989 # file baton members: path, mode_a, mode_b, pool, fh, blob, base
2990 sub new {
2991 my ($class, $git_svn) = @_;
2992 my $self = SVN::Delta::Editor->new;
2993 bless $self, $class;
2994 $self->{c} = $git_svn->{last_commit} if exists $git_svn->{last_commit};
2995 $self->{empty} = {};
2996 $self->{dir_prop} = {};
2997 $self->{file_prop} = {};
2998 $self->{absent_dir} = {};
2999 $self->{absent_file} = {};
3000 $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
3001 $self;
3002 }
3003
3004 sub set_path_strip {
3005 my ($self, $path) = @_;
3006 $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
3007 }
3008
3009 sub open_root {
3010 { path => '' };
3011 }
3012
3013 sub open_directory {
3014 my ($self, $path, $pb, $rev) = @_;
3015 { path => $path };
3016 }
3017
3018 sub git_path {
3019 my ($self, $path) = @_;
3020 if ($self->{path_strip}) {
3021 $path =~ s!$self->{path_strip}!! or
3022 die "Failed to strip path '$path' ($self->{path_strip})\n";
3023 }
3024 $path;
3025 }
3026
3027 sub delete_entry {
3028 my ($self, $path, $rev, $pb) = @_;
3029
3030 my $gpath = $self->git_path($path);
3031 return undef if ($gpath eq '');
3032
3033 # remove entire directories.
3034 if (command('ls-tree', $self->{c}, '--', $gpath) =~ /^040000 tree/) {
3035 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3036 -r --name-only -z/,
3037 $self->{c}, '--', $gpath);
3038 local $/ = "\0";
3039 while (<$ls>) {
3040 chomp;
3041 $self->{gii}->remove($_);
3042 print "\tD\t$_\n" unless $::_q;
3043 }
3044 print "\tD\t$gpath/\n" unless $::_q;
3045 command_close_pipe($ls, $ctx);
3046 $self->{empty}->{$path} = 0
3047 } else {
3048 $self->{gii}->remove($gpath);
3049 print "\tD\t$gpath\n" unless $::_q;
3050 }
3051 undef;
3052 }
3053
3054 sub open_file {
3055 my ($self, $path, $pb, $rev) = @_;
3056 my $gpath = $self->git_path($path);
3057 my ($mode, $blob) = (command('ls-tree', $self->{c}, '--', $gpath)
3058 =~ /^(\d{6}) blob ([a-f\d]{40})\t/);
3059 unless (defined $mode && defined $blob) {
3060 die "$path was not found in commit $self->{c} (r$rev)\n";
3061 }
3062 { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3063 pool => SVN::Pool->new, action => 'M' };
3064 }
3065
3066 sub add_file {
3067 my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3068 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3069 delete $self->{empty}->{$dir};
3070 { path => $path, mode_a => 100644, mode_b => 100644,
3071 pool => SVN::Pool->new, action => 'A' };
3072 }
3073
3074 sub add_directory {
3075 my ($self, $path, $cp_path, $cp_rev) = @_;
3076 my $gpath = $self->git_path($path);
3077 if ($gpath eq '') {
3078 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3079 -r --name-only -z/,
3080 $self->{c});
3081 local $/ = "\0";
3082 while (<$ls>) {
3083 chomp;
3084 $self->{gii}->remove($_);
3085 print "\tD\t$_\n" unless $::_q;
3086 }
3087 command_close_pipe($ls, $ctx);
3088 $self->{empty}->{$path} = 0;
3089 }
3090 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3091 delete $self->{empty}->{$dir};
3092 $self->{empty}->{$path} = 1;
3093 { path => $path };
3094 }
3095
3096 sub change_dir_prop {
3097 my ($self, $db, $prop, $value) = @_;
3098 $self->{dir_prop}->{$db->{path}} ||= {};
3099 $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3100 undef;
3101 }
3102
3103 sub absent_directory {
3104 my ($self, $path, $pb) = @_;
3105 $self->{absent_dir}->{$pb->{path}} ||= [];
3106 push @{$self->{absent_dir}->{$pb->{path}}}, $path;
3107 undef;
3108 }
3109
3110 sub absent_file {
3111 my ($self, $path, $pb) = @_;
3112 $self->{absent_file}->{$pb->{path}} ||= [];
3113 push @{$self->{absent_file}->{$pb->{path}}}, $path;
3114 undef;
3115 }
3116
3117 sub change_file_prop {
3118 my ($self, $fb, $prop, $value) = @_;
3119 if ($prop eq 'svn:executable') {
3120 if ($fb->{mode_b} != 120000) {
3121 $fb->{mode_b} = defined $value ? 100755 : 100644;
3122 }
3123 } elsif ($prop eq 'svn:special') {
3124 $fb->{mode_b} = defined $value ? 120000 : 100644;
3125 } else {
3126 $self->{file_prop}->{$fb->{path}} ||= {};
3127 $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
3128 }
3129 undef;
3130 }
3131
3132 sub apply_textdelta {
3133 my ($self, $fb, $exp) = @_;
3134 my $fh = IO::File->new_tmpfile;
3135 $fh->autoflush(1);
3136 # $fh gets auto-closed() by SVN::TxDelta::apply(),
3137 # (but $base does not,) so dup() it for reading in close_file
3138 open my $dup, '<&', $fh or croak $!;
3139 my $base = IO::File->new_tmpfile;
3140 $base->autoflush(1);
3141 if ($fb->{blob}) {
3142 defined (my $pid = fork) or croak $!;
3143 if (!$pid) {
3144 open STDOUT, '>&', $base or croak $!;
3145 print STDOUT 'link ' if ($fb->{mode_a} == 120000);
3146 exec qw/git-cat-file blob/, $fb->{blob} or croak $!;
3147 }
3148 waitpid $pid, 0;
3149 croak $? if $?;
3150
3151 if (defined $exp) {
3152 seek $base, 0, 0 or croak $!;
3153 my $got = ::md5sum($base);
3154 die "Checksum mismatch: $fb->{path} $fb->{blob}\n",
3155 "expected: $exp\n",
3156 " got: $got\n" if ($got ne $exp);
3157 }
3158 }
3159 seek $base, 0, 0 or croak $!;
3160 $fb->{fh} = $dup;
3161 $fb->{base} = $base;
3162 [ SVN::TxDelta::apply($base, $fh, undef, $fb->{path}, $fb->{pool}) ];
3163 }
3164
3165 sub close_file {
3166 my ($self, $fb, $exp) = @_;
3167 my $hash;
3168 my $path = $self->git_path($fb->{path});
3169 if (my $fh = $fb->{fh}) {
3170 if (defined $exp) {
3171 seek($fh, 0, 0) or croak $!;
3172 my $got = ::md5sum($fh);
3173 if ($got ne $exp) {
3174 die "Checksum mismatch: $path\n",
3175 "expected: $exp\n got: $got\n";
3176 }
3177 }
3178 sysseek($fh, 0, 0) or croak $!;
3179 if ($fb->{mode_b} == 120000) {
3180 sysread($fh, my $buf, 5) == 5 or croak $!;
3181 $buf eq 'link ' or die "$path has mode 120000",
3182 "but is not a link\n";
3183 }
3184 defined(my $pid = open my $out,'-|') or die "Can't fork: $!\n";
3185 if (!$pid) {
3186 open STDIN, '<&', $fh or croak $!;
3187 exec qw/git-hash-object -w --stdin/ or croak $!;
3188 }
3189 chomp($hash = do { local $/; <$out> });
3190 close $out or croak $!;
3191 close $fh or croak $!;
3192 $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
3193 close $fb->{base} or croak $!;
3194 } else {
3195 $hash = $fb->{blob} or die "no blob information\n";
3196 }
3197 $fb->{pool}->clear;
3198 $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
3199 print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
3200 undef;
3201 }
3202
3203 sub abort_edit {
3204 my $self = shift;
3205 $self->{nr} = $self->{gii}->{nr};
3206 delete $self->{gii};
3207 $self->SUPER::abort_edit(@_);
3208 }
3209
3210 sub close_edit {
3211 my $self = shift;
3212 $self->{git_commit_ok} = 1;
3213 $self->{nr} = $self->{gii}->{nr};
3214 delete $self->{gii};
3215 $self->SUPER::close_edit(@_);
3216 }
3217
3218 package SVN::Git::Editor;
3219 use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
3220 use strict;
3221 use warnings;
3222 use Carp qw/croak/;
3223 use IO::File;
3224
3225 sub new {
3226 my ($class, $opts) = @_;
3227 foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
3228 die "$_ required!\n" unless (defined $opts->{$_});
3229 }
3230
3231 my $pool = SVN::Pool->new;
3232 my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
3233 my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
3234 $opts->{r}, $mods);
3235
3236 # $opts->{ra} functions should not be used after this:
3237 my @ce = $opts->{ra}->get_commit_editor($opts->{log},
3238 $opts->{editor_cb}, $pool);
3239 my $self = SVN::Delta::Editor->new(@ce, $pool);
3240 bless $self, $class;
3241 foreach (qw/svn_path r tree_a tree_b/) {
3242 $self->{$_} = $opts->{$_};
3243 }
3244 $self->{url} = $opts->{ra}->{url};
3245 $self->{mods} = $mods;
3246 $self->{types} = $types;
3247 $self->{pool} = $pool;
3248 $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
3249 $self->{rm} = { };
3250 $self->{path_prefix} = length $self->{svn_path} ?
3251 "$self->{svn_path}/" : '';
3252 return $self;
3253 }
3254
3255 sub generate_diff {
3256 my ($tree_a, $tree_b) = @_;
3257 my @diff_tree = qw(diff-tree -z -r);
3258 if ($_cp_similarity) {
3259 push @diff_tree, "-C$_cp_similarity";
3260 } else {
3261 push @diff_tree, '-C';
3262 }
3263 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
3264 push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
3265 push @diff_tree, $tree_a, $tree_b;
3266 my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
3267 local $/ = "\0";
3268 my $state = 'meta';
3269 my @mods;
3270 while (<$diff_fh>) {
3271 chomp $_; # this gets rid of the trailing "\0"
3272 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
3273 $::sha1\s($::sha1)\s
3274 ([MTCRAD])\d*$/xo) {
3275 push @mods, { mode_a => $1, mode_b => $2,
3276 sha1_b => $3, chg => $4 };
3277 if ($4 =~ /^(?:C|R)$/) {
3278 $state = 'file_a';
3279 } else {
3280 $state = 'file_b';
3281 }
3282 } elsif ($state eq 'file_a') {
3283 my $x = $mods[$#mods] or croak "Empty array\n";
3284 if ($x->{chg} !~ /^(?:C|R)$/) {
3285 croak "Error parsing $_, $x->{chg}\n";
3286 }
3287 $x->{file_a} = $_;
3288 $state = 'file_b';
3289 } elsif ($state eq 'file_b') {
3290 my $x = $mods[$#mods] or croak "Empty array\n";
3291 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
3292 croak "Error parsing $_, $x->{chg}\n";
3293 }
3294 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
3295 croak "Error parsing $_, $x->{chg}\n";
3296 }
3297 $x->{file_b} = $_;
3298 $state = 'meta';
3299 } else {
3300 croak "Error parsing $_\n";
3301 }
3302 }
3303 command_close_pipe($diff_fh, $ctx);
3304 \@mods;
3305 }
3306
3307 sub check_diff_paths {
3308 my ($ra, $pfx, $rev, $mods) = @_;
3309 my %types;
3310 $pfx .= '/' if length $pfx;
3311
3312 sub type_diff_paths {
3313 my ($ra, $types, $path, $rev) = @_;
3314 my @p = split m#/+#, $path;
3315 my $c = shift @p;
3316 unless (defined $types->{$c}) {
3317 $types->{$c} = $ra->check_path($c, $rev);
3318 }
3319 while (@p) {
3320 $c .= '/' . shift @p;
3321 next if defined $types->{$c};
3322 $types->{$c} = $ra->check_path($c, $rev);
3323 }
3324 }
3325
3326 foreach my $m (@$mods) {
3327 foreach my $f (qw/file_a file_b/) {
3328 next unless defined $m->{$f};
3329 my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
3330 if (length $pfx.$dir && ! defined $types{$dir}) {
3331 type_diff_paths($ra, \%types, $pfx.$dir, $rev);
3332 }
3333 }
3334 }
3335 \%types;
3336 }
3337
3338 sub split_path {
3339 return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
3340 }
3341
3342 sub repo_path {
3343 my ($self, $path) = @_;
3344 $self->{path_prefix}.(defined $path ? $path : '');
3345 }
3346
3347 sub url_path {
3348 my ($self, $path) = @_;
3349 if ($self->{url} =~ m#^https?://#) {
3350 $path =~ s/([^a-zA-Z0-9_.-])/uc sprintf("%%%02x",ord($1))/eg;
3351 }
3352 $self->{url} . '/' . $self->repo_path($path);
3353 }
3354
3355 sub rmdirs {
3356 my ($self) = @_;
3357 my $rm = $self->{rm};
3358 delete $rm->{''}; # we never delete the url we're tracking
3359 return unless %$rm;
3360
3361 foreach (keys %$rm) {
3362 my @d = split m#/#, $_;
3363 my $c = shift @d;
3364 $rm->{$c} = 1;
3365 while (@d) {
3366 $c .= '/' . shift @d;
3367 $rm->{$c} = 1;
3368 }
3369 }
3370 delete $rm->{$self->{svn_path}};
3371 delete $rm->{''}; # we never delete the url we're tracking
3372 return unless %$rm;
3373
3374 my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
3375 $self->{tree_b});
3376 local $/ = "\0";
3377 while (<$fh>) {
3378 chomp;
3379 my @dn = split m#/#, $_;
3380 while (pop @dn) {
3381 delete $rm->{join '/', @dn};
3382 }
3383 unless (%$rm) {
3384 close $fh;
3385 return;
3386 }
3387 }
3388 command_close_pipe($fh, $ctx);
3389
3390 my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
3391 foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
3392 $self->close_directory($bat->{$d}, $p);
3393 my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
3394 print "\tD+\t$d/\n" unless $::_q;
3395 $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
3396 delete $bat->{$d};
3397 }
3398 }
3399
3400 sub open_or_add_dir {
3401 my ($self, $full_path, $baton) = @_;
3402 my $t = $self->{types}->{$full_path};
3403 if (!defined $t) {
3404 die "$full_path not known in r$self->{r} or we have a bug!\n";
3405 }
3406 {
3407 no warnings 'once';
3408 # SVN::Node::none and SVN::Node::file are used only once,
3409 # so we're shutting up Perl's warnings about them.
3410 if ($t == $SVN::Node::none) {
3411 return $self->add_directory($full_path, $baton,
3412 undef, -1, $self->{pool});
3413 } elsif ($t == $SVN::Node::dir) {
3414 return $self->open_directory($full_path, $baton,
3415 $self->{r}, $self->{pool});
3416 } # no warnings 'once'
3417 print STDERR "$full_path already exists in repository at ",
3418 "r$self->{r} and it is not a directory (",
3419 ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
3420 } # no warnings 'once'
3421 exit 1;
3422 }
3423
3424 sub ensure_path {
3425 my ($self, $path) = @_;
3426 my $bat = $self->{bat};
3427 my $repo_path = $self->repo_path($path);
3428 return $bat->{''} unless (length $repo_path);
3429 my @p = split m#/+#, $repo_path;
3430 my $c = shift @p;
3431 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
3432 while (@p) {
3433 my $c0 = $c;
3434 $c .= '/' . shift @p;
3435 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
3436 }
3437 return $bat->{$c};
3438 }
3439
3440 sub A {
3441 my ($self, $m) = @_;
3442 my ($dir, $file) = split_path($m->{file_b});
3443 my $pbat = $self->ensure_path($dir);
3444 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3445 undef, -1);
3446 print "\tA\t$m->{file_b}\n" unless $::_q;
3447 $self->chg_file($fbat, $m);
3448 $self->close_file($fbat,undef,$self->{pool});
3449 }
3450
3451 sub C {
3452 my ($self, $m) = @_;
3453 my ($dir, $file) = split_path($m->{file_b});
3454 my $pbat = $self->ensure_path($dir);
3455 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3456 $self->url_path($m->{file_a}), $self->{r});
3457 print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3458 $self->chg_file($fbat, $m);
3459 $self->close_file($fbat,undef,$self->{pool});
3460 }
3461
3462 sub delete_entry {
3463 my ($self, $path, $pbat) = @_;
3464 my $rpath = $self->repo_path($path);
3465 my ($dir, $file) = split_path($rpath);
3466 $self->{rm}->{$dir} = 1;
3467 $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
3468 }
3469
3470 sub R {
3471 my ($self, $m) = @_;
3472 my ($dir, $file) = split_path($m->{file_b});
3473 my $pbat = $self->ensure_path($dir);
3474 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3475 $self->url_path($m->{file_a}), $self->{r});
3476 print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3477 $self->chg_file($fbat, $m);
3478 $self->close_file($fbat,undef,$self->{pool});
3479
3480 ($dir, $file) = split_path($m->{file_a});
3481 $pbat = $self->ensure_path($dir);
3482 $self->delete_entry($m->{file_a}, $pbat);
3483 }
3484
3485 sub M {
3486 my ($self, $m) = @_;
3487 my ($dir, $file) = split_path($m->{file_b});
3488 my $pbat = $self->ensure_path($dir);
3489 my $fbat = $self->open_file($self->repo_path($m->{file_b}),
3490 $pbat,$self->{r},$self->{pool});
3491 print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
3492 $self->chg_file($fbat, $m);
3493 $self->close_file($fbat,undef,$self->{pool});
3494 }
3495
3496 sub T { shift->M(@_) }
3497
3498 sub change_file_prop {
3499 my ($self, $fbat, $pname, $pval) = @_;
3500 $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
3501 }
3502
3503 sub chg_file {
3504 my ($self, $fbat, $m) = @_;
3505 if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
3506 $self->change_file_prop($fbat,'svn:executable','*');
3507 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
3508 $self->change_file_prop($fbat,'svn:executable',undef);
3509 }
3510 my $fh = IO::File->new_tmpfile or croak $!;
3511 if ($m->{mode_b} =~ /^120/) {
3512 print $fh 'link ' or croak $!;
3513 $self->change_file_prop($fbat,'svn:special','*');
3514 } elsif ($m->{mode_a} =~ /^120/ && $m->{mode_b} !~ /^120/) {
3515 $self->change_file_prop($fbat,'svn:special',undef);
3516 }
3517 defined(my $pid = fork) or croak $!;
3518 if (!$pid) {
3519 open STDOUT, '>&', $fh or croak $!;
3520 exec qw/git-cat-file blob/, $m->{sha1_b} or croak $!;
3521 }
3522 waitpid $pid, 0;
3523 croak $? if $?;
3524 $fh->flush == 0 or croak $!;
3525 seek $fh, 0, 0 or croak $!;
3526
3527 my $exp = ::md5sum($fh);
3528 seek $fh, 0, 0 or croak $!;
3529
3530 my $pool = SVN::Pool->new;
3531 my $atd = $self->apply_textdelta($fbat, undef, $pool);
3532 my $got = SVN::TxDelta::send_stream($fh, @$atd, $pool);
3533 die "Checksum mismatch\nexpected: $exp\ngot: $got\n" if ($got ne $exp);
3534 $pool->clear;
3535
3536 close $fh or croak $!;
3537 }
3538
3539 sub D {
3540 my ($self, $m) = @_;
3541 my ($dir, $file) = split_path($m->{file_b});
3542 my $pbat = $self->ensure_path($dir);
3543 print "\tD\t$m->{file_b}\n" unless $::_q;
3544 $self->delete_entry($m->{file_b}, $pbat);
3545 }
3546
3547 sub close_edit {
3548 my ($self) = @_;
3549 my ($p,$bat) = ($self->{pool}, $self->{bat});
3550 foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
3551 next if $_ eq '';
3552 $self->close_directory($bat->{$_}, $p);
3553 }
3554 $self->close_directory($bat->{''}, $p);
3555 $self->SUPER::close_edit($p);
3556 $p->clear;
3557 }
3558
3559 sub abort_edit {
3560 my ($self) = @_;
3561 $self->SUPER::abort_edit($self->{pool});
3562 }
3563
3564 sub DESTROY {
3565 my $self = shift;
3566 $self->SUPER::DESTROY(@_);
3567 $self->{pool}->clear;
3568 }
3569
3570 # this drives the editor
3571 sub apply_diff {
3572 my ($self) = @_;
3573 my $mods = $self->{mods};
3574 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
3575 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
3576 my $f = $m->{chg};
3577 if (defined $o{$f}) {
3578 $self->$f($m);
3579 } else {
3580 fatal("Invalid change type: $f");
3581 }
3582 }
3583 $self->rmdirs if $_rmdir;
3584 if (@$mods == 0) {
3585 $self->abort_edit;
3586 } else {
3587 $self->close_edit;
3588 }
3589 return scalar @$mods;
3590 }
3591
3592 package Git::SVN::Ra;
3593 use vars qw/@ISA $config_dir $_log_window_size/;
3594 use strict;
3595 use warnings;
3596 my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
3597
3598 BEGIN {
3599 # enforce temporary pool usage for some simple functions
3600 no strict 'refs';
3601 for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root/) {
3602 my $SUPER = "SUPER::$f";
3603 *$f = sub {
3604 my $self = shift;
3605 my $pool = SVN::Pool->new;
3606 my @ret = $self->$SUPER(@_,$pool);
3607 $pool->clear;
3608 wantarray ? @ret : $ret[0];
3609 };
3610 }
3611 }
3612
3613 sub _auth_providers () {
3614 [
3615 SVN::Client::get_simple_provider(),
3616 SVN::Client::get_ssl_server_trust_file_provider(),
3617 SVN::Client::get_simple_prompt_provider(
3618 \&Git::SVN::Prompt::simple, 2),
3619 SVN::Client::get_ssl_client_cert_file_provider(),
3620 SVN::Client::get_ssl_client_cert_prompt_provider(
3621 \&Git::SVN::Prompt::ssl_client_cert, 2),
3622 SVN::Client::get_ssl_client_cert_pw_prompt_provider(
3623 \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
3624 SVN::Client::get_username_provider(),
3625 SVN::Client::get_ssl_server_trust_prompt_provider(
3626 \&Git::SVN::Prompt::ssl_server_trust),
3627 SVN::Client::get_username_prompt_provider(
3628 \&Git::SVN::Prompt::username, 2)
3629 ]
3630 }
3631
3632 sub escape_uri_only {
3633 my ($uri) = @_;
3634 my @tmp;
3635 foreach (split m{/}, $uri) {
3636 s/([^\w.-])/sprintf("%%%02X",ord($1))/eg;
3637 push @tmp, $_;
3638 }
3639 join('/', @tmp);
3640 }
3641
3642 sub escape_url {
3643 my ($url) = @_;
3644 if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
3645 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
3646 $url = "$scheme://$domain$uri";
3647 }
3648 $url;
3649 }
3650
3651 sub new {
3652 my ($class, $url) = @_;
3653 $url =~ s!/+$!!;
3654 return $RA if ($RA && $RA->{url} eq $url);
3655
3656 SVN::_Core::svn_config_ensure($config_dir, undef);
3657 my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
3658 my $config = SVN::Core::config_get_config($config_dir);
3659 $RA = undef;
3660 my $dont_store_passwords = 1;
3661 my $conf_t = ${$config}{'config'};
3662 {
3663 no warnings 'once';
3664 # The usage of $SVN::_Core::SVN_CONFIG_* variables
3665 # produces warnings that variables are used only once.
3666 # I had not found the better way to shut them up, so
3667 # the warnings of type 'once' are disabled in this block.
3668 if (SVN::_Core::svn_config_get_bool($conf_t,
3669 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3670 $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
3671 1) == 0) {
3672 SVN::_Core::svn_auth_set_parameter($baton,
3673 $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
3674 bless (\$dont_store_passwords, "_p_void"));
3675 }
3676 if (SVN::_Core::svn_config_get_bool($conf_t,
3677 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3678 $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
3679 1) == 0) {
3680 $Git::SVN::Prompt::_no_auth_cache = 1;
3681 }
3682 } # no warnings 'once'
3683 my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
3684 config => $config,
3685 pool => SVN::Pool->new,
3686 auth_provider_callbacks => $callbacks);
3687 $self->{url} = $url;
3688 $self->{svn_path} = $url;
3689 $self->{repos_root} = $self->get_repos_root;
3690 $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
3691 $self->{cache} = { check_path => { r => 0, data => {} },
3692 get_dir => { r => 0, data => {} } };
3693 $RA = bless $self, $class;
3694 }
3695
3696 sub check_path {
3697 my ($self, $path, $r) = @_;
3698 my $cache = $self->{cache}->{check_path};
3699 if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
3700 return $cache->{data}->{$path};
3701 }
3702 my $pool = SVN::Pool->new;
3703 my $t = $self->SUPER::check_path($path, $r, $pool);
3704 $pool->clear;
3705 if ($r != $cache->{r}) {
3706 %{$cache->{data}} = ();
3707 $cache->{r} = $r;
3708 }
3709 $cache->{data}->{$path} = $t;
3710 }
3711
3712 sub get_dir {
3713 my ($self, $dir, $r) = @_;
3714 my $cache = $self->{cache}->{get_dir};
3715 if ($r == $cache->{r}) {
3716 if (my $x = $cache->{data}->{$dir}) {
3717 return wantarray ? @$x : $x->[0];
3718 }
3719 }
3720 my $pool = SVN::Pool->new;
3721 my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
3722 my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
3723 $pool->clear;
3724 if ($r != $cache->{r}) {
3725 %{$cache->{data}} = ();
3726 $cache->{r} = $r;
3727 }
3728 $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
3729 wantarray ? (\%dirents, $r, $props) : \%dirents;
3730 }
3731
3732 sub DESTROY {
3733 # do not call the real DESTROY since we store ourselves in $RA
3734 }
3735
3736 sub get_log {
3737 my ($self, @args) = @_;
3738 my $pool = SVN::Pool->new;
3739 splice(@args, 3, 1) if ($SVN::Core::VERSION le '1.2.0');
3740 my $ret = $self->SUPER::get_log(@args, $pool);
3741 $pool->clear;
3742 $ret;
3743 }
3744
3745 sub trees_match {
3746 my ($self, $url1, $rev1, $url2, $rev2) = @_;
3747 my $ctx = SVN::Client->new(auth => _auth_providers);
3748 my $out = IO::File->new_tmpfile;
3749
3750 # older SVN (1.1.x) doesn't take $pool as the last parameter for
3751 # $ctx->diff(), so we'll create a default one
3752 my $pool = SVN::Pool->new_default_sub;
3753
3754 $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
3755 $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
3756 $out->flush;
3757 my $ret = (($out->stat)[7] == 0);
3758 close $out or croak $!;
3759
3760 $ret;
3761 }
3762
3763 sub get_commit_editor {
3764 my ($self, $log, $cb, $pool) = @_;
3765 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
3766 $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
3767 }
3768
3769 sub gs_do_update {
3770 my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
3771 my $new = ($rev_a == $rev_b);
3772 my $path = $gs->{path};
3773
3774 if ($new && -e $gs->{index}) {
3775 unlink $gs->{index} or die
3776 "Couldn't unlink index: $gs->{index}: $!\n";
3777 }
3778 my $pool = SVN::Pool->new;
3779 $editor->set_path_strip($path);
3780 my (@pc) = split m#/#, $path;
3781 my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
3782 1, $editor, $pool);
3783 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3784
3785 # Since we can't rely on svn_ra_reparent being available, we'll
3786 # just have to do some magic with set_path to make it so
3787 # we only want a partial path.
3788 my $sp = '';
3789 my $final = join('/', @pc);
3790 while (@pc) {
3791 $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
3792 $sp .= '/' if length $sp;
3793 $sp .= shift @pc;
3794 }
3795 die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
3796
3797 $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
3798
3799 $reporter->finish_report($pool);
3800 $pool->clear;
3801 $editor->{git_commit_ok};
3802 }
3803
3804 # this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
3805 # svn_ra_reparent didn't work before 1.4)
3806 sub gs_do_switch {
3807 my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
3808 my $path = $gs->{path};
3809 my $pool = SVN::Pool->new;
3810
3811 my $full_url = $self->{url};
3812 my $old_url = $full_url;
3813 $full_url .= '/' . escape_uri_only($path) if length $path;
3814 my ($ra, $reparented);
3815 if ($old_url ne $full_url) {
3816 if ($old_url !~ m#^svn(\+ssh)?://#) {
3817 SVN::_Ra::svn_ra_reparent($self->{session}, $full_url,
3818 $pool);
3819 $self->{url} = $full_url;
3820 $reparented = 1;
3821 } else {
3822 $_[0] = undef;
3823 $self = undef;
3824 $RA = undef;
3825 $ra = Git::SVN::Ra->new($full_url);
3826 $ra_invalid = 1;
3827 }
3828 }
3829 $ra ||= $self;
3830 my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
3831 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3832 $reporter->set_path('', $rev_a, 0, @lock, $pool);
3833 $reporter->finish_report($pool);
3834
3835 if ($reparented) {
3836 SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
3837 $self->{url} = $old_url;
3838 }
3839
3840 $pool->clear;
3841 $editor->{git_commit_ok};
3842 }
3843
3844 sub longest_common_path {
3845 my ($gsv, $globs) = @_;
3846 my %common;
3847 my $common_max = scalar @$gsv;
3848
3849 foreach my $gs (@$gsv) {
3850 my @tmp = split m#/#, $gs->{path};
3851 my $p = '';
3852 foreach (@tmp) {
3853 $p .= length($p) ? "/$_" : $_;
3854 $common{$p} ||= 0;
3855 $common{$p}++;
3856 }
3857 }
3858 $globs ||= [];
3859 $common_max += scalar @$globs;
3860 foreach my $glob (@$globs) {
3861 my @tmp = split m#/#, $glob->{path}->{left};
3862 my $p = '';
3863 foreach (@tmp) {
3864 $p .= length($p) ? "/$_" : $_;
3865 $common{$p} ||= 0;
3866 $common{$p}++;
3867 }
3868 }
3869
3870 my $longest_path = '';
3871 foreach (sort {length $b <=> length $a} keys %common) {
3872 if ($common{$_} == $common_max) {
3873 $longest_path = $_;
3874 last;
3875 }
3876 }
3877 $longest_path;
3878 }
3879
3880 sub gs_fetch_loop_common {
3881 my ($self, $base, $head, $gsv, $globs) = @_;
3882 return if ($base > $head);
3883 my $inc = $_log_window_size;
3884 my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
3885 my $longest_path = longest_common_path($gsv, $globs);
3886 my $ra_url = $self->{url};
3887 while (1) {
3888 my %revs;
3889 my $err;
3890 my $err_handler = $SVN::Error::handler;
3891 $SVN::Error::handler = sub {
3892 ($err) = @_;
3893 skip_unknown_revs($err);
3894 };
3895 sub _cb {
3896 my ($paths, $r, $author, $date, $log) = @_;
3897 [ dup_changed_paths($paths),
3898 { author => $author, date => $date, log => $log } ];
3899 }
3900 $self->get_log([$longest_path], $min, $max, 0, 1, 1,
3901 sub { $revs{$_[1]} = _cb(@_) });
3902 if ($err && $max >= $head) {
3903 print STDERR "Path '$longest_path' ",
3904 "was probably deleted:\n",
3905 $err->expanded_message,
3906 "\nWill attempt to follow ",
3907 "revisions r$min .. r$max ",
3908 "committed before the deletion\n";
3909 my $hi = $max;
3910 while (--$hi >= $min) {
3911 my $ok;
3912 $self->get_log([$longest_path], $min, $hi,
3913 0, 1, 1, sub {
3914 $ok ||= $_[1];
3915 $revs{$_[1]} = _cb(@_) });
3916 if ($ok) {
3917 print STDERR "r$min .. r$ok OK\n";
3918 last;
3919 }
3920 }
3921 }
3922 $SVN::Error::handler = $err_handler;
3923
3924 my %exists = map { $_->{path} => $_ } @$gsv;
3925 foreach my $r (sort {$a <=> $b} keys %revs) {
3926 my ($paths, $logged) = @{$revs{$r}};
3927
3928 foreach my $gs ($self->match_globs(\%exists, $paths,
3929 $globs, $r)) {
3930 if ($gs->rev_map_max >= $r) {
3931 next;
3932 }
3933 next unless $gs->match_paths($paths, $r);
3934 $gs->{logged_rev_props} = $logged;
3935 if (my $last_commit = $gs->last_commit) {
3936 $gs->assert_index_clean($last_commit);
3937 }
3938 my $log_entry = $gs->do_fetch($paths, $r);
3939 if ($log_entry) {
3940 $gs->do_git_commit($log_entry);
3941 }
3942 }
3943 foreach my $g (@$globs) {
3944 my $k = "svn-remote.$g->{remote}." .
3945 "$g->{t}-maxRev";
3946 Git::SVN::tmp_config($k, $r);
3947 }
3948 if ($ra_invalid) {
3949 $_[0] = undef;
3950 $self = undef;
3951 $RA = undef;
3952 $self = Git::SVN::Ra->new($ra_url);
3953 $ra_invalid = undef;
3954 }
3955 }
3956 # pre-fill the .rev_db since it'll eventually get filled in
3957 # with '0' x40 if something new gets committed
3958 foreach my $gs (@$gsv) {
3959 next if $gs->rev_map_max >= $max;
3960 next if defined $gs->rev_map_get($max);
3961 $gs->rev_map_set($max, 0 x40);
3962 }
3963 foreach my $g (@$globs) {
3964 my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
3965 Git::SVN::tmp_config($k, $max);
3966 }
3967 last if $max >= $head;
3968 $min = $max + 1;
3969 $max += $inc;
3970 $max = $head if ($max > $head);
3971 }
3972 }
3973
3974 sub match_globs {
3975 my ($self, $exists, $paths, $globs, $r) = @_;
3976
3977 sub get_dir_check {
3978 my ($self, $exists, $g, $r) = @_;
3979 my @x = eval { $self->get_dir($g->{path}->{left}, $r) };
3980 return unless scalar @x == 3;
3981 my $dirents = $x[0];
3982 foreach my $de (keys %$dirents) {
3983 next if $dirents->{$de}->{kind} != $SVN::Node::dir;
3984 my $p = $g->{path}->full_path($de);
3985 next if $exists->{$p};
3986 next if (length $g->{path}->{right} &&
3987 ($self->check_path($p, $r) !=
3988 $SVN::Node::dir));
3989 $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
3990 $g->{ref}->full_path($de), 1);
3991 }
3992 }
3993 foreach my $g (@$globs) {
3994 if (my $path = $paths->{"/$g->{path}->{left}"}) {
3995 if ($path->{action} =~ /^[AR]$/) {
3996 get_dir_check($self, $exists, $g, $r);
3997 }
3998 }
3999 foreach (keys %$paths) {
4000 if (/$g->{path}->{left_regex}/ &&
4001 !/$g->{path}->{regex}/) {
4002 next if $paths->{$_}->{action} !~ /^[AR]$/;
4003 get_dir_check($self, $exists, $g, $r);
4004 }
4005 next unless /$g->{path}->{regex}/;
4006 my $p = $1;
4007 my $pathname = $g->{path}->full_path($p);
4008 next if $exists->{$pathname};
4009 next if ($self->check_path($pathname, $r) !=
4010 $SVN::Node::dir);
4011 $exists->{$pathname} = Git::SVN->init(
4012 $self->{url}, $pathname, undef,
4013 $g->{ref}->full_path($p), 1);
4014 }
4015 my $c = '';
4016 foreach (split m#/#, $g->{path}->{left}) {
4017 $c .= "/$_";
4018 next unless ($paths->{$c} &&
4019 ($paths->{$c}->{action} =~ /^[AR]$/));
4020 get_dir_check($self, $exists, $g, $r);
4021 }
4022 }
4023 values %$exists;
4024 }
4025
4026 sub minimize_url {
4027 my ($self) = @_;
4028 return $self->{url} if ($self->{url} eq $self->{repos_root});
4029 my $url = $self->{repos_root};
4030 my @components = split(m!/!, $self->{svn_path});
4031 my $c = '';
4032 do {
4033 $url .= "/$c" if length $c;
4034 eval { (ref $self)->new($url)->get_latest_revnum };
4035 } while ($@ && ($c = shift @components));
4036 $url;
4037 }
4038
4039 sub can_do_switch {
4040 my $self = shift;
4041 unless (defined $can_do_switch) {
4042 my $pool = SVN::Pool->new;
4043 my $rep = eval {
4044 $self->do_switch(1, '', 0, $self->{url},
4045 SVN::Delta::Editor->new, $pool);
4046 };
4047 if ($@) {
4048 $can_do_switch = 0;
4049 } else {
4050 $rep->abort_report($pool);
4051 $can_do_switch = 1;
4052 }
4053 $pool->clear;
4054 }
4055 $can_do_switch;
4056 }
4057
4058 sub skip_unknown_revs {
4059 my ($err) = @_;
4060 my $errno = $err->apr_err();
4061 # Maybe the branch we're tracking didn't
4062 # exist when the repo started, so it's
4063 # not an error if it doesn't, just continue
4064 #
4065 # Wonderfully consistent library, eh?
4066 # 160013 - svn:// and file://
4067 # 175002 - http(s)://
4068 # 175007 - http(s):// (this repo required authorization, too...)
4069 # More codes may be discovered later...
4070 if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
4071 my $err_key = $err->expanded_message;
4072 # revision numbers change every time, filter them out
4073 $err_key =~ s/\d+/\0/g;
4074 $err_key = "$errno\0$err_key";
4075 unless ($ignored_err{$err_key}) {
4076 warn "W: Ignoring error from SVN, path probably ",
4077 "does not exist: ($errno): ",
4078 $err->expanded_message,"\n";
4079 $ignored_err{$err_key} = 1;
4080 }
4081 return;
4082 }
4083 die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
4084 }
4085
4086 # svn_log_changed_path_t objects passed to get_log are likely to be
4087 # overwritten even if only the refs are copied to an external variable,
4088 # so we should dup the structures in their entirety. Using an externally
4089 # passed pool (instead of our temporary and quickly cleared pool in
4090 # Git::SVN::Ra) does not help matters at all...
4091 sub dup_changed_paths {
4092 my ($paths) = @_;
4093 return undef unless $paths;
4094 my %ret;
4095 foreach my $p (keys %$paths) {
4096 my $i = $paths->{$p};
4097 my %s = map { $_ => $i->$_ }
4098 qw/copyfrom_path copyfrom_rev action/;
4099 $ret{$p} = \%s;
4100 }
4101 \%ret;
4102 }
4103
4104 package Git::SVN::Log;
4105 use strict;
4106 use warnings;
4107 use POSIX qw/strftime/;
4108 use constant commit_log_separator => ('-' x 72) . "\n";
4109 use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
4110 %rusers $show_commit $incremental/;
4111 my $l_fmt;
4112
4113 sub cmt_showable {
4114 my ($c) = @_;
4115 return 1 if defined $c->{r};
4116
4117 # big commit message got truncated by the 16k pretty buffer in rev-list
4118 if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
4119 $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
4120 @{$c->{l}} = ();
4121 my @log = command(qw/cat-file commit/, $c->{c});
4122
4123 # shift off the headers
4124 shift @log while ($log[0] ne '');
4125 shift @log;
4126
4127 # TODO: make $c->{l} not have a trailing newline in the future
4128 @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
4129
4130 (undef, $c->{r}, undef) = ::extract_metadata(
4131 (grep(/^git-svn-id: /, @log))[-1]);
4132 }
4133 return defined $c->{r};
4134 }
4135
4136 sub log_use_color {
4137 return $color || Git->repository->get_colorbool('color.diff');
4138 }
4139
4140 sub git_svn_log_cmd {
4141 my ($r_min, $r_max, @args) = @_;
4142 my $head = 'HEAD';
4143 my (@files, @log_opts);
4144 foreach my $x (@args) {
4145 if ($x eq '--' || @files) {
4146 push @files, $x;
4147 } else {
4148 if (::verify_ref("$x^0")) {
4149 $head = $x;
4150 } else {
4151 push @log_opts, $x;
4152 }
4153 }
4154 }
4155
4156 my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
4157 $gs ||= Git::SVN->_new;
4158 my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
4159 $gs->refname);
4160 push @cmd, '-r' unless $non_recursive;
4161 push @cmd, qw/--raw --name-status/ if $verbose;
4162 push @cmd, '--color' if log_use_color();
4163 push @cmd, @log_opts;
4164 if (defined $r_max && $r_max == $r_min) {
4165 push @cmd, '--max-count=1';
4166 if (my $c = $gs->rev_map_get($r_max)) {
4167 push @cmd, $c;
4168 }
4169 } elsif (defined $r_max) {
4170 if ($r_max < $r_min) {
4171 ($r_min, $r_max) = ($r_max, $r_min);
4172 }
4173 my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
4174 my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
4175 # If there are no commits in the range, both $c_max and $c_min
4176 # will be undefined. If there is at least 1 commit in the
4177 # range, both will be defined.
4178 return () if !defined $c_min || !defined $c_max;
4179 if ($c_min eq $c_max) {
4180 push @cmd, '--max-count=1', $c_min;
4181 } else {
4182 push @cmd, '--boundary', "$c_min..$c_max";
4183 }
4184 }
4185 return (@cmd, @files);
4186 }
4187
4188 # adapted from pager.c
4189 sub config_pager {
4190 $pager ||= $ENV{GIT_PAGER} || $ENV{PAGER};
4191 if (!defined $pager) {
4192 $pager = 'less';
4193 } elsif (length $pager == 0 || $pager eq 'cat') {
4194 $pager = undef;
4195 }
4196 $ENV{GIT_PAGER_IN_USE} = defined($pager);
4197 }
4198
4199 sub run_pager {
4200 return unless -t *STDOUT && defined $pager;
4201 pipe my $rfd, my $wfd or return;
4202 defined(my $pid = fork) or ::fatal "Can't fork: $!";
4203 if (!$pid) {
4204 open STDOUT, '>&', $wfd or
4205 ::fatal "Can't redirect to stdout: $!";
4206 return;
4207 }
4208 open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
4209 $ENV{LESS} ||= 'FRSX';
4210 exec $pager or ::fatal "Can't run pager: $! ($pager)";
4211 }
4212
4213 sub format_svn_date {
4214 return strftime("%Y-%m-%d %H:%M:%S %z (%a, %d %b %Y)", localtime(shift));
4215 }
4216
4217 sub parse_git_date {
4218 my ($t, $tz) = @_;
4219 # Date::Parse isn't in the standard Perl distro :(
4220 if ($tz =~ s/^\+//) {
4221 $t += tz_to_s_offset($tz);
4222 } elsif ($tz =~ s/^\-//) {
4223 $t -= tz_to_s_offset($tz);
4224 }
4225 return $t;
4226 }
4227
4228 sub set_local_timezone {
4229 if (defined $TZ) {
4230 $ENV{TZ} = $TZ;
4231 } else {
4232 delete $ENV{TZ};
4233 }
4234 }
4235
4236 sub tz_to_s_offset {
4237 my ($tz) = @_;
4238 $tz =~ s/(\d\d)$//;
4239 return ($1 * 60) + ($tz * 3600);
4240 }
4241
4242 sub get_author_info {
4243 my ($dest, $author, $t, $tz) = @_;
4244 $author =~ s/(?:^\s*|\s*$)//g;
4245 $dest->{a_raw} = $author;
4246 my $au;
4247 if ($::_authors) {
4248 $au = $rusers{$author} || undef;
4249 }
4250 if (!$au) {
4251 ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
4252 }
4253 $dest->{t} = $t;
4254 $dest->{tz} = $tz;
4255 $dest->{a} = $au;
4256 $dest->{t_utc} = parse_git_date($t, $tz);
4257 }
4258
4259 sub process_commit {
4260 my ($c, $r_min, $r_max, $defer) = @_;
4261 if (defined $r_min && defined $r_max) {
4262 if ($r_min == $c->{r} && $r_min == $r_max) {
4263 show_commit($c);
4264 return 0;
4265 }
4266 return 1 if $r_min == $r_max;
4267 if ($r_min < $r_max) {
4268 # we need to reverse the print order
4269 return 0 if (defined $limit && --$limit < 0);
4270 push @$defer, $c;
4271 return 1;
4272 }
4273 if ($r_min != $r_max) {
4274 return 1 if ($r_min < $c->{r});
4275 return 1 if ($r_max > $c->{r});
4276 }
4277 }
4278 return 0 if (defined $limit && --$limit < 0);
4279 show_commit($c);
4280 return 1;
4281 }
4282
4283 sub show_commit {
4284 my $c = shift;
4285 if ($oneline) {
4286 my $x = "\n";
4287 if (my $l = $c->{l}) {
4288 while ($l->[0] =~ /^\s*$/) { shift @$l }
4289 $x = $l->[0];
4290 }
4291 $l_fmt ||= 'A' . length($c->{r});
4292 print 'r',pack($l_fmt, $c->{r}),' | ';
4293 print "$c->{c} | " if $show_commit;
4294 print $x;
4295 } else {
4296 show_commit_normal($c);
4297 }
4298 }
4299
4300 sub show_commit_changed_paths {
4301 my ($c) = @_;
4302 return unless $c->{changed};
4303 print "Changed paths:\n", @{$c->{changed}};
4304 }
4305
4306 sub show_commit_normal {
4307 my ($c) = @_;
4308 print commit_log_separator, "r$c->{r} | ";
4309 print "$c->{c} | " if $show_commit;
4310 print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
4311 my $nr_line = 0;
4312
4313 if (my $l = $c->{l}) {
4314 while ($l->[$#$l] eq "\n" && $#$l > 0
4315 && $l->[($#$l - 1)] eq "\n") {
4316 pop @$l;
4317 }
4318 $nr_line = scalar @$l;
4319 if (!$nr_line) {
4320 print "1 line\n\n\n";
4321 } else {
4322 if ($nr_line == 1) {
4323 $nr_line = '1 line';
4324 } else {
4325 $nr_line .= ' lines';
4326 }
4327 print $nr_line, "\n";
4328 show_commit_changed_paths($c);
4329 print "\n";
4330 print $_ foreach @$l;
4331 }
4332 } else {
4333 print "1 line\n";
4334 show_commit_changed_paths($c);
4335 print "\n";
4336
4337 }
4338 foreach my $x (qw/raw stat diff/) {
4339 if ($c->{$x}) {
4340 print "\n";
4341 print $_ foreach @{$c->{$x}}
4342 }
4343 }
4344 }
4345
4346 sub cmd_show_log {
4347 my (@args) = @_;
4348 my ($r_min, $r_max);
4349 my $r_last = -1; # prevent dupes
4350 set_local_timezone();
4351 if (defined $::_revision) {
4352 if ($::_revision =~ /^(\d+):(\d+)$/) {
4353 ($r_min, $r_max) = ($1, $2);
4354 } elsif ($::_revision =~ /^\d+$/) {
4355 $r_min = $r_max = $::_revision;
4356 } else {
4357 ::fatal "-r$::_revision is not supported, use ",
4358 "standard 'git log' arguments instead";
4359 }
4360 }
4361
4362 config_pager();
4363 @args = git_svn_log_cmd($r_min, $r_max, @args);
4364 if (!@args) {
4365 print commit_log_separator unless $incremental || $oneline;
4366 return;
4367 }
4368 my $log = command_output_pipe(@args);
4369 run_pager();
4370 my (@k, $c, $d, $stat);
4371 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
4372 while (<$log>) {
4373 if (/^${esc_color}commit -?($::sha1_short)/o) {
4374 my $cmt = $1;
4375 if ($c && cmt_showable($c) && $c->{r} != $r_last) {
4376 $r_last = $c->{r};
4377 process_commit($c, $r_min, $r_max, \@k) or
4378 goto out;
4379 }
4380 $d = undef;
4381 $c = { c => $cmt };
4382 } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
4383 get_author_info($c, $1, $2, $3);
4384 } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
4385 # ignore
4386 } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
4387 push @{$c->{raw}}, $_;
4388 } elsif (/^${esc_color}[ACRMDT]\t/) {
4389 # we could add $SVN->{svn_path} here, but that requires
4390 # remote access at the moment (repo_path_split)...
4391 s#^(${esc_color})([ACRMDT])\t#$1 $2 #o;
4392 push @{$c->{changed}}, $_;
4393 } elsif (/^${esc_color}diff /o) {
4394 $d = 1;
4395 push @{$c->{diff}}, $_;
4396 } elsif ($d) {
4397 push @{$c->{diff}}, $_;
4398 } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
4399 $esc_color*[\+\-]*$esc_color$/x) {
4400 $stat = 1;
4401 push @{$c->{stat}}, $_;
4402 } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
4403 push @{$c->{stat}}, $_;
4404 $stat = undef;
4405 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
4406 ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
4407 } elsif (s/^${esc_color} //o) {
4408 push @{$c->{l}}, $_;
4409 }
4410 }
4411 if ($c && defined $c->{r} && $c->{r} != $r_last) {
4412 $r_last = $c->{r};
4413 process_commit($c, $r_min, $r_max, \@k);
4414 }
4415 if (@k) {
4416 ($r_min, $r_max) = ($r_max, $r_min);
4417 process_commit($_, $r_min, $r_max) foreach reverse @k;
4418 }
4419 out:
4420 close $log;
4421 print commit_log_separator unless $incremental || $oneline;
4422 }
4423
4424 package Git::SVN::Migration;
4425 # these version numbers do NOT correspond to actual version numbers
4426 # of git nor git-svn. They are just relative.
4427 #
4428 # v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
4429 #
4430 # v1 layout: .git/$id/info/url, refs/remotes/$id
4431 #
4432 # v2 layout: .git/svn/$id/info/url, refs/remotes/$id
4433 #
4434 # v3 layout: .git/svn/$id, refs/remotes/$id
4435 # - info/url may remain for backwards compatibility
4436 # - this is what we migrate up to this layout automatically,
4437 # - this will be used by git svn init on single branches
4438 # v3.1 layout (auto migrated):
4439 # - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
4440 # for backwards compatibility
4441 #
4442 # v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
4443 # - this is only created for newly multi-init-ed
4444 # repositories. Similar in spirit to the
4445 # --use-separate-remotes option in git-clone (now default)
4446 # - we do not automatically migrate to this (following
4447 # the example set by core git)
4448 #
4449 # v5 layout: .rev_db.$UUID => .rev_map.$UUID
4450 # - newer, more-efficient format that uses 24-bytes per record
4451 # with no filler space.
4452 # - use xxd -c24 < .rev_map.$UUID to view and debug
4453 # - This is a one-way migration, repositories updated to the
4454 # new format will not be able to use old git-svn without
4455 # rebuilding the .rev_db. Rebuilding the rev_db is not
4456 # possible if noMetadata or useSvmProps are set; but should
4457 # be no problem for users that use the (sensible) defaults.
4458 use strict;
4459 use warnings;
4460 use Carp qw/croak/;
4461 use File::Path qw/mkpath/;
4462 use File::Basename qw/dirname basename/;
4463 use vars qw/$_minimize/;
4464
4465 sub migrate_from_v0 {
4466 my $git_dir = $ENV{GIT_DIR};
4467 return undef unless -d $git_dir;
4468 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4469 my $migrated = 0;
4470 while (<$fh>) {
4471 chomp;
4472 my ($id, $orig_ref) = ($_, $_);
4473 next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
4474 next unless -f "$git_dir/$id/info/url";
4475 my $new_ref = "refs/remotes/$id";
4476 if (::verify_ref("$new_ref^0")) {
4477 print STDERR "W: $orig_ref is probably an old ",
4478 "branch used by an ancient version of ",
4479 "git-svn.\n",
4480 "However, $new_ref also exists.\n",
4481 "We will not be able ",
4482 "to use this branch until this ",
4483 "ambiguity is resolved.\n";
4484 next;
4485 }
4486 print STDERR "Migrating from v0 layout...\n" if !$migrated;
4487 print STDERR "Renaming ref: $orig_ref => $new_ref\n";
4488 command_noisy('update-ref', $new_ref, $orig_ref);
4489 command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
4490 $migrated++;
4491 }
4492 command_close_pipe($fh, $ctx);
4493 print STDERR "Done migrating from v0 layout...\n" if $migrated;
4494 $migrated;
4495 }
4496
4497 sub migrate_from_v1 {
4498 my $git_dir = $ENV{GIT_DIR};
4499 my $migrated = 0;
4500 return $migrated unless -d $git_dir;
4501 my $svn_dir = "$git_dir/svn";
4502
4503 # just in case somebody used 'svn' as their $id at some point...
4504 return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
4505
4506 print STDERR "Migrating from a git-svn v1 layout...\n";
4507 mkpath([$svn_dir]);
4508 print STDERR "Data from a previous version of git-svn exists, but\n\t",
4509 "$svn_dir\n\t(required for this version ",
4510 "($::VERSION) of git-svn) does not. exist\n";
4511 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4512 while (<$fh>) {
4513 my $x = $_;
4514 next unless $x =~ s#^refs/remotes/##;
4515 chomp $x;
4516 next unless -f "$git_dir/$x/info/url";
4517 my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
4518 next unless $u;
4519 my $dn = dirname("$git_dir/svn/$x");
4520 mkpath([$dn]) unless -d $dn;
4521 if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
4522 mkpath(["$git_dir/svn/svn"]);
4523 print STDERR " - $git_dir/$x/info => ",
4524 "$git_dir/svn/$x/info\n";
4525 rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
4526 croak "$!: $x";
4527 # don't worry too much about these, they probably
4528 # don't exist with repos this old (save for index,
4529 # and we can easily regenerate that)
4530 foreach my $f (qw/unhandled.log index .rev_db/) {
4531 rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
4532 }
4533 } else {
4534 print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
4535 rename "$git_dir/$x", "$git_dir/svn/$x" or
4536 croak "$!: $x";
4537 }
4538 $migrated++;
4539 }
4540 command_close_pipe($fh, $ctx);
4541 print STDERR "Done migrating from a git-svn v1 layout\n";
4542 $migrated;
4543 }
4544
4545 sub read_old_urls {
4546 my ($l_map, $pfx, $path) = @_;
4547 my @dir;
4548 foreach (<$path/*>) {
4549 if (-r "$_/info/url") {
4550 $pfx .= '/' if $pfx && $pfx !~ m!/$!;
4551 my $ref_id = $pfx . basename $_;
4552 my $url = ::file_to_s("$_/info/url");
4553 $l_map->{$ref_id} = $url;
4554 } elsif (-d $_) {
4555 push @dir, $_;
4556 }
4557 }
4558 foreach (@dir) {
4559 my $x = $_;
4560 $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
4561 read_old_urls($l_map, $x, $_);
4562 }
4563 }
4564
4565 sub migrate_from_v2 {
4566 my @cfg = command(qw/config -l/);
4567 return if grep /^svn-remote\..+\.url=/, @cfg;
4568 my %l_map;
4569 read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
4570 my $migrated = 0;
4571
4572 foreach my $ref_id (sort keys %l_map) {
4573 eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
4574 if ($@) {
4575 Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
4576 }
4577 $migrated++;
4578 }
4579 $migrated;
4580 }
4581
4582 sub minimize_connections {
4583 my $r = Git::SVN::read_all_remotes();
4584 my $new_urls = {};
4585 my $root_repos = {};
4586 foreach my $repo_id (keys %$r) {
4587 my $url = $r->{$repo_id}->{url} or next;
4588 my $fetch = $r->{$repo_id}->{fetch} or next;
4589 my $ra = Git::SVN::Ra->new($url);
4590
4591 # skip existing cases where we already connect to the root
4592 if (($ra->{url} eq $ra->{repos_root}) ||
4593 (Git::SVN::sanitize_remote_name($ra->{repos_root}) eq
4594 $repo_id)) {
4595 $root_repos->{$ra->{url}} = $repo_id;
4596 next;
4597 }
4598
4599 my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
4600 my $root_path = $ra->{url};
4601 $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
4602 foreach my $path (keys %$fetch) {
4603 my $ref_id = $fetch->{$path};
4604 my $gs = Git::SVN->new($ref_id, $repo_id, $path);
4605
4606 # make sure we can read when connecting to
4607 # a higher level of a repository
4608 my ($last_rev, undef) = $gs->last_rev_commit;
4609 if (!defined $last_rev) {
4610 $last_rev = eval {
4611 $root_ra->get_latest_revnum;
4612 };
4613 next if $@;
4614 }
4615 my $new = $root_path;
4616 $new .= length $path ? "/$path" : '';
4617 eval {
4618 $root_ra->get_log([$new], $last_rev, $last_rev,
4619 0, 0, 1, sub { });
4620 };
4621 next if $@;
4622 $new_urls->{$ra->{repos_root}}->{$new} =
4623 { ref_id => $ref_id,
4624 old_repo_id => $repo_id,
4625 old_path => $path };
4626 }
4627 }
4628
4629 my @emptied;
4630 foreach my $url (keys %$new_urls) {
4631 # see if we can re-use an existing [svn-remote "repo_id"]
4632 # instead of creating a(n ugly) new section:
4633 my $repo_id = $root_repos->{$url} ||
4634 Git::SVN::sanitize_remote_name($url);
4635
4636 my $fetch = $new_urls->{$url};
4637 foreach my $path (keys %$fetch) {
4638 my $x = $fetch->{$path};
4639 Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
4640 my $pfx = "svn-remote.$x->{old_repo_id}";
4641
4642 my $old_fetch = quotemeta("$x->{old_path}:".
4643 "refs/remotes/$x->{ref_id}");
4644 command_noisy(qw/config --unset/,
4645 "$pfx.fetch", '^'. $old_fetch . '$');
4646 delete $r->{$x->{old_repo_id}}->
4647 {fetch}->{$x->{old_path}};
4648 if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
4649 command_noisy(qw/config --unset/,
4650 "$pfx.url");
4651 push @emptied, $x->{old_repo_id}
4652 }
4653 }
4654 }
4655 if (@emptied) {
4656 my $file = $ENV{GIT_CONFIG} || $ENV{GIT_CONFIG_LOCAL} ||
4657 "$ENV{GIT_DIR}/config";
4658 print STDERR <<EOF;
4659 The following [svn-remote] sections in your config file ($file) are empty
4660 and can be safely removed:
4661 EOF
4662 print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
4663 }
4664 }
4665
4666 sub migration_check {
4667 migrate_from_v0();
4668 migrate_from_v1();
4669 migrate_from_v2();
4670 minimize_connections() if $_minimize;
4671 }
4672
4673 package Git::IndexInfo;
4674 use strict;
4675 use warnings;
4676 use Git qw/command_input_pipe command_close_pipe/;
4677
4678 sub new {
4679 my ($class) = @_;
4680 my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
4681 bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
4682 }
4683
4684 sub remove {
4685 my ($self, $path) = @_;
4686 if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
4687 return ++$self->{nr};
4688 }
4689 undef;
4690 }
4691
4692 sub update {
4693 my ($self, $mode, $hash, $path) = @_;
4694 if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
4695 return ++$self->{nr};
4696 }
4697 undef;
4698 }
4699
4700 sub DESTROY {
4701 my ($self) = @_;
4702 command_close_pipe($self->{gui}, $self->{ctx});
4703 }
4704
4705 package Git::SVN::GlobSpec;
4706 use strict;
4707 use warnings;
4708
4709 sub new {
4710 my ($class, $glob) = @_;
4711 my $re = $glob;
4712 $re =~ s!/+$!!g; # no need for trailing slashes
4713 my $nr = ($re =~ s!^(.*)\*(.*)$!\(\[^/\]+\)!g);
4714 my ($left, $right) = ($1, $2);
4715 if ($nr > 1) {
4716 die "Only one '*' wildcard expansion ",
4717 "is supported (got $nr): '$glob'\n";
4718 } elsif ($nr == 0) {
4719 die "One '*' is needed for glob: '$glob'\n";
4720 }
4721 $re = quotemeta($left) . $re . quotemeta($right);
4722 if (length $left && !($left =~ s!/+$!!g)) {
4723 die "Missing trailing '/' on left side of: '$glob' ($left)\n";
4724 }
4725 if (length $right && !($right =~ s!^/+!!g)) {
4726 die "Missing leading '/' on right side of: '$glob' ($right)\n";
4727 }
4728 my $left_re = qr/^\/\Q$left\E(\/|$)/;
4729 bless { left => $left, right => $right, left_regex => $left_re,
4730 regex => qr/$re/, glob => $glob }, $class;
4731 }
4732
4733 sub full_path {
4734 my ($self, $path) = @_;
4735 return (length $self->{left} ? "$self->{left}/" : '') .
4736 $path . (length $self->{right} ? "/$self->{right}" : '');
4737 }
4738
4739 __END__
4740
4741 Data structures:
4742
4743
4744 $remotes = { # returned by read_all_remotes()
4745 'svn' => {
4746 # svn-remote.svn.url=https://svn.musicpd.org
4747 url => 'https://svn.musicpd.org',
4748 # svn-remote.svn.fetch=mpd/trunk:trunk
4749 fetch => {
4750 'mpd/trunk' => 'trunk',
4751 },
4752 # svn-remote.svn.tags=mpd/tags/*:tags/*
4753 tags => {
4754 path => {
4755 left => 'mpd/tags',
4756 right => '',
4757 regex => qr!mpd/tags/([^/]+)$!,
4758 glob => 'tags/*',
4759 },
4760 ref => {
4761 left => 'tags',
4762 right => '',
4763 regex => qr!tags/([^/]+)$!,
4764 glob => 'tags/*',
4765 },
4766 }
4767 }
4768 };
4769
4770 $log_entry hashref as returned by libsvn_log_entry()
4771 {
4772 log => 'whitespace-formatted log entry
4773 ', # trailing newline is preserved
4774 revision => '8', # integer
4775 date => '2004-02-24T17:01:44.108345Z', # commit date
4776 author => 'committer name'
4777 };
4778
4779
4780 # this is generated by generate_diff();
4781 @mods = array of diff-index line hashes, each element represents one line
4782 of diff-index output
4783
4784 diff-index line ($m hash)
4785 {
4786 mode_a => first column of diff-index output, no leading ':',
4787 mode_b => second column of diff-index output,
4788 sha1_b => sha1sum of the final blob,
4789 chg => change type [MCRADT],
4790 file_a => original file name of a file (iff chg is 'C' or 'R')
4791 file_b => new/current file name of a file (any chg)
4792 }
4793 ;
4794
4795 # retval of read_url_paths{,_all}();
4796 $l_map = {
4797 # repository root url
4798 'https://svn.musicpd.org' => {
4799 # repository path # GIT_SVN_ID
4800 'mpd/trunk' => 'trunk',
4801 'mpd/tags/0.11.5' => 'tags/0.11.5',
4802 },
4803 }
4804
4805 Notes:
4806 I don't trust the each() function on unless I created %hash myself
4807 because the internal iterator may not have started at base.
kirr's git tree