?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
git-svn: Add --add-author-from option.
[git/dscho.git] / git-svn.perl
blob262acd9b365aba6ce5a0e1623e95cefd9c867a45
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 $Git::SVN::_follow_parent = 1;
70 my %remote_opts = ( 'username=s' => \$Git::SVN::Prompt::_username,
71 'config-dir=s' => \$Git::SVN::Ra::config_dir,
72 'no-auth-cache' => \$Git::SVN::Prompt::_no_auth_cache );
73 my %fc_opts = ( 'follow-parent|follow!' => \$Git::SVN::_follow_parent,
74 'authors-file|A=s' => \$_authors,
75 'repack:i' => \$Git::SVN::_repack,
76 'noMetadata' => \$Git::SVN::_no_metadata,
77 'useSvmProps' => \$Git::SVN::_use_svm_props,
78 'useSvnsyncProps' => \$Git::SVN::_use_svnsync_props,
79 'log-window-size=i' => \$Git::SVN::Ra::_log_window_size,
80 'no-checkout' => \$_no_checkout,
81 'quiet|q' => \$_q,
82 'repack-flags|repack-args|repack-opts=s' =>
83 \$Git::SVN::_repack_flags,
84 'use-log-author' => \$Git::SVN::_use_log_author,
85 'add-author-from' => \$Git::SVN::_add_author_from,
86 %remote_opts );
87
88 my ($_trunk, $_tags, $_branches, $_stdlayout);
89 my %icv;
90 my %init_opts = ( 'template=s' => \$_template, 'shared:s' => \$_shared,
91 'trunk|T=s' => \$_trunk, 'tags|t=s' => \$_tags,
92 'branches|b=s' => \$_branches, 'prefix=s' => \$_prefix,
93 'stdlayout|s' => \$_stdlayout,
94 'minimize-url|m' => \$Git::SVN::_minimize_url,
95 'no-metadata' => sub { $icv{noMetadata} = 1 },
96 'use-svm-props' => sub { $icv{useSvmProps} = 1 },
97 'use-svnsync-props' => sub { $icv{useSvnsyncProps} = 1 },
98 'rewrite-root=s' => sub { $icv{rewriteRoot} = $_[1] },
99 %remote_opts );
100 my %cmt_opts = ( 'edit|e' => \$_edit,
101 'rmdir' => \$SVN::Git::Editor::_rmdir,
102 'find-copies-harder' => \$SVN::Git::Editor::_find_copies_harder,
103 'l=i' => \$SVN::Git::Editor::_rename_limit,
104 'copy-similarity|C=i'=> \$SVN::Git::Editor::_cp_similarity
105 );
106
107 my %cmd = (
108 fetch => [ \&cmd_fetch, "Download new revisions from SVN",
109 { 'revision|r=s' => \$_revision,
110 'fetch-all|all' => \$_fetch_all,
111 %fc_opts } ],
112 clone => [ \&cmd_clone, "Initialize and fetch revisions",
113 { 'revision|r=s' => \$_revision,
114 %fc_opts, %init_opts } ],
115 init => [ \&cmd_init, "Initialize a repo for tracking" .
116 " (requires URL argument)",
117 \%init_opts ],
118 'multi-init' => [ \&cmd_multi_init,
119 "Deprecated alias for ".
120 "'$0 init -T<trunk> -b<branches> -t<tags>'",
121 \%init_opts ],
122 dcommit => [ \&cmd_dcommit,
123 'Commit several diffs to merge with upstream',
124 { 'merge|m|M' => \$_merge,
125 'strategy|s=s' => \$_strategy,
126 'verbose|v' => \$_verbose,
127 'dry-run|n' => \$_dry_run,
128 'fetch-all|all' => \$_fetch_all,
129 'no-rebase' => \$_no_rebase,
130 %cmt_opts, %fc_opts } ],
131 'set-tree' => [ \&cmd_set_tree,
132 "Set an SVN repository to a git tree-ish",
133 { 'stdin|' => \$_stdin, %cmt_opts, %fc_opts, } ],
134 'create-ignore' => [ \&cmd_create_ignore,
135 'Create a .gitignore per svn:ignore',
136 { 'revision|r=i' => \$_revision
137 } ],
138 'propget' => [ \&cmd_propget,
139 'Print the value of a property on a file or directory',
140 { 'revision|r=i' => \$_revision } ],
141 'proplist' => [ \&cmd_proplist,
142 'List all properties of a file or directory',
143 { 'revision|r=i' => \$_revision } ],
144 'show-ignore' => [ \&cmd_show_ignore, "Show svn:ignore listings",
145 { 'revision|r=i' => \$_revision
146 } ],
147 'show-externals' => [ \&cmd_show_externals, "Show svn:externals listings",
148 { 'revision|r=i' => \$_revision
149 } ],
150 'multi-fetch' => [ \&cmd_multi_fetch,
151 "Deprecated alias for $0 fetch --all",
152 { 'revision|r=s' => \$_revision, %fc_opts } ],
153 'migrate' => [ sub { },
154 # no-op, we automatically run this anyways,
155 'Migrate configuration/metadata/layout from
156 previous versions of git-svn',
157 { 'minimize' => \$Git::SVN::Migration::_minimize,
158 %remote_opts } ],
159 'log' => [ \&Git::SVN::Log::cmd_show_log, 'Show commit logs',
160 { 'limit=i' => \$Git::SVN::Log::limit,
161 'revision|r=s' => \$_revision,
162 'verbose|v' => \$Git::SVN::Log::verbose,
163 'incremental' => \$Git::SVN::Log::incremental,
164 'oneline' => \$Git::SVN::Log::oneline,
165 'show-commit' => \$Git::SVN::Log::show_commit,
166 'non-recursive' => \$Git::SVN::Log::non_recursive,
167 'authors-file|A=s' => \$_authors,
168 'color' => \$Git::SVN::Log::color,
169 'pager=s' => \$Git::SVN::Log::pager
170 } ],
171 'find-rev' => [ \&cmd_find_rev, "Translate between SVN revision numbers and tree-ish",
172 {} ],
173 'rebase' => [ \&cmd_rebase, "Fetch and rebase your working directory",
174 { 'merge|m|M' => \$_merge,
175 'verbose|v' => \$_verbose,
176 'strategy|s=s' => \$_strategy,
177 'local|l' => \$_local,
178 'fetch-all|all' => \$_fetch_all,
179 %fc_opts } ],
180 'commit-diff' => [ \&cmd_commit_diff,
181 'Commit a diff between two trees',
182 { 'message|m=s' => \$_message,
183 'file|F=s' => \$_file,
184 'revision|r=s' => \$_revision,
185 %cmt_opts } ],
186 'info' => [ \&cmd_info,
187 "Show info about the latest SVN revision
188 on the current branch",
189 { 'url' => \$_url, } ],
190 'blame' => [ \&Git::SVN::Log::cmd_blame,
191 "Show what revision and author last modified each line of a file",
192 {} ],
193 );
194
195 my $cmd;
196 for (my $i = 0; $i < @ARGV; $i++) {
197 if (defined $cmd{$ARGV[$i]}) {
198 $cmd = $ARGV[$i];
199 splice @ARGV, $i, 1;
200 last;
201 }
202 };
203
204 # make sure we're always running at the top-level working directory
205 unless ($cmd && $cmd =~ /(?:clone|init|multi-init)$/) {
206 unless (-d $ENV{GIT_DIR}) {
207 if ($git_dir_user_set) {
208 die "GIT_DIR=$ENV{GIT_DIR} explicitly set, ",
209 "but it is not a directory\n";
210 }
211 my $git_dir = delete $ENV{GIT_DIR};
212 chomp(my $cdup = command_oneline(qw/rev-parse --show-cdup/));
213 unless (length $cdup) {
214 die "Already at toplevel, but $git_dir ",
215 "not found '$cdup'\n";
216 }
217 chdir $cdup or die "Unable to chdir up to '$cdup'\n";
218 unless (-d $git_dir) {
219 die "$git_dir still not found after going to ",
220 "'$cdup'\n";
221 }
222 $ENV{GIT_DIR} = $git_dir;
223 }
224 }
225
226 my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
227
228 read_repo_config(\%opts);
229 Getopt::Long::Configure('pass_through') if ($cmd && $cmd eq 'log');
230 my $rv = GetOptions(%opts, 'help|H|h' => \$_help, 'version|V' => \$_version,
231 'minimize-connections' => \$Git::SVN::Migration::_minimize,
232 'id|i=s' => \$Git::SVN::default_ref_id,
233 'svn-remote|remote|R=s' => sub {
234 $Git::SVN::no_reuse_existing = 1;
235 $Git::SVN::default_repo_id = $_[1] });
236 exit 1 if (!$rv && $cmd && $cmd ne 'log');
237
238 usage(0) if $_help;
239 version() if $_version;
240 usage(1) unless defined $cmd;
241 load_authors() if $_authors;
242
243 unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
244 Git::SVN::Migration::migration_check();
245 }
246 Git::SVN::init_vars();
247 eval {
248 Git::SVN::verify_remotes_sanity();
249 $cmd{$cmd}->[0]->(@ARGV);
250 };
251 fatal $@ if $@;
252 post_fetch_checkout();
253 exit 0;
254
255 ####################### primary functions ######################
256 sub usage {
257 my $exit = shift || 0;
258 my $fd = $exit ? \*STDERR : \*STDOUT;
259 print $fd <<"";
260 git-svn - bidirectional operations between a single Subversion tree and git
261 Usage: $0 <command> [options] [arguments]\n
262
263 print $fd "Available commands:\n" unless $cmd;
264
265 foreach (sort keys %cmd) {
266 next if $cmd && $cmd ne $_;
267 next if /^multi-/; # don't show deprecated commands
268 print $fd ' ',pack('A17',$_),$cmd{$_}->[1],"\n";
269 foreach (sort keys %{$cmd{$_}->[2]}) {
270 # mixed-case options are for .git/config only
271 next if /[A-Z]/ && /^[a-z]+$/i;
272 # prints out arguments as they should be passed:
273 my $x = s#[:=]s$## ? '<arg>' : s#[:=]i$## ? '<num>' : '';
274 print $fd ' ' x 21, join(', ', map { length $_ > 1 ?
275 "--$_" : "-$_" }
276 split /\|/,$_)," $x\n";
277 }
278 }
279 print $fd <<"";
280 \nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
281 arbitrary identifier if you're tracking multiple SVN branches/repositories in
282 one git repository and want to keep them separate. See git-svn(1) for more
283 information.
284
285 exit $exit;
286 }
287
288 sub version {
289 print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
290 exit 0;
291 }
292
293 sub do_git_init_db {
294 unless (-d $ENV{GIT_DIR}) {
295 my @init_db = ('init');
296 push @init_db, "--template=$_template" if defined $_template;
297 if (defined $_shared) {
298 if ($_shared =~ /[a-z]/) {
299 push @init_db, "--shared=$_shared";
300 } else {
301 push @init_db, "--shared";
302 }
303 }
304 command_noisy(@init_db);
305 }
306 my $set;
307 my $pfx = "svn-remote.$Git::SVN::default_repo_id";
308 foreach my $i (keys %icv) {
309 die "'$set' and '$i' cannot both be set\n" if $set;
310 next unless defined $icv{$i};
311 command_noisy('config', "$pfx.$i", $icv{$i});
312 $set = $i;
313 }
314 }
315
316 sub init_subdir {
317 my $repo_path = shift or return;
318 mkpath([$repo_path]) unless -d $repo_path;
319 chdir $repo_path or die "Couldn't chdir to $repo_path: $!\n";
320 $ENV{GIT_DIR} = '.git';
321 }
322
323 sub cmd_clone {
324 my ($url, $path) = @_;
325 if (!defined $path &&
326 (defined $_trunk || defined $_branches || defined $_tags ||
327 defined $_stdlayout) &&
328 $url !~ m#^[a-z\+]+://#) {
329 $path = $url;
330 }
331 $path = basename($url) if !defined $path || !length $path;
332 cmd_init($url, $path);
333 Git::SVN::fetch_all($Git::SVN::default_repo_id);
334 }
335
336 sub cmd_init {
337 if (defined $_stdlayout) {
338 $_trunk = 'trunk' if (!defined $_trunk);
339 $_tags = 'tags' if (!defined $_tags);
340 $_branches = 'branches' if (!defined $_branches);
341 }
342 if (defined $_trunk || defined $_branches || defined $_tags) {
343 return cmd_multi_init(@_);
344 }
345 my $url = shift or die "SVN repository location required ",
346 "as a command-line argument\n";
347 init_subdir(@_);
348 do_git_init_db();
349
350 Git::SVN->init($url);
351 }
352
353 sub cmd_fetch {
354 if (grep /^\d+=./, @_) {
355 die "'<rev>=<commit>' fetch arguments are ",
356 "no longer supported.\n";
357 }
358 my ($remote) = @_;
359 if (@_ > 1) {
360 die "Usage: $0 fetch [--all] [svn-remote]\n";
361 }
362 $remote ||= $Git::SVN::default_repo_id;
363 if ($_fetch_all) {
364 cmd_multi_fetch();
365 } else {
366 Git::SVN::fetch_all($remote, Git::SVN::read_all_remotes());
367 }
368 }
369
370 sub cmd_set_tree {
371 my (@commits) = @_;
372 if ($_stdin || !@commits) {
373 print "Reading from stdin...\n";
374 @commits = ();
375 while (<STDIN>) {
376 if (/\b($sha1_short)\b/o) {
377 unshift @commits, $1;
378 }
379 }
380 }
381 my @revs;
382 foreach my $c (@commits) {
383 my @tmp = command('rev-parse',$c);
384 if (scalar @tmp == 1) {
385 push @revs, $tmp[0];
386 } elsif (scalar @tmp > 1) {
387 push @revs, reverse(command('rev-list',@tmp));
388 } else {
389 fatal "Failed to rev-parse $c";
390 }
391 }
392 my $gs = Git::SVN->new;
393 my ($r_last, $cmt_last) = $gs->last_rev_commit;
394 $gs->fetch;
395 if (defined $gs->{last_rev} && $r_last != $gs->{last_rev}) {
396 fatal "There are new revisions that were fetched ",
397 "and need to be merged (or acknowledged) ",
398 "before committing.\nlast rev: $r_last\n",
399 " current: $gs->{last_rev}";
400 }
401 $gs->set_tree($_) foreach @revs;
402 print "Done committing ",scalar @revs," revisions to SVN\n";
403 unlink $gs->{index};
404 }
405
406 sub cmd_dcommit {
407 my $head = shift;
408 git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
409 'Cannot dcommit with a dirty index. Commit your changes first, '
410 . "or stash them with `git stash'.\n";
411 $head ||= 'HEAD';
412 my @refs;
413 my ($url, $rev, $uuid, $gs) = working_head_info($head, \@refs);
414 if ($url) {
415 print "Committing to $url ...\n";
416 }
417 unless ($gs) {
418 die "Unable to determine upstream SVN information from ",
419 "$head history.\nPerhaps the repository is empty.";
420 }
421 my $last_rev;
422 my ($linear_refs, $parents) = linearize_history($gs, \@refs);
423 if ($_no_rebase && scalar(@$linear_refs) > 1) {
424 warn "Attempting to commit more than one change while ",
425 "--no-rebase is enabled.\n",
426 "If these changes depend on each other, re-running ",
427 "without --no-rebase may be required."
428 }
429 while (1) {
430 my $d = shift @$linear_refs or last;
431 unless (defined $last_rev) {
432 (undef, $last_rev, undef) = cmt_metadata("$d~1");
433 unless (defined $last_rev) {
434 fatal "Unable to extract revision information ",
435 "from commit $d~1";
436 }
437 }
438 if ($_dry_run) {
439 print "diff-tree $d~1 $d\n";
440 } else {
441 my $cmt_rev;
442 my %ed_opts = ( r => $last_rev,
443 log => get_commit_entry($d)->{log},
444 ra => Git::SVN::Ra->new($gs->full_url),
445 config => SVN::Core::config_get_config(
446 $Git::SVN::Ra::config_dir
447 ),
448 tree_a => "$d~1",
449 tree_b => $d,
450 editor_cb => sub {
451 print "Committed r$_[0]\n";
452 $cmt_rev = $_[0];
453 },
454 svn_path => '');
455 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
456 print "No changes\n$d~1 == $d\n";
457 } elsif ($parents->{$d} && @{$parents->{$d}}) {
458 $gs->{inject_parents_dcommit}->{$cmt_rev} =
459 $parents->{$d};
460 }
461 $_fetch_all ? $gs->fetch_all : $gs->fetch;
462 $last_rev = $cmt_rev;
463 next if $_no_rebase;
464
465 # we always want to rebase against the current HEAD,
466 # not any head that was passed to us
467 my @diff = command('diff-tree', $d,
468 $gs->refname, '--');
469 my @finish;
470 if (@diff) {
471 @finish = rebase_cmd();
472 print STDERR "W: $d and ", $gs->refname,
473 " differ, using @finish:\n",
474 join("\n", @diff), "\n";
475 } else {
476 print "No changes between current HEAD and ",
477 $gs->refname,
478 "\nResetting to the latest ",
479 $gs->refname, "\n";
480 @finish = qw/reset --mixed/;
481 }
482 command_noisy(@finish, $gs->refname);
483 if (@diff) {
484 @refs = ();
485 my ($url_, $rev_, $uuid_, $gs_) =
486 working_head_info($head, \@refs);
487 my ($linear_refs_, $parents_) =
488 linearize_history($gs_, \@refs);
489 if (scalar(@$linear_refs) !=
490 scalar(@$linear_refs_)) {
491 fatal "# of revisions changed ",
492 "\nbefore:\n",
493 join("\n", @$linear_refs),
494 "\n\nafter:\n",
495 join("\n", @$linear_refs_), "\n",
496 'If you are attempting to commit ',
497 "merges, try running:\n\t",
498 'git rebase --interactive',
499 '--preserve-merges ',
500 $gs->refname,
501 "\nBefore dcommitting";
502 }
503 if ($url_ ne $url) {
504 fatal "URL mismatch after rebase: ",
505 "$url_ != $url";
506 }
507 if ($uuid_ ne $uuid) {
508 fatal "uuid mismatch after rebase: ",
509 "$uuid_ != $uuid";
510 }
511 # remap parents
512 my (%p, @l, $i);
513 for ($i = 0; $i < scalar @$linear_refs; $i++) {
514 my $new = $linear_refs_->[$i] or next;
515 $p{$new} =
516 $parents->{$linear_refs->[$i]};
517 push @l, $new;
518 }
519 $parents = \%p;
520 $linear_refs = \@l;
521 }
522 }
523 }
524 unlink $gs->{index};
525 }
526
527 sub cmd_find_rev {
528 my $revision_or_hash = shift or die "SVN or git revision required ",
529 "as a command-line argument\n";
530 my $result;
531 if ($revision_or_hash =~ /^r\d+$/) {
532 my $head = shift;
533 $head ||= 'HEAD';
534 my @refs;
535 my (undef, undef, undef, $gs) = working_head_info($head, \@refs);
536 unless ($gs) {
537 die "Unable to determine upstream SVN information from ",
538 "$head history\n";
539 }
540 my $desired_revision = substr($revision_or_hash, 1);
541 $result = $gs->rev_map_get($desired_revision);
542 } else {
543 my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
544 $result = $rev;
545 }
546 print "$result\n" if $result;
547 }
548
549 sub cmd_rebase {
550 command_noisy(qw/update-index --refresh/);
551 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
552 unless ($gs) {
553 die "Unable to determine upstream SVN information from ",
554 "working tree history\n";
555 }
556 if (command(qw/diff-index HEAD --/)) {
557 print STDERR "Cannot rebase with uncommited changes:\n";
558 command_noisy('status');
559 exit 1;
560 }
561 unless ($_local) {
562 # rebase will checkout for us, so no need to do it explicitly
563 $_no_checkout = 'true';
564 $_fetch_all ? $gs->fetch_all : $gs->fetch;
565 }
566 command_noisy(rebase_cmd(), $gs->refname);
567 }
568
569 sub cmd_show_ignore {
570 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
571 $gs ||= Git::SVN->new;
572 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
573 $gs->prop_walk($gs->{path}, $r, sub {
574 my ($gs, $path, $props) = @_;
575 print STDOUT "\n# $path\n";
576 my $s = $props->{'svn:ignore'} or return;
577 $s =~ s/[\r\n]+/\n/g;
578 chomp $s;
579 $s =~ s#^#$path#gm;
580 print STDOUT "$s\n";
581 });
582 }
583
584 sub cmd_show_externals {
585 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
586 $gs ||= Git::SVN->new;
587 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
588 $gs->prop_walk($gs->{path}, $r, sub {
589 my ($gs, $path, $props) = @_;
590 print STDOUT "\n# $path\n";
591 my $s = $props->{'svn:externals'} or return;
592 $s =~ s/[\r\n]+/\n/g;
593 chomp $s;
594 $s =~ s#^#$path#gm;
595 print STDOUT "$s\n";
596 });
597 }
598
599 sub cmd_create_ignore {
600 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
601 $gs ||= Git::SVN->new;
602 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
603 $gs->prop_walk($gs->{path}, $r, sub {
604 my ($gs, $path, $props) = @_;
605 # $path is of the form /path/to/dir/
606 my $ignore = '.' . $path . '.gitignore';
607 my $s = $props->{'svn:ignore'} or return;
608 open(GITIGNORE, '>', $ignore)
609 or fatal("Failed to open `$ignore' for writing: $!");
610 $s =~ s/[\r\n]+/\n/g;
611 chomp $s;
612 # Prefix all patterns so that the ignore doesn't apply
613 # to sub-directories.
614 $s =~ s#^#/#gm;
615 print GITIGNORE "$s\n";
616 close(GITIGNORE)
617 or fatal("Failed to close `$ignore': $!");
618 command_noisy('add', $ignore);
619 });
620 }
621
622 sub canonicalize_path {
623 my ($path) = @_;
624 my $dot_slash_added = 0;
625 if (substr($path, 0, 1) ne "/") {
626 $path = "./" . $path;
627 $dot_slash_added = 1;
628 }
629 # File::Spec->canonpath doesn't collapse x/../y into y (for a
630 # good reason), so let's do this manually.
631 $path =~ s#/+#/#g;
632 $path =~ s#/\.(?:/|$)#/#g;
633 $path =~ s#/[^/]+/\.\.##g;
634 $path =~ s#/$##g;
635 $path =~ s#^\./## if $dot_slash_added;
636 return $path;
637 }
638
639 # get_svnprops(PATH)
640 # ------------------
641 # Helper for cmd_propget and cmd_proplist below.
642 sub get_svnprops {
643 my $path = shift;
644 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
645 $gs ||= Git::SVN->new;
646
647 # prefix THE PATH by the sub-directory from which the user
648 # invoked us.
649 $path = $cmd_dir_prefix . $path;
650 fatal("No such file or directory: $path") unless -e $path;
651 my $is_dir = -d $path ? 1 : 0;
652 $path = $gs->{path} . '/' . $path;
653
654 # canonicalize the path (otherwise libsvn will abort or fail to
655 # find the file)
656 $path = canonicalize_path($path);
657
658 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
659 my $props;
660 if ($is_dir) {
661 (undef, undef, $props) = $gs->ra->get_dir($path, $r);
662 }
663 else {
664 (undef, $props) = $gs->ra->get_file($path, $r, undef);
665 }
666 return $props;
667 }
668
669 # cmd_propget (PROP, PATH)
670 # ------------------------
671 # Print the SVN property PROP for PATH.
672 sub cmd_propget {
673 my ($prop, $path) = @_;
674 $path = '.' if not defined $path;
675 usage(1) if not defined $prop;
676 my $props = get_svnprops($path);
677 if (not defined $props->{$prop}) {
678 fatal("`$path' does not have a `$prop' SVN property.");
679 }
680 print $props->{$prop} . "\n";
681 }
682
683 # cmd_proplist (PATH)
684 # -------------------
685 # Print the list of SVN properties for PATH.
686 sub cmd_proplist {
687 my $path = shift;
688 $path = '.' if not defined $path;
689 my $props = get_svnprops($path);
690 print "Properties on '$path':\n";
691 foreach (sort keys %{$props}) {
692 print " $_\n";
693 }
694 }
695
696 sub cmd_multi_init {
697 my $url = shift;
698 unless (defined $_trunk || defined $_branches || defined $_tags) {
699 usage(1);
700 }
701
702 # there are currently some bugs that prevent multi-init/multi-fetch
703 # setups from working well without this.
704 $Git::SVN::_minimize_url = 1;
705
706 $_prefix = '' unless defined $_prefix;
707 if (defined $url) {
708 $url =~ s#/+$##;
709 init_subdir(@_);
710 }
711 do_git_init_db();
712 if (defined $_trunk) {
713 my $trunk_ref = $_prefix . 'trunk';
714 # try both old-style and new-style lookups:
715 my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
716 unless ($gs_trunk) {
717 my ($trunk_url, $trunk_path) =
718 complete_svn_url($url, $_trunk);
719 $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
720 undef, $trunk_ref);
721 }
722 }
723 return unless defined $_branches || defined $_tags;
724 my $ra = $url ? Git::SVN::Ra->new($url) : undef;
725 complete_url_ls_init($ra, $_branches, '--branches/-b', $_prefix);
726 complete_url_ls_init($ra, $_tags, '--tags/-t', $_prefix . 'tags/');
727 }
728
729 sub cmd_multi_fetch {
730 my $remotes = Git::SVN::read_all_remotes();
731 foreach my $repo_id (sort keys %$remotes) {
732 if ($remotes->{$repo_id}->{url}) {
733 Git::SVN::fetch_all($repo_id, $remotes);
734 }
735 }
736 }
737
738 # this command is special because it requires no metadata
739 sub cmd_commit_diff {
740 my ($ta, $tb, $url) = @_;
741 my $usage = "Usage: $0 commit-diff -r<revision> ".
742 "<tree-ish> <tree-ish> [<URL>]";
743 fatal($usage) if (!defined $ta || !defined $tb);
744 my $svn_path;
745 if (!defined $url) {
746 my $gs = eval { Git::SVN->new };
747 if (!$gs) {
748 fatal("Needed URL or usable git-svn --id in ",
749 "the command-line\n", $usage);
750 }
751 $url = $gs->{url};
752 $svn_path = $gs->{path};
753 }
754 unless (defined $_revision) {
755 fatal("-r|--revision is a required argument\n", $usage);
756 }
757 if (defined $_message && defined $_file) {
758 fatal("Both --message/-m and --file/-F specified ",
759 "for the commit message.\n",
760 "I have no idea what you mean");
761 }
762 if (defined $_file) {
763 $_message = file_to_s($_file);
764 } else {
765 $_message ||= get_commit_entry($tb)->{log};
766 }
767 my $ra ||= Git::SVN::Ra->new($url);
768 $svn_path ||= $ra->{svn_path};
769 my $r = $_revision;
770 if ($r eq 'HEAD') {
771 $r = $ra->get_latest_revnum;
772 } elsif ($r !~ /^\d+$/) {
773 die "revision argument: $r not understood by git-svn\n";
774 }
775 my %ed_opts = ( r => $r,
776 log => $_message,
777 ra => $ra,
778 tree_a => $ta,
779 tree_b => $tb,
780 editor_cb => sub { print "Committed r$_[0]\n" },
781 svn_path => $svn_path );
782 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
783 print "No changes\n$ta == $tb\n";
784 }
785 }
786
787 sub cmd_info {
788 my $path = canonicalize_path(shift or ".");
789 unless (scalar(@_) == 0) {
790 die "Too many arguments specified\n";
791 }
792
793 my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
794
795 if (!$file_type && !$diff_status) {
796 print STDERR "$path: (Not a versioned resource)\n\n";
797 return;
798 }
799
800 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
801 unless ($gs) {
802 die "Unable to determine upstream SVN information from ",
803 "working tree history\n";
804 }
805 my $full_url = $url . ($path eq "." ? "" : "/$path");
806
807 if ($_url) {
808 print $full_url, "\n";
809 return;
810 }
811
812 my $result = "Path: $path\n";
813 $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
814 $result .= "URL: " . $full_url . "\n";
815
816 eval {
817 my $repos_root = $gs->repos_root;
818 Git::SVN::remove_username($repos_root);
819 $result .= "Repository Root: $repos_root\n";
820 };
821 if ($@) {
822 $result .= "Repository Root: (offline)\n";
823 }
824 $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A";
825 $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
826
827 $result .= "Node Kind: " .
828 ($file_type eq "dir" ? "directory" : "file") . "\n";
829
830 my $schedule = $diff_status eq "A"
831 ? "add"
832 : ($diff_status eq "D" ? "delete" : "normal");
833 $result .= "Schedule: $schedule\n";
834
835 if ($diff_status eq "A") {
836 print $result, "\n";
837 return;
838 }
839
840 my ($lc_author, $lc_rev, $lc_date_utc);
841 my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $path);
842 my $log = command_output_pipe(@args);
843 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
844 while (<$log>) {
845 if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
846 $lc_author = $1;
847 $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
848 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
849 (undef, $lc_rev, undef) = ::extract_metadata($1);
850 }
851 }
852 close $log;
853
854 Git::SVN::Log::set_local_timezone();
855
856 $result .= "Last Changed Author: $lc_author\n";
857 $result .= "Last Changed Rev: $lc_rev\n";
858 $result .= "Last Changed Date: " .
859 Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
860
861 if ($file_type ne "dir") {
862 my $text_last_updated_date =
863 ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
864 $result .=
865 "Text Last Updated: " .
866 Git::SVN::Log::format_svn_date($text_last_updated_date) .
867 "\n";
868 my $checksum;
869 if ($diff_status eq "D") {
870 my ($fh, $ctx) =
871 command_output_pipe(qw(cat-file blob), "HEAD:$path");
872 if ($file_type eq "link") {
873 my $file_name = <$fh>;
874 $checksum = md5sum("link $file_name");
875 } else {
876 $checksum = md5sum($fh);
877 }
878 command_close_pipe($fh, $ctx);
879 } elsif ($file_type eq "link") {
880 my $file_name =
881 command(qw(cat-file blob), "HEAD:$path");
882 $checksum =
883 md5sum("link " . $file_name);
884 } else {
885 open FILE, "<", $path or die $!;
886 $checksum = md5sum(\*FILE);
887 close FILE or die $!;
888 }
889 $result .= "Checksum: " . $checksum . "\n";
890 }
891
892 print $result, "\n";
893 }
894
895 ########################### utility functions #########################
896
897 sub rebase_cmd {
898 my @cmd = qw/rebase/;
899 push @cmd, '-v' if $_verbose;
900 push @cmd, qw/--merge/ if $_merge;
901 push @cmd, "--strategy=$_strategy" if $_strategy;
902 @cmd;
903 }
904
905 sub post_fetch_checkout {
906 return if $_no_checkout;
907 my $gs = $Git::SVN::_head or return;
908 return if verify_ref('refs/heads/master^0');
909
910 my $valid_head = verify_ref('HEAD^0');
911 command_noisy(qw(update-ref refs/heads/master), $gs->refname);
912 return if ($valid_head || !verify_ref('HEAD^0'));
913
914 return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
915 my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
916 return if -f $index;
917
918 return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
919 return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
920 command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
921 print STDERR "Checked out HEAD:\n ",
922 $gs->full_url, " r", $gs->last_rev, "\n";
923 }
924
925 sub complete_svn_url {
926 my ($url, $path) = @_;
927 $path =~ s#/+$##;
928 if ($path !~ m#^[a-z\+]+://#) {
929 if (!defined $url || $url !~ m#^[a-z\+]+://#) {
930 fatal("E: '$path' is not a complete URL ",
931 "and a separate URL is not specified");
932 }
933 return ($url, $path);
934 }
935 return ($path, '');
936 }
937
938 sub complete_url_ls_init {
939 my ($ra, $repo_path, $switch, $pfx) = @_;
940 unless ($repo_path) {
941 print STDERR "W: $switch not specified\n";
942 return;
943 }
944 $repo_path =~ s#/+$##;
945 if ($repo_path =~ m#^[a-z\+]+://#) {
946 $ra = Git::SVN::Ra->new($repo_path);
947 $repo_path = '';
948 } else {
949 $repo_path =~ s#^/+##;
950 unless ($ra) {
951 fatal("E: '$repo_path' is not a complete URL ",
952 "and a separate URL is not specified");
953 }
954 }
955 my $url = $ra->{url};
956 my $gs = Git::SVN->init($url, undef, undef, undef, 1);
957 my $k = "svn-remote.$gs->{repo_id}.url";
958 my $orig_url = eval { command_oneline(qw/config --get/, $k) };
959 if ($orig_url && ($orig_url ne $gs->{url})) {
960 die "$k already set: $orig_url\n",
961 "wanted to set to: $gs->{url}\n";
962 }
963 command_oneline('config', $k, $gs->{url}) unless $orig_url;
964 my $remote_path = "$ra->{svn_path}/$repo_path";
965 $remote_path =~ s#/+#/#g;
966 $remote_path =~ s#^/##g;
967 $remote_path .= "/*" if $remote_path !~ /\*/;
968 my ($n) = ($switch =~ /^--(\w+)/);
969 if (length $pfx && $pfx !~ m#/$#) {
970 die "--prefix='$pfx' must have a trailing slash '/'\n";
971 }
972 command_noisy('config', "svn-remote.$gs->{repo_id}.$n",
973 "$remote_path:refs/remotes/$pfx*");
974 }
975
976 sub verify_ref {
977 my ($ref) = @_;
978 eval { command_oneline([ 'rev-parse', '--verify', $ref ],
979 { STDERR => 0 }); };
980 }
981
982 sub get_tree_from_treeish {
983 my ($treeish) = @_;
984 # $treeish can be a symbolic ref, too:
985 my $type = command_oneline(qw/cat-file -t/, $treeish);
986 my $expected;
987 while ($type eq 'tag') {
988 ($treeish, $type) = command(qw/cat-file tag/, $treeish);
989 }
990 if ($type eq 'commit') {
991 $expected = (grep /^tree /, command(qw/cat-file commit/,
992 $treeish))[0];
993 ($expected) = ($expected =~ /^tree ($sha1)$/o);
994 die "Unable to get tree from $treeish\n" unless $expected;
995 } elsif ($type eq 'tree') {
996 $expected = $treeish;
997 } else {
998 die "$treeish is a $type, expected tree, tag or commit\n";
999 }
1000 return $expected;
1001 }
1002
1003 sub get_commit_entry {
1004 my ($treeish) = shift;
1005 my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
1006 my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
1007 my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
1008 open my $log_fh, '>', $commit_editmsg or croak $!;
1009
1010 my $type = command_oneline(qw/cat-file -t/, $treeish);
1011 if ($type eq 'commit' || $type eq 'tag') {
1012 my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1013 $type, $treeish);
1014 my $in_msg = 0;
1015 my $author;
1016 my $saw_from = 0;
1017 while (<$msg_fh>) {
1018 if (!$in_msg) {
1019 $in_msg = 1 if (/^\s*$/);
1020 $author = $1 if (/^author (.*>)/);
1021 } elsif (/^git-svn-id: /) {
1022 # skip this for now, we regenerate the
1023 # correct one on re-fetch anyways
1024 # TODO: set *:merge properties or like...
1025 } else {
1026 if (/^From:/ || /^Signed-off-by:/) {
1027 $saw_from = 1;
1028 }
1029 print $log_fh $_ or croak $!;
1030 }
1031 }
1032 if ($Git::SVN::_add_author_from && defined($author)
1033 && !$saw_from) {
1034 print $log_fh "\nFrom: $author\n"
1035 or croak $!;
1036 }
1037 command_close_pipe($msg_fh, $ctx);
1038 }
1039 close $log_fh or croak $!;
1040
1041 if ($_edit || ($type eq 'tree')) {
1042 my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
1043 # TODO: strip out spaces, comments, like git-commit.sh
1044 system($editor, $commit_editmsg);
1045 }
1046 rename $commit_editmsg, $commit_msg or croak $!;
1047 open $log_fh, '<', $commit_msg or croak $!;
1048 { local $/; chomp($log_entry{log} = <$log_fh>); }
1049 close $log_fh or croak $!;
1050 unlink $commit_msg;
1051 \%log_entry;
1052 }
1053
1054 sub s_to_file {
1055 my ($str, $file, $mode) = @_;
1056 open my $fd,'>',$file or croak $!;
1057 print $fd $str,"\n" or croak $!;
1058 close $fd or croak $!;
1059 chmod ($mode &~ umask, $file) if (defined $mode);
1060 }
1061
1062 sub file_to_s {
1063 my $file = shift;
1064 open my $fd,'<',$file or croak "$!: file: $file\n";
1065 local $/;
1066 my $ret = <$fd>;
1067 close $fd or croak $!;
1068 $ret =~ s/\s*$//s;
1069 return $ret;
1070 }
1071
1072 # '<svn username> = real-name <email address>' mapping based on git-svnimport:
1073 sub load_authors {
1074 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1075 my $log = $cmd eq 'log';
1076 while (<$authors>) {
1077 chomp;
1078 next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1079 my ($user, $name, $email) = ($1, $2, $3);
1080 if ($log) {
1081 $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1082 } else {
1083 $users{$user} = [$name, $email];
1084 }
1085 }
1086 close $authors or croak $!;
1087 }
1088
1089 # convert GetOpt::Long specs for use by git-config
1090 sub read_repo_config {
1091 return unless -d $ENV{GIT_DIR};
1092 my $opts = shift;
1093 my @config_only;
1094 foreach my $o (keys %$opts) {
1095 # if we have mixedCase and a long option-only, then
1096 # it's a config-only variable that we don't need for
1097 # the command-line.
1098 push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1099 my $v = $opts->{$o};
1100 my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1101 $key =~ s/-//g;
1102 my $arg = 'git-config';
1103 $arg .= ' --int' if ($o =~ /[:=]i$/);
1104 $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1105 if (ref $v eq 'ARRAY') {
1106 chomp(my @tmp = `$arg --get-all svn.$key`);
1107 @$v = @tmp if @tmp;
1108 } else {
1109 chomp(my $tmp = `$arg --get svn.$key`);
1110 if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1111 $$v = $tmp;
1112 }
1113 }
1114 }
1115 delete @$opts{@config_only} if @config_only;
1116 }
1117
1118 sub extract_metadata {
1119 my $id = shift or return (undef, undef, undef);
1120 my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1121 \s([a-f\d\-]+)$/x);
1122 if (!defined $rev || !$uuid || !$url) {
1123 # some of the original repositories I made had
1124 # identifiers like this:
1125 ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/);
1126 }
1127 return ($url, $rev, $uuid);
1128 }
1129
1130 sub cmt_metadata {
1131 return extract_metadata((grep(/^git-svn-id: /,
1132 command(qw/cat-file commit/, shift)))[-1]);
1133 }
1134
1135 sub working_head_info {
1136 my ($head, $refs) = @_;
1137 my @args = ('log', '--no-color', '--first-parent', '--pretty=medium');
1138 my ($fh, $ctx) = command_output_pipe(@args, $head);
1139 my $hash;
1140 my %max;
1141 while (<$fh>) {
1142 if ( m{^commit ($::sha1)$} ) {
1143 unshift @$refs, $hash if $hash and $refs;
1144 $hash = $1;
1145 next;
1146 }
1147 next unless s{^\s*(git-svn-id:)}{$1};
1148 my ($url, $rev, $uuid) = extract_metadata($_);
1149 if (defined $url && defined $rev) {
1150 next if $max{$url} and $max{$url} < $rev;
1151 if (my $gs = Git::SVN->find_by_url($url)) {
1152 my $c = $gs->rev_map_get($rev);
1153 if ($c && $c eq $hash) {
1154 close $fh; # break the pipe
1155 return ($url, $rev, $uuid, $gs);
1156 } else {
1157 $max{$url} ||= $gs->rev_map_max;
1158 }
1159 }
1160 }
1161 }
1162 command_close_pipe($fh, $ctx);
1163 (undef, undef, undef, undef);
1164 }
1165
1166 sub read_commit_parents {
1167 my ($parents, $c) = @_;
1168 chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1169 $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1170 @{$parents->{$c}} = split(/ /, $p);
1171 }
1172
1173 sub linearize_history {
1174 my ($gs, $refs) = @_;
1175 my %parents;
1176 foreach my $c (@$refs) {
1177 read_commit_parents(\%parents, $c);
1178 }
1179
1180 my @linear_refs;
1181 my %skip = ();
1182 my $last_svn_commit = $gs->last_commit;
1183 foreach my $c (reverse @$refs) {
1184 next if $c eq $last_svn_commit;
1185 last if $skip{$c};
1186
1187 unshift @linear_refs, $c;
1188 $skip{$c} = 1;
1189
1190 # we only want the first parent to diff against for linear
1191 # history, we save the rest to inject when we finalize the
1192 # svn commit
1193 my $fp_a = verify_ref("$c~1");
1194 my $fp_b = shift @{$parents{$c}} if $parents{$c};
1195 if (!$fp_a || !$fp_b) {
1196 die "Commit $c\n",
1197 "has no parent commit, and therefore ",
1198 "nothing to diff against.\n",
1199 "You should be working from a repository ",
1200 "originally created by git-svn\n";
1201 }
1202 if ($fp_a ne $fp_b) {
1203 die "$c~1 = $fp_a, however parsing commit $c ",
1204 "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1205 }
1206
1207 foreach my $p (@{$parents{$c}}) {
1208 $skip{$p} = 1;
1209 }
1210 }
1211 (\@linear_refs, \%parents);
1212 }
1213
1214 sub find_file_type_and_diff_status {
1215 my ($path) = @_;
1216 return ('dir', '') if $path eq '.';
1217
1218 my $diff_output =
1219 command_oneline(qw(diff --cached --name-status --), $path) || "";
1220 my $diff_status = (split(' ', $diff_output))[0] || "";
1221
1222 my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1223
1224 return (undef, undef) if !$diff_status && !$ls_tree;
1225
1226 if ($diff_status eq "A") {
1227 return ("link", $diff_status) if -l $path;
1228 return ("dir", $diff_status) if -d $path;
1229 return ("file", $diff_status);
1230 }
1231
1232 my $mode = (split(' ', $ls_tree))[0] || "";
1233
1234 return ("link", $diff_status) if $mode eq "120000";
1235 return ("dir", $diff_status) if $mode eq "040000";
1236 return ("file", $diff_status);
1237 }
1238
1239 sub md5sum {
1240 my $arg = shift;
1241 my $ref = ref $arg;
1242 my $md5 = Digest::MD5->new();
1243 if ($ref eq 'GLOB' || $ref eq 'IO::File') {
1244 $md5->addfile($arg) or croak $!;
1245 } elsif ($ref eq 'SCALAR') {
1246 $md5->add($$arg) or croak $!;
1247 } elsif (!$ref) {
1248 $md5->add($arg) or croak $!;
1249 } else {
1250 ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
1251 }
1252 return $md5->hexdigest();
1253 }
1254
1255 package Git::SVN;
1256 use strict;
1257 use warnings;
1258 use Fcntl qw/:DEFAULT :seek/;
1259 use constant rev_map_fmt => 'NH40';
1260 use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1261 $_repack $_repack_flags $_use_svm_props $_head
1262 $_use_svnsync_props $no_reuse_existing $_minimize_url
1263 $_use_log_author $_add_author_from/;
1264 use Carp qw/croak/;
1265 use File::Path qw/mkpath/;
1266 use File::Copy qw/copy/;
1267 use IPC::Open3;
1268
1269 my ($_gc_nr, $_gc_period);
1270
1271 # properties that we do not log:
1272 my %SKIP_PROP;
1273 BEGIN {
1274 %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1275 svn:special svn:executable
1276 svn:entry:committed-rev
1277 svn:entry:last-author
1278 svn:entry:uuid
1279 svn:entry:committed-date/;
1280
1281 # some options are read globally, but can be overridden locally
1282 # per [svn-remote "..."] section. Command-line options will *NOT*
1283 # override options set in an [svn-remote "..."] section
1284 no strict 'refs';
1285 for my $option (qw/follow_parent no_metadata use_svm_props
1286 use_svnsync_props/) {
1287 my $key = $option;
1288 $key =~ tr/_//d;
1289 my $prop = "-$option";
1290 *$option = sub {
1291 my ($self) = @_;
1292 return $self->{$prop} if exists $self->{$prop};
1293 my $k = "svn-remote.$self->{repo_id}.$key";
1294 eval { command_oneline(qw/config --get/, $k) };
1295 if ($@) {
1296 $self->{$prop} = ${"Git::SVN::_$option"};
1297 } else {
1298 my $v = command_oneline(qw/config --bool/,$k);
1299 $self->{$prop} = $v eq 'false' ? 0 : 1;
1300 }
1301 return $self->{$prop};
1302 }
1303 }
1304 }
1305
1306 my (%LOCKFILES, %INDEX_FILES);
1307 END {
1308 unlink keys %LOCKFILES if %LOCKFILES;
1309 unlink keys %INDEX_FILES if %INDEX_FILES;
1310 }
1311
1312 sub resolve_local_globs {
1313 my ($url, $fetch, $glob_spec) = @_;
1314 return unless defined $glob_spec;
1315 my $ref = $glob_spec->{ref};
1316 my $path = $glob_spec->{path};
1317 foreach (command(qw#for-each-ref --format=%(refname) refs/remotes#)) {
1318 next unless m#^refs/remotes/$ref->{regex}$#;
1319 my $p = $1;
1320 my $pathname = desanitize_refname($path->full_path($p));
1321 my $refname = desanitize_refname($ref->full_path($p));
1322 if (my $existing = $fetch->{$pathname}) {
1323 if ($existing ne $refname) {
1324 die "Refspec conflict:\n",
1325 "existing: refs/remotes/$existing\n",
1326 " globbed: refs/remotes/$refname\n";
1327 }
1328 my $u = (::cmt_metadata("refs/remotes/$refname"))[0];
1329 $u =~ s!^\Q$url\E(/|$)!! or die
1330 "refs/remotes/$refname: '$url' not found in '$u'\n";
1331 if ($pathname ne $u) {
1332 warn "W: Refspec glob conflict ",
1333 "(ref: refs/remotes/$refname):\n",
1334 "expected path: $pathname\n",
1335 " real path: $u\n",
1336 "Continuing ahead with $u\n";
1337 next;
1338 }
1339 } else {
1340 $fetch->{$pathname} = $refname;
1341 }
1342 }
1343 }
1344
1345 sub parse_revision_argument {
1346 my ($base, $head) = @_;
1347 if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1348 return ($base, $head);
1349 }
1350 return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1351 return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1352 return ($head, $head) if ($::_revision eq 'HEAD');
1353 return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1354 return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1355 die "revision argument: $::_revision not understood by git-svn\n";
1356 }
1357
1358 sub fetch_all {
1359 my ($repo_id, $remotes) = @_;
1360 if (ref $repo_id) {
1361 my $gs = $repo_id;
1362 $repo_id = undef;
1363 $repo_id = $gs->{repo_id};
1364 }
1365 $remotes ||= read_all_remotes();
1366 my $remote = $remotes->{$repo_id} or
1367 die "[svn-remote \"$repo_id\"] unknown\n";
1368 my $fetch = $remote->{fetch};
1369 my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1370 my (@gs, @globs);
1371 my $ra = Git::SVN::Ra->new($url);
1372 my $uuid = $ra->get_uuid;
1373 my $head = $ra->get_latest_revnum;
1374 my $base = defined $fetch ? $head : 0;
1375
1376 # read the max revs for wildcard expansion (branches/*, tags/*)
1377 foreach my $t (qw/branches tags/) {
1378 defined $remote->{$t} or next;
1379 push @globs, $remote->{$t};
1380 my $max_rev = eval { tmp_config(qw/--int --get/,
1381 "svn-remote.$repo_id.${t}-maxRev") };
1382 if (defined $max_rev && ($max_rev < $base)) {
1383 $base = $max_rev;
1384 } elsif (!defined $max_rev) {
1385 $base = 0;
1386 }
1387 }
1388
1389 if ($fetch) {
1390 foreach my $p (sort keys %$fetch) {
1391 my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1392 my $lr = $gs->rev_map_max;
1393 if (defined $lr) {
1394 $base = $lr if ($lr < $base);
1395 }
1396 push @gs, $gs;
1397 }
1398 }
1399
1400 ($base, $head) = parse_revision_argument($base, $head);
1401 $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1402 }
1403
1404 sub read_all_remotes {
1405 my $r = {};
1406 foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1407 if (m!^(.+)\.fetch=\s*(.*)\s*:\s*refs/remotes/(.+)\s*$!) {
1408 my ($remote, $local_ref, $remote_ref) = ($1, $2, $3);
1409 $local_ref =~ s{^/}{};
1410 $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1411 } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1412 $r->{$1}->{url} = $2;
1413 } elsif (m!^(.+)\.(branches|tags)=
1414 (.*):refs/remotes/(.+)\s*$/!x) {
1415 my ($p, $g) = ($3, $4);
1416 my $rs = $r->{$1}->{$2} = {
1417 t => $2,
1418 remote => $1,
1419 path => Git::SVN::GlobSpec->new($p),
1420 ref => Git::SVN::GlobSpec->new($g) };
1421 if (length($rs->{ref}->{right}) != 0) {
1422 die "The '*' glob character must be the last ",
1423 "character of '$g'\n";
1424 }
1425 }
1426 }
1427 $r;
1428 }
1429
1430 sub init_vars {
1431 $_gc_nr = $_gc_period = 1000;
1432 if (defined $_repack || defined $_repack_flags) {
1433 warn "Repack options are obsolete; they have no effect.\n";
1434 }
1435 }
1436
1437 sub verify_remotes_sanity {
1438 return unless -d $ENV{GIT_DIR};
1439 my %seen;
1440 foreach (command(qw/config -l/)) {
1441 if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1442 if ($seen{$1}) {
1443 die "Remote ref refs/remote/$1 is tracked by",
1444 "\n \"$_\"\nand\n \"$seen{$1}\"\n",
1445 "Please resolve this ambiguity in ",
1446 "your git configuration file before ",
1447 "continuing\n";
1448 }
1449 $seen{$1} = $_;
1450 }
1451 }
1452 }
1453
1454 # we allow more chars than remotes2config.sh...
1455 sub sanitize_remote_name {
1456 my ($name) = @_;
1457 $name =~ tr{A-Za-z0-9:,/+-}{.}c;
1458 $name;
1459 }
1460
1461 sub find_existing_remote {
1462 my ($url, $remotes) = @_;
1463 return undef if $no_reuse_existing;
1464 my $existing;
1465 foreach my $repo_id (keys %$remotes) {
1466 my $u = $remotes->{$repo_id}->{url} or next;
1467 next if $u ne $url;
1468 $existing = $repo_id;
1469 last;
1470 }
1471 $existing;
1472 }
1473
1474 sub init_remote_config {
1475 my ($self, $url, $no_write) = @_;
1476 $url =~ s!/+$!!; # strip trailing slash
1477 my $r = read_all_remotes();
1478 my $existing = find_existing_remote($url, $r);
1479 if ($existing) {
1480 unless ($no_write) {
1481 print STDERR "Using existing ",
1482 "[svn-remote \"$existing\"]\n";
1483 }
1484 $self->{repo_id} = $existing;
1485 } elsif ($_minimize_url) {
1486 my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1487 $existing = find_existing_remote($min_url, $r);
1488 if ($existing) {
1489 unless ($no_write) {
1490 print STDERR "Using existing ",
1491 "[svn-remote \"$existing\"]\n";
1492 }
1493 $self->{repo_id} = $existing;
1494 }
1495 if ($min_url ne $url) {
1496 unless ($no_write) {
1497 print STDERR "Using higher level of URL: ",
1498 "$url => $min_url\n";
1499 }
1500 my $old_path = $self->{path};
1501 $self->{path} = $url;
1502 $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1503 if (length $old_path) {
1504 $self->{path} .= "/$old_path";
1505 }
1506 $url = $min_url;
1507 }
1508 }
1509 my $orig_url;
1510 if (!$existing) {
1511 # verify that we aren't overwriting anything:
1512 $orig_url = eval {
1513 command_oneline('config', '--get',
1514 "svn-remote.$self->{repo_id}.url")
1515 };
1516 if ($orig_url && ($orig_url ne $url)) {
1517 die "svn-remote.$self->{repo_id}.url already set: ",
1518 "$orig_url\nwanted to set to: $url\n";
1519 }
1520 }
1521 my ($xrepo_id, $xpath) = find_ref($self->refname);
1522 if (defined $xpath) {
1523 die "svn-remote.$xrepo_id.fetch already set to track ",
1524 "$xpath:refs/remotes/", $self->refname, "\n";
1525 }
1526 unless ($no_write) {
1527 command_noisy('config',
1528 "svn-remote.$self->{repo_id}.url", $url);
1529 $self->{path} =~ s{^/}{};
1530 command_noisy('config', '--add',
1531 "svn-remote.$self->{repo_id}.fetch",
1532 "$self->{path}:".$self->refname);
1533 }
1534 $self->{url} = $url;
1535 }
1536
1537 sub find_by_url { # repos_root and, path are optional
1538 my ($class, $full_url, $repos_root, $path) = @_;
1539
1540 return undef unless defined $full_url;
1541 remove_username($full_url);
1542 remove_username($repos_root) if defined $repos_root;
1543 my $remotes = read_all_remotes();
1544 if (defined $full_url && defined $repos_root && !defined $path) {
1545 $path = $full_url;
1546 $path =~ s#^\Q$repos_root\E(?:/|$)##;
1547 }
1548 foreach my $repo_id (keys %$remotes) {
1549 my $u = $remotes->{$repo_id}->{url} or next;
1550 remove_username($u);
1551 next if defined $repos_root && $repos_root ne $u;
1552
1553 my $fetch = $remotes->{$repo_id}->{fetch} || {};
1554 foreach (qw/branches tags/) {
1555 resolve_local_globs($u, $fetch,
1556 $remotes->{$repo_id}->{$_});
1557 }
1558 my $p = $path;
1559 my $rwr = rewrite_root({repo_id => $repo_id});
1560 unless (defined $p) {
1561 $p = $full_url;
1562 my $z = $u;
1563 if ($rwr) {
1564 $z = $rwr;
1565 }
1566 $p =~ s#^\Q$z\E(?:/|$)## or next;
1567 }
1568 foreach my $f (keys %$fetch) {
1569 next if $f ne $p;
1570 return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1571 }
1572 }
1573 undef;
1574 }
1575
1576 sub init {
1577 my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1578 my $self = _new($class, $repo_id, $ref_id, $path);
1579 if (defined $url) {
1580 $self->init_remote_config($url, $no_write);
1581 }
1582 $self;
1583 }
1584
1585 sub find_ref {
1586 my ($ref_id) = @_;
1587 foreach (command(qw/config -l/)) {
1588 next unless m!^svn-remote\.(.+)\.fetch=
1589 \s*(.*)\s*:\s*refs/remotes/(.+)\s*$!x;
1590 my ($repo_id, $path, $ref) = ($1, $2, $3);
1591 if ($ref eq $ref_id) {
1592 $path = '' if ($path =~ m#^\./?#);
1593 return ($repo_id, $path);
1594 }
1595 }
1596 (undef, undef, undef);
1597 }
1598
1599 sub new {
1600 my ($class, $ref_id, $repo_id, $path) = @_;
1601 if (defined $ref_id && !defined $repo_id && !defined $path) {
1602 ($repo_id, $path) = find_ref($ref_id);
1603 if (!defined $repo_id) {
1604 die "Could not find a \"svn-remote.*.fetch\" key ",
1605 "in the repository configuration matching: ",
1606 "refs/remotes/$ref_id\n";
1607 }
1608 }
1609 my $self = _new($class, $repo_id, $ref_id, $path);
1610 if (!defined $self->{path} || !length $self->{path}) {
1611 my $fetch = command_oneline('config', '--get',
1612 "svn-remote.$repo_id.fetch",
1613 ":refs/remotes/$ref_id\$") or
1614 die "Failed to read \"svn-remote.$repo_id.fetch\" ",
1615 "\":refs/remotes/$ref_id\$\" in config\n";
1616 ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
1617 }
1618 $self->{url} = command_oneline('config', '--get',
1619 "svn-remote.$repo_id.url") or
1620 die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
1621 $self->rebuild;
1622 $self;
1623 }
1624
1625 sub refname {
1626 my ($refname) = "refs/remotes/$_[0]->{ref_id}" ;
1627
1628 # It cannot end with a slash /, we'll throw up on this because
1629 # SVN can't have directories with a slash in their name, either:
1630 if ($refname =~ m{/$}) {
1631 die "ref: '$refname' ends with a trailing slash, this is ",
1632 "not permitted by git nor Subversion\n";
1633 }
1634
1635 # It cannot have ASCII control character space, tilde ~, caret ^,
1636 # colon :, question-mark ?, asterisk *, space, or open bracket [
1637 # anywhere.
1638 #
1639 # Additionally, % must be escaped because it is used for escaping
1640 # and we want our escaped refname to be reversible
1641 $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
1642
1643 # no slash-separated component can begin with a dot .
1644 # /.* becomes /%2E*
1645 $refname =~ s{/\.}{/%2E}g;
1646
1647 # It cannot have two consecutive dots .. anywhere
1648 # .. becomes %2E%2E
1649 $refname =~ s{\.\.}{%2E%2E}g;
1650
1651 return $refname;
1652 }
1653
1654 sub desanitize_refname {
1655 my ($refname) = @_;
1656 $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
1657 return $refname;
1658 }
1659
1660 sub svm_uuid {
1661 my ($self) = @_;
1662 return $self->{svm}->{uuid} if $self->svm;
1663 $self->ra;
1664 unless ($self->{svm}) {
1665 die "SVM UUID not cached, and reading remotely failed\n";
1666 }
1667 $self->{svm}->{uuid};
1668 }
1669
1670 sub svm {
1671 my ($self) = @_;
1672 return $self->{svm} if $self->{svm};
1673 my $svm;
1674 # see if we have it in our config, first:
1675 eval {
1676 my $section = "svn-remote.$self->{repo_id}";
1677 $svm = {
1678 source => tmp_config('--get', "$section.svm-source"),
1679 uuid => tmp_config('--get', "$section.svm-uuid"),
1680 replace => tmp_config('--get', "$section.svm-replace"),
1681 }
1682 };
1683 if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
1684 $self->{svm} = $svm;
1685 }
1686 $self->{svm};
1687 }
1688
1689 sub _set_svm_vars {
1690 my ($self, $ra) = @_;
1691 return $ra if $self->svm;
1692
1693 my @err = ( "useSvmProps set, but failed to read SVM properties\n",
1694 "(svm:source, svm:uuid) ",
1695 "from the following URLs:\n" );
1696 sub read_svm_props {
1697 my ($self, $ra, $path, $r) = @_;
1698 my $props = ($ra->get_dir($path, $r))[2];
1699 my $src = $props->{'svm:source'};
1700 my $uuid = $props->{'svm:uuid'};
1701 return undef if (!$src || !$uuid);
1702
1703 chomp($src, $uuid);
1704
1705 $uuid =~ m{^[0-9a-f\-]{30,}$}
1706 or die "doesn't look right - svm:uuid is '$uuid'\n";
1707
1708 # the '!' is used to mark the repos_root!/relative/path
1709 $src =~ s{/?!/?}{/};
1710 $src =~ s{/+$}{}; # no trailing slashes please
1711 # username is of no interest
1712 $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
1713
1714 my $replace = $ra->{url};
1715 $replace .= "/$path" if length $path;
1716
1717 my $section = "svn-remote.$self->{repo_id}";
1718 tmp_config("$section.svm-source", $src);
1719 tmp_config("$section.svm-replace", $replace);
1720 tmp_config("$section.svm-uuid", $uuid);
1721 $self->{svm} = {
1722 source => $src,
1723 uuid => $uuid,
1724 replace => $replace
1725 };
1726 }
1727
1728 my $r = $ra->get_latest_revnum;
1729 my $path = $self->{path};
1730 my %tried;
1731 while (length $path) {
1732 unless ($tried{"$self->{url}/$path"}) {
1733 return $ra if $self->read_svm_props($ra, $path, $r);
1734 $tried{"$self->{url}/$path"} = 1;
1735 }
1736 $path =~ s#/?[^/]+$##;
1737 }
1738 die "Path: '$path' should be ''\n" if $path ne '';
1739 return $ra if $self->read_svm_props($ra, $path, $r);
1740 $tried{"$self->{url}/$path"} = 1;
1741
1742 if ($ra->{repos_root} eq $self->{url}) {
1743 die @err, (map { " $_\n" } keys %tried), "\n";
1744 }
1745
1746 # nope, make sure we're connected to the repository root:
1747 my $ok;
1748 my @tried_b;
1749 $path = $ra->{svn_path};
1750 $ra = Git::SVN::Ra->new($ra->{repos_root});
1751 while (length $path) {
1752 unless ($tried{"$ra->{url}/$path"}) {
1753 $ok = $self->read_svm_props($ra, $path, $r);
1754 last if $ok;
1755 $tried{"$ra->{url}/$path"} = 1;
1756 }
1757 $path =~ s#/?[^/]+$##;
1758 }
1759 die "Path: '$path' should be ''\n" if $path ne '';
1760 $ok ||= $self->read_svm_props($ra, $path, $r);
1761 $tried{"$ra->{url}/$path"} = 1;
1762 if (!$ok) {
1763 die @err, (map { " $_\n" } keys %tried), "\n";
1764 }
1765 Git::SVN::Ra->new($self->{url});
1766 }
1767
1768 sub svnsync {
1769 my ($self) = @_;
1770 return $self->{svnsync} if $self->{svnsync};
1771
1772 if ($self->no_metadata) {
1773 die "Can't have both 'noMetadata' and ",
1774 "'useSvnsyncProps' options set!\n";
1775 }
1776 if ($self->rewrite_root) {
1777 die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
1778 "options set!\n";
1779 }
1780
1781 my $svnsync;
1782 # see if we have it in our config, first:
1783 eval {
1784 my $section = "svn-remote.$self->{repo_id}";
1785
1786 my $url = tmp_config('--get', "$section.svnsync-url");
1787 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1788 die "doesn't look right - svn:sync-from-url is '$url'\n";
1789
1790 my $uuid = tmp_config('--get', "$section.svnsync-uuid");
1791 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1792 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1793
1794 $svnsync = { url => $url, uuid => $uuid }
1795 };
1796 if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
1797 return $self->{svnsync} = $svnsync;
1798 }
1799
1800 my $err = "useSvnsyncProps set, but failed to read " .
1801 "svnsync property: svn:sync-from-";
1802 my $rp = $self->ra->rev_proplist(0);
1803
1804 my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
1805 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1806 die "doesn't look right - svn:sync-from-url is '$url'\n";
1807
1808 my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
1809 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1810 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1811
1812 my $section = "svn-remote.$self->{repo_id}";
1813 tmp_config('--add', "$section.svnsync-uuid", $uuid);
1814 tmp_config('--add', "$section.svnsync-url", $url);
1815 return $self->{svnsync} = { url => $url, uuid => $uuid };
1816 }
1817
1818 # this allows us to memoize our SVN::Ra UUID locally and avoid a
1819 # remote lookup (useful for 'git svn log').
1820 sub ra_uuid {
1821 my ($self) = @_;
1822 unless ($self->{ra_uuid}) {
1823 my $key = "svn-remote.$self->{repo_id}.uuid";
1824 my $uuid = eval { tmp_config('--get', $key) };
1825 if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/) {
1826 $self->{ra_uuid} = $uuid;
1827 } else {
1828 die "ra_uuid called without URL\n" unless $self->{url};
1829 $self->{ra_uuid} = $self->ra->get_uuid;
1830 tmp_config('--add', $key, $self->{ra_uuid});
1831 }
1832 }
1833 $self->{ra_uuid};
1834 }
1835
1836 sub _set_repos_root {
1837 my ($self, $repos_root) = @_;
1838 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1839 $repos_root ||= $self->ra->{repos_root};
1840 tmp_config($k, $repos_root);
1841 $repos_root;
1842 }
1843
1844 sub repos_root {
1845 my ($self) = @_;
1846 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1847 eval { tmp_config('--get', $k) } || $self->_set_repos_root;
1848 }
1849
1850 sub ra {
1851 my ($self) = shift;
1852 my $ra = Git::SVN::Ra->new($self->{url});
1853 $self->_set_repos_root($ra->{repos_root});
1854 if ($self->use_svm_props && !$self->{svm}) {
1855 if ($self->no_metadata) {
1856 die "Can't have both 'noMetadata' and ",
1857 "'useSvmProps' options set!\n";
1858 } elsif ($self->use_svnsync_props) {
1859 die "Can't have both 'useSvnsyncProps' and ",
1860 "'useSvmProps' options set!\n";
1861 }
1862 $ra = $self->_set_svm_vars($ra);
1863 $self->{-want_revprops} = 1;
1864 }
1865 $ra;
1866 }
1867
1868 sub rel_path {
1869 my ($self) = @_;
1870 my $repos_root = $self->ra->{repos_root};
1871 return $self->{path} if ($self->{url} eq $repos_root);
1872 my $url = $self->{url} .
1873 (length $self->{path} ? "/$self->{path}" : $self->{path});
1874 $url =~ s!^\Q$repos_root\E(?:/+|$)!!g;
1875 $url;
1876 }
1877
1878 # prop_walk(PATH, REV, SUB)
1879 # -------------------------
1880 # Recursively traverse PATH at revision REV and invoke SUB for each
1881 # directory that contains a SVN property. SUB will be invoked as
1882 # follows: &SUB(gs, path, props); where `gs' is this instance of
1883 # Git::SVN, `path' the path to the directory where the properties
1884 # `props' were found. The `path' will be relative to point of checkout,
1885 # that is, if url://repo/trunk is the current Git branch, and that
1886 # directory contains a sub-directory `d', SUB will be invoked with `/d/'
1887 # as `path' (note the trailing `/').
1888 sub prop_walk {
1889 my ($self, $path, $rev, $sub) = @_;
1890
1891 $path =~ s#^/##;
1892 my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
1893 $path =~ s#^/*#/#g;
1894 my $p = $path;
1895 # Strip the irrelevant part of the path.
1896 $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
1897 # Ensure the path is terminated by a `/'.
1898 $p =~ s#/*$#/#;
1899
1900 # The properties contain all the internal SVN stuff nobody
1901 # (usually) cares about.
1902 my $interesting_props = 0;
1903 foreach (keys %{$props}) {
1904 # If it doesn't start with `svn:', it must be a
1905 # user-defined property.
1906 ++$interesting_props and next if $_ !~ /^svn:/;
1907 # FIXME: Fragile, if SVN adds new public properties,
1908 # this needs to be updated.
1909 ++$interesting_props if /^svn:(?:ignore|keywords|executable
1910 |eol-style|mime-type
1911 |externals|needs-lock)$/x;
1912 }
1913 &$sub($self, $p, $props) if $interesting_props;
1914
1915 foreach (sort keys %$dirent) {
1916 next if $dirent->{$_}->{kind} != $SVN::Node::dir;
1917 $self->prop_walk($p . $_, $rev, $sub);
1918 }
1919 }
1920
1921 sub last_rev { ($_[0]->last_rev_commit)[0] }
1922 sub last_commit { ($_[0]->last_rev_commit)[1] }
1923
1924 # returns the newest SVN revision number and newest commit SHA1
1925 sub last_rev_commit {
1926 my ($self) = @_;
1927 if (defined $self->{last_rev} && defined $self->{last_commit}) {
1928 return ($self->{last_rev}, $self->{last_commit});
1929 }
1930 my $c = ::verify_ref($self->refname.'^0');
1931 if ($c && !$self->use_svm_props && !$self->no_metadata) {
1932 my $rev = (::cmt_metadata($c))[1];
1933 if (defined $rev) {
1934 ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
1935 return ($rev, $c);
1936 }
1937 }
1938 my $map_path = $self->map_path;
1939 unless (-e $map_path) {
1940 ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
1941 return (undef, undef);
1942 }
1943 my ($rev, $commit) = $self->rev_map_max(1);
1944 ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
1945 return ($rev, $commit);
1946 }
1947
1948 sub get_fetch_range {
1949 my ($self, $min, $max) = @_;
1950 $max ||= $self->ra->get_latest_revnum;
1951 $min ||= $self->rev_map_max;
1952 (++$min, $max);
1953 }
1954
1955 sub tmp_config {
1956 my (@args) = @_;
1957 my $old_def_config = "$ENV{GIT_DIR}/svn/config";
1958 my $config = "$ENV{GIT_DIR}/svn/.metadata";
1959 if (! -f $config && -f $old_def_config) {
1960 rename $old_def_config, $config or
1961 die "Failed rename $old_def_config => $config: $!\n";
1962 }
1963 my $old_config = $ENV{GIT_CONFIG};
1964 $ENV{GIT_CONFIG} = $config;
1965 $@ = undef;
1966 my @ret = eval {
1967 unless (-f $config) {
1968 mkfile($config);
1969 open my $fh, '>', $config or
1970 die "Can't open $config: $!\n";
1971 print $fh "; This file is used internally by ",
1972 "git-svn\n" or die
1973 "Couldn't write to $config: $!\n";
1974 print $fh "; You should not have to edit it\n" or
1975 die "Couldn't write to $config: $!\n";
1976 close $fh or die "Couldn't close $config: $!\n";
1977 }
1978 command('config', @args);
1979 };
1980 my $err = $@;
1981 if (defined $old_config) {
1982 $ENV{GIT_CONFIG} = $old_config;
1983 } else {
1984 delete $ENV{GIT_CONFIG};
1985 }
1986 die $err if $err;
1987 wantarray ? @ret : $ret[0];
1988 }
1989
1990 sub tmp_index_do {
1991 my ($self, $sub) = @_;
1992 my $old_index = $ENV{GIT_INDEX_FILE};
1993 $ENV{GIT_INDEX_FILE} = $self->{index};
1994 $@ = undef;
1995 my @ret = eval {
1996 my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
1997 mkpath([$dir]) unless -d $dir;
1998 &$sub;
1999 };
2000 my $err = $@;
2001 if (defined $old_index) {
2002 $ENV{GIT_INDEX_FILE} = $old_index;
2003 } else {
2004 delete $ENV{GIT_INDEX_FILE};
2005 }
2006 die $err if $err;
2007 wantarray ? @ret : $ret[0];
2008 }
2009
2010 sub assert_index_clean {
2011 my ($self, $treeish) = @_;
2012
2013 $self->tmp_index_do(sub {
2014 command_noisy('read-tree', $treeish) unless -e $self->{index};
2015 my $x = command_oneline('write-tree');
2016 my ($y) = (command(qw/cat-file commit/, $treeish) =~
2017 /^tree ($::sha1)/mo);
2018 return if $y eq $x;
2019
2020 warn "Index mismatch: $y != $x\nrereading $treeish\n";
2021 unlink $self->{index} or die "unlink $self->{index}: $!\n";
2022 command_noisy('read-tree', $treeish);
2023 $x = command_oneline('write-tree');
2024 if ($y ne $x) {
2025 ::fatal "trees ($treeish) $y != $x\n",
2026 "Something is seriously wrong...";
2027 }
2028 });
2029 }
2030
2031 sub get_commit_parents {
2032 my ($self, $log_entry) = @_;
2033 my (%seen, @ret, @tmp);
2034 # legacy support for 'set-tree'; this is only used by set_tree_cb:
2035 if (my $ip = $self->{inject_parents}) {
2036 if (my $commit = delete $ip->{$log_entry->{revision}}) {
2037 push @tmp, $commit;
2038 }
2039 }
2040 if (my $cur = ::verify_ref($self->refname.'^0')) {
2041 push @tmp, $cur;
2042 }
2043 if (my $ipd = $self->{inject_parents_dcommit}) {
2044 if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2045 push @tmp, @$commit;
2046 }
2047 }
2048 push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2049 while (my $p = shift @tmp) {
2050 next if $seen{$p};
2051 $seen{$p} = 1;
2052 push @ret, $p;
2053 # MAXPARENT is defined to 16 in commit-tree.c:
2054 last if @ret >= 16;
2055 }
2056 if (@tmp) {
2057 die "r$log_entry->{revision}: No room for parents:\n\t",
2058 join("\n\t", @tmp), "\n";
2059 }
2060 @ret;
2061 }
2062
2063 sub rewrite_root {
2064 my ($self) = @_;
2065 return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2066 my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2067 my $rwr = eval { command_oneline(qw/config --get/, $k) };
2068 if ($rwr) {
2069 $rwr =~ s#/+$##;
2070 if ($rwr !~ m#^[a-z\+]+://#) {
2071 die "$rwr is not a valid URL (key: $k)\n";
2072 }
2073 }
2074 $self->{-rewrite_root} = $rwr;
2075 }
2076
2077 sub metadata_url {
2078 my ($self) = @_;
2079 ($self->rewrite_root || $self->{url}) .
2080 (length $self->{path} ? '/' . $self->{path} : '');
2081 }
2082
2083 sub full_url {
2084 my ($self) = @_;
2085 $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2086 }
2087
2088
2089 sub set_commit_header_env {
2090 my ($log_entry) = @_;
2091 my %env;
2092 foreach my $ned (qw/NAME EMAIL DATE/) {
2093 foreach my $ac (qw/AUTHOR COMMITTER/) {
2094 $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2095 }
2096 }
2097
2098 $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2099 $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2100 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2101
2102 $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2103 ? $log_entry->{commit_name}
2104 : $log_entry->{name};
2105 $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2106 ? $log_entry->{commit_email}
2107 : $log_entry->{email};
2108 \%env;
2109 }
2110
2111 sub restore_commit_header_env {
2112 my ($env) = @_;
2113 foreach my $ned (qw/NAME EMAIL DATE/) {
2114 foreach my $ac (qw/AUTHOR COMMITTER/) {
2115 my $k = "GIT_${ac}_${ned}";
2116 if (defined $env->{$k}) {
2117 $ENV{$k} = $env->{$k};
2118 } else {
2119 delete $ENV{$k};
2120 }
2121 }
2122 }
2123 }
2124
2125 sub gc {
2126 command_noisy('gc', '--auto');
2127 };
2128
2129 sub do_git_commit {
2130 my ($self, $log_entry) = @_;
2131 my $lr = $self->last_rev;
2132 if (defined $lr && $lr >= $log_entry->{revision}) {
2133 die "Last fetched revision of ", $self->refname,
2134 " was r$lr, but we are about to fetch: ",
2135 "r$log_entry->{revision}!\n";
2136 }
2137 if (my $c = $self->rev_map_get($log_entry->{revision})) {
2138 croak "$log_entry->{revision} = $c already exists! ",
2139 "Why are we refetching it?\n";
2140 }
2141 my $old_env = set_commit_header_env($log_entry);
2142 my $tree = $log_entry->{tree};
2143 if (!defined $tree) {
2144 $tree = $self->tmp_index_do(sub {
2145 command_oneline('write-tree') });
2146 }
2147 die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2148
2149 my @exec = ('git-commit-tree', $tree);
2150 foreach ($self->get_commit_parents($log_entry)) {
2151 push @exec, '-p', $_;
2152 }
2153 defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2154 or croak $!;
2155 print $msg_fh $log_entry->{log} or croak $!;
2156 restore_commit_header_env($old_env);
2157 unless ($self->no_metadata) {
2158 print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2159 or croak $!;
2160 }
2161 $msg_fh->flush == 0 or croak $!;
2162 close $msg_fh or croak $!;
2163 chomp(my $commit = do { local $/; <$out_fh> });
2164 close $out_fh or croak $!;
2165 waitpid $pid, 0;
2166 croak $? if $?;
2167 if ($commit !~ /^$::sha1$/o) {
2168 die "Failed to commit, invalid sha1: $commit\n";
2169 }
2170
2171 $self->rev_map_set($log_entry->{revision}, $commit, 1);
2172
2173 $self->{last_rev} = $log_entry->{revision};
2174 $self->{last_commit} = $commit;
2175 print "r$log_entry->{revision}";
2176 if (defined $log_entry->{svm_revision}) {
2177 print " (\@$log_entry->{svm_revision})";
2178 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2179 0, $self->svm_uuid);
2180 }
2181 print " = $commit ($self->{ref_id})\n";
2182 if (--$_gc_nr == 0) {
2183 $_gc_nr = $_gc_period;
2184 gc();
2185 }
2186 return $commit;
2187 }
2188
2189 sub match_paths {
2190 my ($self, $paths, $r) = @_;
2191 return 1 if $self->{path} eq '';
2192 if (my $path = $paths->{"/$self->{path}"}) {
2193 return ($path->{action} eq 'D') ? 0 : 1;
2194 }
2195 $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
2196 if (grep /$self->{path_regex}/, keys %$paths) {
2197 return 1;
2198 }
2199 my $c = '';
2200 foreach (split m#/#, $self->{path}) {
2201 $c .= "/$_";
2202 next unless ($paths->{$c} &&
2203 ($paths->{$c}->{action} =~ /^[AR]$/));
2204 if ($self->ra->check_path($self->{path}, $r) ==
2205 $SVN::Node::dir) {
2206 return 1;
2207 }
2208 }
2209 return 0;
2210 }
2211
2212 sub find_parent_branch {
2213 my ($self, $paths, $rev) = @_;
2214 return undef unless $self->follow_parent;
2215 unless (defined $paths) {
2216 my $err_handler = $SVN::Error::handler;
2217 $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2218 $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1, sub {
2219 $paths =
2220 Git::SVN::Ra::dup_changed_paths($_[0]) });
2221 $SVN::Error::handler = $err_handler;
2222 }
2223 return undef unless defined $paths;
2224
2225 # look for a parent from another branch:
2226 my @b_path_components = split m#/#, $self->rel_path;
2227 my @a_path_components;
2228 my $i;
2229 while (@b_path_components) {
2230 $i = $paths->{'/'.join('/', @b_path_components)};
2231 last if $i && defined $i->{copyfrom_path};
2232 unshift(@a_path_components, pop(@b_path_components));
2233 }
2234 return undef unless defined $i && defined $i->{copyfrom_path};
2235 my $branch_from = $i->{copyfrom_path};
2236 if (@a_path_components) {
2237 print STDERR "branch_from: $branch_from => ";
2238 $branch_from .= '/'.join('/', @a_path_components);
2239 print STDERR $branch_from, "\n";
2240 }
2241 my $r = $i->{copyfrom_rev};
2242 my $repos_root = $self->ra->{repos_root};
2243 my $url = $self->ra->{url};
2244 my $new_url = $repos_root . $branch_from;
2245 print STDERR "Found possible branch point: ",
2246 "$new_url => ", $self->full_url, ", $r\n";
2247 $branch_from =~ s#^/##;
2248 my $gs = Git::SVN->find_by_url($new_url, $repos_root, $branch_from);
2249 unless ($gs) {
2250 my $ref_id = $self->{ref_id};
2251 $ref_id =~ s/\@\d+$//;
2252 $ref_id .= "\@$r";
2253 # just grow a tail if we're not unique enough :x
2254 $ref_id .= '-' while find_ref($ref_id);
2255 print STDERR "Initializing parent: $ref_id\n";
2256 my ($u, $p, $repo_id) = ($new_url, '', $ref_id);
2257 if ($u =~ s#^\Q$url\E(/|$)##) {
2258 $p = $u;
2259 $u = $url;
2260 $repo_id = $self->{repo_id};
2261 }
2262 $gs = Git::SVN->init($u, $p, $repo_id, $ref_id, 1);
2263 }
2264 my ($r0, $parent) = $gs->find_rev_before($r, 1);
2265 if (!defined $r0 || !defined $parent) {
2266 my ($base, $head) = parse_revision_argument(0, $r);
2267 if ($base <= $r) {
2268 $gs->fetch($base, $r);
2269 }
2270 ($r0, $parent) = $gs->last_rev_commit;
2271 }
2272 if (defined $r0 && defined $parent) {
2273 print STDERR "Found branch parent: ($self->{ref_id}) $parent\n";
2274 my $ed;
2275 if ($self->ra->can_do_switch) {
2276 $self->assert_index_clean($parent);
2277 print STDERR "Following parent with do_switch\n";
2278 # do_switch works with svn/trunk >= r22312, but that
2279 # is not included with SVN 1.4.3 (the latest version
2280 # at the moment), so we can't rely on it
2281 $self->{last_commit} = $parent;
2282 $ed = SVN::Git::Fetcher->new($self);
2283 $gs->ra->gs_do_switch($r0, $rev, $gs,
2284 $self->full_url, $ed)
2285 or die "SVN connection failed somewhere...\n";
2286 } elsif ($self->ra->trees_match($new_url, $r0,
2287 $self->full_url, $rev)) {
2288 print STDERR "Trees match:\n",
2289 " $new_url\@$r0\n",
2290 " ${\$self->full_url}\@$rev\n",
2291 "Following parent with no changes\n";
2292 $self->tmp_index_do(sub {
2293 command_noisy('read-tree', $parent);
2294 });
2295 $self->{last_commit} = $parent;
2296 } else {
2297 print STDERR "Following parent with do_update\n";
2298 $ed = SVN::Git::Fetcher->new($self);
2299 $self->ra->gs_do_update($rev, $rev, $self, $ed)
2300 or die "SVN connection failed somewhere...\n";
2301 }
2302 print STDERR "Successfully followed parent\n";
2303 return $self->make_log_entry($rev, [$parent], $ed);
2304 }
2305 return undef;
2306 }
2307
2308 sub do_fetch {
2309 my ($self, $paths, $rev) = @_;
2310 my $ed;
2311 my ($last_rev, @parents);
2312 if (my $lc = $self->last_commit) {
2313 # we can have a branch that was deleted, then re-added
2314 # under the same name but copied from another path, in
2315 # which case we'll have multiple parents (we don't
2316 # want to break the original ref, nor lose copypath info):
2317 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2318 push @{$log_entry->{parents}}, $lc;
2319 return $log_entry;
2320 }
2321 $ed = SVN::Git::Fetcher->new($self);
2322 $last_rev = $self->{last_rev};
2323 $ed->{c} = $lc;
2324 @parents = ($lc);
2325 } else {
2326 $last_rev = $rev;
2327 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2328 return $log_entry;
2329 }
2330 $ed = SVN::Git::Fetcher->new($self);
2331 }
2332 unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2333 die "SVN connection failed somewhere...\n";
2334 }
2335 $self->make_log_entry($rev, \@parents, $ed);
2336 }
2337
2338 sub get_untracked {
2339 my ($self, $ed) = @_;
2340 my @out;
2341 my $h = $ed->{empty};
2342 foreach (sort keys %$h) {
2343 my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2344 push @out, " $act: " . uri_encode($_);
2345 warn "W: $act: $_\n";
2346 }
2347 foreach my $t (qw/dir_prop file_prop/) {
2348 $h = $ed->{$t} or next;
2349 foreach my $path (sort keys %$h) {
2350 my $ppath = $path eq '' ? '.' : $path;
2351 foreach my $prop (sort keys %{$h->{$path}}) {
2352 next if $SKIP_PROP{$prop};
2353 my $v = $h->{$path}->{$prop};
2354 my $t_ppath_prop = "$t: " .
2355 uri_encode($ppath) . ' ' .
2356 uri_encode($prop);
2357 if (defined $v) {
2358 push @out, " +$t_ppath_prop " .
2359 uri_encode($v);
2360 } else {
2361 push @out, " -$t_ppath_prop";
2362 }
2363 }
2364 }
2365 }
2366 foreach my $t (qw/absent_file absent_directory/) {
2367 $h = $ed->{$t} or next;
2368 foreach my $parent (sort keys %$h) {
2369 foreach my $path (sort @{$h->{$parent}}) {
2370 push @out, " $t: " .
2371 uri_encode("$parent/$path");
2372 warn "W: $t: $parent/$path ",
2373 "Insufficient permissions?\n";
2374 }
2375 }
2376 }
2377 \@out;
2378 }
2379
2380 sub parse_svn_date {
2381 my $date = shift || return '+0000 1970-01-01 00:00:00';
2382 my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2383 (\d\d)\:(\d\d)\:(\d\d).\d+Z$/x) or
2384 croak "Unable to parse date: $date\n";
2385 "+0000 $Y-$m-$d $H:$M:$S";
2386 }
2387
2388 sub check_author {
2389 my ($author) = @_;
2390 if (!defined $author || length $author == 0) {
2391 $author = '(no author)';
2392 } elsif (defined $::_authors && ! defined $::users{$author}) {
2393 die "Author: $author not defined in $::_authors file\n";
2394 }
2395 $author;
2396 }
2397
2398 sub make_log_entry {
2399 my ($self, $rev, $parents, $ed) = @_;
2400 my $untracked = $self->get_untracked($ed);
2401
2402 open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
2403 print $un "r$rev\n" or croak $!;
2404 print $un $_, "\n" foreach @$untracked;
2405 my %log_entry = ( parents => $parents || [], revision => $rev,
2406 log => '');
2407
2408 my $headrev;
2409 my $logged = delete $self->{logged_rev_props};
2410 if (!$logged || $self->{-want_revprops}) {
2411 my $rp = $self->ra->rev_proplist($rev);
2412 foreach (sort keys %$rp) {
2413 my $v = $rp->{$_};
2414 if (/^svn:(author|date|log)$/) {
2415 $log_entry{$1} = $v;
2416 } elsif ($_ eq 'svm:headrev') {
2417 $headrev = $v;
2418 } else {
2419 print $un " rev_prop: ", uri_encode($_), ' ',
2420 uri_encode($v), "\n";
2421 }
2422 }
2423 } else {
2424 map { $log_entry{$_} = $logged->{$_} } keys %$logged;
2425 }
2426 close $un or croak $!;
2427
2428 $log_entry{date} = parse_svn_date($log_entry{date});
2429 $log_entry{log} .= "\n";
2430 my $author = $log_entry{author} = check_author($log_entry{author});
2431 my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
2432 : ($author, undef);
2433
2434 my ($commit_name, $commit_email) = ($name, $email);
2435 if ($_use_log_author) {
2436 my $name_field;
2437 if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
2438 $name_field = $1;
2439 } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
2440 $name_field = $1;
2441 }
2442 if (!defined $name_field) {
2443 if (!defined $email) {
2444 $email = $name;
2445 }
2446 } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
2447 ($name, $email) = ($1, $2);
2448 } elsif ($name_field =~ /(.*)@/) {
2449 ($name, $email) = ($1, $name_field);
2450 } else {
2451 ($name, $email) = ($name_field, $name_field);
2452 }
2453 }
2454 if (defined $headrev && $self->use_svm_props) {
2455 if ($self->rewrite_root) {
2456 die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
2457 "options set!\n";
2458 }
2459 my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$};
2460 # we don't want "SVM: initializing mirror for junk" ...
2461 return undef if $r == 0;
2462 my $svm = $self->svm;
2463 if ($uuid ne $svm->{uuid}) {
2464 die "UUID mismatch on SVM path:\n",
2465 "expected: $svm->{uuid}\n",
2466 " got: $uuid\n";
2467 }
2468 my $full_url = $self->full_url;
2469 $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
2470 die "Failed to replace '$svm->{replace}' with ",
2471 "'$svm->{source}' in $full_url\n";
2472 # throw away username for storing in records
2473 remove_username($full_url);
2474 $log_entry{metadata} = "$full_url\@$r $uuid";
2475 $log_entry{svm_revision} = $r;
2476 $email ||= "$author\@$uuid";
2477 $commit_email ||= "$author\@$uuid";
2478 } elsif ($self->use_svnsync_props) {
2479 my $full_url = $self->svnsync->{url};
2480 $full_url .= "/$self->{path}" if length $self->{path};
2481 remove_username($full_url);
2482 my $uuid = $self->svnsync->{uuid};
2483 $log_entry{metadata} = "$full_url\@$rev $uuid";
2484 $email ||= "$author\@$uuid";
2485 $commit_email ||= "$author\@$uuid";
2486 } else {
2487 my $url = $self->metadata_url;
2488 remove_username($url);
2489 $log_entry{metadata} = "$url\@$rev " .
2490 $self->ra->get_uuid;
2491 $email ||= "$author\@" . $self->ra->get_uuid;
2492 $commit_email ||= "$author\@" . $self->ra->get_uuid;
2493 }
2494 $log_entry{name} = $name;
2495 $log_entry{email} = $email;
2496 $log_entry{commit_name} = $commit_name;
2497 $log_entry{commit_email} = $commit_email;
2498 \%log_entry;
2499 }
2500
2501 sub fetch {
2502 my ($self, $min_rev, $max_rev, @parents) = @_;
2503 my ($last_rev, $last_commit) = $self->last_rev_commit;
2504 my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
2505 $self->ra->gs_fetch_loop_common($base, $head, [$self]);
2506 }
2507
2508 sub set_tree_cb {
2509 my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
2510 $self->{inject_parents} = { $rev => $tree };
2511 $self->fetch(undef, undef);
2512 }
2513
2514 sub set_tree {
2515 my ($self, $tree) = (shift, shift);
2516 my $log_entry = ::get_commit_entry($tree);
2517 unless ($self->{last_rev}) {
2518 fatal("Must have an existing revision to commit");
2519 }
2520 my %ed_opts = ( r => $self->{last_rev},
2521 log => $log_entry->{log},
2522 ra => $self->ra,
2523 tree_a => $self->{last_commit},
2524 tree_b => $tree,
2525 editor_cb => sub {
2526 $self->set_tree_cb($log_entry, $tree, @_) },
2527 svn_path => $self->{path} );
2528 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
2529 print "No changes\nr$self->{last_rev} = $tree\n";
2530 }
2531 }
2532
2533 sub rebuild_from_rev_db {
2534 my ($self, $path) = @_;
2535 my $r = -1;
2536 open my $fh, '<', $path or croak "open: $!";
2537 binmode $fh or croak "binmode: $!";
2538 while (<$fh>) {
2539 length($_) == 41 or croak "inconsistent size in ($_) != 41";
2540 chomp($_);
2541 ++$r;
2542 next if $_ eq ('0' x 40);
2543 $self->rev_map_set($r, $_);
2544 print "r$r = $_\n";
2545 }
2546 close $fh or croak "close: $!";
2547 unlink $path or croak "unlink: $!";
2548 }
2549
2550 sub rebuild {
2551 my ($self) = @_;
2552 my $map_path = $self->map_path;
2553 return if (-e $map_path && ! -z $map_path);
2554 return unless ::verify_ref($self->refname.'^0');
2555 if ($self->use_svm_props || $self->no_metadata) {
2556 my $rev_db = $self->rev_db_path;
2557 $self->rebuild_from_rev_db($rev_db);
2558 if ($self->use_svm_props) {
2559 my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
2560 $self->rebuild_from_rev_db($svm_rev_db);
2561 }
2562 $self->unlink_rev_db_symlink;
2563 return;
2564 }
2565 print "Rebuilding $map_path ...\n";
2566 my ($log, $ctx) =
2567 command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
2568 $self->refname, '--');
2569 my $full_url = $self->full_url;
2570 remove_username($full_url);
2571 my $svn_uuid = $self->ra_uuid;
2572 my $c;
2573 while (<$log>) {
2574 if ( m{^commit ($::sha1)$} ) {
2575 $c = $1;
2576 next;
2577 }
2578 next unless s{^\s*(git-svn-id:)}{$1};
2579 my ($url, $rev, $uuid) = ::extract_metadata($_);
2580 remove_username($url);
2581
2582 # ignore merges (from set-tree)
2583 next if (!defined $rev || !$uuid);
2584
2585 # if we merged or otherwise started elsewhere, this is
2586 # how we break out of it
2587 if (($uuid ne $svn_uuid) ||
2588 ($full_url && $url && ($url ne $full_url))) {
2589 next;
2590 }
2591
2592 $self->rev_map_set($rev, $c);
2593 print "r$rev = $c\n";
2594 }
2595 command_close_pipe($log, $ctx);
2596 print "Done rebuilding $map_path\n";
2597 my $rev_db_path = $self->rev_db_path;
2598 if (-f $self->rev_db_path) {
2599 unlink $self->rev_db_path or croak "unlink: $!";
2600 }
2601 $self->unlink_rev_db_symlink;
2602 }
2603
2604 # rev_map:
2605 # Tie::File seems to be prone to offset errors if revisions get sparse,
2606 # it's not that fast, either. Tie::File is also not in Perl 5.6. So
2607 # one of my favorite modules is out :< Next up would be one of the DBM
2608 # modules, but I'm not sure which is most portable...
2609 #
2610 # This is the replacement for the rev_db format, which was too big
2611 # and inefficient for large repositories with a lot of sparse history
2612 # (mainly tags)
2613 #
2614 # The format is this:
2615 # - 24 bytes for every record,
2616 # * 4 bytes for the integer representing an SVN revision number
2617 # * 20 bytes representing the sha1 of a git commit
2618 # - No empty padding records like the old format
2619 # (except the last record, which can be overwritten)
2620 # - new records are written append-only since SVN revision numbers
2621 # increase monotonically
2622 # - lookups on SVN revision number are done via a binary search
2623 # - Piping the file to xxd -c24 is a good way of dumping it for
2624 # viewing or editing (piped back through xxd -r), should the need
2625 # ever arise.
2626 # - The last record can be padding revision with an all-zero sha1
2627 # This is used to optimize fetch performance when using multiple
2628 # "fetch" directives in .git/config
2629 #
2630 # These files are disposable unless noMetadata or useSvmProps is set
2631
2632 sub _rev_map_set {
2633 my ($fh, $rev, $commit) = @_;
2634
2635 binmode $fh or croak "binmode: $!";
2636 my $size = (stat($fh))[7];
2637 ($size % 24) == 0 or croak "inconsistent size: $size";
2638
2639 my $wr_offset = 0;
2640 if ($size > 0) {
2641 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2642 my $read = sysread($fh, my $buf, 24) or croak "read: $!";
2643 $read == 24 or croak "read only $read bytes (!= 24)";
2644 my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
2645 if ($last_commit eq ('0' x40)) {
2646 if ($size >= 48) {
2647 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2648 $read = sysread($fh, $buf, 24) or
2649 croak "read: $!";
2650 $read == 24 or
2651 croak "read only $read bytes (!= 24)";
2652 ($last_rev, $last_commit) =
2653 unpack(rev_map_fmt, $buf);
2654 if ($last_commit eq ('0' x40)) {
2655 croak "inconsistent .rev_map\n";
2656 }
2657 }
2658 if ($last_rev >= $rev) {
2659 croak "last_rev is higher!: $last_rev >= $rev";
2660 }
2661 $wr_offset = -24;
2662 }
2663 }
2664 sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
2665 syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
2666 croak "write: $!";
2667 }
2668
2669 sub mkfile {
2670 my ($path) = @_;
2671 unless (-e $path) {
2672 my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
2673 mkpath([$dir]) unless -d $dir;
2674 open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
2675 close $fh or die "Couldn't close (create) $path: $!\n";
2676 }
2677 }
2678
2679 sub rev_map_set {
2680 my ($self, $rev, $commit, $update_ref, $uuid) = @_;
2681 length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
2682 my $db = $self->map_path($uuid);
2683 my $db_lock = "$db.lock";
2684 my $sig;
2685 if ($update_ref) {
2686 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2687 $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
2688 }
2689 mkfile($db);
2690
2691 $LOCKFILES{$db_lock} = 1;
2692 my $sync;
2693 # both of these options make our .rev_db file very, very important
2694 # and we can't afford to lose it because rebuild() won't work
2695 if ($self->use_svm_props || $self->no_metadata) {
2696 $sync = 1;
2697 copy($db, $db_lock) or die "rev_map_set(@_): ",
2698 "Failed to copy: ",
2699 "$db => $db_lock ($!)\n";
2700 } else {
2701 rename $db, $db_lock or die "rev_map_set(@_): ",
2702 "Failed to rename: ",
2703 "$db => $db_lock ($!)\n";
2704 }
2705
2706 sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
2707 or croak "Couldn't open $db_lock: $!\n";
2708 _rev_map_set($fh, $rev, $commit);
2709 if ($sync) {
2710 $fh->flush or die "Couldn't flush $db_lock: $!\n";
2711 $fh->sync or die "Couldn't sync $db_lock: $!\n";
2712 }
2713 close $fh or croak $!;
2714 if ($update_ref) {
2715 $_head = $self;
2716 command_noisy('update-ref', '-m', "r$rev",
2717 $self->refname, $commit);
2718 }
2719 rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
2720 "$db_lock => $db ($!)\n";
2721 delete $LOCKFILES{$db_lock};
2722 if ($update_ref) {
2723 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2724 $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
2725 kill $sig, $$ if defined $sig;
2726 }
2727 }
2728
2729 # If want_commit, this will return an array of (rev, commit) where
2730 # commit _must_ be a valid commit in the archive.
2731 # Otherwise, it'll return the max revision (whether or not the
2732 # commit is valid or just a 0x40 placeholder).
2733 sub rev_map_max {
2734 my ($self, $want_commit) = @_;
2735 $self->rebuild;
2736 my $map_path = $self->map_path;
2737 stat $map_path or return $want_commit ? (0, undef) : 0;
2738 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2739 binmode $fh or croak "binmode: $!";
2740 my $size = (stat($fh))[7];
2741 ($size % 24) == 0 or croak "inconsistent size: $size";
2742
2743 if ($size == 0) {
2744 close $fh or croak "close: $!";
2745 return $want_commit ? (0, undef) : 0;
2746 }
2747
2748 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2749 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2750 my ($r, $c) = unpack(rev_map_fmt, $buf);
2751 if ($want_commit && $c eq ('0' x40)) {
2752 if ($size < 48) {
2753 return $want_commit ? (0, undef) : 0;
2754 }
2755 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2756 sysread($fh, $buf, 24) == 24 or croak "read: $!";
2757 ($r, $c) = unpack(rev_map_fmt, $buf);
2758 if ($c eq ('0'x40)) {
2759 croak "Penultimate record is all-zeroes in $map_path";
2760 }
2761 }
2762 close $fh or croak "close: $!";
2763 $want_commit ? ($r, $c) : $r;
2764 }
2765
2766 sub rev_map_get {
2767 my ($self, $rev, $uuid) = @_;
2768 my $map_path = $self->map_path($uuid);
2769 return undef unless -e $map_path;
2770
2771 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2772 binmode $fh or croak "binmode: $!";
2773 my $size = (stat($fh))[7];
2774 ($size % 24) == 0 or croak "inconsistent size: $size";
2775
2776 if ($size == 0) {
2777 close $fh or croak "close: $fh";
2778 return undef;
2779 }
2780
2781 my ($l, $u) = (0, $size - 24);
2782 my ($r, $c, $buf);
2783
2784 while ($l <= $u) {
2785 my $i = int(($l/24 + $u/24) / 2) * 24;
2786 sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
2787 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2788 my ($r, $c) = unpack('NH40', $buf);
2789
2790 if ($r < $rev) {
2791 $l = $i + 24;
2792 } elsif ($r > $rev) {
2793 $u = $i - 24;
2794 } else { # $r == $rev
2795 close($fh) or croak "close: $!";
2796 return $c eq ('0' x 40) ? undef : $c;
2797 }
2798 }
2799 close($fh) or croak "close: $!";
2800 undef;
2801 }
2802
2803 # Finds the first svn revision that exists on (if $eq_ok is true) or
2804 # before $rev for the current branch. It will not search any lower
2805 # than $min_rev. Returns the git commit hash and svn revision number
2806 # if found, else (undef, undef).
2807 sub find_rev_before {
2808 my ($self, $rev, $eq_ok, $min_rev) = @_;
2809 --$rev unless $eq_ok;
2810 $min_rev ||= 1;
2811 while ($rev >= $min_rev) {
2812 if (my $c = $self->rev_map_get($rev)) {
2813 return ($rev, $c);
2814 }
2815 --$rev;
2816 }
2817 return (undef, undef);
2818 }
2819
2820 # Finds the first svn revision that exists on (if $eq_ok is true) or
2821 # after $rev for the current branch. It will not search any higher
2822 # than $max_rev. Returns the git commit hash and svn revision number
2823 # if found, else (undef, undef).
2824 sub find_rev_after {
2825 my ($self, $rev, $eq_ok, $max_rev) = @_;
2826 ++$rev unless $eq_ok;
2827 $max_rev ||= $self->rev_map_max;
2828 while ($rev <= $max_rev) {
2829 if (my $c = $self->rev_map_get($rev)) {
2830 return ($rev, $c);
2831 }
2832 ++$rev;
2833 }
2834 return (undef, undef);
2835 }
2836
2837 sub _new {
2838 my ($class, $repo_id, $ref_id, $path) = @_;
2839 unless (defined $repo_id && length $repo_id) {
2840 $repo_id = $Git::SVN::default_repo_id;
2841 }
2842 unless (defined $ref_id && length $ref_id) {
2843 $_[2] = $ref_id = $Git::SVN::default_ref_id;
2844 }
2845 $_[1] = $repo_id = sanitize_remote_name($repo_id);
2846 my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
2847 $_[3] = $path = '' unless (defined $path);
2848 mkpath(["$ENV{GIT_DIR}/svn"]);
2849 bless {
2850 ref_id => $ref_id, dir => $dir, index => "$dir/index",
2851 path => $path, config => "$ENV{GIT_DIR}/svn/config",
2852 map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
2853 }
2854
2855 # for read-only access of old .rev_db formats
2856 sub unlink_rev_db_symlink {
2857 my ($self) = @_;
2858 my $link = $self->rev_db_path;
2859 $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
2860 if (-l $link) {
2861 unlink $link or croak "unlink: $link failed!";
2862 }
2863 }
2864
2865 sub rev_db_path {
2866 my ($self, $uuid) = @_;
2867 my $db_path = $self->map_path($uuid);
2868 $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
2869 or croak "map_path: $db_path does not contain '/.rev_map.' !";
2870 $db_path;
2871 }
2872
2873 # the new replacement for .rev_db
2874 sub map_path {
2875 my ($self, $uuid) = @_;
2876 $uuid ||= $self->ra_uuid;
2877 "$self->{map_root}.$uuid";
2878 }
2879
2880 sub uri_encode {
2881 my ($f) = @_;
2882 $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
2883 $f
2884 }
2885
2886 sub remove_username {
2887 $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
2888 }
2889
2890 package Git::SVN::Prompt;
2891 use strict;
2892 use warnings;
2893 require SVN::Core;
2894 use vars qw/$_no_auth_cache $_username/;
2895
2896 sub simple {
2897 my ($cred, $realm, $default_username, $may_save, $pool) = @_;
2898 $may_save = undef if $_no_auth_cache;
2899 $default_username = $_username if defined $_username;
2900 if (defined $default_username && length $default_username) {
2901 if (defined $realm && length $realm) {
2902 print STDERR "Authentication realm: $realm\n";
2903 STDERR->flush;
2904 }
2905 $cred->username($default_username);
2906 } else {
2907 username($cred, $realm, $may_save, $pool);
2908 }
2909 $cred->password(_read_password("Password for '" .
2910 $cred->username . "': ", $realm));
2911 $cred->may_save($may_save);
2912 $SVN::_Core::SVN_NO_ERROR;
2913 }
2914
2915 sub ssl_server_trust {
2916 my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
2917 $may_save = undef if $_no_auth_cache;
2918 print STDERR "Error validating server certificate for '$realm':\n";
2919 {
2920 no warnings 'once';
2921 # All variables SVN::Auth::SSL::* are used only once,
2922 # so we're shutting up Perl warnings about this.
2923 if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
2924 print STDERR " - The certificate is not issued ",
2925 "by a trusted authority. Use the\n",
2926 " fingerprint to validate ",
2927 "the certificate manually!\n";
2928 }
2929 if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
2930 print STDERR " - The certificate hostname ",
2931 "does not match.\n";
2932 }
2933 if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
2934 print STDERR " - The certificate is not yet valid.\n";
2935 }
2936 if ($failures & $SVN::Auth::SSL::EXPIRED) {
2937 print STDERR " - The certificate has expired.\n";
2938 }
2939 if ($failures & $SVN::Auth::SSL::OTHER) {
2940 print STDERR " - The certificate has ",
2941 "an unknown error.\n";
2942 }
2943 } # no warnings 'once'
2944 printf STDERR
2945 "Certificate information:\n".
2946 " - Hostname: %s\n".
2947 " - Valid: from %s until %s\n".
2948 " - Issuer: %s\n".
2949 " - Fingerprint: %s\n",
2950 map $cert_info->$_, qw(hostname valid_from valid_until
2951 issuer_dname fingerprint);
2952 my $choice;
2953 prompt:
2954 print STDERR $may_save ?
2955 "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
2956 "(R)eject or accept (t)emporarily? ";
2957 STDERR->flush;
2958 $choice = lc(substr(<STDIN> || 'R', 0, 1));
2959 if ($choice =~ /^t$/i) {
2960 $cred->may_save(undef);
2961 } elsif ($choice =~ /^r$/i) {
2962 return -1;
2963 } elsif ($may_save && $choice =~ /^p$/i) {
2964 $cred->may_save($may_save);
2965 } else {
2966 goto prompt;
2967 }
2968 $cred->accepted_failures($failures);
2969 $SVN::_Core::SVN_NO_ERROR;
2970 }
2971
2972 sub ssl_client_cert {
2973 my ($cred, $realm, $may_save, $pool) = @_;
2974 $may_save = undef if $_no_auth_cache;
2975 print STDERR "Client certificate filename: ";
2976 STDERR->flush;
2977 chomp(my $filename = <STDIN>);
2978 $cred->cert_file($filename);
2979 $cred->may_save($may_save);
2980 $SVN::_Core::SVN_NO_ERROR;
2981 }
2982
2983 sub ssl_client_cert_pw {
2984 my ($cred, $realm, $may_save, $pool) = @_;
2985 $may_save = undef if $_no_auth_cache;
2986 $cred->password(_read_password("Password: ", $realm));
2987 $cred->may_save($may_save);
2988 $SVN::_Core::SVN_NO_ERROR;
2989 }
2990
2991 sub username {
2992 my ($cred, $realm, $may_save, $pool) = @_;
2993 $may_save = undef if $_no_auth_cache;
2994 if (defined $realm && length $realm) {
2995 print STDERR "Authentication realm: $realm\n";
2996 }
2997 my $username;
2998 if (defined $_username) {
2999 $username = $_username;
3000 } else {
3001 print STDERR "Username: ";
3002 STDERR->flush;
3003 chomp($username = <STDIN>);
3004 }
3005 $cred->username($username);
3006 $cred->may_save($may_save);
3007 $SVN::_Core::SVN_NO_ERROR;
3008 }
3009
3010 sub _read_password {
3011 my ($prompt, $realm) = @_;
3012 print STDERR $prompt;
3013 STDERR->flush;
3014 require Term::ReadKey;
3015 Term::ReadKey::ReadMode('noecho');
3016 my $password = '';
3017 while (defined(my $key = Term::ReadKey::ReadKey(0))) {
3018 last if $key =~ /[\012\015]/; # \n\r
3019 $password .= $key;
3020 }
3021 Term::ReadKey::ReadMode('restore');
3022 print STDERR "\n";
3023 STDERR->flush;
3024 $password;
3025 }
3026
3027 package SVN::Git::Fetcher;
3028 use vars qw/@ISA/;
3029 use strict;
3030 use warnings;
3031 use Carp qw/croak/;
3032 use IO::File qw//;
3033
3034 # file baton members: path, mode_a, mode_b, pool, fh, blob, base
3035 sub new {
3036 my ($class, $git_svn) = @_;
3037 my $self = SVN::Delta::Editor->new;
3038 bless $self, $class;
3039 $self->{c} = $git_svn->{last_commit} if exists $git_svn->{last_commit};
3040 $self->{empty} = {};
3041 $self->{dir_prop} = {};
3042 $self->{file_prop} = {};
3043 $self->{absent_dir} = {};
3044 $self->{absent_file} = {};
3045 $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
3046 $self;
3047 }
3048
3049 sub set_path_strip {
3050 my ($self, $path) = @_;
3051 $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
3052 }
3053
3054 sub open_root {
3055 { path => '' };
3056 }
3057
3058 sub open_directory {
3059 my ($self, $path, $pb, $rev) = @_;
3060 { path => $path };
3061 }
3062
3063 sub git_path {
3064 my ($self, $path) = @_;
3065 if ($self->{path_strip}) {
3066 $path =~ s!$self->{path_strip}!! or
3067 die "Failed to strip path '$path' ($self->{path_strip})\n";
3068 }
3069 $path;
3070 }
3071
3072 sub delete_entry {
3073 my ($self, $path, $rev, $pb) = @_;
3074
3075 my $gpath = $self->git_path($path);
3076 return undef if ($gpath eq '');
3077
3078 # remove entire directories.
3079 if (command('ls-tree', $self->{c}, '--', $gpath) =~ /^040000 tree/) {
3080 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3081 -r --name-only -z/,
3082 $self->{c}, '--', $gpath);
3083 local $/ = "\0";
3084 while (<$ls>) {
3085 chomp;
3086 $self->{gii}->remove($_);
3087 print "\tD\t$_\n" unless $::_q;
3088 }
3089 print "\tD\t$gpath/\n" unless $::_q;
3090 command_close_pipe($ls, $ctx);
3091 $self->{empty}->{$path} = 0
3092 } else {
3093 $self->{gii}->remove($gpath);
3094 print "\tD\t$gpath\n" unless $::_q;
3095 }
3096 undef;
3097 }
3098
3099 sub open_file {
3100 my ($self, $path, $pb, $rev) = @_;
3101 my $gpath = $self->git_path($path);
3102 my ($mode, $blob) = (command('ls-tree', $self->{c}, '--', $gpath)
3103 =~ /^(\d{6}) blob ([a-f\d]{40})\t/);
3104 unless (defined $mode && defined $blob) {
3105 die "$path was not found in commit $self->{c} (r$rev)\n";
3106 }
3107 { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3108 pool => SVN::Pool->new, action => 'M' };
3109 }
3110
3111 sub add_file {
3112 my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3113 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3114 delete $self->{empty}->{$dir};
3115 { path => $path, mode_a => 100644, mode_b => 100644,
3116 pool => SVN::Pool->new, action => 'A' };
3117 }
3118
3119 sub add_directory {
3120 my ($self, $path, $cp_path, $cp_rev) = @_;
3121 my $gpath = $self->git_path($path);
3122 if ($gpath eq '') {
3123 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3124 -r --name-only -z/,
3125 $self->{c});
3126 local $/ = "\0";
3127 while (<$ls>) {
3128 chomp;
3129 $self->{gii}->remove($_);
3130 print "\tD\t$_\n" unless $::_q;
3131 }
3132 command_close_pipe($ls, $ctx);
3133 $self->{empty}->{$path} = 0;
3134 }
3135 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3136 delete $self->{empty}->{$dir};
3137 $self->{empty}->{$path} = 1;
3138 { path => $path };
3139 }
3140
3141 sub change_dir_prop {
3142 my ($self, $db, $prop, $value) = @_;
3143 $self->{dir_prop}->{$db->{path}} ||= {};
3144 $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3145 undef;
3146 }
3147
3148 sub absent_directory {
3149 my ($self, $path, $pb) = @_;
3150 $self->{absent_dir}->{$pb->{path}} ||= [];
3151 push @{$self->{absent_dir}->{$pb->{path}}}, $path;
3152 undef;
3153 }
3154
3155 sub absent_file {
3156 my ($self, $path, $pb) = @_;
3157 $self->{absent_file}->{$pb->{path}} ||= [];
3158 push @{$self->{absent_file}->{$pb->{path}}}, $path;
3159 undef;
3160 }
3161
3162 sub change_file_prop {
3163 my ($self, $fb, $prop, $value) = @_;
3164 if ($prop eq 'svn:executable') {
3165 if ($fb->{mode_b} != 120000) {
3166 $fb->{mode_b} = defined $value ? 100755 : 100644;
3167 }
3168 } elsif ($prop eq 'svn:special') {
3169 $fb->{mode_b} = defined $value ? 120000 : 100644;
3170 } else {
3171 $self->{file_prop}->{$fb->{path}} ||= {};
3172 $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
3173 }
3174 undef;
3175 }
3176
3177 sub apply_textdelta {
3178 my ($self, $fb, $exp) = @_;
3179 my $fh = IO::File->new_tmpfile;
3180 $fh->autoflush(1);
3181 # $fh gets auto-closed() by SVN::TxDelta::apply(),
3182 # (but $base does not,) so dup() it for reading in close_file
3183 open my $dup, '<&', $fh or croak $!;
3184 my $base = IO::File->new_tmpfile;
3185 $base->autoflush(1);
3186 if ($fb->{blob}) {
3187 defined (my $pid = fork) or croak $!;
3188 if (!$pid) {
3189 open STDOUT, '>&', $base or croak $!;
3190 print STDOUT 'link ' if ($fb->{mode_a} == 120000);
3191 exec qw/git-cat-file blob/, $fb->{blob} or croak $!;
3192 }
3193 waitpid $pid, 0;
3194 croak $? if $?;
3195
3196 if (defined $exp) {
3197 seek $base, 0, 0 or croak $!;
3198 my $got = ::md5sum($base);
3199 die "Checksum mismatch: $fb->{path} $fb->{blob}\n",
3200 "expected: $exp\n",
3201 " got: $got\n" if ($got ne $exp);
3202 }
3203 }
3204 seek $base, 0, 0 or croak $!;
3205 $fb->{fh} = $dup;
3206 $fb->{base} = $base;
3207 [ SVN::TxDelta::apply($base, $fh, undef, $fb->{path}, $fb->{pool}) ];
3208 }
3209
3210 sub close_file {
3211 my ($self, $fb, $exp) = @_;
3212 my $hash;
3213 my $path = $self->git_path($fb->{path});
3214 if (my $fh = $fb->{fh}) {
3215 if (defined $exp) {
3216 seek($fh, 0, 0) or croak $!;
3217 my $got = ::md5sum($fh);
3218 if ($got ne $exp) {
3219 die "Checksum mismatch: $path\n",
3220 "expected: $exp\n got: $got\n";
3221 }
3222 }
3223 sysseek($fh, 0, 0) or croak $!;
3224 if ($fb->{mode_b} == 120000) {
3225 eval {
3226 sysread($fh, my $buf, 5) == 5 or croak $!;
3227 $buf eq 'link ' or die "$path has mode 120000",
3228 " but is not a link";
3229 };
3230 if ($@) {
3231 warn "$@\n";
3232 sysseek($fh, 0, 0) or croak $!;
3233 }
3234 }
3235 defined(my $pid = open my $out,'-|') or die "Can't fork: $!\n";
3236 if (!$pid) {
3237 open STDIN, '<&', $fh or croak $!;
3238 exec qw/git-hash-object -w --stdin/ or croak $!;
3239 }
3240 chomp($hash = do { local $/; <$out> });
3241 close $out or croak $!;
3242 close $fh or croak $!;
3243 $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
3244 close $fb->{base} or croak $!;
3245 } else {
3246 $hash = $fb->{blob} or die "no blob information\n";
3247 }
3248 $fb->{pool}->clear;
3249 $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
3250 print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
3251 undef;
3252 }
3253
3254 sub abort_edit {
3255 my $self = shift;
3256 $self->{nr} = $self->{gii}->{nr};
3257 delete $self->{gii};
3258 $self->SUPER::abort_edit(@_);
3259 }
3260
3261 sub close_edit {
3262 my $self = shift;
3263 $self->{git_commit_ok} = 1;
3264 $self->{nr} = $self->{gii}->{nr};
3265 delete $self->{gii};
3266 $self->SUPER::close_edit(@_);
3267 }
3268
3269 package SVN::Git::Editor;
3270 use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
3271 use strict;
3272 use warnings;
3273 use Carp qw/croak/;
3274 use IO::File;
3275
3276 sub new {
3277 my ($class, $opts) = @_;
3278 foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
3279 die "$_ required!\n" unless (defined $opts->{$_});
3280 }
3281
3282 my $pool = SVN::Pool->new;
3283 my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
3284 my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
3285 $opts->{r}, $mods);
3286
3287 # $opts->{ra} functions should not be used after this:
3288 my @ce = $opts->{ra}->get_commit_editor($opts->{log},
3289 $opts->{editor_cb}, $pool);
3290 my $self = SVN::Delta::Editor->new(@ce, $pool);
3291 bless $self, $class;
3292 foreach (qw/svn_path r tree_a tree_b/) {
3293 $self->{$_} = $opts->{$_};
3294 }
3295 $self->{url} = $opts->{ra}->{url};
3296 $self->{mods} = $mods;
3297 $self->{types} = $types;
3298 $self->{pool} = $pool;
3299 $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
3300 $self->{rm} = { };
3301 $self->{path_prefix} = length $self->{svn_path} ?
3302 "$self->{svn_path}/" : '';
3303 return $self;
3304 }
3305
3306 sub generate_diff {
3307 my ($tree_a, $tree_b) = @_;
3308 my @diff_tree = qw(diff-tree -z -r);
3309 if ($_cp_similarity) {
3310 push @diff_tree, "-C$_cp_similarity";
3311 } else {
3312 push @diff_tree, '-C';
3313 }
3314 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
3315 push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
3316 push @diff_tree, $tree_a, $tree_b;
3317 my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
3318 local $/ = "\0";
3319 my $state = 'meta';
3320 my @mods;
3321 while (<$diff_fh>) {
3322 chomp $_; # this gets rid of the trailing "\0"
3323 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
3324 $::sha1\s($::sha1)\s
3325 ([MTCRAD])\d*$/xo) {
3326 push @mods, { mode_a => $1, mode_b => $2,
3327 sha1_b => $3, chg => $4 };
3328 if ($4 =~ /^(?:C|R)$/) {
3329 $state = 'file_a';
3330 } else {
3331 $state = 'file_b';
3332 }
3333 } elsif ($state eq 'file_a') {
3334 my $x = $mods[$#mods] or croak "Empty array\n";
3335 if ($x->{chg} !~ /^(?:C|R)$/) {
3336 croak "Error parsing $_, $x->{chg}\n";
3337 }
3338 $x->{file_a} = $_;
3339 $state = 'file_b';
3340 } elsif ($state eq 'file_b') {
3341 my $x = $mods[$#mods] or croak "Empty array\n";
3342 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
3343 croak "Error parsing $_, $x->{chg}\n";
3344 }
3345 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
3346 croak "Error parsing $_, $x->{chg}\n";
3347 }
3348 $x->{file_b} = $_;
3349 $state = 'meta';
3350 } else {
3351 croak "Error parsing $_\n";
3352 }
3353 }
3354 command_close_pipe($diff_fh, $ctx);
3355 \@mods;
3356 }
3357
3358 sub check_diff_paths {
3359 my ($ra, $pfx, $rev, $mods) = @_;
3360 my %types;
3361 $pfx .= '/' if length $pfx;
3362
3363 sub type_diff_paths {
3364 my ($ra, $types, $path, $rev) = @_;
3365 my @p = split m#/+#, $path;
3366 my $c = shift @p;
3367 unless (defined $types->{$c}) {
3368 $types->{$c} = $ra->check_path($c, $rev);
3369 }
3370 while (@p) {
3371 $c .= '/' . shift @p;
3372 next if defined $types->{$c};
3373 $types->{$c} = $ra->check_path($c, $rev);
3374 }
3375 }
3376
3377 foreach my $m (@$mods) {
3378 foreach my $f (qw/file_a file_b/) {
3379 next unless defined $m->{$f};
3380 my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
3381 if (length $pfx.$dir && ! defined $types{$dir}) {
3382 type_diff_paths($ra, \%types, $pfx.$dir, $rev);
3383 }
3384 }
3385 }
3386 \%types;
3387 }
3388
3389 sub split_path {
3390 return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
3391 }
3392
3393 sub repo_path {
3394 my ($self, $path) = @_;
3395 $self->{path_prefix}.(defined $path ? $path : '');
3396 }
3397
3398 sub url_path {
3399 my ($self, $path) = @_;
3400 if ($self->{url} =~ m#^https?://#) {
3401 $path =~ s/([^a-zA-Z0-9_.-])/uc sprintf("%%%02x",ord($1))/eg;
3402 }
3403 $self->{url} . '/' . $self->repo_path($path);
3404 }
3405
3406 sub rmdirs {
3407 my ($self) = @_;
3408 my $rm = $self->{rm};
3409 delete $rm->{''}; # we never delete the url we're tracking
3410 return unless %$rm;
3411
3412 foreach (keys %$rm) {
3413 my @d = split m#/#, $_;
3414 my $c = shift @d;
3415 $rm->{$c} = 1;
3416 while (@d) {
3417 $c .= '/' . shift @d;
3418 $rm->{$c} = 1;
3419 }
3420 }
3421 delete $rm->{$self->{svn_path}};
3422 delete $rm->{''}; # we never delete the url we're tracking
3423 return unless %$rm;
3424
3425 my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
3426 $self->{tree_b});
3427 local $/ = "\0";
3428 while (<$fh>) {
3429 chomp;
3430 my @dn = split m#/#, $_;
3431 while (pop @dn) {
3432 delete $rm->{join '/', @dn};
3433 }
3434 unless (%$rm) {
3435 close $fh;
3436 return;
3437 }
3438 }
3439 command_close_pipe($fh, $ctx);
3440
3441 my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
3442 foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
3443 $self->close_directory($bat->{$d}, $p);
3444 my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
3445 print "\tD+\t$d/\n" unless $::_q;
3446 $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
3447 delete $bat->{$d};
3448 }
3449 }
3450
3451 sub open_or_add_dir {
3452 my ($self, $full_path, $baton) = @_;
3453 my $t = $self->{types}->{$full_path};
3454 if (!defined $t) {
3455 die "$full_path not known in r$self->{r} or we have a bug!\n";
3456 }
3457 {
3458 no warnings 'once';
3459 # SVN::Node::none and SVN::Node::file are used only once,
3460 # so we're shutting up Perl's warnings about them.
3461 if ($t == $SVN::Node::none) {
3462 return $self->add_directory($full_path, $baton,
3463 undef, -1, $self->{pool});
3464 } elsif ($t == $SVN::Node::dir) {
3465 return $self->open_directory($full_path, $baton,
3466 $self->{r}, $self->{pool});
3467 } # no warnings 'once'
3468 print STDERR "$full_path already exists in repository at ",
3469 "r$self->{r} and it is not a directory (",
3470 ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
3471 } # no warnings 'once'
3472 exit 1;
3473 }
3474
3475 sub ensure_path {
3476 my ($self, $path) = @_;
3477 my $bat = $self->{bat};
3478 my $repo_path = $self->repo_path($path);
3479 return $bat->{''} unless (length $repo_path);
3480 my @p = split m#/+#, $repo_path;
3481 my $c = shift @p;
3482 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
3483 while (@p) {
3484 my $c0 = $c;
3485 $c .= '/' . shift @p;
3486 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
3487 }
3488 return $bat->{$c};
3489 }
3490
3491 sub A {
3492 my ($self, $m) = @_;
3493 my ($dir, $file) = split_path($m->{file_b});
3494 my $pbat = $self->ensure_path($dir);
3495 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3496 undef, -1);
3497 print "\tA\t$m->{file_b}\n" unless $::_q;
3498 $self->chg_file($fbat, $m);
3499 $self->close_file($fbat,undef,$self->{pool});
3500 }
3501
3502 sub C {
3503 my ($self, $m) = @_;
3504 my ($dir, $file) = split_path($m->{file_b});
3505 my $pbat = $self->ensure_path($dir);
3506 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3507 $self->url_path($m->{file_a}), $self->{r});
3508 print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3509 $self->chg_file($fbat, $m);
3510 $self->close_file($fbat,undef,$self->{pool});
3511 }
3512
3513 sub delete_entry {
3514 my ($self, $path, $pbat) = @_;
3515 my $rpath = $self->repo_path($path);
3516 my ($dir, $file) = split_path($rpath);
3517 $self->{rm}->{$dir} = 1;
3518 $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
3519 }
3520
3521 sub R {
3522 my ($self, $m) = @_;
3523 my ($dir, $file) = split_path($m->{file_b});
3524 my $pbat = $self->ensure_path($dir);
3525 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3526 $self->url_path($m->{file_a}), $self->{r});
3527 print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3528 $self->chg_file($fbat, $m);
3529 $self->close_file($fbat,undef,$self->{pool});
3530
3531 ($dir, $file) = split_path($m->{file_a});
3532 $pbat = $self->ensure_path($dir);
3533 $self->delete_entry($m->{file_a}, $pbat);
3534 }
3535
3536 sub M {
3537 my ($self, $m) = @_;
3538 my ($dir, $file) = split_path($m->{file_b});
3539 my $pbat = $self->ensure_path($dir);
3540 my $fbat = $self->open_file($self->repo_path($m->{file_b}),
3541 $pbat,$self->{r},$self->{pool});
3542 print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
3543 $self->chg_file($fbat, $m);
3544 $self->close_file($fbat,undef,$self->{pool});
3545 }
3546
3547 sub T { shift->M(@_) }
3548
3549 sub change_file_prop {
3550 my ($self, $fbat, $pname, $pval) = @_;
3551 $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
3552 }
3553
3554 sub chg_file {
3555 my ($self, $fbat, $m) = @_;
3556 if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
3557 $self->change_file_prop($fbat,'svn:executable','*');
3558 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
3559 $self->change_file_prop($fbat,'svn:executable',undef);
3560 }
3561 my $fh = IO::File->new_tmpfile or croak $!;
3562 if ($m->{mode_b} =~ /^120/) {
3563 print $fh 'link ' or croak $!;
3564 $self->change_file_prop($fbat,'svn:special','*');
3565 } elsif ($m->{mode_a} =~ /^120/ && $m->{mode_b} !~ /^120/) {
3566 $self->change_file_prop($fbat,'svn:special',undef);
3567 }
3568 defined(my $pid = fork) or croak $!;
3569 if (!$pid) {
3570 open STDOUT, '>&', $fh or croak $!;
3571 exec qw/git-cat-file blob/, $m->{sha1_b} or croak $!;
3572 }
3573 waitpid $pid, 0;
3574 croak $? if $?;
3575 $fh->flush == 0 or croak $!;
3576 seek $fh, 0, 0 or croak $!;
3577
3578 my $exp = ::md5sum($fh);
3579 seek $fh, 0, 0 or croak $!;
3580
3581 my $pool = SVN::Pool->new;
3582 my $atd = $self->apply_textdelta($fbat, undef, $pool);
3583 my $got = SVN::TxDelta::send_stream($fh, @$atd, $pool);
3584 die "Checksum mismatch\nexpected: $exp\ngot: $got\n" if ($got ne $exp);
3585 $pool->clear;
3586
3587 close $fh or croak $!;
3588 }
3589
3590 sub D {
3591 my ($self, $m) = @_;
3592 my ($dir, $file) = split_path($m->{file_b});
3593 my $pbat = $self->ensure_path($dir);
3594 print "\tD\t$m->{file_b}\n" unless $::_q;
3595 $self->delete_entry($m->{file_b}, $pbat);
3596 }
3597
3598 sub close_edit {
3599 my ($self) = @_;
3600 my ($p,$bat) = ($self->{pool}, $self->{bat});
3601 foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
3602 next if $_ eq '';
3603 $self->close_directory($bat->{$_}, $p);
3604 }
3605 $self->close_directory($bat->{''}, $p);
3606 $self->SUPER::close_edit($p);
3607 $p->clear;
3608 }
3609
3610 sub abort_edit {
3611 my ($self) = @_;
3612 $self->SUPER::abort_edit($self->{pool});
3613 }
3614
3615 sub DESTROY {
3616 my $self = shift;
3617 $self->SUPER::DESTROY(@_);
3618 $self->{pool}->clear;
3619 }
3620
3621 # this drives the editor
3622 sub apply_diff {
3623 my ($self) = @_;
3624 my $mods = $self->{mods};
3625 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
3626 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
3627 my $f = $m->{chg};
3628 if (defined $o{$f}) {
3629 $self->$f($m);
3630 } else {
3631 fatal("Invalid change type: $f");
3632 }
3633 }
3634 $self->rmdirs if $_rmdir;
3635 if (@$mods == 0) {
3636 $self->abort_edit;
3637 } else {
3638 $self->close_edit;
3639 }
3640 return scalar @$mods;
3641 }
3642
3643 package Git::SVN::Ra;
3644 use vars qw/@ISA $config_dir $_log_window_size/;
3645 use strict;
3646 use warnings;
3647 my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
3648
3649 BEGIN {
3650 # enforce temporary pool usage for some simple functions
3651 no strict 'refs';
3652 for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root/) {
3653 my $SUPER = "SUPER::$f";
3654 *$f = sub {
3655 my $self = shift;
3656 my $pool = SVN::Pool->new;
3657 my @ret = $self->$SUPER(@_,$pool);
3658 $pool->clear;
3659 wantarray ? @ret : $ret[0];
3660 };
3661 }
3662 }
3663
3664 sub _auth_providers () {
3665 [
3666 SVN::Client::get_simple_provider(),
3667 SVN::Client::get_ssl_server_trust_file_provider(),
3668 SVN::Client::get_simple_prompt_provider(
3669 \&Git::SVN::Prompt::simple, 2),
3670 SVN::Client::get_ssl_client_cert_file_provider(),
3671 SVN::Client::get_ssl_client_cert_prompt_provider(
3672 \&Git::SVN::Prompt::ssl_client_cert, 2),
3673 SVN::Client::get_ssl_client_cert_pw_file_provider(),
3674 SVN::Client::get_ssl_client_cert_pw_prompt_provider(
3675 \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
3676 SVN::Client::get_username_provider(),
3677 SVN::Client::get_ssl_server_trust_prompt_provider(
3678 \&Git::SVN::Prompt::ssl_server_trust),
3679 SVN::Client::get_username_prompt_provider(
3680 \&Git::SVN::Prompt::username, 2)
3681 ]
3682 }
3683
3684 sub escape_uri_only {
3685 my ($uri) = @_;
3686 my @tmp;
3687 foreach (split m{/}, $uri) {
3688 s/([^\w.%-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
3689 push @tmp, $_;
3690 }
3691 join('/', @tmp);
3692 }
3693
3694 sub escape_url {
3695 my ($url) = @_;
3696 if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
3697 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
3698 $url = "$scheme://$domain$uri";
3699 }
3700 $url;
3701 }
3702
3703 sub new {
3704 my ($class, $url) = @_;
3705 $url =~ s!/+$!!;
3706 return $RA if ($RA && $RA->{url} eq $url);
3707
3708 SVN::_Core::svn_config_ensure($config_dir, undef);
3709 my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
3710 my $config = SVN::Core::config_get_config($config_dir);
3711 $RA = undef;
3712 my $dont_store_passwords = 1;
3713 my $conf_t = ${$config}{'config'};
3714 {
3715 no warnings 'once';
3716 # The usage of $SVN::_Core::SVN_CONFIG_* variables
3717 # produces warnings that variables are used only once.
3718 # I had not found the better way to shut them up, so
3719 # the warnings of type 'once' are disabled in this block.
3720 if (SVN::_Core::svn_config_get_bool($conf_t,
3721 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3722 $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
3723 1) == 0) {
3724 SVN::_Core::svn_auth_set_parameter($baton,
3725 $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
3726 bless (\$dont_store_passwords, "_p_void"));
3727 }
3728 if (SVN::_Core::svn_config_get_bool($conf_t,
3729 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3730 $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
3731 1) == 0) {
3732 $Git::SVN::Prompt::_no_auth_cache = 1;
3733 }
3734 } # no warnings 'once'
3735 my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
3736 config => $config,
3737 pool => SVN::Pool->new,
3738 auth_provider_callbacks => $callbacks);
3739 $self->{url} = $url;
3740 $self->{svn_path} = $url;
3741 $self->{repos_root} = $self->get_repos_root;
3742 $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
3743 $self->{cache} = { check_path => { r => 0, data => {} },
3744 get_dir => { r => 0, data => {} } };
3745 $RA = bless $self, $class;
3746 }
3747
3748 sub check_path {
3749 my ($self, $path, $r) = @_;
3750 my $cache = $self->{cache}->{check_path};
3751 if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
3752 return $cache->{data}->{$path};
3753 }
3754 my $pool = SVN::Pool->new;
3755 my $t = $self->SUPER::check_path($path, $r, $pool);
3756 $pool->clear;
3757 if ($r != $cache->{r}) {
3758 %{$cache->{data}} = ();
3759 $cache->{r} = $r;
3760 }
3761 $cache->{data}->{$path} = $t;
3762 }
3763
3764 sub get_dir {
3765 my ($self, $dir, $r) = @_;
3766 my $cache = $self->{cache}->{get_dir};
3767 if ($r == $cache->{r}) {
3768 if (my $x = $cache->{data}->{$dir}) {
3769 return wantarray ? @$x : $x->[0];
3770 }
3771 }
3772 my $pool = SVN::Pool->new;
3773 my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
3774 my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
3775 $pool->clear;
3776 if ($r != $cache->{r}) {
3777 %{$cache->{data}} = ();
3778 $cache->{r} = $r;
3779 }
3780 $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
3781 wantarray ? (\%dirents, $r, $props) : \%dirents;
3782 }
3783
3784 sub DESTROY {
3785 # do not call the real DESTROY since we store ourselves in $RA
3786 }
3787
3788 sub get_log {
3789 my ($self, @args) = @_;
3790 my $pool = SVN::Pool->new;
3791 splice(@args, 3, 1) if ($SVN::Core::VERSION le '1.2.0');
3792 my $ret = $self->SUPER::get_log(@args, $pool);
3793 $pool->clear;
3794 $ret;
3795 }
3796
3797 sub trees_match {
3798 my ($self, $url1, $rev1, $url2, $rev2) = @_;
3799 my $ctx = SVN::Client->new(auth => _auth_providers);
3800 my $out = IO::File->new_tmpfile;
3801
3802 # older SVN (1.1.x) doesn't take $pool as the last parameter for
3803 # $ctx->diff(), so we'll create a default one
3804 my $pool = SVN::Pool->new_default_sub;
3805
3806 $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
3807 $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
3808 $out->flush;
3809 my $ret = (($out->stat)[7] == 0);
3810 close $out or croak $!;
3811
3812 $ret;
3813 }
3814
3815 sub get_commit_editor {
3816 my ($self, $log, $cb, $pool) = @_;
3817 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
3818 $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
3819 }
3820
3821 sub gs_do_update {
3822 my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
3823 my $new = ($rev_a == $rev_b);
3824 my $path = $gs->{path};
3825
3826 if ($new && -e $gs->{index}) {
3827 unlink $gs->{index} or die
3828 "Couldn't unlink index: $gs->{index}: $!\n";
3829 }
3830 my $pool = SVN::Pool->new;
3831 $editor->set_path_strip($path);
3832 my (@pc) = split m#/#, $path;
3833 my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
3834 1, $editor, $pool);
3835 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3836
3837 # Since we can't rely on svn_ra_reparent being available, we'll
3838 # just have to do some magic with set_path to make it so
3839 # we only want a partial path.
3840 my $sp = '';
3841 my $final = join('/', @pc);
3842 while (@pc) {
3843 $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
3844 $sp .= '/' if length $sp;
3845 $sp .= shift @pc;
3846 }
3847 die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
3848
3849 $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
3850
3851 $reporter->finish_report($pool);
3852 $pool->clear;
3853 $editor->{git_commit_ok};
3854 }
3855
3856 # this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
3857 # svn_ra_reparent didn't work before 1.4)
3858 sub gs_do_switch {
3859 my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
3860 my $path = $gs->{path};
3861 my $pool = SVN::Pool->new;
3862
3863 my $full_url = $self->{url};
3864 my $old_url = $full_url;
3865 $full_url .= '/' . escape_uri_only($path) if length $path;
3866 my ($ra, $reparented);
3867 if ($old_url ne $full_url) {
3868 if ($old_url !~ m#^svn(\+ssh)?://#) {
3869 SVN::_Ra::svn_ra_reparent($self->{session}, $full_url,
3870 $pool);
3871 $self->{url} = $full_url;
3872 $reparented = 1;
3873 } else {
3874 $_[0] = undef;
3875 $self = undef;
3876 $RA = undef;
3877 $ra = Git::SVN::Ra->new($full_url);
3878 $ra_invalid = 1;
3879 }
3880 }
3881 $ra ||= $self;
3882 my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
3883 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3884 $reporter->set_path('', $rev_a, 0, @lock, $pool);
3885 $reporter->finish_report($pool);
3886
3887 if ($reparented) {
3888 SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
3889 $self->{url} = $old_url;
3890 }
3891
3892 $pool->clear;
3893 $editor->{git_commit_ok};
3894 }
3895
3896 sub longest_common_path {
3897 my ($gsv, $globs) = @_;
3898 my %common;
3899 my $common_max = scalar @$gsv;
3900
3901 foreach my $gs (@$gsv) {
3902 my @tmp = split m#/#, $gs->{path};
3903 my $p = '';
3904 foreach (@tmp) {
3905 $p .= length($p) ? "/$_" : $_;
3906 $common{$p} ||= 0;
3907 $common{$p}++;
3908 }
3909 }
3910 $globs ||= [];
3911 $common_max += scalar @$globs;
3912 foreach my $glob (@$globs) {
3913 my @tmp = split m#/#, $glob->{path}->{left};
3914 my $p = '';
3915 foreach (@tmp) {
3916 $p .= length($p) ? "/$_" : $_;
3917 $common{$p} ||= 0;
3918 $common{$p}++;
3919 }
3920 }
3921
3922 my $longest_path = '';
3923 foreach (sort {length $b <=> length $a} keys %common) {
3924 if ($common{$_} == $common_max) {
3925 $longest_path = $_;
3926 last;
3927 }
3928 }
3929 $longest_path;
3930 }
3931
3932 sub gs_fetch_loop_common {
3933 my ($self, $base, $head, $gsv, $globs) = @_;
3934 return if ($base > $head);
3935 my $inc = $_log_window_size;
3936 my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
3937 my $longest_path = longest_common_path($gsv, $globs);
3938 my $ra_url = $self->{url};
3939 while (1) {
3940 my %revs;
3941 my $err;
3942 my $err_handler = $SVN::Error::handler;
3943 $SVN::Error::handler = sub {
3944 ($err) = @_;
3945 skip_unknown_revs($err);
3946 };
3947 sub _cb {
3948 my ($paths, $r, $author, $date, $log) = @_;
3949 [ dup_changed_paths($paths),
3950 { author => $author, date => $date, log => $log } ];
3951 }
3952 $self->get_log([$longest_path], $min, $max, 0, 1, 1,
3953 sub { $revs{$_[1]} = _cb(@_) });
3954 if ($err && $max >= $head) {
3955 print STDERR "Path '$longest_path' ",
3956 "was probably deleted:\n",
3957 $err->expanded_message,
3958 "\nWill attempt to follow ",
3959 "revisions r$min .. r$max ",
3960 "committed before the deletion\n";
3961 my $hi = $max;
3962 while (--$hi >= $min) {
3963 my $ok;
3964 $self->get_log([$longest_path], $min, $hi,
3965 0, 1, 1, sub {
3966 $ok ||= $_[1];
3967 $revs{$_[1]} = _cb(@_) });
3968 if ($ok) {
3969 print STDERR "r$min .. r$ok OK\n";
3970 last;
3971 }
3972 }
3973 }
3974 $SVN::Error::handler = $err_handler;
3975
3976 my %exists = map { $_->{path} => $_ } @$gsv;
3977 foreach my $r (sort {$a <=> $b} keys %revs) {
3978 my ($paths, $logged) = @{$revs{$r}};
3979
3980 foreach my $gs ($self->match_globs(\%exists, $paths,
3981 $globs, $r)) {
3982 if ($gs->rev_map_max >= $r) {
3983 next;
3984 }
3985 next unless $gs->match_paths($paths, $r);
3986 $gs->{logged_rev_props} = $logged;
3987 if (my $last_commit = $gs->last_commit) {
3988 $gs->assert_index_clean($last_commit);
3989 }
3990 my $log_entry = $gs->do_fetch($paths, $r);
3991 if ($log_entry) {
3992 $gs->do_git_commit($log_entry);
3993 }
3994 $INDEX_FILES{$gs->{index}} = 1;
3995 }
3996 foreach my $g (@$globs) {
3997 my $k = "svn-remote.$g->{remote}." .
3998 "$g->{t}-maxRev";
3999 Git::SVN::tmp_config($k, $r);
4000 }
4001 if ($ra_invalid) {
4002 $_[0] = undef;
4003 $self = undef;
4004 $RA = undef;
4005 $self = Git::SVN::Ra->new($ra_url);
4006 $ra_invalid = undef;
4007 }
4008 }
4009 # pre-fill the .rev_db since it'll eventually get filled in
4010 # with '0' x40 if something new gets committed
4011 foreach my $gs (@$gsv) {
4012 next if $gs->rev_map_max >= $max;
4013 next if defined $gs->rev_map_get($max);
4014 $gs->rev_map_set($max, 0 x40);
4015 }
4016 foreach my $g (@$globs) {
4017 my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
4018 Git::SVN::tmp_config($k, $max);
4019 }
4020 last if $max >= $head;
4021 $min = $max + 1;
4022 $max += $inc;
4023 $max = $head if ($max > $head);
4024 }
4025 Git::SVN::gc();
4026 }
4027
4028 sub match_globs {
4029 my ($self, $exists, $paths, $globs, $r) = @_;
4030
4031 sub get_dir_check {
4032 my ($self, $exists, $g, $r) = @_;
4033 my @x = eval { $self->get_dir($g->{path}->{left}, $r) };
4034 return unless scalar @x == 3;
4035 my $dirents = $x[0];
4036 foreach my $de (keys %$dirents) {
4037 next if $dirents->{$de}->{kind} != $SVN::Node::dir;
4038 my $p = $g->{path}->full_path($de);
4039 next if $exists->{$p};
4040 next if (length $g->{path}->{right} &&
4041 ($self->check_path($p, $r) !=
4042 $SVN::Node::dir));
4043 $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
4044 $g->{ref}->full_path($de), 1);
4045 }
4046 }
4047 foreach my $g (@$globs) {
4048 if (my $path = $paths->{"/$g->{path}->{left}"}) {
4049 if ($path->{action} =~ /^[AR]$/) {
4050 get_dir_check($self, $exists, $g, $r);
4051 }
4052 }
4053 foreach (keys %$paths) {
4054 if (/$g->{path}->{left_regex}/ &&
4055 !/$g->{path}->{regex}/) {
4056 next if $paths->{$_}->{action} !~ /^[AR]$/;
4057 get_dir_check($self, $exists, $g, $r);
4058 }
4059 next unless /$g->{path}->{regex}/;
4060 my $p = $1;
4061 my $pathname = $g->{path}->full_path($p);
4062 next if $exists->{$pathname};
4063 next if ($self->check_path($pathname, $r) !=
4064 $SVN::Node::dir);
4065 $exists->{$pathname} = Git::SVN->init(
4066 $self->{url}, $pathname, undef,
4067 $g->{ref}->full_path($p), 1);
4068 }
4069 my $c = '';
4070 foreach (split m#/#, $g->{path}->{left}) {
4071 $c .= "/$_";
4072 next unless ($paths->{$c} &&
4073 ($paths->{$c}->{action} =~ /^[AR]$/));
4074 get_dir_check($self, $exists, $g, $r);
4075 }
4076 }
4077 values %$exists;
4078 }
4079
4080 sub minimize_url {
4081 my ($self) = @_;
4082 return $self->{url} if ($self->{url} eq $self->{repos_root});
4083 my $url = $self->{repos_root};
4084 my @components = split(m!/!, $self->{svn_path});
4085 my $c = '';
4086 do {
4087 $url .= "/$c" if length $c;
4088 eval { (ref $self)->new($url)->get_latest_revnum };
4089 } while ($@ && ($c = shift @components));
4090 $url;
4091 }
4092
4093 sub can_do_switch {
4094 my $self = shift;
4095 unless (defined $can_do_switch) {
4096 my $pool = SVN::Pool->new;
4097 my $rep = eval {
4098 $self->do_switch(1, '', 0, $self->{url},
4099 SVN::Delta::Editor->new, $pool);
4100 };
4101 if ($@) {
4102 $can_do_switch = 0;
4103 } else {
4104 $rep->abort_report($pool);
4105 $can_do_switch = 1;
4106 }
4107 $pool->clear;
4108 }
4109 $can_do_switch;
4110 }
4111
4112 sub skip_unknown_revs {
4113 my ($err) = @_;
4114 my $errno = $err->apr_err();
4115 # Maybe the branch we're tracking didn't
4116 # exist when the repo started, so it's
4117 # not an error if it doesn't, just continue
4118 #
4119 # Wonderfully consistent library, eh?
4120 # 160013 - svn:// and file://
4121 # 175002 - http(s)://
4122 # 175007 - http(s):// (this repo required authorization, too...)
4123 # More codes may be discovered later...
4124 if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
4125 my $err_key = $err->expanded_message;
4126 # revision numbers change every time, filter them out
4127 $err_key =~ s/\d+/\0/g;
4128 $err_key = "$errno\0$err_key";
4129 unless ($ignored_err{$err_key}) {
4130 warn "W: Ignoring error from SVN, path probably ",
4131 "does not exist: ($errno): ",
4132 $err->expanded_message,"\n";
4133 warn "W: Do not be alarmed at the above message ",
4134 "git-svn is just searching aggressively for ",
4135 "old history.\n",
4136 "This may take a while on large repositories\n";
4137 $ignored_err{$err_key} = 1;
4138 }
4139 return;
4140 }
4141 die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
4142 }
4143
4144 # svn_log_changed_path_t objects passed to get_log are likely to be
4145 # overwritten even if only the refs are copied to an external variable,
4146 # so we should dup the structures in their entirety. Using an externally
4147 # passed pool (instead of our temporary and quickly cleared pool in
4148 # Git::SVN::Ra) does not help matters at all...
4149 sub dup_changed_paths {
4150 my ($paths) = @_;
4151 return undef unless $paths;
4152 my %ret;
4153 foreach my $p (keys %$paths) {
4154 my $i = $paths->{$p};
4155 my %s = map { $_ => $i->$_ }
4156 qw/copyfrom_path copyfrom_rev action/;
4157 $ret{$p} = \%s;
4158 }
4159 \%ret;
4160 }
4161
4162 package Git::SVN::Log;
4163 use strict;
4164 use warnings;
4165 use POSIX qw/strftime/;
4166 use constant commit_log_separator => ('-' x 72) . "\n";
4167 use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
4168 %rusers $show_commit $incremental/;
4169 my $l_fmt;
4170
4171 sub cmt_showable {
4172 my ($c) = @_;
4173 return 1 if defined $c->{r};
4174
4175 # big commit message got truncated by the 16k pretty buffer in rev-list
4176 if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
4177 $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
4178 @{$c->{l}} = ();
4179 my @log = command(qw/cat-file commit/, $c->{c});
4180
4181 # shift off the headers
4182 shift @log while ($log[0] ne '');
4183 shift @log;
4184
4185 # TODO: make $c->{l} not have a trailing newline in the future
4186 @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
4187
4188 (undef, $c->{r}, undef) = ::extract_metadata(
4189 (grep(/^git-svn-id: /, @log))[-1]);
4190 }
4191 return defined $c->{r};
4192 }
4193
4194 sub log_use_color {
4195 return $color || Git->repository->get_colorbool('color.diff');
4196 }
4197
4198 sub git_svn_log_cmd {
4199 my ($r_min, $r_max, @args) = @_;
4200 my $head = 'HEAD';
4201 my (@files, @log_opts);
4202 foreach my $x (@args) {
4203 if ($x eq '--' || @files) {
4204 push @files, $x;
4205 } else {
4206 if (::verify_ref("$x^0")) {
4207 $head = $x;
4208 } else {
4209 push @log_opts, $x;
4210 }
4211 }
4212 }
4213
4214 my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
4215 $gs ||= Git::SVN->_new;
4216 my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
4217 $gs->refname);
4218 push @cmd, '-r' unless $non_recursive;
4219 push @cmd, qw/--raw --name-status/ if $verbose;
4220 push @cmd, '--color' if log_use_color();
4221 push @cmd, @log_opts;
4222 if (defined $r_max && $r_max == $r_min) {
4223 push @cmd, '--max-count=1';
4224 if (my $c = $gs->rev_map_get($r_max)) {
4225 push @cmd, $c;
4226 }
4227 } elsif (defined $r_max) {
4228 if ($r_max < $r_min) {
4229 ($r_min, $r_max) = ($r_max, $r_min);
4230 }
4231 my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
4232 my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
4233 # If there are no commits in the range, both $c_max and $c_min
4234 # will be undefined. If there is at least 1 commit in the
4235 # range, both will be defined.
4236 return () if !defined $c_min || !defined $c_max;
4237 if ($c_min eq $c_max) {
4238 push @cmd, '--max-count=1', $c_min;
4239 } else {
4240 push @cmd, '--boundary', "$c_min..$c_max";
4241 }
4242 }
4243 return (@cmd, @files);
4244 }
4245
4246 # adapted from pager.c
4247 sub config_pager {
4248 $pager ||= $ENV{GIT_PAGER} || $ENV{PAGER};
4249 if (!defined $pager) {
4250 $pager = 'less';
4251 } elsif (length $pager == 0 || $pager eq 'cat') {
4252 $pager = undef;
4253 }
4254 $ENV{GIT_PAGER_IN_USE} = defined($pager);
4255 }
4256
4257 sub run_pager {
4258 return unless -t *STDOUT && defined $pager;
4259 pipe my $rfd, my $wfd or return;
4260 defined(my $pid = fork) or ::fatal "Can't fork: $!";
4261 if (!$pid) {
4262 open STDOUT, '>&', $wfd or
4263 ::fatal "Can't redirect to stdout: $!";
4264 return;
4265 }
4266 open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
4267 $ENV{LESS} ||= 'FRSX';
4268 exec $pager or ::fatal "Can't run pager: $! ($pager)";
4269 }
4270
4271 sub format_svn_date {
4272 return strftime("%Y-%m-%d %H:%M:%S %z (%a, %d %b %Y)", localtime(shift));
4273 }
4274
4275 sub parse_git_date {
4276 my ($t, $tz) = @_;
4277 # Date::Parse isn't in the standard Perl distro :(
4278 if ($tz =~ s/^\+//) {
4279 $t += tz_to_s_offset($tz);
4280 } elsif ($tz =~ s/^\-//) {
4281 $t -= tz_to_s_offset($tz);
4282 }
4283 return $t;
4284 }
4285
4286 sub set_local_timezone {
4287 if (defined $TZ) {
4288 $ENV{TZ} = $TZ;
4289 } else {
4290 delete $ENV{TZ};
4291 }
4292 }
4293
4294 sub tz_to_s_offset {
4295 my ($tz) = @_;
4296 $tz =~ s/(\d\d)$//;
4297 return ($1 * 60) + ($tz * 3600);
4298 }
4299
4300 sub get_author_info {
4301 my ($dest, $author, $t, $tz) = @_;
4302 $author =~ s/(?:^\s*|\s*$)//g;
4303 $dest->{a_raw} = $author;
4304 my $au;
4305 if ($::_authors) {
4306 $au = $rusers{$author} || undef;
4307 }
4308 if (!$au) {
4309 ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
4310 }
4311 $dest->{t} = $t;
4312 $dest->{tz} = $tz;
4313 $dest->{a} = $au;
4314 $dest->{t_utc} = parse_git_date($t, $tz);
4315 }
4316
4317 sub process_commit {
4318 my ($c, $r_min, $r_max, $defer) = @_;
4319 if (defined $r_min && defined $r_max) {
4320 if ($r_min == $c->{r} && $r_min == $r_max) {
4321 show_commit($c);
4322 return 0;
4323 }
4324 return 1 if $r_min == $r_max;
4325 if ($r_min < $r_max) {
4326 # we need to reverse the print order
4327 return 0 if (defined $limit && --$limit < 0);
4328 push @$defer, $c;
4329 return 1;
4330 }
4331 if ($r_min != $r_max) {
4332 return 1 if ($r_min < $c->{r});
4333 return 1 if ($r_max > $c->{r});
4334 }
4335 }
4336 return 0 if (defined $limit && --$limit < 0);
4337 show_commit($c);
4338 return 1;
4339 }
4340
4341 sub show_commit {
4342 my $c = shift;
4343 if ($oneline) {
4344 my $x = "\n";
4345 if (my $l = $c->{l}) {
4346 while ($l->[0] =~ /^\s*$/) { shift @$l }
4347 $x = $l->[0];
4348 }
4349 $l_fmt ||= 'A' . length($c->{r});
4350 print 'r',pack($l_fmt, $c->{r}),' | ';
4351 print "$c->{c} | " if $show_commit;
4352 print $x;
4353 } else {
4354 show_commit_normal($c);
4355 }
4356 }
4357
4358 sub show_commit_changed_paths {
4359 my ($c) = @_;
4360 return unless $c->{changed};
4361 print "Changed paths:\n", @{$c->{changed}};
4362 }
4363
4364 sub show_commit_normal {
4365 my ($c) = @_;
4366 print commit_log_separator, "r$c->{r} | ";
4367 print "$c->{c} | " if $show_commit;
4368 print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
4369 my $nr_line = 0;
4370
4371 if (my $l = $c->{l}) {
4372 while ($l->[$#$l] eq "\n" && $#$l > 0
4373 && $l->[($#$l - 1)] eq "\n") {
4374 pop @$l;
4375 }
4376 $nr_line = scalar @$l;
4377 if (!$nr_line) {
4378 print "1 line\n\n\n";
4379 } else {
4380 if ($nr_line == 1) {
4381 $nr_line = '1 line';
4382 } else {
4383 $nr_line .= ' lines';
4384 }
4385 print $nr_line, "\n";
4386 show_commit_changed_paths($c);
4387 print "\n";
4388 print $_ foreach @$l;
4389 }
4390 } else {
4391 print "1 line\n";
4392 show_commit_changed_paths($c);
4393 print "\n";
4394
4395 }
4396 foreach my $x (qw/raw stat diff/) {
4397 if ($c->{$x}) {
4398 print "\n";
4399 print $_ foreach @{$c->{$x}}
4400 }
4401 }
4402 }
4403
4404 sub cmd_show_log {
4405 my (@args) = @_;
4406 my ($r_min, $r_max);
4407 my $r_last = -1; # prevent dupes
4408 set_local_timezone();
4409 if (defined $::_revision) {
4410 if ($::_revision =~ /^(\d+):(\d+)$/) {
4411 ($r_min, $r_max) = ($1, $2);
4412 } elsif ($::_revision =~ /^\d+$/) {
4413 $r_min = $r_max = $::_revision;
4414 } else {
4415 ::fatal "-r$::_revision is not supported, use ",
4416 "standard 'git log' arguments instead";
4417 }
4418 }
4419
4420 config_pager();
4421 @args = git_svn_log_cmd($r_min, $r_max, @args);
4422 if (!@args) {
4423 print commit_log_separator unless $incremental || $oneline;
4424 return;
4425 }
4426 my $log = command_output_pipe(@args);
4427 run_pager();
4428 my (@k, $c, $d, $stat);
4429 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
4430 while (<$log>) {
4431 if (/^${esc_color}commit -?($::sha1_short)/o) {
4432 my $cmt = $1;
4433 if ($c && cmt_showable($c) && $c->{r} != $r_last) {
4434 $r_last = $c->{r};
4435 process_commit($c, $r_min, $r_max, \@k) or
4436 goto out;
4437 }
4438 $d = undef;
4439 $c = { c => $cmt };
4440 } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
4441 get_author_info($c, $1, $2, $3);
4442 } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
4443 # ignore
4444 } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
4445 push @{$c->{raw}}, $_;
4446 } elsif (/^${esc_color}[ACRMDT]\t/) {
4447 # we could add $SVN->{svn_path} here, but that requires
4448 # remote access at the moment (repo_path_split)...
4449 s#^(${esc_color})([ACRMDT])\t#$1 $2 #o;
4450 push @{$c->{changed}}, $_;
4451 } elsif (/^${esc_color}diff /o) {
4452 $d = 1;
4453 push @{$c->{diff}}, $_;
4454 } elsif ($d) {
4455 push @{$c->{diff}}, $_;
4456 } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
4457 $esc_color*[\+\-]*$esc_color$/x) {
4458 $stat = 1;
4459 push @{$c->{stat}}, $_;
4460 } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
4461 push @{$c->{stat}}, $_;
4462 $stat = undef;
4463 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
4464 ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
4465 } elsif (s/^${esc_color} //o) {
4466 push @{$c->{l}}, $_;
4467 }
4468 }
4469 if ($c && defined $c->{r} && $c->{r} != $r_last) {
4470 $r_last = $c->{r};
4471 process_commit($c, $r_min, $r_max, \@k);
4472 }
4473 if (@k) {
4474 ($r_min, $r_max) = ($r_max, $r_min);
4475 process_commit($_, $r_min, $r_max) foreach reverse @k;
4476 }
4477 out:
4478 close $log;
4479 print commit_log_separator unless $incremental || $oneline;
4480 }
4481
4482 sub cmd_blame {
4483 my $path = shift;
4484
4485 config_pager();
4486 run_pager();
4487
4488 my ($fh, $ctx) = command_output_pipe('blame', @_, $path);
4489 while (my $line = <$fh>) {
4490 if ($line =~ /^\^?([[:xdigit:]]+)\s/) {
4491 my (undef, $rev, undef) = ::cmt_metadata($1);
4492 $rev = sprintf('%-10s', $rev);
4493 $line =~ s/^\^?[[:xdigit:]]+(\s)/$rev$1/;
4494 }
4495 print $line;
4496 }
4497 command_close_pipe($fh, $ctx);
4498 }
4499
4500 package Git::SVN::Migration;
4501 # these version numbers do NOT correspond to actual version numbers
4502 # of git nor git-svn. They are just relative.
4503 #
4504 # v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
4505 #
4506 # v1 layout: .git/$id/info/url, refs/remotes/$id
4507 #
4508 # v2 layout: .git/svn/$id/info/url, refs/remotes/$id
4509 #
4510 # v3 layout: .git/svn/$id, refs/remotes/$id
4511 # - info/url may remain for backwards compatibility
4512 # - this is what we migrate up to this layout automatically,
4513 # - this will be used by git svn init on single branches
4514 # v3.1 layout (auto migrated):
4515 # - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
4516 # for backwards compatibility
4517 #
4518 # v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
4519 # - this is only created for newly multi-init-ed
4520 # repositories. Similar in spirit to the
4521 # --use-separate-remotes option in git-clone (now default)
4522 # - we do not automatically migrate to this (following
4523 # the example set by core git)
4524 #
4525 # v5 layout: .rev_db.$UUID => .rev_map.$UUID
4526 # - newer, more-efficient format that uses 24-bytes per record
4527 # with no filler space.
4528 # - use xxd -c24 < .rev_map.$UUID to view and debug
4529 # - This is a one-way migration, repositories updated to the
4530 # new format will not be able to use old git-svn without
4531 # rebuilding the .rev_db. Rebuilding the rev_db is not
4532 # possible if noMetadata or useSvmProps are set; but should
4533 # be no problem for users that use the (sensible) defaults.
4534 use strict;
4535 use warnings;
4536 use Carp qw/croak/;
4537 use File::Path qw/mkpath/;
4538 use File::Basename qw/dirname basename/;
4539 use vars qw/$_minimize/;
4540
4541 sub migrate_from_v0 {
4542 my $git_dir = $ENV{GIT_DIR};
4543 return undef unless -d $git_dir;
4544 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4545 my $migrated = 0;
4546 while (<$fh>) {
4547 chomp;
4548 my ($id, $orig_ref) = ($_, $_);
4549 next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
4550 next unless -f "$git_dir/$id/info/url";
4551 my $new_ref = "refs/remotes/$id";
4552 if (::verify_ref("$new_ref^0")) {
4553 print STDERR "W: $orig_ref is probably an old ",
4554 "branch used by an ancient version of ",
4555 "git-svn.\n",
4556 "However, $new_ref also exists.\n",
4557 "We will not be able ",
4558 "to use this branch until this ",
4559 "ambiguity is resolved.\n";
4560 next;
4561 }
4562 print STDERR "Migrating from v0 layout...\n" if !$migrated;
4563 print STDERR "Renaming ref: $orig_ref => $new_ref\n";
4564 command_noisy('update-ref', $new_ref, $orig_ref);
4565 command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
4566 $migrated++;
4567 }
4568 command_close_pipe($fh, $ctx);
4569 print STDERR "Done migrating from v0 layout...\n" if $migrated;
4570 $migrated;
4571 }
4572
4573 sub migrate_from_v1 {
4574 my $git_dir = $ENV{GIT_DIR};
4575 my $migrated = 0;
4576 return $migrated unless -d $git_dir;
4577 my $svn_dir = "$git_dir/svn";
4578
4579 # just in case somebody used 'svn' as their $id at some point...
4580 return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
4581
4582 print STDERR "Migrating from a git-svn v1 layout...\n";
4583 mkpath([$svn_dir]);
4584 print STDERR "Data from a previous version of git-svn exists, but\n\t",
4585 "$svn_dir\n\t(required for this version ",
4586 "($::VERSION) of git-svn) does not. exist\n";
4587 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4588 while (<$fh>) {
4589 my $x = $_;
4590 next unless $x =~ s#^refs/remotes/##;
4591 chomp $x;
4592 next unless -f "$git_dir/$x/info/url";
4593 my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
4594 next unless $u;
4595 my $dn = dirname("$git_dir/svn/$x");
4596 mkpath([$dn]) unless -d $dn;
4597 if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
4598 mkpath(["$git_dir/svn/svn"]);
4599 print STDERR " - $git_dir/$x/info => ",
4600 "$git_dir/svn/$x/info\n";
4601 rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
4602 croak "$!: $x";
4603 # don't worry too much about these, they probably
4604 # don't exist with repos this old (save for index,
4605 # and we can easily regenerate that)
4606 foreach my $f (qw/unhandled.log index .rev_db/) {
4607 rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
4608 }
4609 } else {
4610 print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
4611 rename "$git_dir/$x", "$git_dir/svn/$x" or
4612 croak "$!: $x";
4613 }
4614 $migrated++;
4615 }
4616 command_close_pipe($fh, $ctx);
4617 print STDERR "Done migrating from a git-svn v1 layout\n";
4618 $migrated;
4619 }
4620
4621 sub read_old_urls {
4622 my ($l_map, $pfx, $path) = @_;
4623 my @dir;
4624 foreach (<$path/*>) {
4625 if (-r "$_/info/url") {
4626 $pfx .= '/' if $pfx && $pfx !~ m!/$!;
4627 my $ref_id = $pfx . basename $_;
4628 my $url = ::file_to_s("$_/info/url");
4629 $l_map->{$ref_id} = $url;
4630 } elsif (-d $_) {
4631 push @dir, $_;
4632 }
4633 }
4634 foreach (@dir) {
4635 my $x = $_;
4636 $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
4637 read_old_urls($l_map, $x, $_);
4638 }
4639 }
4640
4641 sub migrate_from_v2 {
4642 my @cfg = command(qw/config -l/);
4643 return if grep /^svn-remote\..+\.url=/, @cfg;
4644 my %l_map;
4645 read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
4646 my $migrated = 0;
4647
4648 foreach my $ref_id (sort keys %l_map) {
4649 eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
4650 if ($@) {
4651 Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
4652 }
4653 $migrated++;
4654 }
4655 $migrated;
4656 }
4657
4658 sub minimize_connections {
4659 my $r = Git::SVN::read_all_remotes();
4660 my $new_urls = {};
4661 my $root_repos = {};
4662 foreach my $repo_id (keys %$r) {
4663 my $url = $r->{$repo_id}->{url} or next;
4664 my $fetch = $r->{$repo_id}->{fetch} or next;
4665 my $ra = Git::SVN::Ra->new($url);
4666
4667 # skip existing cases where we already connect to the root
4668 if (($ra->{url} eq $ra->{repos_root}) ||
4669 (Git::SVN::sanitize_remote_name($ra->{repos_root}) eq
4670 $repo_id)) {
4671 $root_repos->{$ra->{url}} = $repo_id;
4672 next;
4673 }
4674
4675 my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
4676 my $root_path = $ra->{url};
4677 $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
4678 foreach my $path (keys %$fetch) {
4679 my $ref_id = $fetch->{$path};
4680 my $gs = Git::SVN->new($ref_id, $repo_id, $path);
4681
4682 # make sure we can read when connecting to
4683 # a higher level of a repository
4684 my ($last_rev, undef) = $gs->last_rev_commit;
4685 if (!defined $last_rev) {
4686 $last_rev = eval {
4687 $root_ra->get_latest_revnum;
4688 };
4689 next if $@;
4690 }
4691 my $new = $root_path;
4692 $new .= length $path ? "/$path" : '';
4693 eval {
4694 $root_ra->get_log([$new], $last_rev, $last_rev,
4695 0, 0, 1, sub { });
4696 };
4697 next if $@;
4698 $new_urls->{$ra->{repos_root}}->{$new} =
4699 { ref_id => $ref_id,
4700 old_repo_id => $repo_id,
4701 old_path => $path };
4702 }
4703 }
4704
4705 my @emptied;
4706 foreach my $url (keys %$new_urls) {
4707 # see if we can re-use an existing [svn-remote "repo_id"]
4708 # instead of creating a(n ugly) new section:
4709 my $repo_id = $root_repos->{$url} ||
4710 Git::SVN::sanitize_remote_name($url);
4711
4712 my $fetch = $new_urls->{$url};
4713 foreach my $path (keys %$fetch) {
4714 my $x = $fetch->{$path};
4715 Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
4716 my $pfx = "svn-remote.$x->{old_repo_id}";
4717
4718 my $old_fetch = quotemeta("$x->{old_path}:".
4719 "refs/remotes/$x->{ref_id}");
4720 command_noisy(qw/config --unset/,
4721 "$pfx.fetch", '^'. $old_fetch . '$');
4722 delete $r->{$x->{old_repo_id}}->
4723 {fetch}->{$x->{old_path}};
4724 if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
4725 command_noisy(qw/config --unset/,
4726 "$pfx.url");
4727 push @emptied, $x->{old_repo_id}
4728 }
4729 }
4730 }
4731 if (@emptied) {
4732 my $file = $ENV{GIT_CONFIG} || $ENV{GIT_CONFIG_LOCAL} ||
4733 "$ENV{GIT_DIR}/config";
4734 print STDERR <<EOF;
4735 The following [svn-remote] sections in your config file ($file) are empty
4736 and can be safely removed:
4737 EOF
4738 print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
4739 }
4740 }
4741
4742 sub migration_check {
4743 migrate_from_v0();
4744 migrate_from_v1();
4745 migrate_from_v2();
4746 minimize_connections() if $_minimize;
4747 }
4748
4749 package Git::IndexInfo;
4750 use strict;
4751 use warnings;
4752 use Git qw/command_input_pipe command_close_pipe/;
4753
4754 sub new {
4755 my ($class) = @_;
4756 my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
4757 bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
4758 }
4759
4760 sub remove {
4761 my ($self, $path) = @_;
4762 if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
4763 return ++$self->{nr};
4764 }
4765 undef;
4766 }
4767
4768 sub update {
4769 my ($self, $mode, $hash, $path) = @_;
4770 if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
4771 return ++$self->{nr};
4772 }
4773 undef;
4774 }
4775
4776 sub DESTROY {
4777 my ($self) = @_;
4778 command_close_pipe($self->{gui}, $self->{ctx});
4779 }
4780
4781 package Git::SVN::GlobSpec;
4782 use strict;
4783 use warnings;
4784
4785 sub new {
4786 my ($class, $glob) = @_;
4787 my $re = $glob;
4788 $re =~ s!/+$!!g; # no need for trailing slashes
4789 my $nr = ($re =~ s!^(.*)\*(.*)$!\(\[^/\]+\)!g);
4790 my ($left, $right) = ($1, $2);
4791 if ($nr > 1) {
4792 die "Only one '*' wildcard expansion ",
4793 "is supported (got $nr): '$glob'\n";
4794 } elsif ($nr == 0) {
4795 die "One '*' is needed for glob: '$glob'\n";
4796 }
4797 $re = quotemeta($left) . $re . quotemeta($right);
4798 if (length $left && !($left =~ s!/+$!!g)) {
4799 die "Missing trailing '/' on left side of: '$glob' ($left)\n";
4800 }
4801 if (length $right && !($right =~ s!^/+!!g)) {
4802 die "Missing leading '/' on right side of: '$glob' ($right)\n";
4803 }
4804 my $left_re = qr/^\/\Q$left\E(\/|$)/;
4805 bless { left => $left, right => $right, left_regex => $left_re,
4806 regex => qr/$re/, glob => $glob }, $class;
4807 }
4808
4809 sub full_path {
4810 my ($self, $path) = @_;
4811 return (length $self->{left} ? "$self->{left}/" : '') .
4812 $path . (length $self->{right} ? "/$self->{right}" : '');
4813 }
4814
4815 __END__
4816
4817 Data structures:
4818
4819
4820 $remotes = { # returned by read_all_remotes()
4821 'svn' => {
4822 # svn-remote.svn.url=https://svn.musicpd.org
4823 url => 'https://svn.musicpd.org',
4824 # svn-remote.svn.fetch=mpd/trunk:trunk
4825 fetch => {
4826 'mpd/trunk' => 'trunk',
4827 },
4828 # svn-remote.svn.tags=mpd/tags/*:tags/*
4829 tags => {
4830 path => {
4831 left => 'mpd/tags',
4832 right => '',
4833 regex => qr!mpd/tags/([^/]+)$!,
4834 glob => 'tags/*',
4835 },
4836 ref => {
4837 left => 'tags',
4838 right => '',
4839 regex => qr!tags/([^/]+)$!,
4840 glob => 'tags/*',
4841 },
4842 }
4843 }
4844 };
4845
4846 $log_entry hashref as returned by libsvn_log_entry()
4847 {
4848 log => 'whitespace-formatted log entry
4849 ', # trailing newline is preserved
4850 revision => '8', # integer
4851 date => '2004-02-24T17:01:44.108345Z', # commit date
4852 author => 'committer name'
4853 };
4854
4855
4856 # this is generated by generate_diff();
4857 @mods = array of diff-index line hashes, each element represents one line
4858 of diff-index output
4859
4860 diff-index line ($m hash)
4861 {
4862 mode_a => first column of diff-index output, no leading ':',
4863 mode_b => second column of diff-index output,
4864 sha1_b => sha1sum of the final blob,
4865 chg => change type [MCRADT],
4866 file_a => original file name of a file (iff chg is 'C' or 'R')
4867 file_b => new/current file name of a file (any chg)
4868 }
4869 ;
4870
4871 # retval of read_url_paths{,_all}();
4872 $l_map = {
4873 # repository root url
4874 'https://svn.musicpd.org' => {
4875 # repository path # GIT_SVN_ID
4876 'mpd/trunk' => 'trunk',
4877 'mpd/tags/0.11.5' => 'tags/0.11.5',
4878 },
4879 }
4880
4881 Notes:
4882 I don't trust the each() function on unless I created %hash myself
4883 because the internal iterator may not have started at base.
Dscho's miscellaneous git branches