?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Documentation: help: describe 'man.viewer' config variable
[git/dscho.git] / git-svn.perl
blobd8b38c9a47c0b77dd55781a66e42e4dec26fd4b7
1 #!/usr/bin/env perl
2 # Copyright (C) 2006, Eric Wong <normalperson@yhbt.net>
3 # License: GPL v2 or later
4 use warnings;
5 use strict;
6 use vars qw/ $AUTHOR $VERSION
7 $sha1 $sha1_short $_revision
8 $_q $_authors %users/;
9 $AUTHOR = 'Eric Wong <normalperson@yhbt.net>';
10 $VERSION = '@@GIT_VERSION@@';
11
12 # From which subdir have we been invoked?
13 my $cmd_dir_prefix = eval {
14 command_oneline([qw/rev-parse --show-prefix/], STDERR => 0)
15 } || '';
16
17 my $git_dir_user_set = 1 if defined $ENV{GIT_DIR};
18 $ENV{GIT_DIR} ||= '.git';
19 $Git::SVN::default_repo_id = 'svn';
20 $Git::SVN::default_ref_id = $ENV{GIT_SVN_ID} || 'git-svn';
21 $Git::SVN::Ra::_log_window_size = 100;
22
23 $Git::SVN::Log::TZ = $ENV{TZ};
24 $ENV{TZ} = 'UTC';
25 $| = 1; # unbuffer STDOUT
26
27 sub fatal (@) { print STDERR "@_\n"; exit 1 }
28 require SVN::Core; # use()-ing this causes segfaults for me... *shrug*
29 require SVN::Ra;
30 require SVN::Delta;
31 if ($SVN::Core::VERSION lt '1.1.0') {
32 fatal "Need SVN::Core 1.1.0 or better (got $SVN::Core::VERSION)";
33 }
34 push @Git::SVN::Ra::ISA, 'SVN::Ra';
35 push @SVN::Git::Editor::ISA, 'SVN::Delta::Editor';
36 push @SVN::Git::Fetcher::ISA, 'SVN::Delta::Editor';
37 use Carp qw/croak/;
38 use Digest::MD5;
39 use IO::File qw//;
40 use File::Basename qw/dirname basename/;
41 use File::Path qw/mkpath/;
42 use Getopt::Long qw/:config gnu_getopt no_ignore_case auto_abbrev/;
43 use IPC::Open3;
44 use Git;
45
46 BEGIN {
47 # import functions from Git into our packages, en masse
48 no strict 'refs';
49 foreach (qw/command command_oneline command_noisy command_output_pipe
50 command_input_pipe command_close_pipe/) {
51 for my $package ( qw(SVN::Git::Editor SVN::Git::Fetcher
52 Git::SVN::Migration Git::SVN::Log Git::SVN),
53 __PACKAGE__) {
54 *{"${package}::$_"} = \&{"Git::$_"};
55 }
56 }
57 }
58
59 my ($SVN);
60
61 $sha1 = qr/[a-f\d]{40}/;
62 $sha1_short = qr/[a-f\d]{4,40}/;
63 my ($_stdin, $_help, $_edit,
64 $_message, $_file,
65 $_template, $_shared,
66 $_version, $_fetch_all, $_no_rebase,
67 $_merge, $_strategy, $_dry_run, $_local,
68 $_prefix, $_no_checkout, $_url, $_verbose);
69 $Git::SVN::_follow_parent = 1;
70 my %remote_opts = ( 'username=s' => \$Git::SVN::Prompt::_username,
71 'config-dir=s' => \$Git::SVN::Ra::config_dir,
72 'no-auth-cache' => \$Git::SVN::Prompt::_no_auth_cache );
73 my %fc_opts = ( 'follow-parent|follow!' => \$Git::SVN::_follow_parent,
74 'authors-file|A=s' => \$_authors,
75 'repack:i' => \$Git::SVN::_repack,
76 'noMetadata' => \$Git::SVN::_no_metadata,
77 'useSvmProps' => \$Git::SVN::_use_svm_props,
78 'useSvnsyncProps' => \$Git::SVN::_use_svnsync_props,
79 'log-window-size=i' => \$Git::SVN::Ra::_log_window_size,
80 'no-checkout' => \$_no_checkout,
81 'quiet|q' => \$_q,
82 'repack-flags|repack-args|repack-opts=s' =>
83 \$Git::SVN::_repack_flags,
84 'use-log-author' => \$Git::SVN::_use_log_author,
85 %remote_opts );
86
87 my ($_trunk, $_tags, $_branches, $_stdlayout);
88 my %icv;
89 my %init_opts = ( 'template=s' => \$_template, 'shared:s' => \$_shared,
90 'trunk|T=s' => \$_trunk, 'tags|t=s' => \$_tags,
91 'branches|b=s' => \$_branches, 'prefix=s' => \$_prefix,
92 'stdlayout|s' => \$_stdlayout,
93 'minimize-url|m' => \$Git::SVN::_minimize_url,
94 'no-metadata' => sub { $icv{noMetadata} = 1 },
95 'use-svm-props' => sub { $icv{useSvmProps} = 1 },
96 'use-svnsync-props' => sub { $icv{useSvnsyncProps} = 1 },
97 'rewrite-root=s' => sub { $icv{rewriteRoot} = $_[1] },
98 %remote_opts );
99 my %cmt_opts = ( 'edit|e' => \$_edit,
100 'rmdir' => \$SVN::Git::Editor::_rmdir,
101 'find-copies-harder' => \$SVN::Git::Editor::_find_copies_harder,
102 'l=i' => \$SVN::Git::Editor::_rename_limit,
103 'copy-similarity|C=i'=> \$SVN::Git::Editor::_cp_similarity
104 );
105
106 my %cmd = (
107 fetch => [ \&cmd_fetch, "Download new revisions from SVN",
108 { 'revision|r=s' => \$_revision,
109 'fetch-all|all' => \$_fetch_all,
110 %fc_opts } ],
111 clone => [ \&cmd_clone, "Initialize and fetch revisions",
112 { 'revision|r=s' => \$_revision,
113 %fc_opts, %init_opts } ],
114 init => [ \&cmd_init, "Initialize a repo for tracking" .
115 " (requires URL argument)",
116 \%init_opts ],
117 'multi-init' => [ \&cmd_multi_init,
118 "Deprecated alias for ".
119 "'$0 init -T<trunk> -b<branches> -t<tags>'",
120 \%init_opts ],
121 dcommit => [ \&cmd_dcommit,
122 'Commit several diffs to merge with upstream',
123 { 'merge|m|M' => \$_merge,
124 'strategy|s=s' => \$_strategy,
125 'verbose|v' => \$_verbose,
126 'dry-run|n' => \$_dry_run,
127 'fetch-all|all' => \$_fetch_all,
128 'no-rebase' => \$_no_rebase,
129 %cmt_opts, %fc_opts } ],
130 'set-tree' => [ \&cmd_set_tree,
131 "Set an SVN repository to a git tree-ish",
132 { 'stdin|' => \$_stdin, %cmt_opts, %fc_opts, } ],
133 'create-ignore' => [ \&cmd_create_ignore,
134 'Create a .gitignore per svn:ignore',
135 { 'revision|r=i' => \$_revision
136 } ],
137 'propget' => [ \&cmd_propget,
138 'Print the value of a property on a file or directory',
139 { 'revision|r=i' => \$_revision } ],
140 'proplist' => [ \&cmd_proplist,
141 'List all properties of a file or directory',
142 { 'revision|r=i' => \$_revision } ],
143 'show-ignore' => [ \&cmd_show_ignore, "Show svn:ignore listings",
144 { 'revision|r=i' => \$_revision
145 } ],
146 'show-externals' => [ \&cmd_show_externals, "Show svn:externals listings",
147 { 'revision|r=i' => \$_revision
148 } ],
149 'multi-fetch' => [ \&cmd_multi_fetch,
150 "Deprecated alias for $0 fetch --all",
151 { 'revision|r=s' => \$_revision, %fc_opts } ],
152 'migrate' => [ sub { },
153 # no-op, we automatically run this anyways,
154 'Migrate configuration/metadata/layout from
155 previous versions of git-svn',
156 { 'minimize' => \$Git::SVN::Migration::_minimize,
157 %remote_opts } ],
158 'log' => [ \&Git::SVN::Log::cmd_show_log, 'Show commit logs',
159 { 'limit=i' => \$Git::SVN::Log::limit,
160 'revision|r=s' => \$_revision,
161 'verbose|v' => \$Git::SVN::Log::verbose,
162 'incremental' => \$Git::SVN::Log::incremental,
163 'oneline' => \$Git::SVN::Log::oneline,
164 'show-commit' => \$Git::SVN::Log::show_commit,
165 'non-recursive' => \$Git::SVN::Log::non_recursive,
166 'authors-file|A=s' => \$_authors,
167 'color' => \$Git::SVN::Log::color,
168 'pager=s' => \$Git::SVN::Log::pager
169 } ],
170 'find-rev' => [ \&cmd_find_rev, "Translate between SVN revision numbers and tree-ish",
171 {} ],
172 'rebase' => [ \&cmd_rebase, "Fetch and rebase your working directory",
173 { 'merge|m|M' => \$_merge,
174 'verbose|v' => \$_verbose,
175 'strategy|s=s' => \$_strategy,
176 'local|l' => \$_local,
177 'fetch-all|all' => \$_fetch_all,
178 %fc_opts } ],
179 'commit-diff' => [ \&cmd_commit_diff,
180 'Commit a diff between two trees',
181 { 'message|m=s' => \$_message,
182 'file|F=s' => \$_file,
183 'revision|r=s' => \$_revision,
184 %cmt_opts } ],
185 'info' => [ \&cmd_info,
186 "Show info about the latest SVN revision
187 on the current branch",
188 { 'url' => \$_url, } ],
189 'blame' => [ \&Git::SVN::Log::cmd_blame,
190 "Show what revision and author last modified each line of a file",
191 {} ],
192 );
193
194 my $cmd;
195 for (my $i = 0; $i < @ARGV; $i++) {
196 if (defined $cmd{$ARGV[$i]}) {
197 $cmd = $ARGV[$i];
198 splice @ARGV, $i, 1;
199 last;
200 }
201 };
202
203 # make sure we're always running at the top-level working directory
204 unless ($cmd && $cmd =~ /(?:clone|init|multi-init)$/) {
205 unless (-d $ENV{GIT_DIR}) {
206 if ($git_dir_user_set) {
207 die "GIT_DIR=$ENV{GIT_DIR} explicitly set, ",
208 "but it is not a directory\n";
209 }
210 my $git_dir = delete $ENV{GIT_DIR};
211 chomp(my $cdup = command_oneline(qw/rev-parse --show-cdup/));
212 unless (length $cdup) {
213 die "Already at toplevel, but $git_dir ",
214 "not found '$cdup'\n";
215 }
216 chdir $cdup or die "Unable to chdir up to '$cdup'\n";
217 unless (-d $git_dir) {
218 die "$git_dir still not found after going to ",
219 "'$cdup'\n";
220 }
221 $ENV{GIT_DIR} = $git_dir;
222 }
223 }
224
225 my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
226
227 read_repo_config(\%opts);
228 Getopt::Long::Configure('pass_through') if ($cmd && $cmd eq 'log');
229 my $rv = GetOptions(%opts, 'help|H|h' => \$_help, 'version|V' => \$_version,
230 'minimize-connections' => \$Git::SVN::Migration::_minimize,
231 'id|i=s' => \$Git::SVN::default_ref_id,
232 'svn-remote|remote|R=s' => sub {
233 $Git::SVN::no_reuse_existing = 1;
234 $Git::SVN::default_repo_id = $_[1] });
235 exit 1 if (!$rv && $cmd && $cmd ne 'log');
236
237 usage(0) if $_help;
238 version() if $_version;
239 usage(1) unless defined $cmd;
240 load_authors() if $_authors;
241
242 unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
243 Git::SVN::Migration::migration_check();
244 }
245 Git::SVN::init_vars();
246 eval {
247 Git::SVN::verify_remotes_sanity();
248 $cmd{$cmd}->[0]->(@ARGV);
249 };
250 fatal $@ if $@;
251 post_fetch_checkout();
252 exit 0;
253
254 ####################### primary functions ######################
255 sub usage {
256 my $exit = shift || 0;
257 my $fd = $exit ? \*STDERR : \*STDOUT;
258 print $fd <<"";
259 git-svn - bidirectional operations between a single Subversion tree and git
260 Usage: $0 <command> [options] [arguments]\n
261
262 print $fd "Available commands:\n" unless $cmd;
263
264 foreach (sort keys %cmd) {
265 next if $cmd && $cmd ne $_;
266 next if /^multi-/; # don't show deprecated commands
267 print $fd ' ',pack('A17',$_),$cmd{$_}->[1],"\n";
268 foreach (sort keys %{$cmd{$_}->[2]}) {
269 # mixed-case options are for .git/config only
270 next if /[A-Z]/ && /^[a-z]+$/i;
271 # prints out arguments as they should be passed:
272 my $x = s#[:=]s$## ? '<arg>' : s#[:=]i$## ? '<num>' : '';
273 print $fd ' ' x 21, join(', ', map { length $_ > 1 ?
274 "--$_" : "-$_" }
275 split /\|/,$_)," $x\n";
276 }
277 }
278 print $fd <<"";
279 \nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
280 arbitrary identifier if you're tracking multiple SVN branches/repositories in
281 one git repository and want to keep them separate. See git-svn(1) for more
282 information.
283
284 exit $exit;
285 }
286
287 sub version {
288 print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
289 exit 0;
290 }
291
292 sub do_git_init_db {
293 unless (-d $ENV{GIT_DIR}) {
294 my @init_db = ('init');
295 push @init_db, "--template=$_template" if defined $_template;
296 if (defined $_shared) {
297 if ($_shared =~ /[a-z]/) {
298 push @init_db, "--shared=$_shared";
299 } else {
300 push @init_db, "--shared";
301 }
302 }
303 command_noisy(@init_db);
304 }
305 my $set;
306 my $pfx = "svn-remote.$Git::SVN::default_repo_id";
307 foreach my $i (keys %icv) {
308 die "'$set' and '$i' cannot both be set\n" if $set;
309 next unless defined $icv{$i};
310 command_noisy('config', "$pfx.$i", $icv{$i});
311 $set = $i;
312 }
313 }
314
315 sub init_subdir {
316 my $repo_path = shift or return;
317 mkpath([$repo_path]) unless -d $repo_path;
318 chdir $repo_path or die "Couldn't chdir to $repo_path: $!\n";
319 $ENV{GIT_DIR} = '.git';
320 }
321
322 sub cmd_clone {
323 my ($url, $path) = @_;
324 if (!defined $path &&
325 (defined $_trunk || defined $_branches || defined $_tags ||
326 defined $_stdlayout) &&
327 $url !~ m#^[a-z\+]+://#) {
328 $path = $url;
329 }
330 $path = basename($url) if !defined $path || !length $path;
331 cmd_init($url, $path);
332 Git::SVN::fetch_all($Git::SVN::default_repo_id);
333 }
334
335 sub cmd_init {
336 if (defined $_stdlayout) {
337 $_trunk = 'trunk' if (!defined $_trunk);
338 $_tags = 'tags' if (!defined $_tags);
339 $_branches = 'branches' if (!defined $_branches);
340 }
341 if (defined $_trunk || defined $_branches || defined $_tags) {
342 return cmd_multi_init(@_);
343 }
344 my $url = shift or die "SVN repository location required ",
345 "as a command-line argument\n";
346 init_subdir(@_);
347 do_git_init_db();
348
349 Git::SVN->init($url);
350 }
351
352 sub cmd_fetch {
353 if (grep /^\d+=./, @_) {
354 die "'<rev>=<commit>' fetch arguments are ",
355 "no longer supported.\n";
356 }
357 my ($remote) = @_;
358 if (@_ > 1) {
359 die "Usage: $0 fetch [--all] [svn-remote]\n";
360 }
361 $remote ||= $Git::SVN::default_repo_id;
362 if ($_fetch_all) {
363 cmd_multi_fetch();
364 } else {
365 Git::SVN::fetch_all($remote, Git::SVN::read_all_remotes());
366 }
367 }
368
369 sub cmd_set_tree {
370 my (@commits) = @_;
371 if ($_stdin || !@commits) {
372 print "Reading from stdin...\n";
373 @commits = ();
374 while (<STDIN>) {
375 if (/\b($sha1_short)\b/o) {
376 unshift @commits, $1;
377 }
378 }
379 }
380 my @revs;
381 foreach my $c (@commits) {
382 my @tmp = command('rev-parse',$c);
383 if (scalar @tmp == 1) {
384 push @revs, $tmp[0];
385 } elsif (scalar @tmp > 1) {
386 push @revs, reverse(command('rev-list',@tmp));
387 } else {
388 fatal "Failed to rev-parse $c";
389 }
390 }
391 my $gs = Git::SVN->new;
392 my ($r_last, $cmt_last) = $gs->last_rev_commit;
393 $gs->fetch;
394 if (defined $gs->{last_rev} && $r_last != $gs->{last_rev}) {
395 fatal "There are new revisions that were fetched ",
396 "and need to be merged (or acknowledged) ",
397 "before committing.\nlast rev: $r_last\n",
398 " current: $gs->{last_rev}";
399 }
400 $gs->set_tree($_) foreach @revs;
401 print "Done committing ",scalar @revs," revisions to SVN\n";
402 unlink $gs->{index};
403 }
404
405 sub cmd_dcommit {
406 my $head = shift;
407 git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
408 'Cannot dcommit with a dirty index. Commit your changes first, '
409 . "or stash them with `git stash'.\n";
410 $head ||= 'HEAD';
411 my @refs;
412 my ($url, $rev, $uuid, $gs) = working_head_info($head, \@refs);
413 print "Committing to $url ...\n";
414 unless ($gs) {
415 die "Unable to determine upstream SVN information from ",
416 "$head history\n";
417 }
418 my $last_rev;
419 my ($linear_refs, $parents) = linearize_history($gs, \@refs);
420 if ($_no_rebase && scalar(@$linear_refs) > 1) {
421 warn "Attempting to commit more than one change while ",
422 "--no-rebase is enabled.\n",
423 "If these changes depend on each other, re-running ",
424 "without --no-rebase may be required."
425 }
426 while (1) {
427 my $d = shift @$linear_refs or last;
428 unless (defined $last_rev) {
429 (undef, $last_rev, undef) = cmt_metadata("$d~1");
430 unless (defined $last_rev) {
431 fatal "Unable to extract revision information ",
432 "from commit $d~1";
433 }
434 }
435 if ($_dry_run) {
436 print "diff-tree $d~1 $d\n";
437 } else {
438 my $cmt_rev;
439 my %ed_opts = ( r => $last_rev,
440 log => get_commit_entry($d)->{log},
441 ra => Git::SVN::Ra->new($gs->full_url),
442 config => SVN::Core::config_get_config(
443 $Git::SVN::Ra::config_dir
444 ),
445 tree_a => "$d~1",
446 tree_b => $d,
447 editor_cb => sub {
448 print "Committed r$_[0]\n";
449 $cmt_rev = $_[0];
450 },
451 svn_path => '');
452 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
453 print "No changes\n$d~1 == $d\n";
454 } elsif ($parents->{$d} && @{$parents->{$d}}) {
455 $gs->{inject_parents_dcommit}->{$cmt_rev} =
456 $parents->{$d};
457 }
458 $_fetch_all ? $gs->fetch_all : $gs->fetch;
459 $last_rev = $cmt_rev;
460 next if $_no_rebase;
461
462 # we always want to rebase against the current HEAD,
463 # not any head that was passed to us
464 my @diff = command('diff-tree', $d,
465 $gs->refname, '--');
466 my @finish;
467 if (@diff) {
468 @finish = rebase_cmd();
469 print STDERR "W: $d and ", $gs->refname,
470 " differ, using @finish:\n",
471 join("\n", @diff), "\n";
472 } else {
473 print "No changes between current HEAD and ",
474 $gs->refname,
475 "\nResetting to the latest ",
476 $gs->refname, "\n";
477 @finish = qw/reset --mixed/;
478 }
479 command_noisy(@finish, $gs->refname);
480 if (@diff) {
481 @refs = ();
482 my ($url_, $rev_, $uuid_, $gs_) =
483 working_head_info($head, \@refs);
484 my ($linear_refs_, $parents_) =
485 linearize_history($gs_, \@refs);
486 if (scalar(@$linear_refs) !=
487 scalar(@$linear_refs_)) {
488 fatal "# of revisions changed ",
489 "\nbefore:\n",
490 join("\n", @$linear_refs),
491 "\n\nafter:\n",
492 join("\n", @$linear_refs_), "\n",
493 'If you are attempting to commit ',
494 "merges, try running:\n\t",
495 'git rebase --interactive',
496 '--preserve-merges ',
497 $gs->refname,
498 "\nBefore dcommitting";
499 }
500 if ($url_ ne $url) {
501 fatal "URL mismatch after rebase: ",
502 "$url_ != $url";
503 }
504 if ($uuid_ ne $uuid) {
505 fatal "uuid mismatch after rebase: ",
506 "$uuid_ != $uuid";
507 }
508 # remap parents
509 my (%p, @l, $i);
510 for ($i = 0; $i < scalar @$linear_refs; $i++) {
511 my $new = $linear_refs_->[$i] or next;
512 $p{$new} =
513 $parents->{$linear_refs->[$i]};
514 push @l, $new;
515 }
516 $parents = \%p;
517 $linear_refs = \@l;
518 }
519 }
520 }
521 unlink $gs->{index};
522 }
523
524 sub cmd_find_rev {
525 my $revision_or_hash = shift or die "SVN or git revision required ",
526 "as a command-line argument\n";
527 my $result;
528 if ($revision_or_hash =~ /^r\d+$/) {
529 my $head = shift;
530 $head ||= 'HEAD';
531 my @refs;
532 my (undef, undef, undef, $gs) = working_head_info($head, \@refs);
533 unless ($gs) {
534 die "Unable to determine upstream SVN information from ",
535 "$head history\n";
536 }
537 my $desired_revision = substr($revision_or_hash, 1);
538 $result = $gs->rev_map_get($desired_revision);
539 } else {
540 my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
541 $result = $rev;
542 }
543 print "$result\n" if $result;
544 }
545
546 sub cmd_rebase {
547 command_noisy(qw/update-index --refresh/);
548 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
549 unless ($gs) {
550 die "Unable to determine upstream SVN information from ",
551 "working tree history\n";
552 }
553 if (command(qw/diff-index HEAD --/)) {
554 print STDERR "Cannot rebase with uncommited changes:\n";
555 command_noisy('status');
556 exit 1;
557 }
558 unless ($_local) {
559 # rebase will checkout for us, so no need to do it explicitly
560 $_no_checkout = 'true';
561 $_fetch_all ? $gs->fetch_all : $gs->fetch;
562 }
563 command_noisy(rebase_cmd(), $gs->refname);
564 }
565
566 sub cmd_show_ignore {
567 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
568 $gs ||= Git::SVN->new;
569 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
570 $gs->prop_walk($gs->{path}, $r, sub {
571 my ($gs, $path, $props) = @_;
572 print STDOUT "\n# $path\n";
573 my $s = $props->{'svn:ignore'} or return;
574 $s =~ s/[\r\n]+/\n/g;
575 chomp $s;
576 $s =~ s#^#$path#gm;
577 print STDOUT "$s\n";
578 });
579 }
580
581 sub cmd_show_externals {
582 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
583 $gs ||= Git::SVN->new;
584 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
585 $gs->prop_walk($gs->{path}, $r, sub {
586 my ($gs, $path, $props) = @_;
587 print STDOUT "\n# $path\n";
588 my $s = $props->{'svn:externals'} or return;
589 $s =~ s/[\r\n]+/\n/g;
590 chomp $s;
591 $s =~ s#^#$path#gm;
592 print STDOUT "$s\n";
593 });
594 }
595
596 sub cmd_create_ignore {
597 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
598 $gs ||= Git::SVN->new;
599 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
600 $gs->prop_walk($gs->{path}, $r, sub {
601 my ($gs, $path, $props) = @_;
602 # $path is of the form /path/to/dir/
603 my $ignore = '.' . $path . '.gitignore';
604 my $s = $props->{'svn:ignore'} or return;
605 open(GITIGNORE, '>', $ignore)
606 or fatal("Failed to open `$ignore' for writing: $!");
607 $s =~ s/[\r\n]+/\n/g;
608 chomp $s;
609 # Prefix all patterns so that the ignore doesn't apply
610 # to sub-directories.
611 $s =~ s#^#/#gm;
612 print GITIGNORE "$s\n";
613 close(GITIGNORE)
614 or fatal("Failed to close `$ignore': $!");
615 command_noisy('add', $ignore);
616 });
617 }
618
619 sub canonicalize_path {
620 my ($path) = @_;
621 my $dot_slash_added = 0;
622 if (substr($path, 0, 1) ne "/") {
623 $path = "./" . $path;
624 $dot_slash_added = 1;
625 }
626 # File::Spec->canonpath doesn't collapse x/../y into y (for a
627 # good reason), so let's do this manually.
628 $path =~ s#/+#/#g;
629 $path =~ s#/\.(?:/|$)#/#g;
630 $path =~ s#/[^/]+/\.\.##g;
631 $path =~ s#/$##g;
632 $path =~ s#^\./## if $dot_slash_added;
633 return $path;
634 }
635
636 # get_svnprops(PATH)
637 # ------------------
638 # Helper for cmd_propget and cmd_proplist below.
639 sub get_svnprops {
640 my $path = shift;
641 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
642 $gs ||= Git::SVN->new;
643
644 # prefix THE PATH by the sub-directory from which the user
645 # invoked us.
646 $path = $cmd_dir_prefix . $path;
647 fatal("No such file or directory: $path") unless -e $path;
648 my $is_dir = -d $path ? 1 : 0;
649 $path = $gs->{path} . '/' . $path;
650
651 # canonicalize the path (otherwise libsvn will abort or fail to
652 # find the file)
653 $path = canonicalize_path($path);
654
655 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
656 my $props;
657 if ($is_dir) {
658 (undef, undef, $props) = $gs->ra->get_dir($path, $r);
659 }
660 else {
661 (undef, $props) = $gs->ra->get_file($path, $r, undef);
662 }
663 return $props;
664 }
665
666 # cmd_propget (PROP, PATH)
667 # ------------------------
668 # Print the SVN property PROP for PATH.
669 sub cmd_propget {
670 my ($prop, $path) = @_;
671 $path = '.' if not defined $path;
672 usage(1) if not defined $prop;
673 my $props = get_svnprops($path);
674 if (not defined $props->{$prop}) {
675 fatal("`$path' does not have a `$prop' SVN property.");
676 }
677 print $props->{$prop} . "\n";
678 }
679
680 # cmd_proplist (PATH)
681 # -------------------
682 # Print the list of SVN properties for PATH.
683 sub cmd_proplist {
684 my $path = shift;
685 $path = '.' if not defined $path;
686 my $props = get_svnprops($path);
687 print "Properties on '$path':\n";
688 foreach (sort keys %{$props}) {
689 print " $_\n";
690 }
691 }
692
693 sub cmd_multi_init {
694 my $url = shift;
695 unless (defined $_trunk || defined $_branches || defined $_tags) {
696 usage(1);
697 }
698
699 # there are currently some bugs that prevent multi-init/multi-fetch
700 # setups from working well without this.
701 $Git::SVN::_minimize_url = 1;
702
703 $_prefix = '' unless defined $_prefix;
704 if (defined $url) {
705 $url =~ s#/+$##;
706 init_subdir(@_);
707 }
708 do_git_init_db();
709 if (defined $_trunk) {
710 my $trunk_ref = $_prefix . 'trunk';
711 # try both old-style and new-style lookups:
712 my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
713 unless ($gs_trunk) {
714 my ($trunk_url, $trunk_path) =
715 complete_svn_url($url, $_trunk);
716 $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
717 undef, $trunk_ref);
718 }
719 }
720 return unless defined $_branches || defined $_tags;
721 my $ra = $url ? Git::SVN::Ra->new($url) : undef;
722 complete_url_ls_init($ra, $_branches, '--branches/-b', $_prefix);
723 complete_url_ls_init($ra, $_tags, '--tags/-t', $_prefix . 'tags/');
724 }
725
726 sub cmd_multi_fetch {
727 my $remotes = Git::SVN::read_all_remotes();
728 foreach my $repo_id (sort keys %$remotes) {
729 if ($remotes->{$repo_id}->{url}) {
730 Git::SVN::fetch_all($repo_id, $remotes);
731 }
732 }
733 }
734
735 # this command is special because it requires no metadata
736 sub cmd_commit_diff {
737 my ($ta, $tb, $url) = @_;
738 my $usage = "Usage: $0 commit-diff -r<revision> ".
739 "<tree-ish> <tree-ish> [<URL>]";
740 fatal($usage) if (!defined $ta || !defined $tb);
741 my $svn_path;
742 if (!defined $url) {
743 my $gs = eval { Git::SVN->new };
744 if (!$gs) {
745 fatal("Needed URL or usable git-svn --id in ",
746 "the command-line\n", $usage);
747 }
748 $url = $gs->{url};
749 $svn_path = $gs->{path};
750 }
751 unless (defined $_revision) {
752 fatal("-r|--revision is a required argument\n", $usage);
753 }
754 if (defined $_message && defined $_file) {
755 fatal("Both --message/-m and --file/-F specified ",
756 "for the commit message.\n",
757 "I have no idea what you mean");
758 }
759 if (defined $_file) {
760 $_message = file_to_s($_file);
761 } else {
762 $_message ||= get_commit_entry($tb)->{log};
763 }
764 my $ra ||= Git::SVN::Ra->new($url);
765 $svn_path ||= $ra->{svn_path};
766 my $r = $_revision;
767 if ($r eq 'HEAD') {
768 $r = $ra->get_latest_revnum;
769 } elsif ($r !~ /^\d+$/) {
770 die "revision argument: $r not understood by git-svn\n";
771 }
772 my %ed_opts = ( r => $r,
773 log => $_message,
774 ra => $ra,
775 tree_a => $ta,
776 tree_b => $tb,
777 editor_cb => sub { print "Committed r$_[0]\n" },
778 svn_path => $svn_path );
779 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
780 print "No changes\n$ta == $tb\n";
781 }
782 }
783
784 sub cmd_info {
785 my $path = canonicalize_path(shift or ".");
786 unless (scalar(@_) == 0) {
787 die "Too many arguments specified\n";
788 }
789
790 my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
791
792 if (!$file_type && !$diff_status) {
793 print STDERR "$path: (Not a versioned resource)\n\n";
794 return;
795 }
796
797 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
798 unless ($gs) {
799 die "Unable to determine upstream SVN information from ",
800 "working tree history\n";
801 }
802 my $full_url = $url . ($path eq "." ? "" : "/$path");
803
804 if ($_url) {
805 print $full_url, "\n";
806 return;
807 }
808
809 my $result = "Path: $path\n";
810 $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
811 $result .= "URL: " . $full_url . "\n";
812
813 eval {
814 my $repos_root = $gs->repos_root;
815 Git::SVN::remove_username($repos_root);
816 $result .= "Repository Root: $repos_root\n";
817 };
818 if ($@) {
819 $result .= "Repository Root: (offline)\n";
820 }
821 $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A";
822 $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
823
824 $result .= "Node Kind: " .
825 ($file_type eq "dir" ? "directory" : "file") . "\n";
826
827 my $schedule = $diff_status eq "A"
828 ? "add"
829 : ($diff_status eq "D" ? "delete" : "normal");
830 $result .= "Schedule: $schedule\n";
831
832 if ($diff_status eq "A") {
833 print $result, "\n";
834 return;
835 }
836
837 my ($lc_author, $lc_rev, $lc_date_utc);
838 my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $path);
839 my $log = command_output_pipe(@args);
840 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
841 while (<$log>) {
842 if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
843 $lc_author = $1;
844 $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
845 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
846 (undef, $lc_rev, undef) = ::extract_metadata($1);
847 }
848 }
849 close $log;
850
851 Git::SVN::Log::set_local_timezone();
852
853 $result .= "Last Changed Author: $lc_author\n";
854 $result .= "Last Changed Rev: $lc_rev\n";
855 $result .= "Last Changed Date: " .
856 Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
857
858 if ($file_type ne "dir") {
859 my $text_last_updated_date =
860 ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
861 $result .=
862 "Text Last Updated: " .
863 Git::SVN::Log::format_svn_date($text_last_updated_date) .
864 "\n";
865 my $checksum;
866 if ($diff_status eq "D") {
867 my ($fh, $ctx) =
868 command_output_pipe(qw(cat-file blob), "HEAD:$path");
869 if ($file_type eq "link") {
870 my $file_name = <$fh>;
871 $checksum = md5sum("link $file_name");
872 } else {
873 $checksum = md5sum($fh);
874 }
875 command_close_pipe($fh, $ctx);
876 } elsif ($file_type eq "link") {
877 my $file_name =
878 command(qw(cat-file blob), "HEAD:$path");
879 $checksum =
880 md5sum("link " . $file_name);
881 } else {
882 open FILE, "<", $path or die $!;
883 $checksum = md5sum(\*FILE);
884 close FILE or die $!;
885 }
886 $result .= "Checksum: " . $checksum . "\n";
887 }
888
889 print $result, "\n";
890 }
891
892 ########################### utility functions #########################
893
894 sub rebase_cmd {
895 my @cmd = qw/rebase/;
896 push @cmd, '-v' if $_verbose;
897 push @cmd, qw/--merge/ if $_merge;
898 push @cmd, "--strategy=$_strategy" if $_strategy;
899 @cmd;
900 }
901
902 sub post_fetch_checkout {
903 return if $_no_checkout;
904 my $gs = $Git::SVN::_head or return;
905 return if verify_ref('refs/heads/master^0');
906
907 my $valid_head = verify_ref('HEAD^0');
908 command_noisy(qw(update-ref refs/heads/master), $gs->refname);
909 return if ($valid_head || !verify_ref('HEAD^0'));
910
911 return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
912 my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
913 return if -f $index;
914
915 return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
916 return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
917 command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
918 print STDERR "Checked out HEAD:\n ",
919 $gs->full_url, " r", $gs->last_rev, "\n";
920 }
921
922 sub complete_svn_url {
923 my ($url, $path) = @_;
924 $path =~ s#/+$##;
925 if ($path !~ m#^[a-z\+]+://#) {
926 if (!defined $url || $url !~ m#^[a-z\+]+://#) {
927 fatal("E: '$path' is not a complete URL ",
928 "and a separate URL is not specified");
929 }
930 return ($url, $path);
931 }
932 return ($path, '');
933 }
934
935 sub complete_url_ls_init {
936 my ($ra, $repo_path, $switch, $pfx) = @_;
937 unless ($repo_path) {
938 print STDERR "W: $switch not specified\n";
939 return;
940 }
941 $repo_path =~ s#/+$##;
942 if ($repo_path =~ m#^[a-z\+]+://#) {
943 $ra = Git::SVN::Ra->new($repo_path);
944 $repo_path = '';
945 } else {
946 $repo_path =~ s#^/+##;
947 unless ($ra) {
948 fatal("E: '$repo_path' is not a complete URL ",
949 "and a separate URL is not specified");
950 }
951 }
952 my $url = $ra->{url};
953 my $gs = Git::SVN->init($url, undef, undef, undef, 1);
954 my $k = "svn-remote.$gs->{repo_id}.url";
955 my $orig_url = eval { command_oneline(qw/config --get/, $k) };
956 if ($orig_url && ($orig_url ne $gs->{url})) {
957 die "$k already set: $orig_url\n",
958 "wanted to set to: $gs->{url}\n";
959 }
960 command_oneline('config', $k, $gs->{url}) unless $orig_url;
961 my $remote_path = "$ra->{svn_path}/$repo_path/*";
962 $remote_path =~ s#/+#/#g;
963 $remote_path =~ s#^/##g;
964 my ($n) = ($switch =~ /^--(\w+)/);
965 if (length $pfx && $pfx !~ m#/$#) {
966 die "--prefix='$pfx' must have a trailing slash '/'\n";
967 }
968 command_noisy('config', "svn-remote.$gs->{repo_id}.$n",
969 "$remote_path:refs/remotes/$pfx*");
970 }
971
972 sub verify_ref {
973 my ($ref) = @_;
974 eval { command_oneline([ 'rev-parse', '--verify', $ref ],
975 { STDERR => 0 }); };
976 }
977
978 sub get_tree_from_treeish {
979 my ($treeish) = @_;
980 # $treeish can be a symbolic ref, too:
981 my $type = command_oneline(qw/cat-file -t/, $treeish);
982 my $expected;
983 while ($type eq 'tag') {
984 ($treeish, $type) = command(qw/cat-file tag/, $treeish);
985 }
986 if ($type eq 'commit') {
987 $expected = (grep /^tree /, command(qw/cat-file commit/,
988 $treeish))[0];
989 ($expected) = ($expected =~ /^tree ($sha1)$/o);
990 die "Unable to get tree from $treeish\n" unless $expected;
991 } elsif ($type eq 'tree') {
992 $expected = $treeish;
993 } else {
994 die "$treeish is a $type, expected tree, tag or commit\n";
995 }
996 return $expected;
997 }
998
999 sub get_commit_entry {
1000 my ($treeish) = shift;
1001 my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
1002 my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
1003 my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
1004 open my $log_fh, '>', $commit_editmsg or croak $!;
1005
1006 my $type = command_oneline(qw/cat-file -t/, $treeish);
1007 if ($type eq 'commit' || $type eq 'tag') {
1008 my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1009 $type, $treeish);
1010 my $in_msg = 0;
1011 while (<$msg_fh>) {
1012 if (!$in_msg) {
1013 $in_msg = 1 if (/^\s*$/);
1014 } elsif (/^git-svn-id: /) {
1015 # skip this for now, we regenerate the
1016 # correct one on re-fetch anyways
1017 # TODO: set *:merge properties or like...
1018 } else {
1019 print $log_fh $_ or croak $!;
1020 }
1021 }
1022 command_close_pipe($msg_fh, $ctx);
1023 }
1024 close $log_fh or croak $!;
1025
1026 if ($_edit || ($type eq 'tree')) {
1027 my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
1028 # TODO: strip out spaces, comments, like git-commit.sh
1029 system($editor, $commit_editmsg);
1030 }
1031 rename $commit_editmsg, $commit_msg or croak $!;
1032 open $log_fh, '<', $commit_msg or croak $!;
1033 { local $/; chomp($log_entry{log} = <$log_fh>); }
1034 close $log_fh or croak $!;
1035 unlink $commit_msg;
1036 \%log_entry;
1037 }
1038
1039 sub s_to_file {
1040 my ($str, $file, $mode) = @_;
1041 open my $fd,'>',$file or croak $!;
1042 print $fd $str,"\n" or croak $!;
1043 close $fd or croak $!;
1044 chmod ($mode &~ umask, $file) if (defined $mode);
1045 }
1046
1047 sub file_to_s {
1048 my $file = shift;
1049 open my $fd,'<',$file or croak "$!: file: $file\n";
1050 local $/;
1051 my $ret = <$fd>;
1052 close $fd or croak $!;
1053 $ret =~ s/\s*$//s;
1054 return $ret;
1055 }
1056
1057 # '<svn username> = real-name <email address>' mapping based on git-svnimport:
1058 sub load_authors {
1059 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1060 my $log = $cmd eq 'log';
1061 while (<$authors>) {
1062 chomp;
1063 next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1064 my ($user, $name, $email) = ($1, $2, $3);
1065 if ($log) {
1066 $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1067 } else {
1068 $users{$user} = [$name, $email];
1069 }
1070 }
1071 close $authors or croak $!;
1072 }
1073
1074 # convert GetOpt::Long specs for use by git-config
1075 sub read_repo_config {
1076 return unless -d $ENV{GIT_DIR};
1077 my $opts = shift;
1078 my @config_only;
1079 foreach my $o (keys %$opts) {
1080 # if we have mixedCase and a long option-only, then
1081 # it's a config-only variable that we don't need for
1082 # the command-line.
1083 push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1084 my $v = $opts->{$o};
1085 my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1086 $key =~ s/-//g;
1087 my $arg = 'git-config';
1088 $arg .= ' --int' if ($o =~ /[:=]i$/);
1089 $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1090 if (ref $v eq 'ARRAY') {
1091 chomp(my @tmp = `$arg --get-all svn.$key`);
1092 @$v = @tmp if @tmp;
1093 } else {
1094 chomp(my $tmp = `$arg --get svn.$key`);
1095 if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1096 $$v = $tmp;
1097 }
1098 }
1099 }
1100 delete @$opts{@config_only} if @config_only;
1101 }
1102
1103 sub extract_metadata {
1104 my $id = shift or return (undef, undef, undef);
1105 my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1106 \s([a-f\d\-]+)$/x);
1107 if (!defined $rev || !$uuid || !$url) {
1108 # some of the original repositories I made had
1109 # identifiers like this:
1110 ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/);
1111 }
1112 return ($url, $rev, $uuid);
1113 }
1114
1115 sub cmt_metadata {
1116 return extract_metadata((grep(/^git-svn-id: /,
1117 command(qw/cat-file commit/, shift)))[-1]);
1118 }
1119
1120 sub working_head_info {
1121 my ($head, $refs) = @_;
1122 my @args = ('log', '--no-color', '--first-parent');
1123 my ($fh, $ctx) = command_output_pipe(@args, $head);
1124 my $hash;
1125 my %max;
1126 while (<$fh>) {
1127 if ( m{^commit ($::sha1)$} ) {
1128 unshift @$refs, $hash if $hash and $refs;
1129 $hash = $1;
1130 next;
1131 }
1132 next unless s{^\s*(git-svn-id:)}{$1};
1133 my ($url, $rev, $uuid) = extract_metadata($_);
1134 if (defined $url && defined $rev) {
1135 next if $max{$url} and $max{$url} < $rev;
1136 if (my $gs = Git::SVN->find_by_url($url)) {
1137 my $c = $gs->rev_map_get($rev);
1138 if ($c && $c eq $hash) {
1139 close $fh; # break the pipe
1140 return ($url, $rev, $uuid, $gs);
1141 } else {
1142 $max{$url} ||= $gs->rev_map_max;
1143 }
1144 }
1145 }
1146 }
1147 command_close_pipe($fh, $ctx);
1148 (undef, undef, undef, undef);
1149 }
1150
1151 sub read_commit_parents {
1152 my ($parents, $c) = @_;
1153 chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1154 $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1155 @{$parents->{$c}} = split(/ /, $p);
1156 }
1157
1158 sub linearize_history {
1159 my ($gs, $refs) = @_;
1160 my %parents;
1161 foreach my $c (@$refs) {
1162 read_commit_parents(\%parents, $c);
1163 }
1164
1165 my @linear_refs;
1166 my %skip = ();
1167 my $last_svn_commit = $gs->last_commit;
1168 foreach my $c (reverse @$refs) {
1169 next if $c eq $last_svn_commit;
1170 last if $skip{$c};
1171
1172 unshift @linear_refs, $c;
1173 $skip{$c} = 1;
1174
1175 # we only want the first parent to diff against for linear
1176 # history, we save the rest to inject when we finalize the
1177 # svn commit
1178 my $fp_a = verify_ref("$c~1");
1179 my $fp_b = shift @{$parents{$c}} if $parents{$c};
1180 if (!$fp_a || !$fp_b) {
1181 die "Commit $c\n",
1182 "has no parent commit, and therefore ",
1183 "nothing to diff against.\n",
1184 "You should be working from a repository ",
1185 "originally created by git-svn\n";
1186 }
1187 if ($fp_a ne $fp_b) {
1188 die "$c~1 = $fp_a, however parsing commit $c ",
1189 "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1190 }
1191
1192 foreach my $p (@{$parents{$c}}) {
1193 $skip{$p} = 1;
1194 }
1195 }
1196 (\@linear_refs, \%parents);
1197 }
1198
1199 sub find_file_type_and_diff_status {
1200 my ($path) = @_;
1201 return ('dir', '') if $path eq '.';
1202
1203 my $diff_output =
1204 command_oneline(qw(diff --cached --name-status --), $path) || "";
1205 my $diff_status = (split(' ', $diff_output))[0] || "";
1206
1207 my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1208
1209 return (undef, undef) if !$diff_status && !$ls_tree;
1210
1211 if ($diff_status eq "A") {
1212 return ("link", $diff_status) if -l $path;
1213 return ("dir", $diff_status) if -d $path;
1214 return ("file", $diff_status);
1215 }
1216
1217 my $mode = (split(' ', $ls_tree))[0] || "";
1218
1219 return ("link", $diff_status) if $mode eq "120000";
1220 return ("dir", $diff_status) if $mode eq "040000";
1221 return ("file", $diff_status);
1222 }
1223
1224 sub md5sum {
1225 my $arg = shift;
1226 my $ref = ref $arg;
1227 my $md5 = Digest::MD5->new();
1228 if ($ref eq 'GLOB' || $ref eq 'IO::File') {
1229 $md5->addfile($arg) or croak $!;
1230 } elsif ($ref eq 'SCALAR') {
1231 $md5->add($$arg) or croak $!;
1232 } elsif (!$ref) {
1233 $md5->add($arg) or croak $!;
1234 } else {
1235 ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
1236 }
1237 return $md5->hexdigest();
1238 }
1239
1240 package Git::SVN;
1241 use strict;
1242 use warnings;
1243 use Fcntl qw/:DEFAULT :seek/;
1244 use constant rev_map_fmt => 'NH40';
1245 use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1246 $_repack $_repack_flags $_use_svm_props $_head
1247 $_use_svnsync_props $no_reuse_existing $_minimize_url
1248 $_use_log_author/;
1249 use Carp qw/croak/;
1250 use File::Path qw/mkpath/;
1251 use File::Copy qw/copy/;
1252 use IPC::Open3;
1253
1254 my ($_gc_nr, $_gc_period);
1255
1256 # properties that we do not log:
1257 my %SKIP_PROP;
1258 BEGIN {
1259 %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1260 svn:special svn:executable
1261 svn:entry:committed-rev
1262 svn:entry:last-author
1263 svn:entry:uuid
1264 svn:entry:committed-date/;
1265
1266 # some options are read globally, but can be overridden locally
1267 # per [svn-remote "..."] section. Command-line options will *NOT*
1268 # override options set in an [svn-remote "..."] section
1269 no strict 'refs';
1270 for my $option (qw/follow_parent no_metadata use_svm_props
1271 use_svnsync_props/) {
1272 my $key = $option;
1273 $key =~ tr/_//d;
1274 my $prop = "-$option";
1275 *$option = sub {
1276 my ($self) = @_;
1277 return $self->{$prop} if exists $self->{$prop};
1278 my $k = "svn-remote.$self->{repo_id}.$key";
1279 eval { command_oneline(qw/config --get/, $k) };
1280 if ($@) {
1281 $self->{$prop} = ${"Git::SVN::_$option"};
1282 } else {
1283 my $v = command_oneline(qw/config --bool/,$k);
1284 $self->{$prop} = $v eq 'false' ? 0 : 1;
1285 }
1286 return $self->{$prop};
1287 }
1288 }
1289 }
1290
1291 my (%LOCKFILES, %INDEX_FILES);
1292 END {
1293 unlink keys %LOCKFILES if %LOCKFILES;
1294 unlink keys %INDEX_FILES if %INDEX_FILES;
1295 }
1296
1297 sub resolve_local_globs {
1298 my ($url, $fetch, $glob_spec) = @_;
1299 return unless defined $glob_spec;
1300 my $ref = $glob_spec->{ref};
1301 my $path = $glob_spec->{path};
1302 foreach (command(qw#for-each-ref --format=%(refname) refs/remotes#)) {
1303 next unless m#^refs/remotes/$ref->{regex}$#;
1304 my $p = $1;
1305 my $pathname = desanitize_refname($path->full_path($p));
1306 my $refname = desanitize_refname($ref->full_path($p));
1307 if (my $existing = $fetch->{$pathname}) {
1308 if ($existing ne $refname) {
1309 die "Refspec conflict:\n",
1310 "existing: refs/remotes/$existing\n",
1311 " globbed: refs/remotes/$refname\n";
1312 }
1313 my $u = (::cmt_metadata("refs/remotes/$refname"))[0];
1314 $u =~ s!^\Q$url\E(/|$)!! or die
1315 "refs/remotes/$refname: '$url' not found in '$u'\n";
1316 if ($pathname ne $u) {
1317 warn "W: Refspec glob conflict ",
1318 "(ref: refs/remotes/$refname):\n",
1319 "expected path: $pathname\n",
1320 " real path: $u\n",
1321 "Continuing ahead with $u\n";
1322 next;
1323 }
1324 } else {
1325 $fetch->{$pathname} = $refname;
1326 }
1327 }
1328 }
1329
1330 sub parse_revision_argument {
1331 my ($base, $head) = @_;
1332 if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1333 return ($base, $head);
1334 }
1335 return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1336 return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1337 return ($head, $head) if ($::_revision eq 'HEAD');
1338 return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1339 return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1340 die "revision argument: $::_revision not understood by git-svn\n";
1341 }
1342
1343 sub fetch_all {
1344 my ($repo_id, $remotes) = @_;
1345 if (ref $repo_id) {
1346 my $gs = $repo_id;
1347 $repo_id = undef;
1348 $repo_id = $gs->{repo_id};
1349 }
1350 $remotes ||= read_all_remotes();
1351 my $remote = $remotes->{$repo_id} or
1352 die "[svn-remote \"$repo_id\"] unknown\n";
1353 my $fetch = $remote->{fetch};
1354 my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1355 my (@gs, @globs);
1356 my $ra = Git::SVN::Ra->new($url);
1357 my $uuid = $ra->get_uuid;
1358 my $head = $ra->get_latest_revnum;
1359 my $base = defined $fetch ? $head : 0;
1360
1361 # read the max revs for wildcard expansion (branches/*, tags/*)
1362 foreach my $t (qw/branches tags/) {
1363 defined $remote->{$t} or next;
1364 push @globs, $remote->{$t};
1365 my $max_rev = eval { tmp_config(qw/--int --get/,
1366 "svn-remote.$repo_id.${t}-maxRev") };
1367 if (defined $max_rev && ($max_rev < $base)) {
1368 $base = $max_rev;
1369 } elsif (!defined $max_rev) {
1370 $base = 0;
1371 }
1372 }
1373
1374 if ($fetch) {
1375 foreach my $p (sort keys %$fetch) {
1376 my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1377 my $lr = $gs->rev_map_max;
1378 if (defined $lr) {
1379 $base = $lr if ($lr < $base);
1380 }
1381 push @gs, $gs;
1382 }
1383 }
1384
1385 ($base, $head) = parse_revision_argument($base, $head);
1386 $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1387 }
1388
1389 sub read_all_remotes {
1390 my $r = {};
1391 foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1392 if (m!^(.+)\.fetch=\s*(.*)\s*:\s*refs/remotes/(.+)\s*$!) {
1393 my ($remote, $local_ref, $remote_ref) = ($1, $2, $3);
1394 $local_ref =~ s{^/}{};
1395 $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1396 } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1397 $r->{$1}->{url} = $2;
1398 } elsif (m!^(.+)\.(branches|tags)=
1399 (.*):refs/remotes/(.+)\s*$/!x) {
1400 my ($p, $g) = ($3, $4);
1401 my $rs = $r->{$1}->{$2} = {
1402 t => $2,
1403 remote => $1,
1404 path => Git::SVN::GlobSpec->new($p),
1405 ref => Git::SVN::GlobSpec->new($g) };
1406 if (length($rs->{ref}->{right}) != 0) {
1407 die "The '*' glob character must be the last ",
1408 "character of '$g'\n";
1409 }
1410 }
1411 }
1412 $r;
1413 }
1414
1415 sub init_vars {
1416 $_gc_nr = $_gc_period = 1000;
1417 if (defined $_repack || defined $_repack_flags) {
1418 warn "Repack options are obsolete; they have no effect.\n";
1419 }
1420 }
1421
1422 sub verify_remotes_sanity {
1423 return unless -d $ENV{GIT_DIR};
1424 my %seen;
1425 foreach (command(qw/config -l/)) {
1426 if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1427 if ($seen{$1}) {
1428 die "Remote ref refs/remote/$1 is tracked by",
1429 "\n \"$_\"\nand\n \"$seen{$1}\"\n",
1430 "Please resolve this ambiguity in ",
1431 "your git configuration file before ",
1432 "continuing\n";
1433 }
1434 $seen{$1} = $_;
1435 }
1436 }
1437 }
1438
1439 # we allow more chars than remotes2config.sh...
1440 sub sanitize_remote_name {
1441 my ($name) = @_;
1442 $name =~ tr{A-Za-z0-9:,/+-}{.}c;
1443 $name;
1444 }
1445
1446 sub find_existing_remote {
1447 my ($url, $remotes) = @_;
1448 return undef if $no_reuse_existing;
1449 my $existing;
1450 foreach my $repo_id (keys %$remotes) {
1451 my $u = $remotes->{$repo_id}->{url} or next;
1452 next if $u ne $url;
1453 $existing = $repo_id;
1454 last;
1455 }
1456 $existing;
1457 }
1458
1459 sub init_remote_config {
1460 my ($self, $url, $no_write) = @_;
1461 $url =~ s!/+$!!; # strip trailing slash
1462 my $r = read_all_remotes();
1463 my $existing = find_existing_remote($url, $r);
1464 if ($existing) {
1465 unless ($no_write) {
1466 print STDERR "Using existing ",
1467 "[svn-remote \"$existing\"]\n";
1468 }
1469 $self->{repo_id} = $existing;
1470 } elsif ($_minimize_url) {
1471 my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1472 $existing = find_existing_remote($min_url, $r);
1473 if ($existing) {
1474 unless ($no_write) {
1475 print STDERR "Using existing ",
1476 "[svn-remote \"$existing\"]\n";
1477 }
1478 $self->{repo_id} = $existing;
1479 }
1480 if ($min_url ne $url) {
1481 unless ($no_write) {
1482 print STDERR "Using higher level of URL: ",
1483 "$url => $min_url\n";
1484 }
1485 my $old_path = $self->{path};
1486 $self->{path} = $url;
1487 $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1488 if (length $old_path) {
1489 $self->{path} .= "/$old_path";
1490 }
1491 $url = $min_url;
1492 }
1493 }
1494 my $orig_url;
1495 if (!$existing) {
1496 # verify that we aren't overwriting anything:
1497 $orig_url = eval {
1498 command_oneline('config', '--get',
1499 "svn-remote.$self->{repo_id}.url")
1500 };
1501 if ($orig_url && ($orig_url ne $url)) {
1502 die "svn-remote.$self->{repo_id}.url already set: ",
1503 "$orig_url\nwanted to set to: $url\n";
1504 }
1505 }
1506 my ($xrepo_id, $xpath) = find_ref($self->refname);
1507 if (defined $xpath) {
1508 die "svn-remote.$xrepo_id.fetch already set to track ",
1509 "$xpath:refs/remotes/", $self->refname, "\n";
1510 }
1511 unless ($no_write) {
1512 command_noisy('config',
1513 "svn-remote.$self->{repo_id}.url", $url);
1514 $self->{path} =~ s{^/}{};
1515 command_noisy('config', '--add',
1516 "svn-remote.$self->{repo_id}.fetch",
1517 "$self->{path}:".$self->refname);
1518 }
1519 $self->{url} = $url;
1520 }
1521
1522 sub find_by_url { # repos_root and, path are optional
1523 my ($class, $full_url, $repos_root, $path) = @_;
1524
1525 return undef unless defined $full_url;
1526 remove_username($full_url);
1527 remove_username($repos_root) if defined $repos_root;
1528 my $remotes = read_all_remotes();
1529 if (defined $full_url && defined $repos_root && !defined $path) {
1530 $path = $full_url;
1531 $path =~ s#^\Q$repos_root\E(?:/|$)##;
1532 }
1533 foreach my $repo_id (keys %$remotes) {
1534 my $u = $remotes->{$repo_id}->{url} or next;
1535 remove_username($u);
1536 next if defined $repos_root && $repos_root ne $u;
1537
1538 my $fetch = $remotes->{$repo_id}->{fetch} || {};
1539 foreach (qw/branches tags/) {
1540 resolve_local_globs($u, $fetch,
1541 $remotes->{$repo_id}->{$_});
1542 }
1543 my $p = $path;
1544 my $rwr = rewrite_root({repo_id => $repo_id});
1545 unless (defined $p) {
1546 $p = $full_url;
1547 my $z = $u;
1548 if ($rwr) {
1549 $z = $rwr;
1550 }
1551 $p =~ s#^\Q$z\E(?:/|$)## or next;
1552 }
1553 foreach my $f (keys %$fetch) {
1554 next if $f ne $p;
1555 return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1556 }
1557 }
1558 undef;
1559 }
1560
1561 sub init {
1562 my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1563 my $self = _new($class, $repo_id, $ref_id, $path);
1564 if (defined $url) {
1565 $self->init_remote_config($url, $no_write);
1566 }
1567 $self;
1568 }
1569
1570 sub find_ref {
1571 my ($ref_id) = @_;
1572 foreach (command(qw/config -l/)) {
1573 next unless m!^svn-remote\.(.+)\.fetch=
1574 \s*(.*)\s*:\s*refs/remotes/(.+)\s*$!x;
1575 my ($repo_id, $path, $ref) = ($1, $2, $3);
1576 if ($ref eq $ref_id) {
1577 $path = '' if ($path =~ m#^\./?#);
1578 return ($repo_id, $path);
1579 }
1580 }
1581 (undef, undef, undef);
1582 }
1583
1584 sub new {
1585 my ($class, $ref_id, $repo_id, $path) = @_;
1586 if (defined $ref_id && !defined $repo_id && !defined $path) {
1587 ($repo_id, $path) = find_ref($ref_id);
1588 if (!defined $repo_id) {
1589 die "Could not find a \"svn-remote.*.fetch\" key ",
1590 "in the repository configuration matching: ",
1591 "refs/remotes/$ref_id\n";
1592 }
1593 }
1594 my $self = _new($class, $repo_id, $ref_id, $path);
1595 if (!defined $self->{path} || !length $self->{path}) {
1596 my $fetch = command_oneline('config', '--get',
1597 "svn-remote.$repo_id.fetch",
1598 ":refs/remotes/$ref_id\$") or
1599 die "Failed to read \"svn-remote.$repo_id.fetch\" ",
1600 "\":refs/remotes/$ref_id\$\" in config\n";
1601 ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
1602 }
1603 $self->{url} = command_oneline('config', '--get',
1604 "svn-remote.$repo_id.url") or
1605 die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
1606 $self->rebuild;
1607 $self;
1608 }
1609
1610 sub refname {
1611 my ($refname) = "refs/remotes/$_[0]->{ref_id}" ;
1612
1613 # It cannot end with a slash /, we'll throw up on this because
1614 # SVN can't have directories with a slash in their name, either:
1615 if ($refname =~ m{/$}) {
1616 die "ref: '$refname' ends with a trailing slash, this is ",
1617 "not permitted by git nor Subversion\n";
1618 }
1619
1620 # It cannot have ASCII control character space, tilde ~, caret ^,
1621 # colon :, question-mark ?, asterisk *, space, or open bracket [
1622 # anywhere.
1623 #
1624 # Additionally, % must be escaped because it is used for escaping
1625 # and we want our escaped refname to be reversible
1626 $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
1627
1628 # no slash-separated component can begin with a dot .
1629 # /.* becomes /%2E*
1630 $refname =~ s{/\.}{/%2E}g;
1631
1632 # It cannot have two consecutive dots .. anywhere
1633 # .. becomes %2E%2E
1634 $refname =~ s{\.\.}{%2E%2E}g;
1635
1636 return $refname;
1637 }
1638
1639 sub desanitize_refname {
1640 my ($refname) = @_;
1641 $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
1642 return $refname;
1643 }
1644
1645 sub svm_uuid {
1646 my ($self) = @_;
1647 return $self->{svm}->{uuid} if $self->svm;
1648 $self->ra;
1649 unless ($self->{svm}) {
1650 die "SVM UUID not cached, and reading remotely failed\n";
1651 }
1652 $self->{svm}->{uuid};
1653 }
1654
1655 sub svm {
1656 my ($self) = @_;
1657 return $self->{svm} if $self->{svm};
1658 my $svm;
1659 # see if we have it in our config, first:
1660 eval {
1661 my $section = "svn-remote.$self->{repo_id}";
1662 $svm = {
1663 source => tmp_config('--get', "$section.svm-source"),
1664 uuid => tmp_config('--get', "$section.svm-uuid"),
1665 replace => tmp_config('--get', "$section.svm-replace"),
1666 }
1667 };
1668 if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
1669 $self->{svm} = $svm;
1670 }
1671 $self->{svm};
1672 }
1673
1674 sub _set_svm_vars {
1675 my ($self, $ra) = @_;
1676 return $ra if $self->svm;
1677
1678 my @err = ( "useSvmProps set, but failed to read SVM properties\n",
1679 "(svm:source, svm:uuid) ",
1680 "from the following URLs:\n" );
1681 sub read_svm_props {
1682 my ($self, $ra, $path, $r) = @_;
1683 my $props = ($ra->get_dir($path, $r))[2];
1684 my $src = $props->{'svm:source'};
1685 my $uuid = $props->{'svm:uuid'};
1686 return undef if (!$src || !$uuid);
1687
1688 chomp($src, $uuid);
1689
1690 $uuid =~ m{^[0-9a-f\-]{30,}$}
1691 or die "doesn't look right - svm:uuid is '$uuid'\n";
1692
1693 # the '!' is used to mark the repos_root!/relative/path
1694 $src =~ s{/?!/?}{/};
1695 $src =~ s{/+$}{}; # no trailing slashes please
1696 # username is of no interest
1697 $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
1698
1699 my $replace = $ra->{url};
1700 $replace .= "/$path" if length $path;
1701
1702 my $section = "svn-remote.$self->{repo_id}";
1703 tmp_config("$section.svm-source", $src);
1704 tmp_config("$section.svm-replace", $replace);
1705 tmp_config("$section.svm-uuid", $uuid);
1706 $self->{svm} = {
1707 source => $src,
1708 uuid => $uuid,
1709 replace => $replace
1710 };
1711 }
1712
1713 my $r = $ra->get_latest_revnum;
1714 my $path = $self->{path};
1715 my %tried;
1716 while (length $path) {
1717 unless ($tried{"$self->{url}/$path"}) {
1718 return $ra if $self->read_svm_props($ra, $path, $r);
1719 $tried{"$self->{url}/$path"} = 1;
1720 }
1721 $path =~ s#/?[^/]+$##;
1722 }
1723 die "Path: '$path' should be ''\n" if $path ne '';
1724 return $ra if $self->read_svm_props($ra, $path, $r);
1725 $tried{"$self->{url}/$path"} = 1;
1726
1727 if ($ra->{repos_root} eq $self->{url}) {
1728 die @err, (map { " $_\n" } keys %tried), "\n";
1729 }
1730
1731 # nope, make sure we're connected to the repository root:
1732 my $ok;
1733 my @tried_b;
1734 $path = $ra->{svn_path};
1735 $ra = Git::SVN::Ra->new($ra->{repos_root});
1736 while (length $path) {
1737 unless ($tried{"$ra->{url}/$path"}) {
1738 $ok = $self->read_svm_props($ra, $path, $r);
1739 last if $ok;
1740 $tried{"$ra->{url}/$path"} = 1;
1741 }
1742 $path =~ s#/?[^/]+$##;
1743 }
1744 die "Path: '$path' should be ''\n" if $path ne '';
1745 $ok ||= $self->read_svm_props($ra, $path, $r);
1746 $tried{"$ra->{url}/$path"} = 1;
1747 if (!$ok) {
1748 die @err, (map { " $_\n" } keys %tried), "\n";
1749 }
1750 Git::SVN::Ra->new($self->{url});
1751 }
1752
1753 sub svnsync {
1754 my ($self) = @_;
1755 return $self->{svnsync} if $self->{svnsync};
1756
1757 if ($self->no_metadata) {
1758 die "Can't have both 'noMetadata' and ",
1759 "'useSvnsyncProps' options set!\n";
1760 }
1761 if ($self->rewrite_root) {
1762 die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
1763 "options set!\n";
1764 }
1765
1766 my $svnsync;
1767 # see if we have it in our config, first:
1768 eval {
1769 my $section = "svn-remote.$self->{repo_id}";
1770
1771 my $url = tmp_config('--get', "$section.svnsync-url");
1772 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1773 die "doesn't look right - svn:sync-from-url is '$url'\n";
1774
1775 my $uuid = tmp_config('--get', "$section.svnsync-uuid");
1776 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1777 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1778
1779 $svnsync = { url => $url, uuid => $uuid }
1780 };
1781 if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
1782 return $self->{svnsync} = $svnsync;
1783 }
1784
1785 my $err = "useSvnsyncProps set, but failed to read " .
1786 "svnsync property: svn:sync-from-";
1787 my $rp = $self->ra->rev_proplist(0);
1788
1789 my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
1790 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1791 die "doesn't look right - svn:sync-from-url is '$url'\n";
1792
1793 my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
1794 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1795 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1796
1797 my $section = "svn-remote.$self->{repo_id}";
1798 tmp_config('--add', "$section.svnsync-uuid", $uuid);
1799 tmp_config('--add', "$section.svnsync-url", $url);
1800 return $self->{svnsync} = { url => $url, uuid => $uuid };
1801 }
1802
1803 # this allows us to memoize our SVN::Ra UUID locally and avoid a
1804 # remote lookup (useful for 'git svn log').
1805 sub ra_uuid {
1806 my ($self) = @_;
1807 unless ($self->{ra_uuid}) {
1808 my $key = "svn-remote.$self->{repo_id}.uuid";
1809 my $uuid = eval { tmp_config('--get', $key) };
1810 if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/) {
1811 $self->{ra_uuid} = $uuid;
1812 } else {
1813 die "ra_uuid called without URL\n" unless $self->{url};
1814 $self->{ra_uuid} = $self->ra->get_uuid;
1815 tmp_config('--add', $key, $self->{ra_uuid});
1816 }
1817 }
1818 $self->{ra_uuid};
1819 }
1820
1821 sub _set_repos_root {
1822 my ($self, $repos_root) = @_;
1823 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1824 $repos_root ||= $self->ra->{repos_root};
1825 tmp_config($k, $repos_root);
1826 $repos_root;
1827 }
1828
1829 sub repos_root {
1830 my ($self) = @_;
1831 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1832 eval { tmp_config('--get', $k) } || $self->_set_repos_root;
1833 }
1834
1835 sub ra {
1836 my ($self) = shift;
1837 my $ra = Git::SVN::Ra->new($self->{url});
1838 $self->_set_repos_root($ra->{repos_root});
1839 if ($self->use_svm_props && !$self->{svm}) {
1840 if ($self->no_metadata) {
1841 die "Can't have both 'noMetadata' and ",
1842 "'useSvmProps' options set!\n";
1843 } elsif ($self->use_svnsync_props) {
1844 die "Can't have both 'useSvnsyncProps' and ",
1845 "'useSvmProps' options set!\n";
1846 }
1847 $ra = $self->_set_svm_vars($ra);
1848 $self->{-want_revprops} = 1;
1849 }
1850 $ra;
1851 }
1852
1853 sub rel_path {
1854 my ($self) = @_;
1855 my $repos_root = $self->ra->{repos_root};
1856 return $self->{path} if ($self->{url} eq $repos_root);
1857 my $url = $self->{url} .
1858 (length $self->{path} ? "/$self->{path}" : $self->{path});
1859 $url =~ s!^\Q$repos_root\E(?:/+|$)!!g;
1860 $url;
1861 }
1862
1863 # prop_walk(PATH, REV, SUB)
1864 # -------------------------
1865 # Recursively traverse PATH at revision REV and invoke SUB for each
1866 # directory that contains a SVN property. SUB will be invoked as
1867 # follows: &SUB(gs, path, props); where `gs' is this instance of
1868 # Git::SVN, `path' the path to the directory where the properties
1869 # `props' were found. The `path' will be relative to point of checkout,
1870 # that is, if url://repo/trunk is the current Git branch, and that
1871 # directory contains a sub-directory `d', SUB will be invoked with `/d/'
1872 # as `path' (note the trailing `/').
1873 sub prop_walk {
1874 my ($self, $path, $rev, $sub) = @_;
1875
1876 $path =~ s#^/##;
1877 my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
1878 $path =~ s#^/*#/#g;
1879 my $p = $path;
1880 # Strip the irrelevant part of the path.
1881 $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
1882 # Ensure the path is terminated by a `/'.
1883 $p =~ s#/*$#/#;
1884
1885 # The properties contain all the internal SVN stuff nobody
1886 # (usually) cares about.
1887 my $interesting_props = 0;
1888 foreach (keys %{$props}) {
1889 # If it doesn't start with `svn:', it must be a
1890 # user-defined property.
1891 ++$interesting_props and next if $_ !~ /^svn:/;
1892 # FIXME: Fragile, if SVN adds new public properties,
1893 # this needs to be updated.
1894 ++$interesting_props if /^svn:(?:ignore|keywords|executable
1895 |eol-style|mime-type
1896 |externals|needs-lock)$/x;
1897 }
1898 &$sub($self, $p, $props) if $interesting_props;
1899
1900 foreach (sort keys %$dirent) {
1901 next if $dirent->{$_}->{kind} != $SVN::Node::dir;
1902 $self->prop_walk($path . '/' . $_, $rev, $sub);
1903 }
1904 }
1905
1906 sub last_rev { ($_[0]->last_rev_commit)[0] }
1907 sub last_commit { ($_[0]->last_rev_commit)[1] }
1908
1909 # returns the newest SVN revision number and newest commit SHA1
1910 sub last_rev_commit {
1911 my ($self) = @_;
1912 if (defined $self->{last_rev} && defined $self->{last_commit}) {
1913 return ($self->{last_rev}, $self->{last_commit});
1914 }
1915 my $c = ::verify_ref($self->refname.'^0');
1916 if ($c && !$self->use_svm_props && !$self->no_metadata) {
1917 my $rev = (::cmt_metadata($c))[1];
1918 if (defined $rev) {
1919 ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
1920 return ($rev, $c);
1921 }
1922 }
1923 my $map_path = $self->map_path;
1924 unless (-e $map_path) {
1925 ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
1926 return (undef, undef);
1927 }
1928 my ($rev, $commit) = $self->rev_map_max(1);
1929 ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
1930 return ($rev, $commit);
1931 }
1932
1933 sub get_fetch_range {
1934 my ($self, $min, $max) = @_;
1935 $max ||= $self->ra->get_latest_revnum;
1936 $min ||= $self->rev_map_max;
1937 (++$min, $max);
1938 }
1939
1940 sub tmp_config {
1941 my (@args) = @_;
1942 my $old_def_config = "$ENV{GIT_DIR}/svn/config";
1943 my $config = "$ENV{GIT_DIR}/svn/.metadata";
1944 if (! -f $config && -f $old_def_config) {
1945 rename $old_def_config, $config or
1946 die "Failed rename $old_def_config => $config: $!\n";
1947 }
1948 my $old_config = $ENV{GIT_CONFIG};
1949 $ENV{GIT_CONFIG} = $config;
1950 $@ = undef;
1951 my @ret = eval {
1952 unless (-f $config) {
1953 mkfile($config);
1954 open my $fh, '>', $config or
1955 die "Can't open $config: $!\n";
1956 print $fh "; This file is used internally by ",
1957 "git-svn\n" or die
1958 "Couldn't write to $config: $!\n";
1959 print $fh "; You should not have to edit it\n" or
1960 die "Couldn't write to $config: $!\n";
1961 close $fh or die "Couldn't close $config: $!\n";
1962 }
1963 command('config', @args);
1964 };
1965 my $err = $@;
1966 if (defined $old_config) {
1967 $ENV{GIT_CONFIG} = $old_config;
1968 } else {
1969 delete $ENV{GIT_CONFIG};
1970 }
1971 die $err if $err;
1972 wantarray ? @ret : $ret[0];
1973 }
1974
1975 sub tmp_index_do {
1976 my ($self, $sub) = @_;
1977 my $old_index = $ENV{GIT_INDEX_FILE};
1978 $ENV{GIT_INDEX_FILE} = $self->{index};
1979 $@ = undef;
1980 my @ret = eval {
1981 my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
1982 mkpath([$dir]) unless -d $dir;
1983 &$sub;
1984 };
1985 my $err = $@;
1986 if (defined $old_index) {
1987 $ENV{GIT_INDEX_FILE} = $old_index;
1988 } else {
1989 delete $ENV{GIT_INDEX_FILE};
1990 }
1991 die $err if $err;
1992 wantarray ? @ret : $ret[0];
1993 }
1994
1995 sub assert_index_clean {
1996 my ($self, $treeish) = @_;
1997
1998 $self->tmp_index_do(sub {
1999 command_noisy('read-tree', $treeish) unless -e $self->{index};
2000 my $x = command_oneline('write-tree');
2001 my ($y) = (command(qw/cat-file commit/, $treeish) =~
2002 /^tree ($::sha1)/mo);
2003 return if $y eq $x;
2004
2005 warn "Index mismatch: $y != $x\nrereading $treeish\n";
2006 unlink $self->{index} or die "unlink $self->{index}: $!\n";
2007 command_noisy('read-tree', $treeish);
2008 $x = command_oneline('write-tree');
2009 if ($y ne $x) {
2010 ::fatal "trees ($treeish) $y != $x\n",
2011 "Something is seriously wrong...";
2012 }
2013 });
2014 }
2015
2016 sub get_commit_parents {
2017 my ($self, $log_entry) = @_;
2018 my (%seen, @ret, @tmp);
2019 # legacy support for 'set-tree'; this is only used by set_tree_cb:
2020 if (my $ip = $self->{inject_parents}) {
2021 if (my $commit = delete $ip->{$log_entry->{revision}}) {
2022 push @tmp, $commit;
2023 }
2024 }
2025 if (my $cur = ::verify_ref($self->refname.'^0')) {
2026 push @tmp, $cur;
2027 }
2028 if (my $ipd = $self->{inject_parents_dcommit}) {
2029 if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2030 push @tmp, @$commit;
2031 }
2032 }
2033 push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2034 while (my $p = shift @tmp) {
2035 next if $seen{$p};
2036 $seen{$p} = 1;
2037 push @ret, $p;
2038 # MAXPARENT is defined to 16 in commit-tree.c:
2039 last if @ret >= 16;
2040 }
2041 if (@tmp) {
2042 die "r$log_entry->{revision}: No room for parents:\n\t",
2043 join("\n\t", @tmp), "\n";
2044 }
2045 @ret;
2046 }
2047
2048 sub rewrite_root {
2049 my ($self) = @_;
2050 return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2051 my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2052 my $rwr = eval { command_oneline(qw/config --get/, $k) };
2053 if ($rwr) {
2054 $rwr =~ s#/+$##;
2055 if ($rwr !~ m#^[a-z\+]+://#) {
2056 die "$rwr is not a valid URL (key: $k)\n";
2057 }
2058 }
2059 $self->{-rewrite_root} = $rwr;
2060 }
2061
2062 sub metadata_url {
2063 my ($self) = @_;
2064 ($self->rewrite_root || $self->{url}) .
2065 (length $self->{path} ? '/' . $self->{path} : '');
2066 }
2067
2068 sub full_url {
2069 my ($self) = @_;
2070 $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2071 }
2072
2073
2074 sub set_commit_header_env {
2075 my ($log_entry) = @_;
2076 my %env;
2077 foreach my $ned (qw/NAME EMAIL DATE/) {
2078 foreach my $ac (qw/AUTHOR COMMITTER/) {
2079 $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2080 }
2081 }
2082
2083 $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2084 $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2085 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2086
2087 $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2088 ? $log_entry->{commit_name}
2089 : $log_entry->{name};
2090 $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2091 ? $log_entry->{commit_email}
2092 : $log_entry->{email};
2093 \%env;
2094 }
2095
2096 sub restore_commit_header_env {
2097 my ($env) = @_;
2098 foreach my $ned (qw/NAME EMAIL DATE/) {
2099 foreach my $ac (qw/AUTHOR COMMITTER/) {
2100 my $k = "GIT_${ac}_${ned}";
2101 if (defined $env->{$k}) {
2102 $ENV{$k} = $env->{$k};
2103 } else {
2104 delete $ENV{$k};
2105 }
2106 }
2107 }
2108 }
2109
2110 sub gc {
2111 command_noisy('gc', '--auto');
2112 };
2113
2114 sub do_git_commit {
2115 my ($self, $log_entry) = @_;
2116 my $lr = $self->last_rev;
2117 if (defined $lr && $lr >= $log_entry->{revision}) {
2118 die "Last fetched revision of ", $self->refname,
2119 " was r$lr, but we are about to fetch: ",
2120 "r$log_entry->{revision}!\n";
2121 }
2122 if (my $c = $self->rev_map_get($log_entry->{revision})) {
2123 croak "$log_entry->{revision} = $c already exists! ",
2124 "Why are we refetching it?\n";
2125 }
2126 my $old_env = set_commit_header_env($log_entry);
2127 my $tree = $log_entry->{tree};
2128 if (!defined $tree) {
2129 $tree = $self->tmp_index_do(sub {
2130 command_oneline('write-tree') });
2131 }
2132 die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2133
2134 my @exec = ('git-commit-tree', $tree);
2135 foreach ($self->get_commit_parents($log_entry)) {
2136 push @exec, '-p', $_;
2137 }
2138 defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2139 or croak $!;
2140 print $msg_fh $log_entry->{log} or croak $!;
2141 restore_commit_header_env($old_env);
2142 unless ($self->no_metadata) {
2143 print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2144 or croak $!;
2145 }
2146 $msg_fh->flush == 0 or croak $!;
2147 close $msg_fh or croak $!;
2148 chomp(my $commit = do { local $/; <$out_fh> });
2149 close $out_fh or croak $!;
2150 waitpid $pid, 0;
2151 croak $? if $?;
2152 if ($commit !~ /^$::sha1$/o) {
2153 die "Failed to commit, invalid sha1: $commit\n";
2154 }
2155
2156 $self->rev_map_set($log_entry->{revision}, $commit, 1);
2157
2158 $self->{last_rev} = $log_entry->{revision};
2159 $self->{last_commit} = $commit;
2160 print "r$log_entry->{revision}";
2161 if (defined $log_entry->{svm_revision}) {
2162 print " (\@$log_entry->{svm_revision})";
2163 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2164 0, $self->svm_uuid);
2165 }
2166 print " = $commit ($self->{ref_id})\n";
2167 if (--$_gc_nr == 0) {
2168 $_gc_nr = $_gc_period;
2169 gc();
2170 }
2171 return $commit;
2172 }
2173
2174 sub match_paths {
2175 my ($self, $paths, $r) = @_;
2176 return 1 if $self->{path} eq '';
2177 if (my $path = $paths->{"/$self->{path}"}) {
2178 return ($path->{action} eq 'D') ? 0 : 1;
2179 }
2180 $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
2181 if (grep /$self->{path_regex}/, keys %$paths) {
2182 return 1;
2183 }
2184 my $c = '';
2185 foreach (split m#/#, $self->{path}) {
2186 $c .= "/$_";
2187 next unless ($paths->{$c} &&
2188 ($paths->{$c}->{action} =~ /^[AR]$/));
2189 if ($self->ra->check_path($self->{path}, $r) ==
2190 $SVN::Node::dir) {
2191 return 1;
2192 }
2193 }
2194 return 0;
2195 }
2196
2197 sub find_parent_branch {
2198 my ($self, $paths, $rev) = @_;
2199 return undef unless $self->follow_parent;
2200 unless (defined $paths) {
2201 my $err_handler = $SVN::Error::handler;
2202 $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2203 $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1, sub {
2204 $paths =
2205 Git::SVN::Ra::dup_changed_paths($_[0]) });
2206 $SVN::Error::handler = $err_handler;
2207 }
2208 return undef unless defined $paths;
2209
2210 # look for a parent from another branch:
2211 my @b_path_components = split m#/#, $self->rel_path;
2212 my @a_path_components;
2213 my $i;
2214 while (@b_path_components) {
2215 $i = $paths->{'/'.join('/', @b_path_components)};
2216 last if $i && defined $i->{copyfrom_path};
2217 unshift(@a_path_components, pop(@b_path_components));
2218 }
2219 return undef unless defined $i && defined $i->{copyfrom_path};
2220 my $branch_from = $i->{copyfrom_path};
2221 if (@a_path_components) {
2222 print STDERR "branch_from: $branch_from => ";
2223 $branch_from .= '/'.join('/', @a_path_components);
2224 print STDERR $branch_from, "\n";
2225 }
2226 my $r = $i->{copyfrom_rev};
2227 my $repos_root = $self->ra->{repos_root};
2228 my $url = $self->ra->{url};
2229 my $new_url = $repos_root . $branch_from;
2230 print STDERR "Found possible branch point: ",
2231 "$new_url => ", $self->full_url, ", $r\n";
2232 $branch_from =~ s#^/##;
2233 my $gs = Git::SVN->find_by_url($new_url, $repos_root, $branch_from);
2234 unless ($gs) {
2235 my $ref_id = $self->{ref_id};
2236 $ref_id =~ s/\@\d+$//;
2237 $ref_id .= "\@$r";
2238 # just grow a tail if we're not unique enough :x
2239 $ref_id .= '-' while find_ref($ref_id);
2240 print STDERR "Initializing parent: $ref_id\n";
2241 my ($u, $p) = ($new_url, '');
2242 if ($u =~ s#^\Q$url\E(/|$)##) {
2243 $p = $u;
2244 $u = $url;
2245 }
2246 $gs = Git::SVN->init($u, $p, $self->{repo_id}, $ref_id, 1);
2247 }
2248 my ($r0, $parent) = $gs->find_rev_before($r, 1);
2249 if (!defined $r0 || !defined $parent) {
2250 my ($base, $head) = parse_revision_argument(0, $r);
2251 if ($base <= $r) {
2252 $gs->fetch($base, $r);
2253 }
2254 ($r0, $parent) = $gs->last_rev_commit;
2255 }
2256 if (defined $r0 && defined $parent) {
2257 print STDERR "Found branch parent: ($self->{ref_id}) $parent\n";
2258 my $ed;
2259 if ($self->ra->can_do_switch) {
2260 $self->assert_index_clean($parent);
2261 print STDERR "Following parent with do_switch\n";
2262 # do_switch works with svn/trunk >= r22312, but that
2263 # is not included with SVN 1.4.3 (the latest version
2264 # at the moment), so we can't rely on it
2265 $self->{last_commit} = $parent;
2266 $ed = SVN::Git::Fetcher->new($self);
2267 $gs->ra->gs_do_switch($r0, $rev, $gs,
2268 $self->full_url, $ed)
2269 or die "SVN connection failed somewhere...\n";
2270 } elsif ($self->ra->trees_match($new_url, $r0,
2271 $self->full_url, $rev)) {
2272 print STDERR "Trees match:\n",
2273 " $new_url\@$r0\n",
2274 " ${\$self->full_url}\@$rev\n",
2275 "Following parent with no changes\n";
2276 $self->tmp_index_do(sub {
2277 command_noisy('read-tree', $parent);
2278 });
2279 $self->{last_commit} = $parent;
2280 } else {
2281 print STDERR "Following parent with do_update\n";
2282 $ed = SVN::Git::Fetcher->new($self);
2283 $self->ra->gs_do_update($rev, $rev, $self, $ed)
2284 or die "SVN connection failed somewhere...\n";
2285 }
2286 print STDERR "Successfully followed parent\n";
2287 return $self->make_log_entry($rev, [$parent], $ed);
2288 }
2289 return undef;
2290 }
2291
2292 sub do_fetch {
2293 my ($self, $paths, $rev) = @_;
2294 my $ed;
2295 my ($last_rev, @parents);
2296 if (my $lc = $self->last_commit) {
2297 # we can have a branch that was deleted, then re-added
2298 # under the same name but copied from another path, in
2299 # which case we'll have multiple parents (we don't
2300 # want to break the original ref, nor lose copypath info):
2301 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2302 push @{$log_entry->{parents}}, $lc;
2303 return $log_entry;
2304 }
2305 $ed = SVN::Git::Fetcher->new($self);
2306 $last_rev = $self->{last_rev};
2307 $ed->{c} = $lc;
2308 @parents = ($lc);
2309 } else {
2310 $last_rev = $rev;
2311 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2312 return $log_entry;
2313 }
2314 $ed = SVN::Git::Fetcher->new($self);
2315 }
2316 unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2317 die "SVN connection failed somewhere...\n";
2318 }
2319 $self->make_log_entry($rev, \@parents, $ed);
2320 }
2321
2322 sub get_untracked {
2323 my ($self, $ed) = @_;
2324 my @out;
2325 my $h = $ed->{empty};
2326 foreach (sort keys %$h) {
2327 my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2328 push @out, " $act: " . uri_encode($_);
2329 warn "W: $act: $_\n";
2330 }
2331 foreach my $t (qw/dir_prop file_prop/) {
2332 $h = $ed->{$t} or next;
2333 foreach my $path (sort keys %$h) {
2334 my $ppath = $path eq '' ? '.' : $path;
2335 foreach my $prop (sort keys %{$h->{$path}}) {
2336 next if $SKIP_PROP{$prop};
2337 my $v = $h->{$path}->{$prop};
2338 my $t_ppath_prop = "$t: " .
2339 uri_encode($ppath) . ' ' .
2340 uri_encode($prop);
2341 if (defined $v) {
2342 push @out, " +$t_ppath_prop " .
2343 uri_encode($v);
2344 } else {
2345 push @out, " -$t_ppath_prop";
2346 }
2347 }
2348 }
2349 }
2350 foreach my $t (qw/absent_file absent_directory/) {
2351 $h = $ed->{$t} or next;
2352 foreach my $parent (sort keys %$h) {
2353 foreach my $path (sort @{$h->{$parent}}) {
2354 push @out, " $t: " .
2355 uri_encode("$parent/$path");
2356 warn "W: $t: $parent/$path ",
2357 "Insufficient permissions?\n";
2358 }
2359 }
2360 }
2361 \@out;
2362 }
2363
2364 sub parse_svn_date {
2365 my $date = shift || return '+0000 1970-01-01 00:00:00';
2366 my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2367 (\d\d)\:(\d\d)\:(\d\d).\d+Z$/x) or
2368 croak "Unable to parse date: $date\n";
2369 "+0000 $Y-$m-$d $H:$M:$S";
2370 }
2371
2372 sub check_author {
2373 my ($author) = @_;
2374 if (!defined $author || length $author == 0) {
2375 $author = '(no author)';
2376 }
2377 if (defined $::_authors && ! defined $::users{$author}) {
2378 die "Author: $author not defined in $::_authors file\n";
2379 }
2380 $author;
2381 }
2382
2383 sub make_log_entry {
2384 my ($self, $rev, $parents, $ed) = @_;
2385 my $untracked = $self->get_untracked($ed);
2386
2387 open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
2388 print $un "r$rev\n" or croak $!;
2389 print $un $_, "\n" foreach @$untracked;
2390 my %log_entry = ( parents => $parents || [], revision => $rev,
2391 log => '');
2392
2393 my $headrev;
2394 my $logged = delete $self->{logged_rev_props};
2395 if (!$logged || $self->{-want_revprops}) {
2396 my $rp = $self->ra->rev_proplist($rev);
2397 foreach (sort keys %$rp) {
2398 my $v = $rp->{$_};
2399 if (/^svn:(author|date|log)$/) {
2400 $log_entry{$1} = $v;
2401 } elsif ($_ eq 'svm:headrev') {
2402 $headrev = $v;
2403 } else {
2404 print $un " rev_prop: ", uri_encode($_), ' ',
2405 uri_encode($v), "\n";
2406 }
2407 }
2408 } else {
2409 map { $log_entry{$_} = $logged->{$_} } keys %$logged;
2410 }
2411 close $un or croak $!;
2412
2413 $log_entry{date} = parse_svn_date($log_entry{date});
2414 $log_entry{log} .= "\n";
2415 my $author = $log_entry{author} = check_author($log_entry{author});
2416 my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
2417 : ($author, undef);
2418
2419 my ($commit_name, $commit_email) = ($name, $email);
2420 if ($_use_log_author) {
2421 my $name_field;
2422 if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
2423 $name_field = $1;
2424 } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
2425 $name_field = $1;
2426 }
2427 if (!defined $name_field) {
2428 #
2429 } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
2430 ($name, $email) = ($1, $2);
2431 } elsif ($name_field =~ /(.*)@/) {
2432 ($name, $email) = ($1, $name_field);
2433 } else {
2434 ($name, $email) = ($name_field, 'unknown');
2435 }
2436 }
2437 if (defined $headrev && $self->use_svm_props) {
2438 if ($self->rewrite_root) {
2439 die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
2440 "options set!\n";
2441 }
2442 my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$};
2443 # we don't want "SVM: initializing mirror for junk" ...
2444 return undef if $r == 0;
2445 my $svm = $self->svm;
2446 if ($uuid ne $svm->{uuid}) {
2447 die "UUID mismatch on SVM path:\n",
2448 "expected: $svm->{uuid}\n",
2449 " got: $uuid\n";
2450 }
2451 my $full_url = $self->full_url;
2452 $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
2453 die "Failed to replace '$svm->{replace}' with ",
2454 "'$svm->{source}' in $full_url\n";
2455 # throw away username for storing in records
2456 remove_username($full_url);
2457 $log_entry{metadata} = "$full_url\@$r $uuid";
2458 $log_entry{svm_revision} = $r;
2459 $email ||= "$author\@$uuid";
2460 $commit_email ||= "$author\@$uuid";
2461 } elsif ($self->use_svnsync_props) {
2462 my $full_url = $self->svnsync->{url};
2463 $full_url .= "/$self->{path}" if length $self->{path};
2464 remove_username($full_url);
2465 my $uuid = $self->svnsync->{uuid};
2466 $log_entry{metadata} = "$full_url\@$rev $uuid";
2467 $email ||= "$author\@$uuid";
2468 $commit_email ||= "$author\@$uuid";
2469 } else {
2470 my $url = $self->metadata_url;
2471 remove_username($url);
2472 $log_entry{metadata} = "$url\@$rev " .
2473 $self->ra->get_uuid;
2474 $email ||= "$author\@" . $self->ra->get_uuid;
2475 $commit_email ||= "$author\@" . $self->ra->get_uuid;
2476 }
2477 $log_entry{name} = $name;
2478 $log_entry{email} = $email;
2479 $log_entry{commit_name} = $commit_name;
2480 $log_entry{commit_email} = $commit_email;
2481 \%log_entry;
2482 }
2483
2484 sub fetch {
2485 my ($self, $min_rev, $max_rev, @parents) = @_;
2486 my ($last_rev, $last_commit) = $self->last_rev_commit;
2487 my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
2488 $self->ra->gs_fetch_loop_common($base, $head, [$self]);
2489 }
2490
2491 sub set_tree_cb {
2492 my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
2493 $self->{inject_parents} = { $rev => $tree };
2494 $self->fetch(undef, undef);
2495 }
2496
2497 sub set_tree {
2498 my ($self, $tree) = (shift, shift);
2499 my $log_entry = ::get_commit_entry($tree);
2500 unless ($self->{last_rev}) {
2501 fatal("Must have an existing revision to commit");
2502 }
2503 my %ed_opts = ( r => $self->{last_rev},
2504 log => $log_entry->{log},
2505 ra => $self->ra,
2506 tree_a => $self->{last_commit},
2507 tree_b => $tree,
2508 editor_cb => sub {
2509 $self->set_tree_cb($log_entry, $tree, @_) },
2510 svn_path => $self->{path} );
2511 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
2512 print "No changes\nr$self->{last_rev} = $tree\n";
2513 }
2514 }
2515
2516 sub rebuild_from_rev_db {
2517 my ($self, $path) = @_;
2518 my $r = -1;
2519 open my $fh, '<', $path or croak "open: $!";
2520 while (<$fh>) {
2521 length($_) == 41 or croak "inconsistent size in ($_) != 41";
2522 chomp($_);
2523 ++$r;
2524 next if $_ eq ('0' x 40);
2525 $self->rev_map_set($r, $_);
2526 print "r$r = $_\n";
2527 }
2528 close $fh or croak "close: $!";
2529 unlink $path or croak "unlink: $!";
2530 }
2531
2532 sub rebuild {
2533 my ($self) = @_;
2534 my $map_path = $self->map_path;
2535 return if (-e $map_path && ! -z $map_path);
2536 return unless ::verify_ref($self->refname.'^0');
2537 if ($self->use_svm_props || $self->no_metadata) {
2538 my $rev_db = $self->rev_db_path;
2539 $self->rebuild_from_rev_db($rev_db);
2540 if ($self->use_svm_props) {
2541 my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
2542 $self->rebuild_from_rev_db($svm_rev_db);
2543 }
2544 $self->unlink_rev_db_symlink;
2545 return;
2546 }
2547 print "Rebuilding $map_path ...\n";
2548 my ($log, $ctx) =
2549 command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
2550 $self->refname, '--');
2551 my $full_url = $self->full_url;
2552 remove_username($full_url);
2553 my $svn_uuid = $self->ra_uuid;
2554 my $c;
2555 while (<$log>) {
2556 if ( m{^commit ($::sha1)$} ) {
2557 $c = $1;
2558 next;
2559 }
2560 next unless s{^\s*(git-svn-id:)}{$1};
2561 my ($url, $rev, $uuid) = ::extract_metadata($_);
2562 remove_username($url);
2563
2564 # ignore merges (from set-tree)
2565 next if (!defined $rev || !$uuid);
2566
2567 # if we merged or otherwise started elsewhere, this is
2568 # how we break out of it
2569 if (($uuid ne $svn_uuid) ||
2570 ($full_url && $url && ($url ne $full_url))) {
2571 next;
2572 }
2573
2574 $self->rev_map_set($rev, $c);
2575 print "r$rev = $c\n";
2576 }
2577 command_close_pipe($log, $ctx);
2578 print "Done rebuilding $map_path\n";
2579 my $rev_db_path = $self->rev_db_path;
2580 if (-f $self->rev_db_path) {
2581 unlink $self->rev_db_path or croak "unlink: $!";
2582 }
2583 $self->unlink_rev_db_symlink;
2584 }
2585
2586 # rev_map:
2587 # Tie::File seems to be prone to offset errors if revisions get sparse,
2588 # it's not that fast, either. Tie::File is also not in Perl 5.6. So
2589 # one of my favorite modules is out :< Next up would be one of the DBM
2590 # modules, but I'm not sure which is most portable...
2591 #
2592 # This is the replacement for the rev_db format, which was too big
2593 # and inefficient for large repositories with a lot of sparse history
2594 # (mainly tags)
2595 #
2596 # The format is this:
2597 # - 24 bytes for every record,
2598 # * 4 bytes for the integer representing an SVN revision number
2599 # * 20 bytes representing the sha1 of a git commit
2600 # - No empty padding records like the old format
2601 # (except the last record, which can be overwritten)
2602 # - new records are written append-only since SVN revision numbers
2603 # increase monotonically
2604 # - lookups on SVN revision number are done via a binary search
2605 # - Piping the file to xxd -c24 is a good way of dumping it for
2606 # viewing or editing (piped back through xxd -r), should the need
2607 # ever arise.
2608 # - The last record can be padding revision with an all-zero sha1
2609 # This is used to optimize fetch performance when using multiple
2610 # "fetch" directives in .git/config
2611 #
2612 # These files are disposable unless noMetadata or useSvmProps is set
2613
2614 sub _rev_map_set {
2615 my ($fh, $rev, $commit) = @_;
2616
2617 my $size = (stat($fh))[7];
2618 ($size % 24) == 0 or croak "inconsistent size: $size";
2619
2620 my $wr_offset = 0;
2621 if ($size > 0) {
2622 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2623 my $read = sysread($fh, my $buf, 24) or croak "read: $!";
2624 $read == 24 or croak "read only $read bytes (!= 24)";
2625 my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
2626 if ($last_commit eq ('0' x40)) {
2627 if ($size >= 48) {
2628 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2629 $read = sysread($fh, $buf, 24) or
2630 croak "read: $!";
2631 $read == 24 or
2632 croak "read only $read bytes (!= 24)";
2633 ($last_rev, $last_commit) =
2634 unpack(rev_map_fmt, $buf);
2635 if ($last_commit eq ('0' x40)) {
2636 croak "inconsistent .rev_map\n";
2637 }
2638 }
2639 if ($last_rev >= $rev) {
2640 croak "last_rev is higher!: $last_rev >= $rev";
2641 }
2642 $wr_offset = -24;
2643 }
2644 }
2645 sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
2646 syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
2647 croak "write: $!";
2648 }
2649
2650 sub mkfile {
2651 my ($path) = @_;
2652 unless (-e $path) {
2653 my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
2654 mkpath([$dir]) unless -d $dir;
2655 open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
2656 close $fh or die "Couldn't close (create) $path: $!\n";
2657 }
2658 }
2659
2660 sub rev_map_set {
2661 my ($self, $rev, $commit, $update_ref, $uuid) = @_;
2662 length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
2663 my $db = $self->map_path($uuid);
2664 my $db_lock = "$db.lock";
2665 my $sig;
2666 if ($update_ref) {
2667 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2668 $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
2669 }
2670 mkfile($db);
2671
2672 $LOCKFILES{$db_lock} = 1;
2673 my $sync;
2674 # both of these options make our .rev_db file very, very important
2675 # and we can't afford to lose it because rebuild() won't work
2676 if ($self->use_svm_props || $self->no_metadata) {
2677 $sync = 1;
2678 copy($db, $db_lock) or die "rev_map_set(@_): ",
2679 "Failed to copy: ",
2680 "$db => $db_lock ($!)\n";
2681 } else {
2682 rename $db, $db_lock or die "rev_map_set(@_): ",
2683 "Failed to rename: ",
2684 "$db => $db_lock ($!)\n";
2685 }
2686
2687 sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
2688 or croak "Couldn't open $db_lock: $!\n";
2689 _rev_map_set($fh, $rev, $commit);
2690 if ($sync) {
2691 $fh->flush or die "Couldn't flush $db_lock: $!\n";
2692 $fh->sync or die "Couldn't sync $db_lock: $!\n";
2693 }
2694 close $fh or croak $!;
2695 if ($update_ref) {
2696 $_head = $self;
2697 command_noisy('update-ref', '-m', "r$rev",
2698 $self->refname, $commit);
2699 }
2700 rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
2701 "$db_lock => $db ($!)\n";
2702 delete $LOCKFILES{$db_lock};
2703 if ($update_ref) {
2704 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2705 $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
2706 kill $sig, $$ if defined $sig;
2707 }
2708 }
2709
2710 # If want_commit, this will return an array of (rev, commit) where
2711 # commit _must_ be a valid commit in the archive.
2712 # Otherwise, it'll return the max revision (whether or not the
2713 # commit is valid or just a 0x40 placeholder).
2714 sub rev_map_max {
2715 my ($self, $want_commit) = @_;
2716 $self->rebuild;
2717 my $map_path = $self->map_path;
2718 stat $map_path or return $want_commit ? (0, undef) : 0;
2719 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2720 my $size = (stat($fh))[7];
2721 ($size % 24) == 0 or croak "inconsistent size: $size";
2722
2723 if ($size == 0) {
2724 close $fh or croak "close: $!";
2725 return $want_commit ? (0, undef) : 0;
2726 }
2727
2728 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2729 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2730 my ($r, $c) = unpack(rev_map_fmt, $buf);
2731 if ($want_commit && $c eq ('0' x40)) {
2732 if ($size < 48) {
2733 return $want_commit ? (0, undef) : 0;
2734 }
2735 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2736 sysread($fh, $buf, 24) == 24 or croak "read: $!";
2737 ($r, $c) = unpack(rev_map_fmt, $buf);
2738 if ($c eq ('0'x40)) {
2739 croak "Penultimate record is all-zeroes in $map_path";
2740 }
2741 }
2742 close $fh or croak "close: $!";
2743 $want_commit ? ($r, $c) : $r;
2744 }
2745
2746 sub rev_map_get {
2747 my ($self, $rev, $uuid) = @_;
2748 my $map_path = $self->map_path($uuid);
2749 return undef unless -e $map_path;
2750
2751 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2752 my $size = (stat($fh))[7];
2753 ($size % 24) == 0 or croak "inconsistent size: $size";
2754
2755 if ($size == 0) {
2756 close $fh or croak "close: $fh";
2757 return undef;
2758 }
2759
2760 my ($l, $u) = (0, $size - 24);
2761 my ($r, $c, $buf);
2762
2763 while ($l <= $u) {
2764 my $i = int(($l/24 + $u/24) / 2) * 24;
2765 sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
2766 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2767 my ($r, $c) = unpack('NH40', $buf);
2768
2769 if ($r < $rev) {
2770 $l = $i + 24;
2771 } elsif ($r > $rev) {
2772 $u = $i - 24;
2773 } else { # $r == $rev
2774 close($fh) or croak "close: $!";
2775 return $c eq ('0' x 40) ? undef : $c;
2776 }
2777 }
2778 close($fh) or croak "close: $!";
2779 undef;
2780 }
2781
2782 # Finds the first svn revision that exists on (if $eq_ok is true) or
2783 # before $rev for the current branch. It will not search any lower
2784 # than $min_rev. Returns the git commit hash and svn revision number
2785 # if found, else (undef, undef).
2786 sub find_rev_before {
2787 my ($self, $rev, $eq_ok, $min_rev) = @_;
2788 --$rev unless $eq_ok;
2789 $min_rev ||= 1;
2790 while ($rev >= $min_rev) {
2791 if (my $c = $self->rev_map_get($rev)) {
2792 return ($rev, $c);
2793 }
2794 --$rev;
2795 }
2796 return (undef, undef);
2797 }
2798
2799 # Finds the first svn revision that exists on (if $eq_ok is true) or
2800 # after $rev for the current branch. It will not search any higher
2801 # than $max_rev. Returns the git commit hash and svn revision number
2802 # if found, else (undef, undef).
2803 sub find_rev_after {
2804 my ($self, $rev, $eq_ok, $max_rev) = @_;
2805 ++$rev unless $eq_ok;
2806 $max_rev ||= $self->rev_map_max;
2807 while ($rev <= $max_rev) {
2808 if (my $c = $self->rev_map_get($rev)) {
2809 return ($rev, $c);
2810 }
2811 ++$rev;
2812 }
2813 return (undef, undef);
2814 }
2815
2816 sub _new {
2817 my ($class, $repo_id, $ref_id, $path) = @_;
2818 unless (defined $repo_id && length $repo_id) {
2819 $repo_id = $Git::SVN::default_repo_id;
2820 }
2821 unless (defined $ref_id && length $ref_id) {
2822 $_[2] = $ref_id = $Git::SVN::default_ref_id;
2823 }
2824 $_[1] = $repo_id = sanitize_remote_name($repo_id);
2825 my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
2826 $_[3] = $path = '' unless (defined $path);
2827 mkpath(["$ENV{GIT_DIR}/svn"]);
2828 bless {
2829 ref_id => $ref_id, dir => $dir, index => "$dir/index",
2830 path => $path, config => "$ENV{GIT_DIR}/svn/config",
2831 map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
2832 }
2833
2834 # for read-only access of old .rev_db formats
2835 sub unlink_rev_db_symlink {
2836 my ($self) = @_;
2837 my $link = $self->rev_db_path;
2838 $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
2839 if (-l $link) {
2840 unlink $link or croak "unlink: $link failed!";
2841 }
2842 }
2843
2844 sub rev_db_path {
2845 my ($self, $uuid) = @_;
2846 my $db_path = $self->map_path($uuid);
2847 $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
2848 or croak "map_path: $db_path does not contain '/.rev_map.' !";
2849 $db_path;
2850 }
2851
2852 # the new replacement for .rev_db
2853 sub map_path {
2854 my ($self, $uuid) = @_;
2855 $uuid ||= $self->ra_uuid;
2856 "$self->{map_root}.$uuid";
2857 }
2858
2859 sub uri_encode {
2860 my ($f) = @_;
2861 $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
2862 $f
2863 }
2864
2865 sub remove_username {
2866 $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
2867 }
2868
2869 package Git::SVN::Prompt;
2870 use strict;
2871 use warnings;
2872 require SVN::Core;
2873 use vars qw/$_no_auth_cache $_username/;
2874
2875 sub simple {
2876 my ($cred, $realm, $default_username, $may_save, $pool) = @_;
2877 $may_save = undef if $_no_auth_cache;
2878 $default_username = $_username if defined $_username;
2879 if (defined $default_username && length $default_username) {
2880 if (defined $realm && length $realm) {
2881 print STDERR "Authentication realm: $realm\n";
2882 STDERR->flush;
2883 }
2884 $cred->username($default_username);
2885 } else {
2886 username($cred, $realm, $may_save, $pool);
2887 }
2888 $cred->password(_read_password("Password for '" .
2889 $cred->username . "': ", $realm));
2890 $cred->may_save($may_save);
2891 $SVN::_Core::SVN_NO_ERROR;
2892 }
2893
2894 sub ssl_server_trust {
2895 my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
2896 $may_save = undef if $_no_auth_cache;
2897 print STDERR "Error validating server certificate for '$realm':\n";
2898 {
2899 no warnings 'once';
2900 # All variables SVN::Auth::SSL::* are used only once,
2901 # so we're shutting up Perl warnings about this.
2902 if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
2903 print STDERR " - The certificate is not issued ",
2904 "by a trusted authority. Use the\n",
2905 " fingerprint to validate ",
2906 "the certificate manually!\n";
2907 }
2908 if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
2909 print STDERR " - The certificate hostname ",
2910 "does not match.\n";
2911 }
2912 if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
2913 print STDERR " - The certificate is not yet valid.\n";
2914 }
2915 if ($failures & $SVN::Auth::SSL::EXPIRED) {
2916 print STDERR " - The certificate has expired.\n";
2917 }
2918 if ($failures & $SVN::Auth::SSL::OTHER) {
2919 print STDERR " - The certificate has ",
2920 "an unknown error.\n";
2921 }
2922 } # no warnings 'once'
2923 printf STDERR
2924 "Certificate information:\n".
2925 " - Hostname: %s\n".
2926 " - Valid: from %s until %s\n".
2927 " - Issuer: %s\n".
2928 " - Fingerprint: %s\n",
2929 map $cert_info->$_, qw(hostname valid_from valid_until
2930 issuer_dname fingerprint);
2931 my $choice;
2932 prompt:
2933 print STDERR $may_save ?
2934 "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
2935 "(R)eject or accept (t)emporarily? ";
2936 STDERR->flush;
2937 $choice = lc(substr(<STDIN> || 'R', 0, 1));
2938 if ($choice =~ /^t$/i) {
2939 $cred->may_save(undef);
2940 } elsif ($choice =~ /^r$/i) {
2941 return -1;
2942 } elsif ($may_save && $choice =~ /^p$/i) {
2943 $cred->may_save($may_save);
2944 } else {
2945 goto prompt;
2946 }
2947 $cred->accepted_failures($failures);
2948 $SVN::_Core::SVN_NO_ERROR;
2949 }
2950
2951 sub ssl_client_cert {
2952 my ($cred, $realm, $may_save, $pool) = @_;
2953 $may_save = undef if $_no_auth_cache;
2954 print STDERR "Client certificate filename: ";
2955 STDERR->flush;
2956 chomp(my $filename = <STDIN>);
2957 $cred->cert_file($filename);
2958 $cred->may_save($may_save);
2959 $SVN::_Core::SVN_NO_ERROR;
2960 }
2961
2962 sub ssl_client_cert_pw {
2963 my ($cred, $realm, $may_save, $pool) = @_;
2964 $may_save = undef if $_no_auth_cache;
2965 $cred->password(_read_password("Password: ", $realm));
2966 $cred->may_save($may_save);
2967 $SVN::_Core::SVN_NO_ERROR;
2968 }
2969
2970 sub username {
2971 my ($cred, $realm, $may_save, $pool) = @_;
2972 $may_save = undef if $_no_auth_cache;
2973 if (defined $realm && length $realm) {
2974 print STDERR "Authentication realm: $realm\n";
2975 }
2976 my $username;
2977 if (defined $_username) {
2978 $username = $_username;
2979 } else {
2980 print STDERR "Username: ";
2981 STDERR->flush;
2982 chomp($username = <STDIN>);
2983 }
2984 $cred->username($username);
2985 $cred->may_save($may_save);
2986 $SVN::_Core::SVN_NO_ERROR;
2987 }
2988
2989 sub _read_password {
2990 my ($prompt, $realm) = @_;
2991 print STDERR $prompt;
2992 STDERR->flush;
2993 require Term::ReadKey;
2994 Term::ReadKey::ReadMode('noecho');
2995 my $password = '';
2996 while (defined(my $key = Term::ReadKey::ReadKey(0))) {
2997 last if $key =~ /[\012\015]/; # \n\r
2998 $password .= $key;
2999 }
3000 Term::ReadKey::ReadMode('restore');
3001 print STDERR "\n";
3002 STDERR->flush;
3003 $password;
3004 }
3005
3006 package SVN::Git::Fetcher;
3007 use vars qw/@ISA/;
3008 use strict;
3009 use warnings;
3010 use Carp qw/croak/;
3011 use IO::File qw//;
3012
3013 # file baton members: path, mode_a, mode_b, pool, fh, blob, base
3014 sub new {
3015 my ($class, $git_svn) = @_;
3016 my $self = SVN::Delta::Editor->new;
3017 bless $self, $class;
3018 $self->{c} = $git_svn->{last_commit} if exists $git_svn->{last_commit};
3019 $self->{empty} = {};
3020 $self->{dir_prop} = {};
3021 $self->{file_prop} = {};
3022 $self->{absent_dir} = {};
3023 $self->{absent_file} = {};
3024 $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
3025 $self;
3026 }
3027
3028 sub set_path_strip {
3029 my ($self, $path) = @_;
3030 $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
3031 }
3032
3033 sub open_root {
3034 { path => '' };
3035 }
3036
3037 sub open_directory {
3038 my ($self, $path, $pb, $rev) = @_;
3039 { path => $path };
3040 }
3041
3042 sub git_path {
3043 my ($self, $path) = @_;
3044 if ($self->{path_strip}) {
3045 $path =~ s!$self->{path_strip}!! or
3046 die "Failed to strip path '$path' ($self->{path_strip})\n";
3047 }
3048 $path;
3049 }
3050
3051 sub delete_entry {
3052 my ($self, $path, $rev, $pb) = @_;
3053
3054 my $gpath = $self->git_path($path);
3055 return undef if ($gpath eq '');
3056
3057 # remove entire directories.
3058 if (command('ls-tree', $self->{c}, '--', $gpath) =~ /^040000 tree/) {
3059 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3060 -r --name-only -z/,
3061 $self->{c}, '--', $gpath);
3062 local $/ = "\0";
3063 while (<$ls>) {
3064 chomp;
3065 $self->{gii}->remove($_);
3066 print "\tD\t$_\n" unless $::_q;
3067 }
3068 print "\tD\t$gpath/\n" unless $::_q;
3069 command_close_pipe($ls, $ctx);
3070 $self->{empty}->{$path} = 0
3071 } else {
3072 $self->{gii}->remove($gpath);
3073 print "\tD\t$gpath\n" unless $::_q;
3074 }
3075 undef;
3076 }
3077
3078 sub open_file {
3079 my ($self, $path, $pb, $rev) = @_;
3080 my $gpath = $self->git_path($path);
3081 my ($mode, $blob) = (command('ls-tree', $self->{c}, '--', $gpath)
3082 =~ /^(\d{6}) blob ([a-f\d]{40})\t/);
3083 unless (defined $mode && defined $blob) {
3084 die "$path was not found in commit $self->{c} (r$rev)\n";
3085 }
3086 { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3087 pool => SVN::Pool->new, action => 'M' };
3088 }
3089
3090 sub add_file {
3091 my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3092 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3093 delete $self->{empty}->{$dir};
3094 { path => $path, mode_a => 100644, mode_b => 100644,
3095 pool => SVN::Pool->new, action => 'A' };
3096 }
3097
3098 sub add_directory {
3099 my ($self, $path, $cp_path, $cp_rev) = @_;
3100 my $gpath = $self->git_path($path);
3101 if ($gpath eq '') {
3102 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3103 -r --name-only -z/,
3104 $self->{c});
3105 local $/ = "\0";
3106 while (<$ls>) {
3107 chomp;
3108 $self->{gii}->remove($_);
3109 print "\tD\t$_\n" unless $::_q;
3110 }
3111 command_close_pipe($ls, $ctx);
3112 $self->{empty}->{$path} = 0;
3113 }
3114 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3115 delete $self->{empty}->{$dir};
3116 $self->{empty}->{$path} = 1;
3117 { path => $path };
3118 }
3119
3120 sub change_dir_prop {
3121 my ($self, $db, $prop, $value) = @_;
3122 $self->{dir_prop}->{$db->{path}} ||= {};
3123 $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3124 undef;
3125 }
3126
3127 sub absent_directory {
3128 my ($self, $path, $pb) = @_;
3129 $self->{absent_dir}->{$pb->{path}} ||= [];
3130 push @{$self->{absent_dir}->{$pb->{path}}}, $path;
3131 undef;
3132 }
3133
3134 sub absent_file {
3135 my ($self, $path, $pb) = @_;
3136 $self->{absent_file}->{$pb->{path}} ||= [];
3137 push @{$self->{absent_file}->{$pb->{path}}}, $path;
3138 undef;
3139 }
3140
3141 sub change_file_prop {
3142 my ($self, $fb, $prop, $value) = @_;
3143 if ($prop eq 'svn:executable') {
3144 if ($fb->{mode_b} != 120000) {
3145 $fb->{mode_b} = defined $value ? 100755 : 100644;
3146 }
3147 } elsif ($prop eq 'svn:special') {
3148 $fb->{mode_b} = defined $value ? 120000 : 100644;
3149 } else {
3150 $self->{file_prop}->{$fb->{path}} ||= {};
3151 $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
3152 }
3153 undef;
3154 }
3155
3156 sub apply_textdelta {
3157 my ($self, $fb, $exp) = @_;
3158 my $fh = IO::File->new_tmpfile;
3159 $fh->autoflush(1);
3160 # $fh gets auto-closed() by SVN::TxDelta::apply(),
3161 # (but $base does not,) so dup() it for reading in close_file
3162 open my $dup, '<&', $fh or croak $!;
3163 my $base = IO::File->new_tmpfile;
3164 $base->autoflush(1);
3165 if ($fb->{blob}) {
3166 defined (my $pid = fork) or croak $!;
3167 if (!$pid) {
3168 open STDOUT, '>&', $base or croak $!;
3169 print STDOUT 'link ' if ($fb->{mode_a} == 120000);
3170 exec qw/git-cat-file blob/, $fb->{blob} or croak $!;
3171 }
3172 waitpid $pid, 0;
3173 croak $? if $?;
3174
3175 if (defined $exp) {
3176 seek $base, 0, 0 or croak $!;
3177 my $got = ::md5sum($base);
3178 die "Checksum mismatch: $fb->{path} $fb->{blob}\n",
3179 "expected: $exp\n",
3180 " got: $got\n" if ($got ne $exp);
3181 }
3182 }
3183 seek $base, 0, 0 or croak $!;
3184 $fb->{fh} = $dup;
3185 $fb->{base} = $base;
3186 [ SVN::TxDelta::apply($base, $fh, undef, $fb->{path}, $fb->{pool}) ];
3187 }
3188
3189 sub close_file {
3190 my ($self, $fb, $exp) = @_;
3191 my $hash;
3192 my $path = $self->git_path($fb->{path});
3193 if (my $fh = $fb->{fh}) {
3194 if (defined $exp) {
3195 seek($fh, 0, 0) or croak $!;
3196 my $got = ::md5sum($fh);
3197 if ($got ne $exp) {
3198 die "Checksum mismatch: $path\n",
3199 "expected: $exp\n got: $got\n";
3200 }
3201 }
3202 sysseek($fh, 0, 0) or croak $!;
3203 if ($fb->{mode_b} == 120000) {
3204 eval {
3205 sysread($fh, my $buf, 5) == 5 or croak $!;
3206 $buf eq 'link ' or die "$path has mode 120000",
3207 " but is not a link";
3208 };
3209 if ($@) {
3210 warn "$@\n";
3211 sysseek($fh, 0, 0) or croak $!;
3212 }
3213 }
3214 defined(my $pid = open my $out,'-|') or die "Can't fork: $!\n";
3215 if (!$pid) {
3216 open STDIN, '<&', $fh or croak $!;
3217 exec qw/git-hash-object -w --stdin/ or croak $!;
3218 }
3219 chomp($hash = do { local $/; <$out> });
3220 close $out or croak $!;
3221 close $fh or croak $!;
3222 $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
3223 close $fb->{base} or croak $!;
3224 } else {
3225 $hash = $fb->{blob} or die "no blob information\n";
3226 }
3227 $fb->{pool}->clear;
3228 $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
3229 print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
3230 undef;
3231 }
3232
3233 sub abort_edit {
3234 my $self = shift;
3235 $self->{nr} = $self->{gii}->{nr};
3236 delete $self->{gii};
3237 $self->SUPER::abort_edit(@_);
3238 }
3239
3240 sub close_edit {
3241 my $self = shift;
3242 $self->{git_commit_ok} = 1;
3243 $self->{nr} = $self->{gii}->{nr};
3244 delete $self->{gii};
3245 $self->SUPER::close_edit(@_);
3246 }
3247
3248 package SVN::Git::Editor;
3249 use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
3250 use strict;
3251 use warnings;
3252 use Carp qw/croak/;
3253 use IO::File;
3254
3255 sub new {
3256 my ($class, $opts) = @_;
3257 foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
3258 die "$_ required!\n" unless (defined $opts->{$_});
3259 }
3260
3261 my $pool = SVN::Pool->new;
3262 my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
3263 my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
3264 $opts->{r}, $mods);
3265
3266 # $opts->{ra} functions should not be used after this:
3267 my @ce = $opts->{ra}->get_commit_editor($opts->{log},
3268 $opts->{editor_cb}, $pool);
3269 my $self = SVN::Delta::Editor->new(@ce, $pool);
3270 bless $self, $class;
3271 foreach (qw/svn_path r tree_a tree_b/) {
3272 $self->{$_} = $opts->{$_};
3273 }
3274 $self->{url} = $opts->{ra}->{url};
3275 $self->{mods} = $mods;
3276 $self->{types} = $types;
3277 $self->{pool} = $pool;
3278 $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
3279 $self->{rm} = { };
3280 $self->{path_prefix} = length $self->{svn_path} ?
3281 "$self->{svn_path}/" : '';
3282 return $self;
3283 }
3284
3285 sub generate_diff {
3286 my ($tree_a, $tree_b) = @_;
3287 my @diff_tree = qw(diff-tree -z -r);
3288 if ($_cp_similarity) {
3289 push @diff_tree, "-C$_cp_similarity";
3290 } else {
3291 push @diff_tree, '-C';
3292 }
3293 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
3294 push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
3295 push @diff_tree, $tree_a, $tree_b;
3296 my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
3297 local $/ = "\0";
3298 my $state = 'meta';
3299 my @mods;
3300 while (<$diff_fh>) {
3301 chomp $_; # this gets rid of the trailing "\0"
3302 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
3303 $::sha1\s($::sha1)\s
3304 ([MTCRAD])\d*$/xo) {
3305 push @mods, { mode_a => $1, mode_b => $2,
3306 sha1_b => $3, chg => $4 };
3307 if ($4 =~ /^(?:C|R)$/) {
3308 $state = 'file_a';
3309 } else {
3310 $state = 'file_b';
3311 }
3312 } elsif ($state eq 'file_a') {
3313 my $x = $mods[$#mods] or croak "Empty array\n";
3314 if ($x->{chg} !~ /^(?:C|R)$/) {
3315 croak "Error parsing $_, $x->{chg}\n";
3316 }
3317 $x->{file_a} = $_;
3318 $state = 'file_b';
3319 } elsif ($state eq 'file_b') {
3320 my $x = $mods[$#mods] or croak "Empty array\n";
3321 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
3322 croak "Error parsing $_, $x->{chg}\n";
3323 }
3324 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
3325 croak "Error parsing $_, $x->{chg}\n";
3326 }
3327 $x->{file_b} = $_;
3328 $state = 'meta';
3329 } else {
3330 croak "Error parsing $_\n";
3331 }
3332 }
3333 command_close_pipe($diff_fh, $ctx);
3334 \@mods;
3335 }
3336
3337 sub check_diff_paths {
3338 my ($ra, $pfx, $rev, $mods) = @_;
3339 my %types;
3340 $pfx .= '/' if length $pfx;
3341
3342 sub type_diff_paths {
3343 my ($ra, $types, $path, $rev) = @_;
3344 my @p = split m#/+#, $path;
3345 my $c = shift @p;
3346 unless (defined $types->{$c}) {
3347 $types->{$c} = $ra->check_path($c, $rev);
3348 }
3349 while (@p) {
3350 $c .= '/' . shift @p;
3351 next if defined $types->{$c};
3352 $types->{$c} = $ra->check_path($c, $rev);
3353 }
3354 }
3355
3356 foreach my $m (@$mods) {
3357 foreach my $f (qw/file_a file_b/) {
3358 next unless defined $m->{$f};
3359 my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
3360 if (length $pfx.$dir && ! defined $types{$dir}) {
3361 type_diff_paths($ra, \%types, $pfx.$dir, $rev);
3362 }
3363 }
3364 }
3365 \%types;
3366 }
3367
3368 sub split_path {
3369 return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
3370 }
3371
3372 sub repo_path {
3373 my ($self, $path) = @_;
3374 $self->{path_prefix}.(defined $path ? $path : '');
3375 }
3376
3377 sub url_path {
3378 my ($self, $path) = @_;
3379 if ($self->{url} =~ m#^https?://#) {
3380 $path =~ s/([^a-zA-Z0-9_.-])/uc sprintf("%%%02x",ord($1))/eg;
3381 }
3382 $self->{url} . '/' . $self->repo_path($path);
3383 }
3384
3385 sub rmdirs {
3386 my ($self) = @_;
3387 my $rm = $self->{rm};
3388 delete $rm->{''}; # we never delete the url we're tracking
3389 return unless %$rm;
3390
3391 foreach (keys %$rm) {
3392 my @d = split m#/#, $_;
3393 my $c = shift @d;
3394 $rm->{$c} = 1;
3395 while (@d) {
3396 $c .= '/' . shift @d;
3397 $rm->{$c} = 1;
3398 }
3399 }
3400 delete $rm->{$self->{svn_path}};
3401 delete $rm->{''}; # we never delete the url we're tracking
3402 return unless %$rm;
3403
3404 my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
3405 $self->{tree_b});
3406 local $/ = "\0";
3407 while (<$fh>) {
3408 chomp;
3409 my @dn = split m#/#, $_;
3410 while (pop @dn) {
3411 delete $rm->{join '/', @dn};
3412 }
3413 unless (%$rm) {
3414 close $fh;
3415 return;
3416 }
3417 }
3418 command_close_pipe($fh, $ctx);
3419
3420 my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
3421 foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
3422 $self->close_directory($bat->{$d}, $p);
3423 my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
3424 print "\tD+\t$d/\n" unless $::_q;
3425 $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
3426 delete $bat->{$d};
3427 }
3428 }
3429
3430 sub open_or_add_dir {
3431 my ($self, $full_path, $baton) = @_;
3432 my $t = $self->{types}->{$full_path};
3433 if (!defined $t) {
3434 die "$full_path not known in r$self->{r} or we have a bug!\n";
3435 }
3436 {
3437 no warnings 'once';
3438 # SVN::Node::none and SVN::Node::file are used only once,
3439 # so we're shutting up Perl's warnings about them.
3440 if ($t == $SVN::Node::none) {
3441 return $self->add_directory($full_path, $baton,
3442 undef, -1, $self->{pool});
3443 } elsif ($t == $SVN::Node::dir) {
3444 return $self->open_directory($full_path, $baton,
3445 $self->{r}, $self->{pool});
3446 } # no warnings 'once'
3447 print STDERR "$full_path already exists in repository at ",
3448 "r$self->{r} and it is not a directory (",
3449 ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
3450 } # no warnings 'once'
3451 exit 1;
3452 }
3453
3454 sub ensure_path {
3455 my ($self, $path) = @_;
3456 my $bat = $self->{bat};
3457 my $repo_path = $self->repo_path($path);
3458 return $bat->{''} unless (length $repo_path);
3459 my @p = split m#/+#, $repo_path;
3460 my $c = shift @p;
3461 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
3462 while (@p) {
3463 my $c0 = $c;
3464 $c .= '/' . shift @p;
3465 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
3466 }
3467 return $bat->{$c};
3468 }
3469
3470 sub A {
3471 my ($self, $m) = @_;
3472 my ($dir, $file) = split_path($m->{file_b});
3473 my $pbat = $self->ensure_path($dir);
3474 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3475 undef, -1);
3476 print "\tA\t$m->{file_b}\n" unless $::_q;
3477 $self->chg_file($fbat, $m);
3478 $self->close_file($fbat,undef,$self->{pool});
3479 }
3480
3481 sub C {
3482 my ($self, $m) = @_;
3483 my ($dir, $file) = split_path($m->{file_b});
3484 my $pbat = $self->ensure_path($dir);
3485 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3486 $self->url_path($m->{file_a}), $self->{r});
3487 print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3488 $self->chg_file($fbat, $m);
3489 $self->close_file($fbat,undef,$self->{pool});
3490 }
3491
3492 sub delete_entry {
3493 my ($self, $path, $pbat) = @_;
3494 my $rpath = $self->repo_path($path);
3495 my ($dir, $file) = split_path($rpath);
3496 $self->{rm}->{$dir} = 1;
3497 $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
3498 }
3499
3500 sub R {
3501 my ($self, $m) = @_;
3502 my ($dir, $file) = split_path($m->{file_b});
3503 my $pbat = $self->ensure_path($dir);
3504 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3505 $self->url_path($m->{file_a}), $self->{r});
3506 print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3507 $self->chg_file($fbat, $m);
3508 $self->close_file($fbat,undef,$self->{pool});
3509
3510 ($dir, $file) = split_path($m->{file_a});
3511 $pbat = $self->ensure_path($dir);
3512 $self->delete_entry($m->{file_a}, $pbat);
3513 }
3514
3515 sub M {
3516 my ($self, $m) = @_;
3517 my ($dir, $file) = split_path($m->{file_b});
3518 my $pbat = $self->ensure_path($dir);
3519 my $fbat = $self->open_file($self->repo_path($m->{file_b}),
3520 $pbat,$self->{r},$self->{pool});
3521 print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
3522 $self->chg_file($fbat, $m);
3523 $self->close_file($fbat,undef,$self->{pool});
3524 }
3525
3526 sub T { shift->M(@_) }
3527
3528 sub change_file_prop {
3529 my ($self, $fbat, $pname, $pval) = @_;
3530 $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
3531 }
3532
3533 sub chg_file {
3534 my ($self, $fbat, $m) = @_;
3535 if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
3536 $self->change_file_prop($fbat,'svn:executable','*');
3537 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
3538 $self->change_file_prop($fbat,'svn:executable',undef);
3539 }
3540 my $fh = IO::File->new_tmpfile or croak $!;
3541 if ($m->{mode_b} =~ /^120/) {
3542 print $fh 'link ' or croak $!;
3543 $self->change_file_prop($fbat,'svn:special','*');
3544 } elsif ($m->{mode_a} =~ /^120/ && $m->{mode_b} !~ /^120/) {
3545 $self->change_file_prop($fbat,'svn:special',undef);
3546 }
3547 defined(my $pid = fork) or croak $!;
3548 if (!$pid) {
3549 open STDOUT, '>&', $fh or croak $!;
3550 exec qw/git-cat-file blob/, $m->{sha1_b} or croak $!;
3551 }
3552 waitpid $pid, 0;
3553 croak $? if $?;
3554 $fh->flush == 0 or croak $!;
3555 seek $fh, 0, 0 or croak $!;
3556
3557 my $exp = ::md5sum($fh);
3558 seek $fh, 0, 0 or croak $!;
3559
3560 my $pool = SVN::Pool->new;
3561 my $atd = $self->apply_textdelta($fbat, undef, $pool);
3562 my $got = SVN::TxDelta::send_stream($fh, @$atd, $pool);
3563 die "Checksum mismatch\nexpected: $exp\ngot: $got\n" if ($got ne $exp);
3564 $pool->clear;
3565
3566 close $fh or croak $!;
3567 }
3568
3569 sub D {
3570 my ($self, $m) = @_;
3571 my ($dir, $file) = split_path($m->{file_b});
3572 my $pbat = $self->ensure_path($dir);
3573 print "\tD\t$m->{file_b}\n" unless $::_q;
3574 $self->delete_entry($m->{file_b}, $pbat);
3575 }
3576
3577 sub close_edit {
3578 my ($self) = @_;
3579 my ($p,$bat) = ($self->{pool}, $self->{bat});
3580 foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
3581 next if $_ eq '';
3582 $self->close_directory($bat->{$_}, $p);
3583 }
3584 $self->close_directory($bat->{''}, $p);
3585 $self->SUPER::close_edit($p);
3586 $p->clear;
3587 }
3588
3589 sub abort_edit {
3590 my ($self) = @_;
3591 $self->SUPER::abort_edit($self->{pool});
3592 }
3593
3594 sub DESTROY {
3595 my $self = shift;
3596 $self->SUPER::DESTROY(@_);
3597 $self->{pool}->clear;
3598 }
3599
3600 # this drives the editor
3601 sub apply_diff {
3602 my ($self) = @_;
3603 my $mods = $self->{mods};
3604 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
3605 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
3606 my $f = $m->{chg};
3607 if (defined $o{$f}) {
3608 $self->$f($m);
3609 } else {
3610 fatal("Invalid change type: $f");
3611 }
3612 }
3613 $self->rmdirs if $_rmdir;
3614 if (@$mods == 0) {
3615 $self->abort_edit;
3616 } else {
3617 $self->close_edit;
3618 }
3619 return scalar @$mods;
3620 }
3621
3622 package Git::SVN::Ra;
3623 use vars qw/@ISA $config_dir $_log_window_size/;
3624 use strict;
3625 use warnings;
3626 my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
3627
3628 BEGIN {
3629 # enforce temporary pool usage for some simple functions
3630 no strict 'refs';
3631 for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root/) {
3632 my $SUPER = "SUPER::$f";
3633 *$f = sub {
3634 my $self = shift;
3635 my $pool = SVN::Pool->new;
3636 my @ret = $self->$SUPER(@_,$pool);
3637 $pool->clear;
3638 wantarray ? @ret : $ret[0];
3639 };
3640 }
3641 }
3642
3643 sub _auth_providers () {
3644 [
3645 SVN::Client::get_simple_provider(),
3646 SVN::Client::get_ssl_server_trust_file_provider(),
3647 SVN::Client::get_simple_prompt_provider(
3648 \&Git::SVN::Prompt::simple, 2),
3649 SVN::Client::get_ssl_client_cert_file_provider(),
3650 SVN::Client::get_ssl_client_cert_prompt_provider(
3651 \&Git::SVN::Prompt::ssl_client_cert, 2),
3652 SVN::Client::get_ssl_client_cert_pw_file_provider(),
3653 SVN::Client::get_ssl_client_cert_pw_prompt_provider(
3654 \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
3655 SVN::Client::get_username_provider(),
3656 SVN::Client::get_ssl_server_trust_prompt_provider(
3657 \&Git::SVN::Prompt::ssl_server_trust),
3658 SVN::Client::get_username_prompt_provider(
3659 \&Git::SVN::Prompt::username, 2)
3660 ]
3661 }
3662
3663 sub escape_uri_only {
3664 my ($uri) = @_;
3665 my @tmp;
3666 foreach (split m{/}, $uri) {
3667 s/([^\w.-])/sprintf("%%%02X",ord($1))/eg;
3668 push @tmp, $_;
3669 }
3670 join('/', @tmp);
3671 }
3672
3673 sub escape_url {
3674 my ($url) = @_;
3675 if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
3676 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
3677 $url = "$scheme://$domain$uri";
3678 }
3679 $url;
3680 }
3681
3682 sub new {
3683 my ($class, $url) = @_;
3684 $url =~ s!/+$!!;
3685 return $RA if ($RA && $RA->{url} eq $url);
3686
3687 SVN::_Core::svn_config_ensure($config_dir, undef);
3688 my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
3689 my $config = SVN::Core::config_get_config($config_dir);
3690 $RA = undef;
3691 my $dont_store_passwords = 1;
3692 my $conf_t = ${$config}{'config'};
3693 {
3694 no warnings 'once';
3695 # The usage of $SVN::_Core::SVN_CONFIG_* variables
3696 # produces warnings that variables are used only once.
3697 # I had not found the better way to shut them up, so
3698 # the warnings of type 'once' are disabled in this block.
3699 if (SVN::_Core::svn_config_get_bool($conf_t,
3700 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3701 $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
3702 1) == 0) {
3703 SVN::_Core::svn_auth_set_parameter($baton,
3704 $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
3705 bless (\$dont_store_passwords, "_p_void"));
3706 }
3707 if (SVN::_Core::svn_config_get_bool($conf_t,
3708 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3709 $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
3710 1) == 0) {
3711 $Git::SVN::Prompt::_no_auth_cache = 1;
3712 }
3713 } # no warnings 'once'
3714 my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
3715 config => $config,
3716 pool => SVN::Pool->new,
3717 auth_provider_callbacks => $callbacks);
3718 $self->{url} = $url;
3719 $self->{svn_path} = $url;
3720 $self->{repos_root} = $self->get_repos_root;
3721 $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
3722 $self->{cache} = { check_path => { r => 0, data => {} },
3723 get_dir => { r => 0, data => {} } };
3724 $RA = bless $self, $class;
3725 }
3726
3727 sub check_path {
3728 my ($self, $path, $r) = @_;
3729 my $cache = $self->{cache}->{check_path};
3730 if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
3731 return $cache->{data}->{$path};
3732 }
3733 my $pool = SVN::Pool->new;
3734 my $t = $self->SUPER::check_path($path, $r, $pool);
3735 $pool->clear;
3736 if ($r != $cache->{r}) {
3737 %{$cache->{data}} = ();
3738 $cache->{r} = $r;
3739 }
3740 $cache->{data}->{$path} = $t;
3741 }
3742
3743 sub get_dir {
3744 my ($self, $dir, $r) = @_;
3745 my $cache = $self->{cache}->{get_dir};
3746 if ($r == $cache->{r}) {
3747 if (my $x = $cache->{data}->{$dir}) {
3748 return wantarray ? @$x : $x->[0];
3749 }
3750 }
3751 my $pool = SVN::Pool->new;
3752 my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
3753 my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
3754 $pool->clear;
3755 if ($r != $cache->{r}) {
3756 %{$cache->{data}} = ();
3757 $cache->{r} = $r;
3758 }
3759 $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
3760 wantarray ? (\%dirents, $r, $props) : \%dirents;
3761 }
3762
3763 sub DESTROY {
3764 # do not call the real DESTROY since we store ourselves in $RA
3765 }
3766
3767 sub get_log {
3768 my ($self, @args) = @_;
3769 my $pool = SVN::Pool->new;
3770 splice(@args, 3, 1) if ($SVN::Core::VERSION le '1.2.0');
3771 my $ret = $self->SUPER::get_log(@args, $pool);
3772 $pool->clear;
3773 $ret;
3774 }
3775
3776 sub trees_match {
3777 my ($self, $url1, $rev1, $url2, $rev2) = @_;
3778 my $ctx = SVN::Client->new(auth => _auth_providers);
3779 my $out = IO::File->new_tmpfile;
3780
3781 # older SVN (1.1.x) doesn't take $pool as the last parameter for
3782 # $ctx->diff(), so we'll create a default one
3783 my $pool = SVN::Pool->new_default_sub;
3784
3785 $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
3786 $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
3787 $out->flush;
3788 my $ret = (($out->stat)[7] == 0);
3789 close $out or croak $!;
3790
3791 $ret;
3792 }
3793
3794 sub get_commit_editor {
3795 my ($self, $log, $cb, $pool) = @_;
3796 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
3797 $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
3798 }
3799
3800 sub gs_do_update {
3801 my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
3802 my $new = ($rev_a == $rev_b);
3803 my $path = $gs->{path};
3804
3805 if ($new && -e $gs->{index}) {
3806 unlink $gs->{index} or die
3807 "Couldn't unlink index: $gs->{index}: $!\n";
3808 }
3809 my $pool = SVN::Pool->new;
3810 $editor->set_path_strip($path);
3811 my (@pc) = split m#/#, $path;
3812 my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
3813 1, $editor, $pool);
3814 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3815
3816 # Since we can't rely on svn_ra_reparent being available, we'll
3817 # just have to do some magic with set_path to make it so
3818 # we only want a partial path.
3819 my $sp = '';
3820 my $final = join('/', @pc);
3821 while (@pc) {
3822 $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
3823 $sp .= '/' if length $sp;
3824 $sp .= shift @pc;
3825 }
3826 die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
3827
3828 $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
3829
3830 $reporter->finish_report($pool);
3831 $pool->clear;
3832 $editor->{git_commit_ok};
3833 }
3834
3835 # this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
3836 # svn_ra_reparent didn't work before 1.4)
3837 sub gs_do_switch {
3838 my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
3839 my $path = $gs->{path};
3840 my $pool = SVN::Pool->new;
3841
3842 my $full_url = $self->{url};
3843 my $old_url = $full_url;
3844 $full_url .= '/' . escape_uri_only($path) if length $path;
3845 my ($ra, $reparented);
3846 if ($old_url ne $full_url) {
3847 if ($old_url !~ m#^svn(\+ssh)?://#) {
3848 SVN::_Ra::svn_ra_reparent($self->{session}, $full_url,
3849 $pool);
3850 $self->{url} = $full_url;
3851 $reparented = 1;
3852 } else {
3853 $_[0] = undef;
3854 $self = undef;
3855 $RA = undef;
3856 $ra = Git::SVN::Ra->new($full_url);
3857 $ra_invalid = 1;
3858 }
3859 }
3860 $ra ||= $self;
3861 my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
3862 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3863 $reporter->set_path('', $rev_a, 0, @lock, $pool);
3864 $reporter->finish_report($pool);
3865
3866 if ($reparented) {
3867 SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
3868 $self->{url} = $old_url;
3869 }
3870
3871 $pool->clear;
3872 $editor->{git_commit_ok};
3873 }
3874
3875 sub longest_common_path {
3876 my ($gsv, $globs) = @_;
3877 my %common;
3878 my $common_max = scalar @$gsv;
3879
3880 foreach my $gs (@$gsv) {
3881 my @tmp = split m#/#, $gs->{path};
3882 my $p = '';
3883 foreach (@tmp) {
3884 $p .= length($p) ? "/$_" : $_;
3885 $common{$p} ||= 0;
3886 $common{$p}++;
3887 }
3888 }
3889 $globs ||= [];
3890 $common_max += scalar @$globs;
3891 foreach my $glob (@$globs) {
3892 my @tmp = split m#/#, $glob->{path}->{left};
3893 my $p = '';
3894 foreach (@tmp) {
3895 $p .= length($p) ? "/$_" : $_;
3896 $common{$p} ||= 0;
3897 $common{$p}++;
3898 }
3899 }
3900
3901 my $longest_path = '';
3902 foreach (sort {length $b <=> length $a} keys %common) {
3903 if ($common{$_} == $common_max) {
3904 $longest_path = $_;
3905 last;
3906 }
3907 }
3908 $longest_path;
3909 }
3910
3911 sub gs_fetch_loop_common {
3912 my ($self, $base, $head, $gsv, $globs) = @_;
3913 return if ($base > $head);
3914 my $inc = $_log_window_size;
3915 my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
3916 my $longest_path = longest_common_path($gsv, $globs);
3917 my $ra_url = $self->{url};
3918 while (1) {
3919 my %revs;
3920 my $err;
3921 my $err_handler = $SVN::Error::handler;
3922 $SVN::Error::handler = sub {
3923 ($err) = @_;
3924 skip_unknown_revs($err);
3925 };
3926 sub _cb {
3927 my ($paths, $r, $author, $date, $log) = @_;
3928 [ dup_changed_paths($paths),
3929 { author => $author, date => $date, log => $log } ];
3930 }
3931 $self->get_log([$longest_path], $min, $max, 0, 1, 1,
3932 sub { $revs{$_[1]} = _cb(@_) });
3933 if ($err && $max >= $head) {
3934 print STDERR "Path '$longest_path' ",
3935 "was probably deleted:\n",
3936 $err->expanded_message,
3937 "\nWill attempt to follow ",
3938 "revisions r$min .. r$max ",
3939 "committed before the deletion\n";
3940 my $hi = $max;
3941 while (--$hi >= $min) {
3942 my $ok;
3943 $self->get_log([$longest_path], $min, $hi,
3944 0, 1, 1, sub {
3945 $ok ||= $_[1];
3946 $revs{$_[1]} = _cb(@_) });
3947 if ($ok) {
3948 print STDERR "r$min .. r$ok OK\n";
3949 last;
3950 }
3951 }
3952 }
3953 $SVN::Error::handler = $err_handler;
3954
3955 my %exists = map { $_->{path} => $_ } @$gsv;
3956 foreach my $r (sort {$a <=> $b} keys %revs) {
3957 my ($paths, $logged) = @{$revs{$r}};
3958
3959 foreach my $gs ($self->match_globs(\%exists, $paths,
3960 $globs, $r)) {
3961 if ($gs->rev_map_max >= $r) {
3962 next;
3963 }
3964 next unless $gs->match_paths($paths, $r);
3965 $gs->{logged_rev_props} = $logged;
3966 if (my $last_commit = $gs->last_commit) {
3967 $gs->assert_index_clean($last_commit);
3968 }
3969 my $log_entry = $gs->do_fetch($paths, $r);
3970 if ($log_entry) {
3971 $gs->do_git_commit($log_entry);
3972 }
3973 $INDEX_FILES{$gs->{index}} = 1;
3974 }
3975 foreach my $g (@$globs) {
3976 my $k = "svn-remote.$g->{remote}." .
3977 "$g->{t}-maxRev";
3978 Git::SVN::tmp_config($k, $r);
3979 }
3980 if ($ra_invalid) {
3981 $_[0] = undef;
3982 $self = undef;
3983 $RA = undef;
3984 $self = Git::SVN::Ra->new($ra_url);
3985 $ra_invalid = undef;
3986 }
3987 }
3988 # pre-fill the .rev_db since it'll eventually get filled in
3989 # with '0' x40 if something new gets committed
3990 foreach my $gs (@$gsv) {
3991 next if $gs->rev_map_max >= $max;
3992 next if defined $gs->rev_map_get($max);
3993 $gs->rev_map_set($max, 0 x40);
3994 }
3995 foreach my $g (@$globs) {
3996 my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
3997 Git::SVN::tmp_config($k, $max);
3998 }
3999 last if $max >= $head;
4000 $min = $max + 1;
4001 $max += $inc;
4002 $max = $head if ($max > $head);
4003 }
4004 Git::SVN::gc();
4005 }
4006
4007 sub match_globs {
4008 my ($self, $exists, $paths, $globs, $r) = @_;
4009
4010 sub get_dir_check {
4011 my ($self, $exists, $g, $r) = @_;
4012 my @x = eval { $self->get_dir($g->{path}->{left}, $r) };
4013 return unless scalar @x == 3;
4014 my $dirents = $x[0];
4015 foreach my $de (keys %$dirents) {
4016 next if $dirents->{$de}->{kind} != $SVN::Node::dir;
4017 my $p = $g->{path}->full_path($de);
4018 next if $exists->{$p};
4019 next if (length $g->{path}->{right} &&
4020 ($self->check_path($p, $r) !=
4021 $SVN::Node::dir));
4022 $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
4023 $g->{ref}->full_path($de), 1);
4024 }
4025 }
4026 foreach my $g (@$globs) {
4027 if (my $path = $paths->{"/$g->{path}->{left}"}) {
4028 if ($path->{action} =~ /^[AR]$/) {
4029 get_dir_check($self, $exists, $g, $r);
4030 }
4031 }
4032 foreach (keys %$paths) {
4033 if (/$g->{path}->{left_regex}/ &&
4034 !/$g->{path}->{regex}/) {
4035 next if $paths->{$_}->{action} !~ /^[AR]$/;
4036 get_dir_check($self, $exists, $g, $r);
4037 }
4038 next unless /$g->{path}->{regex}/;
4039 my $p = $1;
4040 my $pathname = $g->{path}->full_path($p);
4041 next if $exists->{$pathname};
4042 next if ($self->check_path($pathname, $r) !=
4043 $SVN::Node::dir);
4044 $exists->{$pathname} = Git::SVN->init(
4045 $self->{url}, $pathname, undef,
4046 $g->{ref}->full_path($p), 1);
4047 }
4048 my $c = '';
4049 foreach (split m#/#, $g->{path}->{left}) {
4050 $c .= "/$_";
4051 next unless ($paths->{$c} &&
4052 ($paths->{$c}->{action} =~ /^[AR]$/));
4053 get_dir_check($self, $exists, $g, $r);
4054 }
4055 }
4056 values %$exists;
4057 }
4058
4059 sub minimize_url {
4060 my ($self) = @_;
4061 return $self->{url} if ($self->{url} eq $self->{repos_root});
4062 my $url = $self->{repos_root};
4063 my @components = split(m!/!, $self->{svn_path});
4064 my $c = '';
4065 do {
4066 $url .= "/$c" if length $c;
4067 eval { (ref $self)->new($url)->get_latest_revnum };
4068 } while ($@ && ($c = shift @components));
4069 $url;
4070 }
4071
4072 sub can_do_switch {
4073 my $self = shift;
4074 unless (defined $can_do_switch) {
4075 my $pool = SVN::Pool->new;
4076 my $rep = eval {
4077 $self->do_switch(1, '', 0, $self->{url},
4078 SVN::Delta::Editor->new, $pool);
4079 };
4080 if ($@) {
4081 $can_do_switch = 0;
4082 } else {
4083 $rep->abort_report($pool);
4084 $can_do_switch = 1;
4085 }
4086 $pool->clear;
4087 }
4088 $can_do_switch;
4089 }
4090
4091 sub skip_unknown_revs {
4092 my ($err) = @_;
4093 my $errno = $err->apr_err();
4094 # Maybe the branch we're tracking didn't
4095 # exist when the repo started, so it's
4096 # not an error if it doesn't, just continue
4097 #
4098 # Wonderfully consistent library, eh?
4099 # 160013 - svn:// and file://
4100 # 175002 - http(s)://
4101 # 175007 - http(s):// (this repo required authorization, too...)
4102 # More codes may be discovered later...
4103 if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
4104 my $err_key = $err->expanded_message;
4105 # revision numbers change every time, filter them out
4106 $err_key =~ s/\d+/\0/g;
4107 $err_key = "$errno\0$err_key";
4108 unless ($ignored_err{$err_key}) {
4109 warn "W: Ignoring error from SVN, path probably ",
4110 "does not exist: ($errno): ",
4111 $err->expanded_message,"\n";
4112 warn "W: Do not be alarmed at the above message ",
4113 "git-svn is just searching aggressively for ",
4114 "old history.\n",
4115 "This may take a while on large repositories\n";
4116 $ignored_err{$err_key} = 1;
4117 }
4118 return;
4119 }
4120 die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
4121 }
4122
4123 # svn_log_changed_path_t objects passed to get_log are likely to be
4124 # overwritten even if only the refs are copied to an external variable,
4125 # so we should dup the structures in their entirety. Using an externally
4126 # passed pool (instead of our temporary and quickly cleared pool in
4127 # Git::SVN::Ra) does not help matters at all...
4128 sub dup_changed_paths {
4129 my ($paths) = @_;
4130 return undef unless $paths;
4131 my %ret;
4132 foreach my $p (keys %$paths) {
4133 my $i = $paths->{$p};
4134 my %s = map { $_ => $i->$_ }
4135 qw/copyfrom_path copyfrom_rev action/;
4136 $ret{$p} = \%s;
4137 }
4138 \%ret;
4139 }
4140
4141 package Git::SVN::Log;
4142 use strict;
4143 use warnings;
4144 use POSIX qw/strftime/;
4145 use constant commit_log_separator => ('-' x 72) . "\n";
4146 use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
4147 %rusers $show_commit $incremental/;
4148 my $l_fmt;
4149
4150 sub cmt_showable {
4151 my ($c) = @_;
4152 return 1 if defined $c->{r};
4153
4154 # big commit message got truncated by the 16k pretty buffer in rev-list
4155 if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
4156 $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
4157 @{$c->{l}} = ();
4158 my @log = command(qw/cat-file commit/, $c->{c});
4159
4160 # shift off the headers
4161 shift @log while ($log[0] ne '');
4162 shift @log;
4163
4164 # TODO: make $c->{l} not have a trailing newline in the future
4165 @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
4166
4167 (undef, $c->{r}, undef) = ::extract_metadata(
4168 (grep(/^git-svn-id: /, @log))[-1]);
4169 }
4170 return defined $c->{r};
4171 }
4172
4173 sub log_use_color {
4174 return $color || Git->repository->get_colorbool('color.diff');
4175 }
4176
4177 sub git_svn_log_cmd {
4178 my ($r_min, $r_max, @args) = @_;
4179 my $head = 'HEAD';
4180 my (@files, @log_opts);
4181 foreach my $x (@args) {
4182 if ($x eq '--' || @files) {
4183 push @files, $x;
4184 } else {
4185 if (::verify_ref("$x^0")) {
4186 $head = $x;
4187 } else {
4188 push @log_opts, $x;
4189 }
4190 }
4191 }
4192
4193 my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
4194 $gs ||= Git::SVN->_new;
4195 my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
4196 $gs->refname);
4197 push @cmd, '-r' unless $non_recursive;
4198 push @cmd, qw/--raw --name-status/ if $verbose;
4199 push @cmd, '--color' if log_use_color();
4200 push @cmd, @log_opts;
4201 if (defined $r_max && $r_max == $r_min) {
4202 push @cmd, '--max-count=1';
4203 if (my $c = $gs->rev_map_get($r_max)) {
4204 push @cmd, $c;
4205 }
4206 } elsif (defined $r_max) {
4207 if ($r_max < $r_min) {
4208 ($r_min, $r_max) = ($r_max, $r_min);
4209 }
4210 my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
4211 my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
4212 # If there are no commits in the range, both $c_max and $c_min
4213 # will be undefined. If there is at least 1 commit in the
4214 # range, both will be defined.
4215 return () if !defined $c_min || !defined $c_max;
4216 if ($c_min eq $c_max) {
4217 push @cmd, '--max-count=1', $c_min;
4218 } else {
4219 push @cmd, '--boundary', "$c_min..$c_max";
4220 }
4221 }
4222 return (@cmd, @files);
4223 }
4224
4225 # adapted from pager.c
4226 sub config_pager {
4227 $pager ||= $ENV{GIT_PAGER} || $ENV{PAGER};
4228 if (!defined $pager) {
4229 $pager = 'less';
4230 } elsif (length $pager == 0 || $pager eq 'cat') {
4231 $pager = undef;
4232 }
4233 $ENV{GIT_PAGER_IN_USE} = defined($pager);
4234 }
4235
4236 sub run_pager {
4237 return unless -t *STDOUT && defined $pager;
4238 pipe my $rfd, my $wfd or return;
4239 defined(my $pid = fork) or ::fatal "Can't fork: $!";
4240 if (!$pid) {
4241 open STDOUT, '>&', $wfd or
4242 ::fatal "Can't redirect to stdout: $!";
4243 return;
4244 }
4245 open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
4246 $ENV{LESS} ||= 'FRSX';
4247 exec $pager or ::fatal "Can't run pager: $! ($pager)";
4248 }
4249
4250 sub format_svn_date {
4251 return strftime("%Y-%m-%d %H:%M:%S %z (%a, %d %b %Y)", localtime(shift));
4252 }
4253
4254 sub parse_git_date {
4255 my ($t, $tz) = @_;
4256 # Date::Parse isn't in the standard Perl distro :(
4257 if ($tz =~ s/^\+//) {
4258 $t += tz_to_s_offset($tz);
4259 } elsif ($tz =~ s/^\-//) {
4260 $t -= tz_to_s_offset($tz);
4261 }
4262 return $t;
4263 }
4264
4265 sub set_local_timezone {
4266 if (defined $TZ) {
4267 $ENV{TZ} = $TZ;
4268 } else {
4269 delete $ENV{TZ};
4270 }
4271 }
4272
4273 sub tz_to_s_offset {
4274 my ($tz) = @_;
4275 $tz =~ s/(\d\d)$//;
4276 return ($1 * 60) + ($tz * 3600);
4277 }
4278
4279 sub get_author_info {
4280 my ($dest, $author, $t, $tz) = @_;
4281 $author =~ s/(?:^\s*|\s*$)//g;
4282 $dest->{a_raw} = $author;
4283 my $au;
4284 if ($::_authors) {
4285 $au = $rusers{$author} || undef;
4286 }
4287 if (!$au) {
4288 ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
4289 }
4290 $dest->{t} = $t;
4291 $dest->{tz} = $tz;
4292 $dest->{a} = $au;
4293 $dest->{t_utc} = parse_git_date($t, $tz);
4294 }
4295
4296 sub process_commit {
4297 my ($c, $r_min, $r_max, $defer) = @_;
4298 if (defined $r_min && defined $r_max) {
4299 if ($r_min == $c->{r} && $r_min == $r_max) {
4300 show_commit($c);
4301 return 0;
4302 }
4303 return 1 if $r_min == $r_max;
4304 if ($r_min < $r_max) {
4305 # we need to reverse the print order
4306 return 0 if (defined $limit && --$limit < 0);
4307 push @$defer, $c;
4308 return 1;
4309 }
4310 if ($r_min != $r_max) {
4311 return 1 if ($r_min < $c->{r});
4312 return 1 if ($r_max > $c->{r});
4313 }
4314 }
4315 return 0 if (defined $limit && --$limit < 0);
4316 show_commit($c);
4317 return 1;
4318 }
4319
4320 sub show_commit {
4321 my $c = shift;
4322 if ($oneline) {
4323 my $x = "\n";
4324 if (my $l = $c->{l}) {
4325 while ($l->[0] =~ /^\s*$/) { shift @$l }
4326 $x = $l->[0];
4327 }
4328 $l_fmt ||= 'A' . length($c->{r});
4329 print 'r',pack($l_fmt, $c->{r}),' | ';
4330 print "$c->{c} | " if $show_commit;
4331 print $x;
4332 } else {
4333 show_commit_normal($c);
4334 }
4335 }
4336
4337 sub show_commit_changed_paths {
4338 my ($c) = @_;
4339 return unless $c->{changed};
4340 print "Changed paths:\n", @{$c->{changed}};
4341 }
4342
4343 sub show_commit_normal {
4344 my ($c) = @_;
4345 print commit_log_separator, "r$c->{r} | ";
4346 print "$c->{c} | " if $show_commit;
4347 print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
4348 my $nr_line = 0;
4349
4350 if (my $l = $c->{l}) {
4351 while ($l->[$#$l] eq "\n" && $#$l > 0
4352 && $l->[($#$l - 1)] eq "\n") {
4353 pop @$l;
4354 }
4355 $nr_line = scalar @$l;
4356 if (!$nr_line) {
4357 print "1 line\n\n\n";
4358 } else {
4359 if ($nr_line == 1) {
4360 $nr_line = '1 line';
4361 } else {
4362 $nr_line .= ' lines';
4363 }
4364 print $nr_line, "\n";
4365 show_commit_changed_paths($c);
4366 print "\n";
4367 print $_ foreach @$l;
4368 }
4369 } else {
4370 print "1 line\n";
4371 show_commit_changed_paths($c);
4372 print "\n";
4373
4374 }
4375 foreach my $x (qw/raw stat diff/) {
4376 if ($c->{$x}) {
4377 print "\n";
4378 print $_ foreach @{$c->{$x}}
4379 }
4380 }
4381 }
4382
4383 sub cmd_show_log {
4384 my (@args) = @_;
4385 my ($r_min, $r_max);
4386 my $r_last = -1; # prevent dupes
4387 set_local_timezone();
4388 if (defined $::_revision) {
4389 if ($::_revision =~ /^(\d+):(\d+)$/) {
4390 ($r_min, $r_max) = ($1, $2);
4391 } elsif ($::_revision =~ /^\d+$/) {
4392 $r_min = $r_max = $::_revision;
4393 } else {
4394 ::fatal "-r$::_revision is not supported, use ",
4395 "standard 'git log' arguments instead";
4396 }
4397 }
4398
4399 config_pager();
4400 @args = git_svn_log_cmd($r_min, $r_max, @args);
4401 if (!@args) {
4402 print commit_log_separator unless $incremental || $oneline;
4403 return;
4404 }
4405 my $log = command_output_pipe(@args);
4406 run_pager();
4407 my (@k, $c, $d, $stat);
4408 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
4409 while (<$log>) {
4410 if (/^${esc_color}commit -?($::sha1_short)/o) {
4411 my $cmt = $1;
4412 if ($c && cmt_showable($c) && $c->{r} != $r_last) {
4413 $r_last = $c->{r};
4414 process_commit($c, $r_min, $r_max, \@k) or
4415 goto out;
4416 }
4417 $d = undef;
4418 $c = { c => $cmt };
4419 } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
4420 get_author_info($c, $1, $2, $3);
4421 } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
4422 # ignore
4423 } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
4424 push @{$c->{raw}}, $_;
4425 } elsif (/^${esc_color}[ACRMDT]\t/) {
4426 # we could add $SVN->{svn_path} here, but that requires
4427 # remote access at the moment (repo_path_split)...
4428 s#^(${esc_color})([ACRMDT])\t#$1 $2 #o;
4429 push @{$c->{changed}}, $_;
4430 } elsif (/^${esc_color}diff /o) {
4431 $d = 1;
4432 push @{$c->{diff}}, $_;
4433 } elsif ($d) {
4434 push @{$c->{diff}}, $_;
4435 } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
4436 $esc_color*[\+\-]*$esc_color$/x) {
4437 $stat = 1;
4438 push @{$c->{stat}}, $_;
4439 } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
4440 push @{$c->{stat}}, $_;
4441 $stat = undef;
4442 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
4443 ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
4444 } elsif (s/^${esc_color} //o) {
4445 push @{$c->{l}}, $_;
4446 }
4447 }
4448 if ($c && defined $c->{r} && $c->{r} != $r_last) {
4449 $r_last = $c->{r};
4450 process_commit($c, $r_min, $r_max, \@k);
4451 }
4452 if (@k) {
4453 ($r_min, $r_max) = ($r_max, $r_min);
4454 process_commit($_, $r_min, $r_max) foreach reverse @k;
4455 }
4456 out:
4457 close $log;
4458 print commit_log_separator unless $incremental || $oneline;
4459 }
4460
4461 sub cmd_blame {
4462 my $path = shift;
4463
4464 config_pager();
4465 run_pager();
4466
4467 my ($fh, $ctx) = command_output_pipe('blame', @_, $path);
4468 while (my $line = <$fh>) {
4469 if ($line =~ /^\^?([[:xdigit:]]+)\s/) {
4470 my (undef, $rev, undef) = ::cmt_metadata($1);
4471 $rev = sprintf('%-10s', $rev);
4472 $line =~ s/^\^?[[:xdigit:]]+(\s)/$rev$1/;
4473 }
4474 print $line;
4475 }
4476 command_close_pipe($fh, $ctx);
4477 }
4478
4479 package Git::SVN::Migration;
4480 # these version numbers do NOT correspond to actual version numbers
4481 # of git nor git-svn. They are just relative.
4482 #
4483 # v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
4484 #
4485 # v1 layout: .git/$id/info/url, refs/remotes/$id
4486 #
4487 # v2 layout: .git/svn/$id/info/url, refs/remotes/$id
4488 #
4489 # v3 layout: .git/svn/$id, refs/remotes/$id
4490 # - info/url may remain for backwards compatibility
4491 # - this is what we migrate up to this layout automatically,
4492 # - this will be used by git svn init on single branches
4493 # v3.1 layout (auto migrated):
4494 # - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
4495 # for backwards compatibility
4496 #
4497 # v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
4498 # - this is only created for newly multi-init-ed
4499 # repositories. Similar in spirit to the
4500 # --use-separate-remotes option in git-clone (now default)
4501 # - we do not automatically migrate to this (following
4502 # the example set by core git)
4503 #
4504 # v5 layout: .rev_db.$UUID => .rev_map.$UUID
4505 # - newer, more-efficient format that uses 24-bytes per record
4506 # with no filler space.
4507 # - use xxd -c24 < .rev_map.$UUID to view and debug
4508 # - This is a one-way migration, repositories updated to the
4509 # new format will not be able to use old git-svn without
4510 # rebuilding the .rev_db. Rebuilding the rev_db is not
4511 # possible if noMetadata or useSvmProps are set; but should
4512 # be no problem for users that use the (sensible) defaults.
4513 use strict;
4514 use warnings;
4515 use Carp qw/croak/;
4516 use File::Path qw/mkpath/;
4517 use File::Basename qw/dirname basename/;
4518 use vars qw/$_minimize/;
4519
4520 sub migrate_from_v0 {
4521 my $git_dir = $ENV{GIT_DIR};
4522 return undef unless -d $git_dir;
4523 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4524 my $migrated = 0;
4525 while (<$fh>) {
4526 chomp;
4527 my ($id, $orig_ref) = ($_, $_);
4528 next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
4529 next unless -f "$git_dir/$id/info/url";
4530 my $new_ref = "refs/remotes/$id";
4531 if (::verify_ref("$new_ref^0")) {
4532 print STDERR "W: $orig_ref is probably an old ",
4533 "branch used by an ancient version of ",
4534 "git-svn.\n",
4535 "However, $new_ref also exists.\n",
4536 "We will not be able ",
4537 "to use this branch until this ",
4538 "ambiguity is resolved.\n";
4539 next;
4540 }
4541 print STDERR "Migrating from v0 layout...\n" if !$migrated;
4542 print STDERR "Renaming ref: $orig_ref => $new_ref\n";
4543 command_noisy('update-ref', $new_ref, $orig_ref);
4544 command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
4545 $migrated++;
4546 }
4547 command_close_pipe($fh, $ctx);
4548 print STDERR "Done migrating from v0 layout...\n" if $migrated;
4549 $migrated;
4550 }
4551
4552 sub migrate_from_v1 {
4553 my $git_dir = $ENV{GIT_DIR};
4554 my $migrated = 0;
4555 return $migrated unless -d $git_dir;
4556 my $svn_dir = "$git_dir/svn";
4557
4558 # just in case somebody used 'svn' as their $id at some point...
4559 return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
4560
4561 print STDERR "Migrating from a git-svn v1 layout...\n";
4562 mkpath([$svn_dir]);
4563 print STDERR "Data from a previous version of git-svn exists, but\n\t",
4564 "$svn_dir\n\t(required for this version ",
4565 "($::VERSION) of git-svn) does not. exist\n";
4566 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4567 while (<$fh>) {
4568 my $x = $_;
4569 next unless $x =~ s#^refs/remotes/##;
4570 chomp $x;
4571 next unless -f "$git_dir/$x/info/url";
4572 my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
4573 next unless $u;
4574 my $dn = dirname("$git_dir/svn/$x");
4575 mkpath([$dn]) unless -d $dn;
4576 if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
4577 mkpath(["$git_dir/svn/svn"]);
4578 print STDERR " - $git_dir/$x/info => ",
4579 "$git_dir/svn/$x/info\n";
4580 rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
4581 croak "$!: $x";
4582 # don't worry too much about these, they probably
4583 # don't exist with repos this old (save for index,
4584 # and we can easily regenerate that)
4585 foreach my $f (qw/unhandled.log index .rev_db/) {
4586 rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
4587 }
4588 } else {
4589 print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
4590 rename "$git_dir/$x", "$git_dir/svn/$x" or
4591 croak "$!: $x";
4592 }
4593 $migrated++;
4594 }
4595 command_close_pipe($fh, $ctx);
4596 print STDERR "Done migrating from a git-svn v1 layout\n";
4597 $migrated;
4598 }
4599
4600 sub read_old_urls {
4601 my ($l_map, $pfx, $path) = @_;
4602 my @dir;
4603 foreach (<$path/*>) {
4604 if (-r "$_/info/url") {
4605 $pfx .= '/' if $pfx && $pfx !~ m!/$!;
4606 my $ref_id = $pfx . basename $_;
4607 my $url = ::file_to_s("$_/info/url");
4608 $l_map->{$ref_id} = $url;
4609 } elsif (-d $_) {
4610 push @dir, $_;
4611 }
4612 }
4613 foreach (@dir) {
4614 my $x = $_;
4615 $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
4616 read_old_urls($l_map, $x, $_);
4617 }
4618 }
4619
4620 sub migrate_from_v2 {
4621 my @cfg = command(qw/config -l/);
4622 return if grep /^svn-remote\..+\.url=/, @cfg;
4623 my %l_map;
4624 read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
4625 my $migrated = 0;
4626
4627 foreach my $ref_id (sort keys %l_map) {
4628 eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
4629 if ($@) {
4630 Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
4631 }
4632 $migrated++;
4633 }
4634 $migrated;
4635 }
4636
4637 sub minimize_connections {
4638 my $r = Git::SVN::read_all_remotes();
4639 my $new_urls = {};
4640 my $root_repos = {};
4641 foreach my $repo_id (keys %$r) {
4642 my $url = $r->{$repo_id}->{url} or next;
4643 my $fetch = $r->{$repo_id}->{fetch} or next;
4644 my $ra = Git::SVN::Ra->new($url);
4645
4646 # skip existing cases where we already connect to the root
4647 if (($ra->{url} eq $ra->{repos_root}) ||
4648 (Git::SVN::sanitize_remote_name($ra->{repos_root}) eq
4649 $repo_id)) {
4650 $root_repos->{$ra->{url}} = $repo_id;
4651 next;
4652 }
4653
4654 my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
4655 my $root_path = $ra->{url};
4656 $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
4657 foreach my $path (keys %$fetch) {
4658 my $ref_id = $fetch->{$path};
4659 my $gs = Git::SVN->new($ref_id, $repo_id, $path);
4660
4661 # make sure we can read when connecting to
4662 # a higher level of a repository
4663 my ($last_rev, undef) = $gs->last_rev_commit;
4664 if (!defined $last_rev) {
4665 $last_rev = eval {
4666 $root_ra->get_latest_revnum;
4667 };
4668 next if $@;
4669 }
4670 my $new = $root_path;
4671 $new .= length $path ? "/$path" : '';
4672 eval {
4673 $root_ra->get_log([$new], $last_rev, $last_rev,
4674 0, 0, 1, sub { });
4675 };
4676 next if $@;
4677 $new_urls->{$ra->{repos_root}}->{$new} =
4678 { ref_id => $ref_id,
4679 old_repo_id => $repo_id,
4680 old_path => $path };
4681 }
4682 }
4683
4684 my @emptied;
4685 foreach my $url (keys %$new_urls) {
4686 # see if we can re-use an existing [svn-remote "repo_id"]
4687 # instead of creating a(n ugly) new section:
4688 my $repo_id = $root_repos->{$url} ||
4689 Git::SVN::sanitize_remote_name($url);
4690
4691 my $fetch = $new_urls->{$url};
4692 foreach my $path (keys %$fetch) {
4693 my $x = $fetch->{$path};
4694 Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
4695 my $pfx = "svn-remote.$x->{old_repo_id}";
4696
4697 my $old_fetch = quotemeta("$x->{old_path}:".
4698 "refs/remotes/$x->{ref_id}");
4699 command_noisy(qw/config --unset/,
4700 "$pfx.fetch", '^'. $old_fetch . '$');
4701 delete $r->{$x->{old_repo_id}}->
4702 {fetch}->{$x->{old_path}};
4703 if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
4704 command_noisy(qw/config --unset/,
4705 "$pfx.url");
4706 push @emptied, $x->{old_repo_id}
4707 }
4708 }
4709 }
4710 if (@emptied) {
4711 my $file = $ENV{GIT_CONFIG} || $ENV{GIT_CONFIG_LOCAL} ||
4712 "$ENV{GIT_DIR}/config";
4713 print STDERR <<EOF;
4714 The following [svn-remote] sections in your config file ($file) are empty
4715 and can be safely removed:
4716 EOF
4717 print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
4718 }
4719 }
4720
4721 sub migration_check {
4722 migrate_from_v0();
4723 migrate_from_v1();
4724 migrate_from_v2();
4725 minimize_connections() if $_minimize;
4726 }
4727
4728 package Git::IndexInfo;
4729 use strict;
4730 use warnings;
4731 use Git qw/command_input_pipe command_close_pipe/;
4732
4733 sub new {
4734 my ($class) = @_;
4735 my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
4736 bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
4737 }
4738
4739 sub remove {
4740 my ($self, $path) = @_;
4741 if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
4742 return ++$self->{nr};
4743 }
4744 undef;
4745 }
4746
4747 sub update {
4748 my ($self, $mode, $hash, $path) = @_;
4749 if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
4750 return ++$self->{nr};
4751 }
4752 undef;
4753 }
4754
4755 sub DESTROY {
4756 my ($self) = @_;
4757 command_close_pipe($self->{gui}, $self->{ctx});
4758 }
4759
4760 package Git::SVN::GlobSpec;
4761 use strict;
4762 use warnings;
4763
4764 sub new {
4765 my ($class, $glob) = @_;
4766 my $re = $glob;
4767 $re =~ s!/+$!!g; # no need for trailing slashes
4768 my $nr = ($re =~ s!^(.*)\*(.*)$!\(\[^/\]+\)!g);
4769 my ($left, $right) = ($1, $2);
4770 if ($nr > 1) {
4771 die "Only one '*' wildcard expansion ",
4772 "is supported (got $nr): '$glob'\n";
4773 } elsif ($nr == 0) {
4774 die "One '*' is needed for glob: '$glob'\n";
4775 }
4776 $re = quotemeta($left) . $re . quotemeta($right);
4777 if (length $left && !($left =~ s!/+$!!g)) {
4778 die "Missing trailing '/' on left side of: '$glob' ($left)\n";
4779 }
4780 if (length $right && !($right =~ s!^/+!!g)) {
4781 die "Missing leading '/' on right side of: '$glob' ($right)\n";
4782 }
4783 my $left_re = qr/^\/\Q$left\E(\/|$)/;
4784 bless { left => $left, right => $right, left_regex => $left_re,
4785 regex => qr/$re/, glob => $glob }, $class;
4786 }
4787
4788 sub full_path {
4789 my ($self, $path) = @_;
4790 return (length $self->{left} ? "$self->{left}/" : '') .
4791 $path . (length $self->{right} ? "/$self->{right}" : '');
4792 }
4793
4794 __END__
4795
4796 Data structures:
4797
4798
4799 $remotes = { # returned by read_all_remotes()
4800 'svn' => {
4801 # svn-remote.svn.url=https://svn.musicpd.org
4802 url => 'https://svn.musicpd.org',
4803 # svn-remote.svn.fetch=mpd/trunk:trunk
4804 fetch => {
4805 'mpd/trunk' => 'trunk',
4806 },
4807 # svn-remote.svn.tags=mpd/tags/*:tags/*
4808 tags => {
4809 path => {
4810 left => 'mpd/tags',
4811 right => '',
4812 regex => qr!mpd/tags/([^/]+)$!,
4813 glob => 'tags/*',
4814 },
4815 ref => {
4816 left => 'tags',
4817 right => '',
4818 regex => qr!tags/([^/]+)$!,
4819 glob => 'tags/*',
4820 },
4821 }
4822 }
4823 };
4824
4825 $log_entry hashref as returned by libsvn_log_entry()
4826 {
4827 log => 'whitespace-formatted log entry
4828 ', # trailing newline is preserved
4829 revision => '8', # integer
4830 date => '2004-02-24T17:01:44.108345Z', # commit date
4831 author => 'committer name'
4832 };
4833
4834
4835 # this is generated by generate_diff();
4836 @mods = array of diff-index line hashes, each element represents one line
4837 of diff-index output
4838
4839 diff-index line ($m hash)
4840 {
4841 mode_a => first column of diff-index output, no leading ':',
4842 mode_b => second column of diff-index output,
4843 sha1_b => sha1sum of the final blob,
4844 chg => change type [MCRADT],
4845 file_a => original file name of a file (iff chg is 'C' or 'R')
4846 file_b => new/current file name of a file (any chg)
4847 }
4848 ;
4849
4850 # retval of read_url_paths{,_all}();
4851 $l_map = {
4852 # repository root url
4853 'https://svn.musicpd.org' => {
4854 # repository path # GIT_SVN_ID
4855 'mpd/trunk' => 'trunk',
4856 'mpd/tags/0.11.5' => 'tags/0.11.5',
4857 },
4858 }
4859
4860 Notes:
4861 I don't trust the each() function on unless I created %hash myself
4862 because the internal iterator may not have started at base.
Dscho's miscellaneous git branches