?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
format-patch: support deep threading
[git/jrn.git] / git-svn.perl
blobcbc5211d58a37f71b5c4566b9375ee398a1612ba
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 $_repository
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_format, $_commit_url, $_tag);
70 $Git::SVN::_follow_parent = 1;
71 my %remote_opts = ( 'username=s' => \$Git::SVN::Prompt::_username,
72 'config-dir=s' => \$Git::SVN::Ra::config_dir,
73 'no-auth-cache' => \$Git::SVN::Prompt::_no_auth_cache,
74 'ignore-paths=s' => \$SVN::Git::Fetcher::_ignore_regex );
75 my %fc_opts = ( 'follow-parent|follow!' => \$Git::SVN::_follow_parent,
76 'authors-file|A=s' => \$_authors,
77 'repack:i' => \$Git::SVN::_repack,
78 'noMetadata' => \$Git::SVN::_no_metadata,
79 'useSvmProps' => \$Git::SVN::_use_svm_props,
80 'useSvnsyncProps' => \$Git::SVN::_use_svnsync_props,
81 'log-window-size=i' => \$Git::SVN::Ra::_log_window_size,
82 'no-checkout' => \$_no_checkout,
83 'quiet|q' => \$_q,
84 'repack-flags|repack-args|repack-opts=s' =>
85 \$Git::SVN::_repack_flags,
86 'use-log-author' => \$Git::SVN::_use_log_author,
87 'add-author-from' => \$Git::SVN::_add_author_from,
88 'localtime' => \$Git::SVN::_localtime,
89 %remote_opts );
90
91 my ($_trunk, $_tags, $_branches, $_stdlayout);
92 my %icv;
93 my %init_opts = ( 'template=s' => \$_template, 'shared:s' => \$_shared,
94 'trunk|T=s' => \$_trunk, 'tags|t=s' => \$_tags,
95 'branches|b=s' => \$_branches, 'prefix=s' => \$_prefix,
96 'stdlayout|s' => \$_stdlayout,
97 'minimize-url|m' => \$Git::SVN::_minimize_url,
98 'no-metadata' => sub { $icv{noMetadata} = 1 },
99 'use-svm-props' => sub { $icv{useSvmProps} = 1 },
100 'use-svnsync-props' => sub { $icv{useSvnsyncProps} = 1 },
101 'rewrite-root=s' => sub { $icv{rewriteRoot} = $_[1] },
102 %remote_opts );
103 my %cmt_opts = ( 'edit|e' => \$_edit,
104 'rmdir' => \$SVN::Git::Editor::_rmdir,
105 'find-copies-harder' => \$SVN::Git::Editor::_find_copies_harder,
106 'l=i' => \$SVN::Git::Editor::_rename_limit,
107 'copy-similarity|C=i'=> \$SVN::Git::Editor::_cp_similarity
108 );
109
110 my %cmd = (
111 fetch => [ \&cmd_fetch, "Download new revisions from SVN",
112 { 'revision|r=s' => \$_revision,
113 'fetch-all|all' => \$_fetch_all,
114 %fc_opts } ],
115 clone => [ \&cmd_clone, "Initialize and fetch revisions",
116 { 'revision|r=s' => \$_revision,
117 %fc_opts, %init_opts } ],
118 init => [ \&cmd_init, "Initialize a repo for tracking" .
119 " (requires URL argument)",
120 \%init_opts ],
121 'multi-init' => [ \&cmd_multi_init,
122 "Deprecated alias for ".
123 "'$0 init -T<trunk> -b<branches> -t<tags>'",
124 \%init_opts ],
125 dcommit => [ \&cmd_dcommit,
126 'Commit several diffs to merge with upstream',
127 { 'merge|m|M' => \$_merge,
128 'strategy|s=s' => \$_strategy,
129 'verbose|v' => \$_verbose,
130 'dry-run|n' => \$_dry_run,
131 'fetch-all|all' => \$_fetch_all,
132 'commit-url=s' => \$_commit_url,
133 'revision|r=i' => \$_revision,
134 'no-rebase' => \$_no_rebase,
135 %cmt_opts, %fc_opts } ],
136 branch => [ \&cmd_branch,
137 'Create a branch in the SVN repository',
138 { 'message|m=s' => \$_message,
139 'dry-run|n' => \$_dry_run,
140 'tag|t' => \$_tag } ],
141 tag => [ sub { $_tag = 1; cmd_branch(@_) },
142 'Create a tag in the SVN repository',
143 { 'message|m=s' => \$_message,
144 'dry-run|n' => \$_dry_run } ],
145 'set-tree' => [ \&cmd_set_tree,
146 "Set an SVN repository to a git tree-ish",
147 { 'stdin|' => \$_stdin, %cmt_opts, %fc_opts, } ],
148 'create-ignore' => [ \&cmd_create_ignore,
149 'Create a .gitignore per svn:ignore',
150 { 'revision|r=i' => \$_revision
151 } ],
152 'propget' => [ \&cmd_propget,
153 'Print the value of a property on a file or directory',
154 { 'revision|r=i' => \$_revision } ],
155 'proplist' => [ \&cmd_proplist,
156 'List all properties of a file or directory',
157 { 'revision|r=i' => \$_revision } ],
158 'show-ignore' => [ \&cmd_show_ignore, "Show svn:ignore listings",
159 { 'revision|r=i' => \$_revision
160 } ],
161 'show-externals' => [ \&cmd_show_externals, "Show svn:externals listings",
162 { 'revision|r=i' => \$_revision
163 } ],
164 'multi-fetch' => [ \&cmd_multi_fetch,
165 "Deprecated alias for $0 fetch --all",
166 { 'revision|r=s' => \$_revision, %fc_opts } ],
167 'migrate' => [ sub { },
168 # no-op, we automatically run this anyways,
169 'Migrate configuration/metadata/layout from
170 previous versions of git-svn',
171 { 'minimize' => \$Git::SVN::Migration::_minimize,
172 %remote_opts } ],
173 'log' => [ \&Git::SVN::Log::cmd_show_log, 'Show commit logs',
174 { 'limit=i' => \$Git::SVN::Log::limit,
175 'revision|r=s' => \$_revision,
176 'verbose|v' => \$Git::SVN::Log::verbose,
177 'incremental' => \$Git::SVN::Log::incremental,
178 'oneline' => \$Git::SVN::Log::oneline,
179 'show-commit' => \$Git::SVN::Log::show_commit,
180 'non-recursive' => \$Git::SVN::Log::non_recursive,
181 'authors-file|A=s' => \$_authors,
182 'color' => \$Git::SVN::Log::color,
183 'pager=s' => \$Git::SVN::Log::pager
184 } ],
185 'find-rev' => [ \&cmd_find_rev,
186 "Translate between SVN revision numbers and tree-ish",
187 {} ],
188 'rebase' => [ \&cmd_rebase, "Fetch and rebase your working directory",
189 { 'merge|m|M' => \$_merge,
190 'verbose|v' => \$_verbose,
191 'strategy|s=s' => \$_strategy,
192 'local|l' => \$_local,
193 'fetch-all|all' => \$_fetch_all,
194 'dry-run|n' => \$_dry_run,
195 %fc_opts } ],
196 'commit-diff' => [ \&cmd_commit_diff,
197 'Commit a diff between two trees',
198 { 'message|m=s' => \$_message,
199 'file|F=s' => \$_file,
200 'revision|r=s' => \$_revision,
201 %cmt_opts } ],
202 'info' => [ \&cmd_info,
203 "Show info about the latest SVN revision
204 on the current branch",
205 { 'url' => \$_url, } ],
206 'blame' => [ \&Git::SVN::Log::cmd_blame,
207 "Show what revision and author last modified each line of a file",
208 { 'git-format' => \$_git_format } ],
209 );
210
211 my $cmd;
212 for (my $i = 0; $i < @ARGV; $i++) {
213 if (defined $cmd{$ARGV[$i]}) {
214 $cmd = $ARGV[$i];
215 splice @ARGV, $i, 1;
216 last;
217 }
218 };
219
220 # make sure we're always running at the top-level working directory
221 unless ($cmd && $cmd =~ /(?:clone|init|multi-init)$/) {
222 unless (-d $ENV{GIT_DIR}) {
223 if ($git_dir_user_set) {
224 die "GIT_DIR=$ENV{GIT_DIR} explicitly set, ",
225 "but it is not a directory\n";
226 }
227 my $git_dir = delete $ENV{GIT_DIR};
228 my $cdup = undef;
229 git_cmd_try {
230 $cdup = command_oneline(qw/rev-parse --show-cdup/);
231 $git_dir = '.' unless ($cdup);
232 chomp $cdup if ($cdup);
233 $cdup = "." unless ($cdup && length $cdup);
234 } "Already at toplevel, but $git_dir not found\n";
235 chdir $cdup or die "Unable to chdir up to '$cdup'\n";
236 unless (-d $git_dir) {
237 die "$git_dir still not found after going to ",
238 "'$cdup'\n";
239 }
240 $ENV{GIT_DIR} = $git_dir;
241 }
242 $_repository = Git->repository(Repository => $ENV{GIT_DIR});
243 }
244
245 my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
246
247 read_repo_config(\%opts);
248 if ($cmd && ($cmd eq 'log' || $cmd eq 'blame')) {
249 Getopt::Long::Configure('pass_through');
250 }
251 my $rv = GetOptions(%opts, 'help|H|h' => \$_help, 'version|V' => \$_version,
252 'minimize-connections' => \$Git::SVN::Migration::_minimize,
253 'id|i=s' => \$Git::SVN::default_ref_id,
254 'svn-remote|remote|R=s' => sub {
255 $Git::SVN::no_reuse_existing = 1;
256 $Git::SVN::default_repo_id = $_[1] });
257 exit 1 if (!$rv && $cmd && $cmd ne 'log');
258
259 usage(0) if $_help;
260 version() if $_version;
261 usage(1) unless defined $cmd;
262 load_authors() if $_authors;
263
264 unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
265 Git::SVN::Migration::migration_check();
266 }
267 Git::SVN::init_vars();
268 eval {
269 Git::SVN::verify_remotes_sanity();
270 $cmd{$cmd}->[0]->(@ARGV);
271 };
272 fatal $@ if $@;
273 post_fetch_checkout();
274 exit 0;
275
276 ####################### primary functions ######################
277 sub usage {
278 my $exit = shift || 0;
279 my $fd = $exit ? \*STDERR : \*STDOUT;
280 print $fd <<"";
281 git-svn - bidirectional operations between a single Subversion tree and git
282 Usage: git svn <command> [options] [arguments]\n
283
284 print $fd "Available commands:\n" unless $cmd;
285
286 foreach (sort keys %cmd) {
287 next if $cmd && $cmd ne $_;
288 next if /^multi-/; # don't show deprecated commands
289 print $fd ' ',pack('A17',$_),$cmd{$_}->[1],"\n";
290 foreach (sort keys %{$cmd{$_}->[2]}) {
291 # mixed-case options are for .git/config only
292 next if /[A-Z]/ && /^[a-z]+$/i;
293 # prints out arguments as they should be passed:
294 my $x = s#[:=]s$## ? '<arg>' : s#[:=]i$## ? '<num>' : '';
295 print $fd ' ' x 21, join(', ', map { length $_ > 1 ?
296 "--$_" : "-$_" }
297 split /\|/,$_)," $x\n";
298 }
299 }
300 print $fd <<"";
301 \nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
302 arbitrary identifier if you're tracking multiple SVN branches/repositories in
303 one git repository and want to keep them separate. See git-svn(1) for more
304 information.
305
306 exit $exit;
307 }
308
309 sub version {
310 print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
311 exit 0;
312 }
313
314 sub do_git_init_db {
315 unless (-d $ENV{GIT_DIR}) {
316 my @init_db = ('init');
317 push @init_db, "--template=$_template" if defined $_template;
318 if (defined $_shared) {
319 if ($_shared =~ /[a-z]/) {
320 push @init_db, "--shared=$_shared";
321 } else {
322 push @init_db, "--shared";
323 }
324 }
325 command_noisy(@init_db);
326 $_repository = Git->repository(Repository => ".git");
327 }
328 my $set;
329 my $pfx = "svn-remote.$Git::SVN::default_repo_id";
330 foreach my $i (keys %icv) {
331 die "'$set' and '$i' cannot both be set\n" if $set;
332 next unless defined $icv{$i};
333 command_noisy('config', "$pfx.$i", $icv{$i});
334 $set = $i;
335 }
336 }
337
338 sub init_subdir {
339 my $repo_path = shift or return;
340 mkpath([$repo_path]) unless -d $repo_path;
341 chdir $repo_path or die "Couldn't chdir to $repo_path: $!\n";
342 $ENV{GIT_DIR} = '.git';
343 $_repository = Git->repository(Repository => $ENV{GIT_DIR});
344 }
345
346 sub cmd_clone {
347 my ($url, $path) = @_;
348 if (!defined $path &&
349 (defined $_trunk || defined $_branches || defined $_tags ||
350 defined $_stdlayout) &&
351 $url !~ m#^[a-z\+]+://#) {
352 $path = $url;
353 }
354 $path = basename($url) if !defined $path || !length $path;
355 cmd_init($url, $path);
356 Git::SVN::fetch_all($Git::SVN::default_repo_id);
357 }
358
359 sub cmd_init {
360 if (defined $_stdlayout) {
361 $_trunk = 'trunk' if (!defined $_trunk);
362 $_tags = 'tags' if (!defined $_tags);
363 $_branches = 'branches' if (!defined $_branches);
364 }
365 if (defined $_trunk || defined $_branches || defined $_tags) {
366 return cmd_multi_init(@_);
367 }
368 my $url = shift or die "SVN repository location required ",
369 "as a command-line argument\n";
370 init_subdir(@_);
371 do_git_init_db();
372
373 Git::SVN->init($url);
374 }
375
376 sub cmd_fetch {
377 if (grep /^\d+=./, @_) {
378 die "'<rev>=<commit>' fetch arguments are ",
379 "no longer supported.\n";
380 }
381 my ($remote) = @_;
382 if (@_ > 1) {
383 die "Usage: $0 fetch [--all] [svn-remote]\n";
384 }
385 $remote ||= $Git::SVN::default_repo_id;
386 if ($_fetch_all) {
387 cmd_multi_fetch();
388 } else {
389 Git::SVN::fetch_all($remote, Git::SVN::read_all_remotes());
390 }
391 }
392
393 sub cmd_set_tree {
394 my (@commits) = @_;
395 if ($_stdin || !@commits) {
396 print "Reading from stdin...\n";
397 @commits = ();
398 while (<STDIN>) {
399 if (/\b($sha1_short)\b/o) {
400 unshift @commits, $1;
401 }
402 }
403 }
404 my @revs;
405 foreach my $c (@commits) {
406 my @tmp = command('rev-parse',$c);
407 if (scalar @tmp == 1) {
408 push @revs, $tmp[0];
409 } elsif (scalar @tmp > 1) {
410 push @revs, reverse(command('rev-list',@tmp));
411 } else {
412 fatal "Failed to rev-parse $c";
413 }
414 }
415 my $gs = Git::SVN->new;
416 my ($r_last, $cmt_last) = $gs->last_rev_commit;
417 $gs->fetch;
418 if (defined $gs->{last_rev} && $r_last != $gs->{last_rev}) {
419 fatal "There are new revisions that were fetched ",
420 "and need to be merged (or acknowledged) ",
421 "before committing.\nlast rev: $r_last\n",
422 " current: $gs->{last_rev}";
423 }
424 $gs->set_tree($_) foreach @revs;
425 print "Done committing ",scalar @revs," revisions to SVN\n";
426 unlink $gs->{index};
427 }
428
429 sub cmd_dcommit {
430 my $head = shift;
431 git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
432 'Cannot dcommit with a dirty index. Commit your changes first, '
433 . "or stash them with `git stash'.\n";
434 $head ||= 'HEAD';
435 my @refs;
436 my ($url, $rev, $uuid, $gs) = working_head_info($head, \@refs);
437 unless ($gs) {
438 die "Unable to determine upstream SVN information from ",
439 "$head history.\nPerhaps the repository is empty.";
440 }
441 $url = defined $_commit_url ? $_commit_url : $gs->full_url;
442 my $last_rev = $_revision if defined $_revision;
443 if ($url) {
444 print "Committing to $url ...\n";
445 }
446 my ($linear_refs, $parents) = linearize_history($gs, \@refs);
447 if ($_no_rebase && scalar(@$linear_refs) > 1) {
448 warn "Attempting to commit more than one change while ",
449 "--no-rebase is enabled.\n",
450 "If these changes depend on each other, re-running ",
451 "without --no-rebase may be required."
452 }
453 my $expect_url = $url;
454 Git::SVN::remove_username($expect_url);
455 while (1) {
456 my $d = shift @$linear_refs or last;
457 unless (defined $last_rev) {
458 (undef, $last_rev, undef) = cmt_metadata("$d~1");
459 unless (defined $last_rev) {
460 fatal "Unable to extract revision information ",
461 "from commit $d~1";
462 }
463 }
464 if ($_dry_run) {
465 print "diff-tree $d~1 $d\n";
466 } else {
467 my $cmt_rev;
468 my %ed_opts = ( r => $last_rev,
469 log => get_commit_entry($d)->{log},
470 ra => Git::SVN::Ra->new($url),
471 config => SVN::Core::config_get_config(
472 $Git::SVN::Ra::config_dir
473 ),
474 tree_a => "$d~1",
475 tree_b => $d,
476 editor_cb => sub {
477 print "Committed r$_[0]\n";
478 $cmt_rev = $_[0];
479 },
480 svn_path => '');
481 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
482 print "No changes\n$d~1 == $d\n";
483 } elsif ($parents->{$d} && @{$parents->{$d}}) {
484 $gs->{inject_parents_dcommit}->{$cmt_rev} =
485 $parents->{$d};
486 }
487 $_fetch_all ? $gs->fetch_all : $gs->fetch;
488 $last_rev = $cmt_rev;
489 next if $_no_rebase;
490
491 # we always want to rebase against the current HEAD,
492 # not any head that was passed to us
493 my @diff = command('diff-tree', $d,
494 $gs->refname, '--');
495 my @finish;
496 if (@diff) {
497 @finish = rebase_cmd();
498 print STDERR "W: $d and ", $gs->refname,
499 " differ, using @finish:\n",
500 join("\n", @diff), "\n";
501 } else {
502 print "No changes between current HEAD and ",
503 $gs->refname,
504 "\nResetting to the latest ",
505 $gs->refname, "\n";
506 @finish = qw/reset --mixed/;
507 }
508 command_noisy(@finish, $gs->refname);
509 if (@diff) {
510 @refs = ();
511 my ($url_, $rev_, $uuid_, $gs_) =
512 working_head_info($head, \@refs);
513 my ($linear_refs_, $parents_) =
514 linearize_history($gs_, \@refs);
515 if (scalar(@$linear_refs) !=
516 scalar(@$linear_refs_)) {
517 fatal "# of revisions changed ",
518 "\nbefore:\n",
519 join("\n", @$linear_refs),
520 "\n\nafter:\n",
521 join("\n", @$linear_refs_), "\n",
522 'If you are attempting to commit ',
523 "merges, try running:\n\t",
524 'git rebase --interactive',
525 '--preserve-merges ',
526 $gs->refname,
527 "\nBefore dcommitting";
528 }
529 if ($url_ ne $expect_url) {
530 fatal "URL mismatch after rebase: ",
531 "$url_ != $expect_url";
532 }
533 if ($uuid_ ne $uuid) {
534 fatal "uuid mismatch after rebase: ",
535 "$uuid_ != $uuid";
536 }
537 # remap parents
538 my (%p, @l, $i);
539 for ($i = 0; $i < scalar @$linear_refs; $i++) {
540 my $new = $linear_refs_->[$i] or next;
541 $p{$new} =
542 $parents->{$linear_refs->[$i]};
543 push @l, $new;
544 }
545 $parents = \%p;
546 $linear_refs = \@l;
547 }
548 }
549 }
550 unlink $gs->{index};
551 }
552
553 sub cmd_branch {
554 my ($branch_name, $head) = @_;
555
556 unless (defined $branch_name && length $branch_name) {
557 die(($_tag ? "tag" : "branch") . " name required\n");
558 }
559 $head ||= 'HEAD';
560
561 my ($src, $rev, undef, $gs) = working_head_info($head);
562
563 my $remote = Git::SVN::read_all_remotes()->{$gs->{repo_id}};
564 my $glob = $remote->{ $_tag ? 'tags' : 'branches' };
565 my ($lft, $rgt) = @{ $glob->{path} }{qw/left right/};
566 my $dst = join '/', $remote->{url}, $lft, $branch_name, ($rgt || ());
567
568 my $ctx = SVN::Client->new(
569 auth => Git::SVN::Ra::_auth_providers(),
570 log_msg => sub {
571 ${ $_[0] } = defined $_message
572 ? $_message
573 : 'Create ' . ($_tag ? 'tag ' : 'branch ' )
574 . $branch_name;
575 },
576 );
577
578 eval {
579 $ctx->ls($dst, 'HEAD', 0);
580 } and die "branch ${branch_name} already exists\n";
581
582 print "Copying ${src} at r${rev} to ${dst}...\n";
583 $ctx->copy($src, $rev, $dst)
584 unless $_dry_run;
585
586 $gs->fetch_all;
587 }
588
589 sub cmd_find_rev {
590 my $revision_or_hash = shift or die "SVN or git revision required ",
591 "as a command-line argument\n";
592 my $result;
593 if ($revision_or_hash =~ /^r\d+$/) {
594 my $head = shift;
595 $head ||= 'HEAD';
596 my @refs;
597 my (undef, undef, $uuid, $gs) = working_head_info($head, \@refs);
598 unless ($gs) {
599 die "Unable to determine upstream SVN information from ",
600 "$head history\n";
601 }
602 my $desired_revision = substr($revision_or_hash, 1);
603 $result = $gs->rev_map_get($desired_revision, $uuid);
604 } else {
605 my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
606 $result = $rev;
607 }
608 print "$result\n" if $result;
609 }
610
611 sub cmd_rebase {
612 command_noisy(qw/update-index --refresh/);
613 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
614 unless ($gs) {
615 die "Unable to determine upstream SVN information from ",
616 "working tree history\n";
617 }
618 if ($_dry_run) {
619 print "Remote Branch: " . $gs->refname . "\n";
620 print "SVN URL: " . $url . "\n";
621 return;
622 }
623 if (command(qw/diff-index HEAD --/)) {
624 print STDERR "Cannot rebase with uncommited changes:\n";
625 command_noisy('status');
626 exit 1;
627 }
628 unless ($_local) {
629 # rebase will checkout for us, so no need to do it explicitly
630 $_no_checkout = 'true';
631 $_fetch_all ? $gs->fetch_all : $gs->fetch;
632 }
633 command_noisy(rebase_cmd(), $gs->refname);
634 }
635
636 sub cmd_show_ignore {
637 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
638 $gs ||= Git::SVN->new;
639 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
640 $gs->prop_walk($gs->{path}, $r, sub {
641 my ($gs, $path, $props) = @_;
642 print STDOUT "\n# $path\n";
643 my $s = $props->{'svn:ignore'} or return;
644 $s =~ s/[\r\n]+/\n/g;
645 chomp $s;
646 $s =~ s#^#$path#gm;
647 print STDOUT "$s\n";
648 });
649 }
650
651 sub cmd_show_externals {
652 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
653 $gs ||= Git::SVN->new;
654 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
655 $gs->prop_walk($gs->{path}, $r, sub {
656 my ($gs, $path, $props) = @_;
657 print STDOUT "\n# $path\n";
658 my $s = $props->{'svn:externals'} or return;
659 $s =~ s/[\r\n]+/\n/g;
660 chomp $s;
661 $s =~ s#^#$path#gm;
662 print STDOUT "$s\n";
663 });
664 }
665
666 sub cmd_create_ignore {
667 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
668 $gs ||= Git::SVN->new;
669 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
670 $gs->prop_walk($gs->{path}, $r, sub {
671 my ($gs, $path, $props) = @_;
672 # $path is of the form /path/to/dir/
673 my $ignore = '.' . $path . '.gitignore';
674 my $s = $props->{'svn:ignore'} or return;
675 open(GITIGNORE, '>', $ignore)
676 or fatal("Failed to open `$ignore' for writing: $!");
677 $s =~ s/[\r\n]+/\n/g;
678 chomp $s;
679 # Prefix all patterns so that the ignore doesn't apply
680 # to sub-directories.
681 $s =~ s#^#/#gm;
682 print GITIGNORE "$s\n";
683 close(GITIGNORE)
684 or fatal("Failed to close `$ignore': $!");
685 command_noisy('add', '-f', $ignore);
686 });
687 }
688
689 sub canonicalize_path {
690 my ($path) = @_;
691 my $dot_slash_added = 0;
692 if (substr($path, 0, 1) ne "/") {
693 $path = "./" . $path;
694 $dot_slash_added = 1;
695 }
696 # File::Spec->canonpath doesn't collapse x/../y into y (for a
697 # good reason), so let's do this manually.
698 $path =~ s#/+#/#g;
699 $path =~ s#/\.(?:/|$)#/#g;
700 $path =~ s#/[^/]+/\.\.##g;
701 $path =~ s#/$##g;
702 $path =~ s#^\./## if $dot_slash_added;
703 $path =~ s#^/##;
704 $path =~ s#^\.$##;
705 return $path;
706 }
707
708 # get_svnprops(PATH)
709 # ------------------
710 # Helper for cmd_propget and cmd_proplist below.
711 sub get_svnprops {
712 my $path = shift;
713 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
714 $gs ||= Git::SVN->new;
715
716 # prefix THE PATH by the sub-directory from which the user
717 # invoked us.
718 $path = $cmd_dir_prefix . $path;
719 fatal("No such file or directory: $path") unless -e $path;
720 my $is_dir = -d $path ? 1 : 0;
721 $path = $gs->{path} . '/' . $path;
722
723 # canonicalize the path (otherwise libsvn will abort or fail to
724 # find the file)
725 $path = canonicalize_path($path);
726
727 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
728 my $props;
729 if ($is_dir) {
730 (undef, undef, $props) = $gs->ra->get_dir($path, $r);
731 }
732 else {
733 (undef, $props) = $gs->ra->get_file($path, $r, undef);
734 }
735 return $props;
736 }
737
738 # cmd_propget (PROP, PATH)
739 # ------------------------
740 # Print the SVN property PROP for PATH.
741 sub cmd_propget {
742 my ($prop, $path) = @_;
743 $path = '.' if not defined $path;
744 usage(1) if not defined $prop;
745 my $props = get_svnprops($path);
746 if (not defined $props->{$prop}) {
747 fatal("`$path' does not have a `$prop' SVN property.");
748 }
749 print $props->{$prop} . "\n";
750 }
751
752 # cmd_proplist (PATH)
753 # -------------------
754 # Print the list of SVN properties for PATH.
755 sub cmd_proplist {
756 my $path = shift;
757 $path = '.' if not defined $path;
758 my $props = get_svnprops($path);
759 print "Properties on '$path':\n";
760 foreach (sort keys %{$props}) {
761 print " $_\n";
762 }
763 }
764
765 sub cmd_multi_init {
766 my $url = shift;
767 unless (defined $_trunk || defined $_branches || defined $_tags) {
768 usage(1);
769 }
770
771 # there are currently some bugs that prevent multi-init/multi-fetch
772 # setups from working well without this.
773 $Git::SVN::_minimize_url = 1;
774
775 $_prefix = '' unless defined $_prefix;
776 if (defined $url) {
777 $url =~ s#/+$##;
778 init_subdir(@_);
779 }
780 do_git_init_db();
781 if (defined $_trunk) {
782 my $trunk_ref = $_prefix . 'trunk';
783 # try both old-style and new-style lookups:
784 my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
785 unless ($gs_trunk) {
786 my ($trunk_url, $trunk_path) =
787 complete_svn_url($url, $_trunk);
788 $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
789 undef, $trunk_ref);
790 }
791 }
792 return unless defined $_branches || defined $_tags;
793 my $ra = $url ? Git::SVN::Ra->new($url) : undef;
794 complete_url_ls_init($ra, $_branches, '--branches/-b', $_prefix);
795 complete_url_ls_init($ra, $_tags, '--tags/-t', $_prefix . 'tags/');
796 }
797
798 sub cmd_multi_fetch {
799 my $remotes = Git::SVN::read_all_remotes();
800 foreach my $repo_id (sort keys %$remotes) {
801 if ($remotes->{$repo_id}->{url}) {
802 Git::SVN::fetch_all($repo_id, $remotes);
803 }
804 }
805 }
806
807 # this command is special because it requires no metadata
808 sub cmd_commit_diff {
809 my ($ta, $tb, $url) = @_;
810 my $usage = "Usage: $0 commit-diff -r<revision> ".
811 "<tree-ish> <tree-ish> [<URL>]";
812 fatal($usage) if (!defined $ta || !defined $tb);
813 my $svn_path = '';
814 if (!defined $url) {
815 my $gs = eval { Git::SVN->new };
816 if (!$gs) {
817 fatal("Needed URL or usable git-svn --id in ",
818 "the command-line\n", $usage);
819 }
820 $url = $gs->{url};
821 $svn_path = $gs->{path};
822 }
823 unless (defined $_revision) {
824 fatal("-r|--revision is a required argument\n", $usage);
825 }
826 if (defined $_message && defined $_file) {
827 fatal("Both --message/-m and --file/-F specified ",
828 "for the commit message.\n",
829 "I have no idea what you mean");
830 }
831 if (defined $_file) {
832 $_message = file_to_s($_file);
833 } else {
834 $_message ||= get_commit_entry($tb)->{log};
835 }
836 my $ra ||= Git::SVN::Ra->new($url);
837 my $r = $_revision;
838 if ($r eq 'HEAD') {
839 $r = $ra->get_latest_revnum;
840 } elsif ($r !~ /^\d+$/) {
841 die "revision argument: $r not understood by git-svn\n";
842 }
843 my %ed_opts = ( r => $r,
844 log => $_message,
845 ra => $ra,
846 tree_a => $ta,
847 tree_b => $tb,
848 editor_cb => sub { print "Committed r$_[0]\n" },
849 svn_path => $svn_path );
850 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
851 print "No changes\n$ta == $tb\n";
852 }
853 }
854
855 sub escape_uri_only {
856 my ($uri) = @_;
857 my @tmp;
858 foreach (split m{/}, $uri) {
859 s/([^~\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
860 push @tmp, $_;
861 }
862 join('/', @tmp);
863 }
864
865 sub escape_url {
866 my ($url) = @_;
867 if ($url =~ m#^([^:]+)://([^/]*)(.*)$#) {
868 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
869 $url = "$scheme://$domain$uri";
870 }
871 $url;
872 }
873
874 sub cmd_info {
875 my $path = canonicalize_path(defined($_[0]) ? $_[0] : ".");
876 my $fullpath = canonicalize_path($cmd_dir_prefix . $path);
877 if (exists $_[1]) {
878 die "Too many arguments specified\n";
879 }
880
881 my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
882
883 if (!$file_type && !$diff_status) {
884 print STDERR "svn: '$path' is not under version control\n";
885 exit 1;
886 }
887
888 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
889 unless ($gs) {
890 die "Unable to determine upstream SVN information from ",
891 "working tree history\n";
892 }
893
894 # canonicalize_path() will return "" to make libsvn 1.5.x happy,
895 $path = "." if $path eq "";
896
897 my $full_url = $url . ($fullpath eq "" ? "" : "/$fullpath");
898
899 if ($_url) {
900 print escape_url($full_url), "\n";
901 return;
902 }
903
904 my $result = "Path: $path\n";
905 $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
906 $result .= "URL: " . escape_url($full_url) . "\n";
907
908 eval {
909 my $repos_root = $gs->repos_root;
910 Git::SVN::remove_username($repos_root);
911 $result .= "Repository Root: " . escape_url($repos_root) . "\n";
912 };
913 if ($@) {
914 $result .= "Repository Root: (offline)\n";
915 }
916 $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A" &&
917 ($SVN::Core::VERSION le '1.5.4' || $file_type ne "dir");
918 $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
919
920 $result .= "Node Kind: " .
921 ($file_type eq "dir" ? "directory" : "file") . "\n";
922
923 my $schedule = $diff_status eq "A"
924 ? "add"
925 : ($diff_status eq "D" ? "delete" : "normal");
926 $result .= "Schedule: $schedule\n";
927
928 if ($diff_status eq "A") {
929 print $result, "\n";
930 return;
931 }
932
933 my ($lc_author, $lc_rev, $lc_date_utc);
934 my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $fullpath);
935 my $log = command_output_pipe(@args);
936 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
937 while (<$log>) {
938 if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
939 $lc_author = $1;
940 $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
941 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
942 (undef, $lc_rev, undef) = ::extract_metadata($1);
943 }
944 }
945 close $log;
946
947 Git::SVN::Log::set_local_timezone();
948
949 $result .= "Last Changed Author: $lc_author\n";
950 $result .= "Last Changed Rev: $lc_rev\n";
951 $result .= "Last Changed Date: " .
952 Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
953
954 if ($file_type ne "dir") {
955 my $text_last_updated_date =
956 ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
957 $result .=
958 "Text Last Updated: " .
959 Git::SVN::Log::format_svn_date($text_last_updated_date) .
960 "\n";
961 my $checksum;
962 if ($diff_status eq "D") {
963 my ($fh, $ctx) =
964 command_output_pipe(qw(cat-file blob), "HEAD:$path");
965 if ($file_type eq "link") {
966 my $file_name = <$fh>;
967 $checksum = md5sum("link $file_name");
968 } else {
969 $checksum = md5sum($fh);
970 }
971 command_close_pipe($fh, $ctx);
972 } elsif ($file_type eq "link") {
973 my $file_name =
974 command(qw(cat-file blob), "HEAD:$path");
975 $checksum =
976 md5sum("link " . $file_name);
977 } else {
978 open FILE, "<", $path or die $!;
979 $checksum = md5sum(\*FILE);
980 close FILE or die $!;
981 }
982 $result .= "Checksum: " . $checksum . "\n";
983 }
984
985 print $result, "\n";
986 }
987
988 ########################### utility functions #########################
989
990 sub rebase_cmd {
991 my @cmd = qw/rebase/;
992 push @cmd, '-v' if $_verbose;
993 push @cmd, qw/--merge/ if $_merge;
994 push @cmd, "--strategy=$_strategy" if $_strategy;
995 @cmd;
996 }
997
998 sub post_fetch_checkout {
999 return if $_no_checkout;
1000 my $gs = $Git::SVN::_head or return;
1001 return if verify_ref('refs/heads/master^0');
1002
1003 my $valid_head = verify_ref('HEAD^0');
1004 command_noisy(qw(update-ref refs/heads/master), $gs->refname);
1005 return if ($valid_head || !verify_ref('HEAD^0'));
1006
1007 return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
1008 my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
1009 return if -f $index;
1010
1011 return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
1012 return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
1013 command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
1014 print STDERR "Checked out HEAD:\n ",
1015 $gs->full_url, " r", $gs->last_rev, "\n";
1016 }
1017
1018 sub complete_svn_url {
1019 my ($url, $path) = @_;
1020 $path =~ s#/+$##;
1021 if ($path !~ m#^[a-z\+]+://#) {
1022 if (!defined $url || $url !~ m#^[a-z\+]+://#) {
1023 fatal("E: '$path' is not a complete URL ",
1024 "and a separate URL is not specified");
1025 }
1026 return ($url, $path);
1027 }
1028 return ($path, '');
1029 }
1030
1031 sub complete_url_ls_init {
1032 my ($ra, $repo_path, $switch, $pfx) = @_;
1033 unless ($repo_path) {
1034 print STDERR "W: $switch not specified\n";
1035 return;
1036 }
1037 $repo_path =~ s#/+$##;
1038 if ($repo_path =~ m#^[a-z\+]+://#) {
1039 $ra = Git::SVN::Ra->new($repo_path);
1040 $repo_path = '';
1041 } else {
1042 $repo_path =~ s#^/+##;
1043 unless ($ra) {
1044 fatal("E: '$repo_path' is not a complete URL ",
1045 "and a separate URL is not specified");
1046 }
1047 }
1048 my $url = $ra->{url};
1049 my $gs = Git::SVN->init($url, undef, undef, undef, 1);
1050 my $k = "svn-remote.$gs->{repo_id}.url";
1051 my $orig_url = eval { command_oneline(qw/config --get/, $k) };
1052 if ($orig_url && ($orig_url ne $gs->{url})) {
1053 die "$k already set: $orig_url\n",
1054 "wanted to set to: $gs->{url}\n";
1055 }
1056 command_oneline('config', $k, $gs->{url}) unless $orig_url;
1057 my $remote_path = "$ra->{svn_path}/$repo_path";
1058 $remote_path =~ s#/+#/#g;
1059 $remote_path =~ s#^/##g;
1060 $remote_path .= "/*" if $remote_path !~ /\*/;
1061 my ($n) = ($switch =~ /^--(\w+)/);
1062 if (length $pfx && $pfx !~ m#/$#) {
1063 die "--prefix='$pfx' must have a trailing slash '/'\n";
1064 }
1065 command_noisy('config',
1066 "svn-remote.$gs->{repo_id}.$n",
1067 "$remote_path:refs/remotes/$pfx*" .
1068 ('/*' x (($remote_path =~ tr/*/*/) - 1)) );
1069 }
1070
1071 sub verify_ref {
1072 my ($ref) = @_;
1073 eval { command_oneline([ 'rev-parse', '--verify', $ref ],
1074 { STDERR => 0 }); };
1075 }
1076
1077 sub get_tree_from_treeish {
1078 my ($treeish) = @_;
1079 # $treeish can be a symbolic ref, too:
1080 my $type = command_oneline(qw/cat-file -t/, $treeish);
1081 my $expected;
1082 while ($type eq 'tag') {
1083 ($treeish, $type) = command(qw/cat-file tag/, $treeish);
1084 }
1085 if ($type eq 'commit') {
1086 $expected = (grep /^tree /, command(qw/cat-file commit/,
1087 $treeish))[0];
1088 ($expected) = ($expected =~ /^tree ($sha1)$/o);
1089 die "Unable to get tree from $treeish\n" unless $expected;
1090 } elsif ($type eq 'tree') {
1091 $expected = $treeish;
1092 } else {
1093 die "$treeish is a $type, expected tree, tag or commit\n";
1094 }
1095 return $expected;
1096 }
1097
1098 sub get_commit_entry {
1099 my ($treeish) = shift;
1100 my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
1101 my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
1102 my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
1103 open my $log_fh, '>', $commit_editmsg or croak $!;
1104
1105 my $type = command_oneline(qw/cat-file -t/, $treeish);
1106 if ($type eq 'commit' || $type eq 'tag') {
1107 my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1108 $type, $treeish);
1109 my $in_msg = 0;
1110 my $author;
1111 my $saw_from = 0;
1112 my $msgbuf = "";
1113 while (<$msg_fh>) {
1114 if (!$in_msg) {
1115 $in_msg = 1 if (/^\s*$/);
1116 $author = $1 if (/^author (.*>)/);
1117 } elsif (/^git-svn-id: /) {
1118 # skip this for now, we regenerate the
1119 # correct one on re-fetch anyways
1120 # TODO: set *:merge properties or like...
1121 } else {
1122 if (/^From:/ || /^Signed-off-by:/) {
1123 $saw_from = 1;
1124 }
1125 $msgbuf .= $_;
1126 }
1127 }
1128 $msgbuf =~ s/\s+$//s;
1129 if ($Git::SVN::_add_author_from && defined($author)
1130 && !$saw_from) {
1131 $msgbuf .= "\n\nFrom: $author";
1132 }
1133 print $log_fh $msgbuf or croak $!;
1134 command_close_pipe($msg_fh, $ctx);
1135 }
1136 close $log_fh or croak $!;
1137
1138 if ($_edit || ($type eq 'tree')) {
1139 my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
1140 # TODO: strip out spaces, comments, like git-commit.sh
1141 system($editor, $commit_editmsg);
1142 }
1143 rename $commit_editmsg, $commit_msg or croak $!;
1144 {
1145 # SVN requires messages to be UTF-8 when entering the repo
1146 local $/;
1147 open $log_fh, '<', $commit_msg or croak $!;
1148 binmode $log_fh;
1149 chomp($log_entry{log} = <$log_fh>);
1150
1151 if (my $enc = Git::config('i18n.commitencoding')) {
1152 require Encode;
1153 Encode::from_to($log_entry{log}, $enc, 'UTF-8');
1154 }
1155 close $log_fh or croak $!;
1156 }
1157 unlink $commit_msg;
1158 \%log_entry;
1159 }
1160
1161 sub s_to_file {
1162 my ($str, $file, $mode) = @_;
1163 open my $fd,'>',$file or croak $!;
1164 print $fd $str,"\n" or croak $!;
1165 close $fd or croak $!;
1166 chmod ($mode &~ umask, $file) if (defined $mode);
1167 }
1168
1169 sub file_to_s {
1170 my $file = shift;
1171 open my $fd,'<',$file or croak "$!: file: $file\n";
1172 local $/;
1173 my $ret = <$fd>;
1174 close $fd or croak $!;
1175 $ret =~ s/\s*$//s;
1176 return $ret;
1177 }
1178
1179 # '<svn username> = real-name <email address>' mapping based on git-svnimport:
1180 sub load_authors {
1181 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1182 my $log = $cmd eq 'log';
1183 while (<$authors>) {
1184 chomp;
1185 next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1186 my ($user, $name, $email) = ($1, $2, $3);
1187 if ($log) {
1188 $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1189 } else {
1190 $users{$user} = [$name, $email];
1191 }
1192 }
1193 close $authors or croak $!;
1194 }
1195
1196 # convert GetOpt::Long specs for use by git-config
1197 sub read_repo_config {
1198 return unless -d $ENV{GIT_DIR};
1199 my $opts = shift;
1200 my @config_only;
1201 foreach my $o (keys %$opts) {
1202 # if we have mixedCase and a long option-only, then
1203 # it's a config-only variable that we don't need for
1204 # the command-line.
1205 push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1206 my $v = $opts->{$o};
1207 my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1208 $key =~ s/-//g;
1209 my $arg = 'git config';
1210 $arg .= ' --int' if ($o =~ /[:=]i$/);
1211 $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1212 if (ref $v eq 'ARRAY') {
1213 chomp(my @tmp = `$arg --get-all svn.$key`);
1214 @$v = @tmp if @tmp;
1215 } else {
1216 chomp(my $tmp = `$arg --get svn.$key`);
1217 if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1218 $$v = $tmp;
1219 }
1220 }
1221 }
1222 delete @$opts{@config_only} if @config_only;
1223 }
1224
1225 sub extract_metadata {
1226 my $id = shift or return (undef, undef, undef);
1227 my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1228 \s([a-f\d\-]+)$/x);
1229 if (!defined $rev || !$uuid || !$url) {
1230 # some of the original repositories I made had
1231 # identifiers like this:
1232 ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/);
1233 }
1234 return ($url, $rev, $uuid);
1235 }
1236
1237 sub cmt_metadata {
1238 return extract_metadata((grep(/^git-svn-id: /,
1239 command(qw/cat-file commit/, shift)))[-1]);
1240 }
1241
1242 sub working_head_info {
1243 my ($head, $refs) = @_;
1244 my @args = ('log', '--no-color', '--first-parent', '--pretty=medium');
1245 my ($fh, $ctx) = command_output_pipe(@args, $head);
1246 my $hash;
1247 my %max;
1248 while (<$fh>) {
1249 if ( m{^commit ($::sha1)$} ) {
1250 unshift @$refs, $hash if $hash and $refs;
1251 $hash = $1;
1252 next;
1253 }
1254 next unless s{^\s*(git-svn-id:)}{$1};
1255 my ($url, $rev, $uuid) = extract_metadata($_);
1256 if (defined $url && defined $rev) {
1257 next if $max{$url} and $max{$url} < $rev;
1258 if (my $gs = Git::SVN->find_by_url($url)) {
1259 my $c = $gs->rev_map_get($rev, $uuid);
1260 if ($c && $c eq $hash) {
1261 close $fh; # break the pipe
1262 return ($url, $rev, $uuid, $gs);
1263 } else {
1264 $max{$url} ||= $gs->rev_map_max;
1265 }
1266 }
1267 }
1268 }
1269 command_close_pipe($fh, $ctx);
1270 (undef, undef, undef, undef);
1271 }
1272
1273 sub read_commit_parents {
1274 my ($parents, $c) = @_;
1275 chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1276 $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1277 @{$parents->{$c}} = split(/ /, $p);
1278 }
1279
1280 sub linearize_history {
1281 my ($gs, $refs) = @_;
1282 my %parents;
1283 foreach my $c (@$refs) {
1284 read_commit_parents(\%parents, $c);
1285 }
1286
1287 my @linear_refs;
1288 my %skip = ();
1289 my $last_svn_commit = $gs->last_commit;
1290 foreach my $c (reverse @$refs) {
1291 next if $c eq $last_svn_commit;
1292 last if $skip{$c};
1293
1294 unshift @linear_refs, $c;
1295 $skip{$c} = 1;
1296
1297 # we only want the first parent to diff against for linear
1298 # history, we save the rest to inject when we finalize the
1299 # svn commit
1300 my $fp_a = verify_ref("$c~1");
1301 my $fp_b = shift @{$parents{$c}} if $parents{$c};
1302 if (!$fp_a || !$fp_b) {
1303 die "Commit $c\n",
1304 "has no parent commit, and therefore ",
1305 "nothing to diff against.\n",
1306 "You should be working from a repository ",
1307 "originally created by git-svn\n";
1308 }
1309 if ($fp_a ne $fp_b) {
1310 die "$c~1 = $fp_a, however parsing commit $c ",
1311 "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1312 }
1313
1314 foreach my $p (@{$parents{$c}}) {
1315 $skip{$p} = 1;
1316 }
1317 }
1318 (\@linear_refs, \%parents);
1319 }
1320
1321 sub find_file_type_and_diff_status {
1322 my ($path) = @_;
1323 return ('dir', '') if $path eq '';
1324
1325 my $diff_output =
1326 command_oneline(qw(diff --cached --name-status --), $path) || "";
1327 my $diff_status = (split(' ', $diff_output))[0] || "";
1328
1329 my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1330
1331 return (undef, undef) if !$diff_status && !$ls_tree;
1332
1333 if ($diff_status eq "A") {
1334 return ("link", $diff_status) if -l $path;
1335 return ("dir", $diff_status) if -d $path;
1336 return ("file", $diff_status);
1337 }
1338
1339 my $mode = (split(' ', $ls_tree))[0] || "";
1340
1341 return ("link", $diff_status) if $mode eq "120000";
1342 return ("dir", $diff_status) if $mode eq "040000";
1343 return ("file", $diff_status);
1344 }
1345
1346 sub md5sum {
1347 my $arg = shift;
1348 my $ref = ref $arg;
1349 my $md5 = Digest::MD5->new();
1350 if ($ref eq 'GLOB' || $ref eq 'IO::File' || $ref eq 'File::Temp') {
1351 $md5->addfile($arg) or croak $!;
1352 } elsif ($ref eq 'SCALAR') {
1353 $md5->add($$arg) or croak $!;
1354 } elsif (!$ref) {
1355 $md5->add($arg) or croak $!;
1356 } else {
1357 ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
1358 }
1359 return $md5->hexdigest();
1360 }
1361
1362 package Git::SVN;
1363 use strict;
1364 use warnings;
1365 use Fcntl qw/:DEFAULT :seek/;
1366 use constant rev_map_fmt => 'NH40';
1367 use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1368 $_repack $_repack_flags $_use_svm_props $_head
1369 $_use_svnsync_props $no_reuse_existing $_minimize_url
1370 $_use_log_author $_add_author_from $_localtime/;
1371 use Carp qw/croak/;
1372 use File::Path qw/mkpath/;
1373 use File::Copy qw/copy/;
1374 use IPC::Open3;
1375
1376 my ($_gc_nr, $_gc_period);
1377
1378 # properties that we do not log:
1379 my %SKIP_PROP;
1380 BEGIN {
1381 %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1382 svn:special svn:executable
1383 svn:entry:committed-rev
1384 svn:entry:last-author
1385 svn:entry:uuid
1386 svn:entry:committed-date/;
1387
1388 # some options are read globally, but can be overridden locally
1389 # per [svn-remote "..."] section. Command-line options will *NOT*
1390 # override options set in an [svn-remote "..."] section
1391 no strict 'refs';
1392 for my $option (qw/follow_parent no_metadata use_svm_props
1393 use_svnsync_props/) {
1394 my $key = $option;
1395 $key =~ tr/_//d;
1396 my $prop = "-$option";
1397 *$option = sub {
1398 my ($self) = @_;
1399 return $self->{$prop} if exists $self->{$prop};
1400 my $k = "svn-remote.$self->{repo_id}.$key";
1401 eval { command_oneline(qw/config --get/, $k) };
1402 if ($@) {
1403 $self->{$prop} = ${"Git::SVN::_$option"};
1404 } else {
1405 my $v = command_oneline(qw/config --bool/,$k);
1406 $self->{$prop} = $v eq 'false' ? 0 : 1;
1407 }
1408 return $self->{$prop};
1409 }
1410 }
1411 }
1412
1413
1414 my (%LOCKFILES, %INDEX_FILES);
1415 END {
1416 unlink keys %LOCKFILES if %LOCKFILES;
1417 unlink keys %INDEX_FILES if %INDEX_FILES;
1418 }
1419
1420 sub resolve_local_globs {
1421 my ($url, $fetch, $glob_spec) = @_;
1422 return unless defined $glob_spec;
1423 my $ref = $glob_spec->{ref};
1424 my $path = $glob_spec->{path};
1425 foreach (command(qw#for-each-ref --format=%(refname) refs/remotes#)) {
1426 next unless m#^refs/remotes/$ref->{regex}$#;
1427 my $p = $1;
1428 my $pathname = desanitize_refname($path->full_path($p));
1429 my $refname = desanitize_refname($ref->full_path($p));
1430 if (my $existing = $fetch->{$pathname}) {
1431 if ($existing ne $refname) {
1432 die "Refspec conflict:\n",
1433 "existing: refs/remotes/$existing\n",
1434 " globbed: refs/remotes/$refname\n";
1435 }
1436 my $u = (::cmt_metadata("refs/remotes/$refname"))[0];
1437 $u =~ s!^\Q$url\E(/|$)!! or die
1438 "refs/remotes/$refname: '$url' not found in '$u'\n";
1439 if ($pathname ne $u) {
1440 warn "W: Refspec glob conflict ",
1441 "(ref: refs/remotes/$refname):\n",
1442 "expected path: $pathname\n",
1443 " real path: $u\n",
1444 "Continuing ahead with $u\n";
1445 next;
1446 }
1447 } else {
1448 $fetch->{$pathname} = $refname;
1449 }
1450 }
1451 }
1452
1453 sub parse_revision_argument {
1454 my ($base, $head) = @_;
1455 if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1456 return ($base, $head);
1457 }
1458 return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1459 return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1460 return ($head, $head) if ($::_revision eq 'HEAD');
1461 return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1462 return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1463 die "revision argument: $::_revision not understood by git-svn\n";
1464 }
1465
1466 sub fetch_all {
1467 my ($repo_id, $remotes) = @_;
1468 if (ref $repo_id) {
1469 my $gs = $repo_id;
1470 $repo_id = undef;
1471 $repo_id = $gs->{repo_id};
1472 }
1473 $remotes ||= read_all_remotes();
1474 my $remote = $remotes->{$repo_id} or
1475 die "[svn-remote \"$repo_id\"] unknown\n";
1476 my $fetch = $remote->{fetch};
1477 my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1478 my (@gs, @globs);
1479 my $ra = Git::SVN::Ra->new($url);
1480 my $uuid = $ra->get_uuid;
1481 my $head = $ra->get_latest_revnum;
1482 my $base = defined $fetch ? $head : 0;
1483
1484 # read the max revs for wildcard expansion (branches/*, tags/*)
1485 foreach my $t (qw/branches tags/) {
1486 defined $remote->{$t} or next;
1487 push @globs, $remote->{$t};
1488 my $max_rev = eval { tmp_config(qw/--int --get/,
1489 "svn-remote.$repo_id.${t}-maxRev") };
1490 if (defined $max_rev && ($max_rev < $base)) {
1491 $base = $max_rev;
1492 } elsif (!defined $max_rev) {
1493 $base = 0;
1494 }
1495 }
1496
1497 if ($fetch) {
1498 foreach my $p (sort keys %$fetch) {
1499 my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1500 my $lr = $gs->rev_map_max;
1501 if (defined $lr) {
1502 $base = $lr if ($lr < $base);
1503 }
1504 push @gs, $gs;
1505 }
1506 }
1507
1508 ($base, $head) = parse_revision_argument($base, $head);
1509 $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1510 }
1511
1512 sub read_all_remotes {
1513 my $r = {};
1514 my $use_svm_props = eval { command_oneline(qw/config --bool
1515 svn.useSvmProps/) };
1516 $use_svm_props = $use_svm_props eq 'true' if $use_svm_props;
1517 foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1518 if (m!^(.+)\.fetch=\s*(.*)\s*:\s*(.+)\s*$!) {
1519 my ($remote, $local_ref, $_remote_ref) = ($1, $2, $3);
1520 die("svn-remote.$remote: remote ref '$_remote_ref' "
1521 . "must start with 'refs/remotes/'\n")
1522 unless $_remote_ref =~ m{^refs/remotes/(.+)};
1523 my $remote_ref = $1;
1524 $local_ref =~ s{^/}{};
1525 $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1526 $r->{$remote}->{svm} = {} if $use_svm_props;
1527 } elsif (m!^(.+)\.usesvmprops=\s*(.*)\s*$!) {
1528 $r->{$1}->{svm} = {};
1529 } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1530 $r->{$1}->{url} = $2;
1531 } elsif (m!^(.+)\.(branches|tags)=
1532 (.*):refs/remotes/(.+)\s*$/!x) {
1533 my ($p, $g) = ($3, $4);
1534 my $rs = $r->{$1}->{$2} = {
1535 t => $2,
1536 remote => $1,
1537 path => Git::SVN::GlobSpec->new($p),
1538 ref => Git::SVN::GlobSpec->new($g) };
1539 if (length($rs->{ref}->{right}) != 0) {
1540 die "The '*' glob character must be the last ",
1541 "character of '$g'\n";
1542 }
1543 }
1544 }
1545
1546 map {
1547 if (defined $r->{$_}->{svm}) {
1548 my $svm;
1549 eval {
1550 my $section = "svn-remote.$_";
1551 $svm = {
1552 source => tmp_config('--get',
1553 "$section.svm-source"),
1554 replace => tmp_config('--get',
1555 "$section.svm-replace"),
1556 }
1557 };
1558 $r->{$_}->{svm} = $svm;
1559 }
1560 } keys %$r;
1561
1562 $r;
1563 }
1564
1565 sub init_vars {
1566 $_gc_nr = $_gc_period = 1000;
1567 if (defined $_repack || defined $_repack_flags) {
1568 warn "Repack options are obsolete; they have no effect.\n";
1569 }
1570 }
1571
1572 sub verify_remotes_sanity {
1573 return unless -d $ENV{GIT_DIR};
1574 my %seen;
1575 foreach (command(qw/config -l/)) {
1576 if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1577 if ($seen{$1}) {
1578 die "Remote ref refs/remote/$1 is tracked by",
1579 "\n \"$_\"\nand\n \"$seen{$1}\"\n",
1580 "Please resolve this ambiguity in ",
1581 "your git configuration file before ",
1582 "continuing\n";
1583 }
1584 $seen{$1} = $_;
1585 }
1586 }
1587 }
1588
1589 sub find_existing_remote {
1590 my ($url, $remotes) = @_;
1591 return undef if $no_reuse_existing;
1592 my $existing;
1593 foreach my $repo_id (keys %$remotes) {
1594 my $u = $remotes->{$repo_id}->{url} or next;
1595 next if $u ne $url;
1596 $existing = $repo_id;
1597 last;
1598 }
1599 $existing;
1600 }
1601
1602 sub init_remote_config {
1603 my ($self, $url, $no_write) = @_;
1604 $url =~ s!/+$!!; # strip trailing slash
1605 my $r = read_all_remotes();
1606 my $existing = find_existing_remote($url, $r);
1607 if ($existing) {
1608 unless ($no_write) {
1609 print STDERR "Using existing ",
1610 "[svn-remote \"$existing\"]\n";
1611 }
1612 $self->{repo_id} = $existing;
1613 } elsif ($_minimize_url) {
1614 my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1615 $existing = find_existing_remote($min_url, $r);
1616 if ($existing) {
1617 unless ($no_write) {
1618 print STDERR "Using existing ",
1619 "[svn-remote \"$existing\"]\n";
1620 }
1621 $self->{repo_id} = $existing;
1622 }
1623 if ($min_url ne $url) {
1624 unless ($no_write) {
1625 print STDERR "Using higher level of URL: ",
1626 "$url => $min_url\n";
1627 }
1628 my $old_path = $self->{path};
1629 $self->{path} = $url;
1630 $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1631 if (length $old_path) {
1632 $self->{path} .= "/$old_path";
1633 }
1634 $url = $min_url;
1635 }
1636 }
1637 my $orig_url;
1638 if (!$existing) {
1639 # verify that we aren't overwriting anything:
1640 $orig_url = eval {
1641 command_oneline('config', '--get',
1642 "svn-remote.$self->{repo_id}.url")
1643 };
1644 if ($orig_url && ($orig_url ne $url)) {
1645 die "svn-remote.$self->{repo_id}.url already set: ",
1646 "$orig_url\nwanted to set to: $url\n";
1647 }
1648 }
1649 my ($xrepo_id, $xpath) = find_ref($self->refname);
1650 if (defined $xpath) {
1651 die "svn-remote.$xrepo_id.fetch already set to track ",
1652 "$xpath:refs/remotes/", $self->refname, "\n";
1653 }
1654 unless ($no_write) {
1655 command_noisy('config',
1656 "svn-remote.$self->{repo_id}.url", $url);
1657 $self->{path} =~ s{^/}{};
1658 command_noisy('config', '--add',
1659 "svn-remote.$self->{repo_id}.fetch",
1660 "$self->{path}:".$self->refname);
1661 }
1662 $self->{url} = $url;
1663 }
1664
1665 sub find_by_url { # repos_root and, path are optional
1666 my ($class, $full_url, $repos_root, $path) = @_;
1667
1668 return undef unless defined $full_url;
1669 remove_username($full_url);
1670 remove_username($repos_root) if defined $repos_root;
1671 my $remotes = read_all_remotes();
1672 if (defined $full_url && defined $repos_root && !defined $path) {
1673 $path = $full_url;
1674 $path =~ s#^\Q$repos_root\E(?:/|$)##;
1675 }
1676 foreach my $repo_id (keys %$remotes) {
1677 my $u = $remotes->{$repo_id}->{url} or next;
1678 remove_username($u);
1679 next if defined $repos_root && $repos_root ne $u;
1680
1681 my $fetch = $remotes->{$repo_id}->{fetch} || {};
1682 foreach (qw/branches tags/) {
1683 resolve_local_globs($u, $fetch,
1684 $remotes->{$repo_id}->{$_});
1685 }
1686 my $p = $path;
1687 my $rwr = rewrite_root({repo_id => $repo_id});
1688 my $svm = $remotes->{$repo_id}->{svm}
1689 if defined $remotes->{$repo_id}->{svm};
1690 unless (defined $p) {
1691 $p = $full_url;
1692 my $z = $u;
1693 my $prefix = '';
1694 if ($rwr) {
1695 $z = $rwr;
1696 remove_username($z);
1697 } elsif (defined $svm) {
1698 $z = $svm->{source};
1699 $prefix = $svm->{replace};
1700 $prefix =~ s#^\Q$u\E(?:/|$)##;
1701 $prefix =~ s#/$##;
1702 }
1703 $p =~ s#^\Q$z\E(?:/|$)#$prefix# or next;
1704 }
1705 foreach my $f (keys %$fetch) {
1706 next if $f ne $p;
1707 return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1708 }
1709 }
1710 undef;
1711 }
1712
1713 sub init {
1714 my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1715 my $self = _new($class, $repo_id, $ref_id, $path);
1716 if (defined $url) {
1717 $self->init_remote_config($url, $no_write);
1718 }
1719 $self;
1720 }
1721
1722 sub find_ref {
1723 my ($ref_id) = @_;
1724 foreach (command(qw/config -l/)) {
1725 next unless m!^svn-remote\.(.+)\.fetch=
1726 \s*(.*)\s*:\s*refs/remotes/(.+)\s*$!x;
1727 my ($repo_id, $path, $ref) = ($1, $2, $3);
1728 if ($ref eq $ref_id) {
1729 $path = '' if ($path =~ m#^\./?#);
1730 return ($repo_id, $path);
1731 }
1732 }
1733 (undef, undef, undef);
1734 }
1735
1736 sub new {
1737 my ($class, $ref_id, $repo_id, $path) = @_;
1738 if (defined $ref_id && !defined $repo_id && !defined $path) {
1739 ($repo_id, $path) = find_ref($ref_id);
1740 if (!defined $repo_id) {
1741 die "Could not find a \"svn-remote.*.fetch\" key ",
1742 "in the repository configuration matching: ",
1743 "refs/remotes/$ref_id\n";
1744 }
1745 }
1746 my $self = _new($class, $repo_id, $ref_id, $path);
1747 if (!defined $self->{path} || !length $self->{path}) {
1748 my $fetch = command_oneline('config', '--get',
1749 "svn-remote.$repo_id.fetch",
1750 ":refs/remotes/$ref_id\$") or
1751 die "Failed to read \"svn-remote.$repo_id.fetch\" ",
1752 "\":refs/remotes/$ref_id\$\" in config\n";
1753 ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
1754 }
1755 $self->{url} = command_oneline('config', '--get',
1756 "svn-remote.$repo_id.url") or
1757 die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
1758 $self->rebuild;
1759 $self;
1760 }
1761
1762 sub refname {
1763 my ($refname) = "refs/remotes/$_[0]->{ref_id}" ;
1764
1765 # It cannot end with a slash /, we'll throw up on this because
1766 # SVN can't have directories with a slash in their name, either:
1767 if ($refname =~ m{/$}) {
1768 die "ref: '$refname' ends with a trailing slash, this is ",
1769 "not permitted by git nor Subversion\n";
1770 }
1771
1772 # It cannot have ASCII control character space, tilde ~, caret ^,
1773 # colon :, question-mark ?, asterisk *, space, or open bracket [
1774 # anywhere.
1775 #
1776 # Additionally, % must be escaped because it is used for escaping
1777 # and we want our escaped refname to be reversible
1778 $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
1779
1780 # no slash-separated component can begin with a dot .
1781 # /.* becomes /%2E*
1782 $refname =~ s{/\.}{/%2E}g;
1783
1784 # It cannot have two consecutive dots .. anywhere
1785 # .. becomes %2E%2E
1786 $refname =~ s{\.\.}{%2E%2E}g;
1787
1788 return $refname;
1789 }
1790
1791 sub desanitize_refname {
1792 my ($refname) = @_;
1793 $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
1794 return $refname;
1795 }
1796
1797 sub svm_uuid {
1798 my ($self) = @_;
1799 return $self->{svm}->{uuid} if $self->svm;
1800 $self->ra;
1801 unless ($self->{svm}) {
1802 die "SVM UUID not cached, and reading remotely failed\n";
1803 }
1804 $self->{svm}->{uuid};
1805 }
1806
1807 sub svm {
1808 my ($self) = @_;
1809 return $self->{svm} if $self->{svm};
1810 my $svm;
1811 # see if we have it in our config, first:
1812 eval {
1813 my $section = "svn-remote.$self->{repo_id}";
1814 $svm = {
1815 source => tmp_config('--get', "$section.svm-source"),
1816 uuid => tmp_config('--get', "$section.svm-uuid"),
1817 replace => tmp_config('--get', "$section.svm-replace"),
1818 }
1819 };
1820 if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
1821 $self->{svm} = $svm;
1822 }
1823 $self->{svm};
1824 }
1825
1826 sub _set_svm_vars {
1827 my ($self, $ra) = @_;
1828 return $ra if $self->svm;
1829
1830 my @err = ( "useSvmProps set, but failed to read SVM properties\n",
1831 "(svm:source, svm:uuid) ",
1832 "from the following URLs:\n" );
1833 sub read_svm_props {
1834 my ($self, $ra, $path, $r) = @_;
1835 my $props = ($ra->get_dir($path, $r))[2];
1836 my $src = $props->{'svm:source'};
1837 my $uuid = $props->{'svm:uuid'};
1838 return undef if (!$src || !$uuid);
1839
1840 chomp($src, $uuid);
1841
1842 $uuid =~ m{^[0-9a-f\-]{30,}$}
1843 or die "doesn't look right - svm:uuid is '$uuid'\n";
1844
1845 # the '!' is used to mark the repos_root!/relative/path
1846 $src =~ s{/?!/?}{/};
1847 $src =~ s{/+$}{}; # no trailing slashes please
1848 # username is of no interest
1849 $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
1850
1851 my $replace = $ra->{url};
1852 $replace .= "/$path" if length $path;
1853
1854 my $section = "svn-remote.$self->{repo_id}";
1855 tmp_config("$section.svm-source", $src);
1856 tmp_config("$section.svm-replace", $replace);
1857 tmp_config("$section.svm-uuid", $uuid);
1858 $self->{svm} = {
1859 source => $src,
1860 uuid => $uuid,
1861 replace => $replace
1862 };
1863 }
1864
1865 my $r = $ra->get_latest_revnum;
1866 my $path = $self->{path};
1867 my %tried;
1868 while (length $path) {
1869 unless ($tried{"$self->{url}/$path"}) {
1870 return $ra if $self->read_svm_props($ra, $path, $r);
1871 $tried{"$self->{url}/$path"} = 1;
1872 }
1873 $path =~ s#/?[^/]+$##;
1874 }
1875 die "Path: '$path' should be ''\n" if $path ne '';
1876 return $ra if $self->read_svm_props($ra, $path, $r);
1877 $tried{"$self->{url}/$path"} = 1;
1878
1879 if ($ra->{repos_root} eq $self->{url}) {
1880 die @err, (map { " $_\n" } keys %tried), "\n";
1881 }
1882
1883 # nope, make sure we're connected to the repository root:
1884 my $ok;
1885 my @tried_b;
1886 $path = $ra->{svn_path};
1887 $ra = Git::SVN::Ra->new($ra->{repos_root});
1888 while (length $path) {
1889 unless ($tried{"$ra->{url}/$path"}) {
1890 $ok = $self->read_svm_props($ra, $path, $r);
1891 last if $ok;
1892 $tried{"$ra->{url}/$path"} = 1;
1893 }
1894 $path =~ s#/?[^/]+$##;
1895 }
1896 die "Path: '$path' should be ''\n" if $path ne '';
1897 $ok ||= $self->read_svm_props($ra, $path, $r);
1898 $tried{"$ra->{url}/$path"} = 1;
1899 if (!$ok) {
1900 die @err, (map { " $_\n" } keys %tried), "\n";
1901 }
1902 Git::SVN::Ra->new($self->{url});
1903 }
1904
1905 sub svnsync {
1906 my ($self) = @_;
1907 return $self->{svnsync} if $self->{svnsync};
1908
1909 if ($self->no_metadata) {
1910 die "Can't have both 'noMetadata' and ",
1911 "'useSvnsyncProps' options set!\n";
1912 }
1913 if ($self->rewrite_root) {
1914 die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
1915 "options set!\n";
1916 }
1917
1918 my $svnsync;
1919 # see if we have it in our config, first:
1920 eval {
1921 my $section = "svn-remote.$self->{repo_id}";
1922
1923 my $url = tmp_config('--get', "$section.svnsync-url");
1924 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1925 die "doesn't look right - svn:sync-from-url is '$url'\n";
1926
1927 my $uuid = tmp_config('--get', "$section.svnsync-uuid");
1928 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1929 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1930
1931 $svnsync = { url => $url, uuid => $uuid }
1932 };
1933 if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
1934 return $self->{svnsync} = $svnsync;
1935 }
1936
1937 my $err = "useSvnsyncProps set, but failed to read " .
1938 "svnsync property: svn:sync-from-";
1939 my $rp = $self->ra->rev_proplist(0);
1940
1941 my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
1942 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1943 die "doesn't look right - svn:sync-from-url is '$url'\n";
1944
1945 my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
1946 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1947 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1948
1949 my $section = "svn-remote.$self->{repo_id}";
1950 tmp_config('--add', "$section.svnsync-uuid", $uuid);
1951 tmp_config('--add', "$section.svnsync-url", $url);
1952 return $self->{svnsync} = { url => $url, uuid => $uuid };
1953 }
1954
1955 # this allows us to memoize our SVN::Ra UUID locally and avoid a
1956 # remote lookup (useful for 'git svn log').
1957 sub ra_uuid {
1958 my ($self) = @_;
1959 unless ($self->{ra_uuid}) {
1960 my $key = "svn-remote.$self->{repo_id}.uuid";
1961 my $uuid = eval { tmp_config('--get', $key) };
1962 if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/) {
1963 $self->{ra_uuid} = $uuid;
1964 } else {
1965 die "ra_uuid called without URL\n" unless $self->{url};
1966 $self->{ra_uuid} = $self->ra->get_uuid;
1967 tmp_config('--add', $key, $self->{ra_uuid});
1968 }
1969 }
1970 $self->{ra_uuid};
1971 }
1972
1973 sub _set_repos_root {
1974 my ($self, $repos_root) = @_;
1975 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1976 $repos_root ||= $self->ra->{repos_root};
1977 tmp_config($k, $repos_root);
1978 $repos_root;
1979 }
1980
1981 sub repos_root {
1982 my ($self) = @_;
1983 my $k = "svn-remote.$self->{repo_id}.reposRoot";
1984 eval { tmp_config('--get', $k) } || $self->_set_repos_root;
1985 }
1986
1987 sub ra {
1988 my ($self) = shift;
1989 my $ra = Git::SVN::Ra->new($self->{url});
1990 $self->_set_repos_root($ra->{repos_root});
1991 if ($self->use_svm_props && !$self->{svm}) {
1992 if ($self->no_metadata) {
1993 die "Can't have both 'noMetadata' and ",
1994 "'useSvmProps' options set!\n";
1995 } elsif ($self->use_svnsync_props) {
1996 die "Can't have both 'useSvnsyncProps' and ",
1997 "'useSvmProps' options set!\n";
1998 }
1999 $ra = $self->_set_svm_vars($ra);
2000 $self->{-want_revprops} = 1;
2001 }
2002 $ra;
2003 }
2004
2005 sub rel_path {
2006 my ($self) = @_;
2007 my $repos_root = $self->ra->{repos_root};
2008 return $self->{path} if ($self->{url} eq $repos_root);
2009 my $url = $self->{url} .
2010 (length $self->{path} ? "/$self->{path}" : $self->{path});
2011 $url =~ s!^\Q$repos_root\E(?:/+|$)!!g;
2012 $url;
2013 }
2014
2015 # prop_walk(PATH, REV, SUB)
2016 # -------------------------
2017 # Recursively traverse PATH at revision REV and invoke SUB for each
2018 # directory that contains a SVN property. SUB will be invoked as
2019 # follows: &SUB(gs, path, props); where `gs' is this instance of
2020 # Git::SVN, `path' the path to the directory where the properties
2021 # `props' were found. The `path' will be relative to point of checkout,
2022 # that is, if url://repo/trunk is the current Git branch, and that
2023 # directory contains a sub-directory `d', SUB will be invoked with `/d/'
2024 # as `path' (note the trailing `/').
2025 sub prop_walk {
2026 my ($self, $path, $rev, $sub) = @_;
2027
2028 $path =~ s#^/##;
2029 my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
2030 $path =~ s#^/*#/#g;
2031 my $p = $path;
2032 # Strip the irrelevant part of the path.
2033 $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
2034 # Ensure the path is terminated by a `/'.
2035 $p =~ s#/*$#/#;
2036
2037 # The properties contain all the internal SVN stuff nobody
2038 # (usually) cares about.
2039 my $interesting_props = 0;
2040 foreach (keys %{$props}) {
2041 # If it doesn't start with `svn:', it must be a
2042 # user-defined property.
2043 ++$interesting_props and next if $_ !~ /^svn:/;
2044 # FIXME: Fragile, if SVN adds new public properties,
2045 # this needs to be updated.
2046 ++$interesting_props if /^svn:(?:ignore|keywords|executable
2047 |eol-style|mime-type
2048 |externals|needs-lock)$/x;
2049 }
2050 &$sub($self, $p, $props) if $interesting_props;
2051
2052 foreach (sort keys %$dirent) {
2053 next if $dirent->{$_}->{kind} != $SVN::Node::dir;
2054 $self->prop_walk($self->{path} . $p . $_, $rev, $sub);
2055 }
2056 }
2057
2058 sub last_rev { ($_[0]->last_rev_commit)[0] }
2059 sub last_commit { ($_[0]->last_rev_commit)[1] }
2060
2061 # returns the newest SVN revision number and newest commit SHA1
2062 sub last_rev_commit {
2063 my ($self) = @_;
2064 if (defined $self->{last_rev} && defined $self->{last_commit}) {
2065 return ($self->{last_rev}, $self->{last_commit});
2066 }
2067 my $c = ::verify_ref($self->refname.'^0');
2068 if ($c && !$self->use_svm_props && !$self->no_metadata) {
2069 my $rev = (::cmt_metadata($c))[1];
2070 if (defined $rev) {
2071 ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
2072 return ($rev, $c);
2073 }
2074 }
2075 my $map_path = $self->map_path;
2076 unless (-e $map_path) {
2077 ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
2078 return (undef, undef);
2079 }
2080 my ($rev, $commit) = $self->rev_map_max(1);
2081 ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
2082 return ($rev, $commit);
2083 }
2084
2085 sub get_fetch_range {
2086 my ($self, $min, $max) = @_;
2087 $max ||= $self->ra->get_latest_revnum;
2088 $min ||= $self->rev_map_max;
2089 (++$min, $max);
2090 }
2091
2092 sub tmp_config {
2093 my (@args) = @_;
2094 my $old_def_config = "$ENV{GIT_DIR}/svn/config";
2095 my $config = "$ENV{GIT_DIR}/svn/.metadata";
2096 if (! -f $config && -f $old_def_config) {
2097 rename $old_def_config, $config or
2098 die "Failed rename $old_def_config => $config: $!\n";
2099 }
2100 my $old_config = $ENV{GIT_CONFIG};
2101 $ENV{GIT_CONFIG} = $config;
2102 $@ = undef;
2103 my @ret = eval {
2104 unless (-f $config) {
2105 mkfile($config);
2106 open my $fh, '>', $config or
2107 die "Can't open $config: $!\n";
2108 print $fh "; This file is used internally by ",
2109 "git-svn\n" or die
2110 "Couldn't write to $config: $!\n";
2111 print $fh "; You should not have to edit it\n" or
2112 die "Couldn't write to $config: $!\n";
2113 close $fh or die "Couldn't close $config: $!\n";
2114 }
2115 command('config', @args);
2116 };
2117 my $err = $@;
2118 if (defined $old_config) {
2119 $ENV{GIT_CONFIG} = $old_config;
2120 } else {
2121 delete $ENV{GIT_CONFIG};
2122 }
2123 die $err if $err;
2124 wantarray ? @ret : $ret[0];
2125 }
2126
2127 sub tmp_index_do {
2128 my ($self, $sub) = @_;
2129 my $old_index = $ENV{GIT_INDEX_FILE};
2130 $ENV{GIT_INDEX_FILE} = $self->{index};
2131 $@ = undef;
2132 my @ret = eval {
2133 my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
2134 mkpath([$dir]) unless -d $dir;
2135 &$sub;
2136 };
2137 my $err = $@;
2138 if (defined $old_index) {
2139 $ENV{GIT_INDEX_FILE} = $old_index;
2140 } else {
2141 delete $ENV{GIT_INDEX_FILE};
2142 }
2143 die $err if $err;
2144 wantarray ? @ret : $ret[0];
2145 }
2146
2147 sub assert_index_clean {
2148 my ($self, $treeish) = @_;
2149
2150 $self->tmp_index_do(sub {
2151 command_noisy('read-tree', $treeish) unless -e $self->{index};
2152 my $x = command_oneline('write-tree');
2153 my ($y) = (command(qw/cat-file commit/, $treeish) =~
2154 /^tree ($::sha1)/mo);
2155 return if $y eq $x;
2156
2157 warn "Index mismatch: $y != $x\nrereading $treeish\n";
2158 unlink $self->{index} or die "unlink $self->{index}: $!\n";
2159 command_noisy('read-tree', $treeish);
2160 $x = command_oneline('write-tree');
2161 if ($y ne $x) {
2162 ::fatal "trees ($treeish) $y != $x\n",
2163 "Something is seriously wrong...";
2164 }
2165 });
2166 }
2167
2168 sub get_commit_parents {
2169 my ($self, $log_entry) = @_;
2170 my (%seen, @ret, @tmp);
2171 # legacy support for 'set-tree'; this is only used by set_tree_cb:
2172 if (my $ip = $self->{inject_parents}) {
2173 if (my $commit = delete $ip->{$log_entry->{revision}}) {
2174 push @tmp, $commit;
2175 }
2176 }
2177 if (my $cur = ::verify_ref($self->refname.'^0')) {
2178 push @tmp, $cur;
2179 }
2180 if (my $ipd = $self->{inject_parents_dcommit}) {
2181 if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2182 push @tmp, @$commit;
2183 }
2184 }
2185 push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2186 while (my $p = shift @tmp) {
2187 next if $seen{$p};
2188 $seen{$p} = 1;
2189 push @ret, $p;
2190 # MAXPARENT is defined to 16 in commit-tree.c:
2191 last if @ret >= 16;
2192 }
2193 if (@tmp) {
2194 die "r$log_entry->{revision}: No room for parents:\n\t",
2195 join("\n\t", @tmp), "\n";
2196 }
2197 @ret;
2198 }
2199
2200 sub rewrite_root {
2201 my ($self) = @_;
2202 return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2203 my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2204 my $rwr = eval { command_oneline(qw/config --get/, $k) };
2205 if ($rwr) {
2206 $rwr =~ s#/+$##;
2207 if ($rwr !~ m#^[a-z\+]+://#) {
2208 die "$rwr is not a valid URL (key: $k)\n";
2209 }
2210 }
2211 $self->{-rewrite_root} = $rwr;
2212 }
2213
2214 sub metadata_url {
2215 my ($self) = @_;
2216 ($self->rewrite_root || $self->{url}) .
2217 (length $self->{path} ? '/' . $self->{path} : '');
2218 }
2219
2220 sub full_url {
2221 my ($self) = @_;
2222 $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2223 }
2224
2225
2226 sub set_commit_header_env {
2227 my ($log_entry) = @_;
2228 my %env;
2229 foreach my $ned (qw/NAME EMAIL DATE/) {
2230 foreach my $ac (qw/AUTHOR COMMITTER/) {
2231 $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2232 }
2233 }
2234
2235 $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2236 $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2237 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2238
2239 $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2240 ? $log_entry->{commit_name}
2241 : $log_entry->{name};
2242 $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2243 ? $log_entry->{commit_email}
2244 : $log_entry->{email};
2245 \%env;
2246 }
2247
2248 sub restore_commit_header_env {
2249 my ($env) = @_;
2250 foreach my $ned (qw/NAME EMAIL DATE/) {
2251 foreach my $ac (qw/AUTHOR COMMITTER/) {
2252 my $k = "GIT_${ac}_${ned}";
2253 if (defined $env->{$k}) {
2254 $ENV{$k} = $env->{$k};
2255 } else {
2256 delete $ENV{$k};
2257 }
2258 }
2259 }
2260 }
2261
2262 sub gc {
2263 command_noisy('gc', '--auto');
2264 };
2265
2266 sub do_git_commit {
2267 my ($self, $log_entry) = @_;
2268 my $lr = $self->last_rev;
2269 if (defined $lr && $lr >= $log_entry->{revision}) {
2270 die "Last fetched revision of ", $self->refname,
2271 " was r$lr, but we are about to fetch: ",
2272 "r$log_entry->{revision}!\n";
2273 }
2274 if (my $c = $self->rev_map_get($log_entry->{revision})) {
2275 croak "$log_entry->{revision} = $c already exists! ",
2276 "Why are we refetching it?\n";
2277 }
2278 my $old_env = set_commit_header_env($log_entry);
2279 my $tree = $log_entry->{tree};
2280 if (!defined $tree) {
2281 $tree = $self->tmp_index_do(sub {
2282 command_oneline('write-tree') });
2283 }
2284 die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2285
2286 my @exec = ('git', 'commit-tree', $tree);
2287 foreach ($self->get_commit_parents($log_entry)) {
2288 push @exec, '-p', $_;
2289 }
2290 defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2291 or croak $!;
2292 binmode $msg_fh;
2293
2294 # we always get UTF-8 from SVN, but we may want our commits in
2295 # a different encoding.
2296 if (my $enc = Git::config('i18n.commitencoding')) {
2297 require Encode;
2298 Encode::from_to($log_entry->{log}, 'UTF-8', $enc);
2299 }
2300 print $msg_fh $log_entry->{log} or croak $!;
2301 restore_commit_header_env($old_env);
2302 unless ($self->no_metadata) {
2303 print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2304 or croak $!;
2305 }
2306 $msg_fh->flush == 0 or croak $!;
2307 close $msg_fh or croak $!;
2308 chomp(my $commit = do { local $/; <$out_fh> });
2309 close $out_fh or croak $!;
2310 waitpid $pid, 0;
2311 croak $? if $?;
2312 if ($commit !~ /^$::sha1$/o) {
2313 die "Failed to commit, invalid sha1: $commit\n";
2314 }
2315
2316 $self->rev_map_set($log_entry->{revision}, $commit, 1);
2317
2318 $self->{last_rev} = $log_entry->{revision};
2319 $self->{last_commit} = $commit;
2320 print "r$log_entry->{revision}";
2321 if (defined $log_entry->{svm_revision}) {
2322 print " (\@$log_entry->{svm_revision})";
2323 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2324 0, $self->svm_uuid);
2325 }
2326 print " = $commit ($self->{ref_id})\n";
2327 if (--$_gc_nr == 0) {
2328 $_gc_nr = $_gc_period;
2329 gc();
2330 }
2331 return $commit;
2332 }
2333
2334 sub match_paths {
2335 my ($self, $paths, $r) = @_;
2336 return 1 if $self->{path} eq '';
2337 if (my $path = $paths->{"/$self->{path}"}) {
2338 return ($path->{action} eq 'D') ? 0 : 1;
2339 }
2340 $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
2341 if (grep /$self->{path_regex}/, keys %$paths) {
2342 return 1;
2343 }
2344 my $c = '';
2345 foreach (split m#/#, $self->{path}) {
2346 $c .= "/$_";
2347 next unless ($paths->{$c} &&
2348 ($paths->{$c}->{action} =~ /^[AR]$/));
2349 if ($self->ra->check_path($self->{path}, $r) ==
2350 $SVN::Node::dir) {
2351 return 1;
2352 }
2353 }
2354 return 0;
2355 }
2356
2357 sub find_parent_branch {
2358 my ($self, $paths, $rev) = @_;
2359 return undef unless $self->follow_parent;
2360 unless (defined $paths) {
2361 my $err_handler = $SVN::Error::handler;
2362 $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2363 $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1, sub {
2364 $paths =
2365 Git::SVN::Ra::dup_changed_paths($_[0]) });
2366 $SVN::Error::handler = $err_handler;
2367 }
2368 return undef unless defined $paths;
2369
2370 # look for a parent from another branch:
2371 my @b_path_components = split m#/#, $self->rel_path;
2372 my @a_path_components;
2373 my $i;
2374 while (@b_path_components) {
2375 $i = $paths->{'/'.join('/', @b_path_components)};
2376 last if $i && defined $i->{copyfrom_path};
2377 unshift(@a_path_components, pop(@b_path_components));
2378 }
2379 return undef unless defined $i && defined $i->{copyfrom_path};
2380 my $branch_from = $i->{copyfrom_path};
2381 if (@a_path_components) {
2382 print STDERR "branch_from: $branch_from => ";
2383 $branch_from .= '/'.join('/', @a_path_components);
2384 print STDERR $branch_from, "\n";
2385 }
2386 my $r = $i->{copyfrom_rev};
2387 my $repos_root = $self->ra->{repos_root};
2388 my $url = $self->ra->{url};
2389 my $new_url = $repos_root . $branch_from;
2390 print STDERR "Found possible branch point: ",
2391 "$new_url => ", $self->full_url, ", $r\n";
2392 $branch_from =~ s#^/##;
2393 my $gs = $self->other_gs($new_url, $url, $repos_root,
2394 $branch_from, $r, $self->{ref_id});
2395 my ($r0, $parent) = $gs->find_rev_before($r, 1);
2396 {
2397 my ($base, $head);
2398 if (!defined $r0 || !defined $parent) {
2399 ($base, $head) = parse_revision_argument(0, $r);
2400 } else {
2401 if ($r0 < $r) {
2402 $gs->ra->get_log([$gs->{path}], $r0 + 1, $r, 1,
2403 0, 1, sub { $base = $_[1] - 1 });
2404 }
2405 }
2406 if (defined $base && $base <= $r) {
2407 $gs->fetch($base, $r);
2408 }
2409 ($r0, $parent) = $gs->find_rev_before($r, 1);
2410 }
2411 if (defined $r0 && defined $parent) {
2412 print STDERR "Found branch parent: ($self->{ref_id}) $parent\n";
2413 my $ed;
2414 if ($self->ra->can_do_switch) {
2415 $self->assert_index_clean($parent);
2416 print STDERR "Following parent with do_switch\n";
2417 # do_switch works with svn/trunk >= r22312, but that
2418 # is not included with SVN 1.4.3 (the latest version
2419 # at the moment), so we can't rely on it
2420 $self->{last_commit} = $parent;
2421 $ed = SVN::Git::Fetcher->new($self, $gs->{path});
2422 $gs->ra->gs_do_switch($r0, $rev, $gs,
2423 $self->full_url, $ed)
2424 or die "SVN connection failed somewhere...\n";
2425 } elsif ($self->ra->trees_match($new_url, $r0,
2426 $self->full_url, $rev)) {
2427 print STDERR "Trees match:\n",
2428 " $new_url\@$r0\n",
2429 " ${\$self->full_url}\@$rev\n",
2430 "Following parent with no changes\n";
2431 $self->tmp_index_do(sub {
2432 command_noisy('read-tree', $parent);
2433 });
2434 $self->{last_commit} = $parent;
2435 } else {
2436 print STDERR "Following parent with do_update\n";
2437 $ed = SVN::Git::Fetcher->new($self);
2438 $self->ra->gs_do_update($rev, $rev, $self, $ed)
2439 or die "SVN connection failed somewhere...\n";
2440 }
2441 print STDERR "Successfully followed parent\n";
2442 return $self->make_log_entry($rev, [$parent], $ed);
2443 }
2444 return undef;
2445 }
2446
2447 sub do_fetch {
2448 my ($self, $paths, $rev) = @_;
2449 my $ed;
2450 my ($last_rev, @parents);
2451 if (my $lc = $self->last_commit) {
2452 # we can have a branch that was deleted, then re-added
2453 # under the same name but copied from another path, in
2454 # which case we'll have multiple parents (we don't
2455 # want to break the original ref, nor lose copypath info):
2456 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2457 push @{$log_entry->{parents}}, $lc;
2458 return $log_entry;
2459 }
2460 $ed = SVN::Git::Fetcher->new($self);
2461 $last_rev = $self->{last_rev};
2462 $ed->{c} = $lc;
2463 @parents = ($lc);
2464 } else {
2465 $last_rev = $rev;
2466 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2467 return $log_entry;
2468 }
2469 $ed = SVN::Git::Fetcher->new($self);
2470 }
2471 unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2472 die "SVN connection failed somewhere...\n";
2473 }
2474 $self->make_log_entry($rev, \@parents, $ed);
2475 }
2476
2477 sub get_untracked {
2478 my ($self, $ed) = @_;
2479 my @out;
2480 my $h = $ed->{empty};
2481 foreach (sort keys %$h) {
2482 my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2483 push @out, " $act: " . uri_encode($_);
2484 warn "W: $act: $_\n";
2485 }
2486 foreach my $t (qw/dir_prop file_prop/) {
2487 $h = $ed->{$t} or next;
2488 foreach my $path (sort keys %$h) {
2489 my $ppath = $path eq '' ? '.' : $path;
2490 foreach my $prop (sort keys %{$h->{$path}}) {
2491 next if $SKIP_PROP{$prop};
2492 my $v = $h->{$path}->{$prop};
2493 my $t_ppath_prop = "$t: " .
2494 uri_encode($ppath) . ' ' .
2495 uri_encode($prop);
2496 if (defined $v) {
2497 push @out, " +$t_ppath_prop " .
2498 uri_encode($v);
2499 } else {
2500 push @out, " -$t_ppath_prop";
2501 }
2502 }
2503 }
2504 }
2505 foreach my $t (qw/absent_file absent_directory/) {
2506 $h = $ed->{$t} or next;
2507 foreach my $parent (sort keys %$h) {
2508 foreach my $path (sort @{$h->{$parent}}) {
2509 push @out, " $t: " .
2510 uri_encode("$parent/$path");
2511 warn "W: $t: $parent/$path ",
2512 "Insufficient permissions?\n";
2513 }
2514 }
2515 }
2516 \@out;
2517 }
2518
2519 # parse_svn_date(DATE)
2520 # --------------------
2521 # Given a date (in UTC) from Subversion, return a string in the format
2522 # "<TZ Offset> <local date/time>" that Git will use.
2523 #
2524 # By default the parsed date will be in UTC; if $Git::SVN::_localtime
2525 # is true we'll convert it to the local timezone instead.
2526 sub parse_svn_date {
2527 my $date = shift || return '+0000 1970-01-01 00:00:00';
2528 my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2529 (\d\d)\:(\d\d)\:(\d\d)\.\d*Z$/x) or
2530 croak "Unable to parse date: $date\n";
2531 my $parsed_date; # Set next.
2532
2533 if ($Git::SVN::_localtime) {
2534 # Translate the Subversion datetime to an epoch time.
2535 # Begin by switching ourselves to $date's timezone, UTC.
2536 my $old_env_TZ = $ENV{TZ};
2537 $ENV{TZ} = 'UTC';
2538
2539 my $epoch_in_UTC =
2540 POSIX::strftime('%s', $S, $M, $H, $d, $m - 1, $Y - 1900);
2541
2542 # Determine our local timezone (including DST) at the
2543 # time of $epoch_in_UTC. $Git::SVN::Log::TZ stored the
2544 # value of TZ, if any, at the time we were run.
2545 if (defined $Git::SVN::Log::TZ) {
2546 $ENV{TZ} = $Git::SVN::Log::TZ;
2547 } else {
2548 delete $ENV{TZ};
2549 }
2550
2551 my $our_TZ =
2552 POSIX::strftime('%Z', $S, $M, $H, $d, $m - 1, $Y - 1900);
2553
2554 # This converts $epoch_in_UTC into our local timezone.
2555 my ($sec, $min, $hour, $mday, $mon, $year,
2556 $wday, $yday, $isdst) = localtime($epoch_in_UTC);
2557
2558 $parsed_date = sprintf('%s %04d-%02d-%02d %02d:%02d:%02d',
2559 $our_TZ, $year + 1900, $mon + 1,
2560 $mday, $hour, $min, $sec);
2561
2562 # Reset us to the timezone in effect when we entered
2563 # this routine.
2564 if (defined $old_env_TZ) {
2565 $ENV{TZ} = $old_env_TZ;
2566 } else {
2567 delete $ENV{TZ};
2568 }
2569 } else {
2570 $parsed_date = "+0000 $Y-$m-$d $H:$M:$S";
2571 }
2572
2573 return $parsed_date;
2574 }
2575
2576 sub other_gs {
2577 my ($self, $new_url, $url, $repos_root,
2578 $branch_from, $r, $old_ref_id) = @_;
2579 my $gs = Git::SVN->find_by_url($new_url, $repos_root, $branch_from);
2580 unless ($gs) {
2581 my $ref_id = $old_ref_id;
2582 $ref_id =~ s/\@\d+$//;
2583 $ref_id .= "\@$r";
2584 # just grow a tail if we're not unique enough :x
2585 $ref_id .= '-' while find_ref($ref_id);
2586 print STDERR "Initializing parent: $ref_id\n";
2587 my ($u, $p, $repo_id) = ($new_url, '', $ref_id);
2588 if ($u =~ s#^\Q$url\E(/|$)##) {
2589 $p = $u;
2590 $u = $url;
2591 $repo_id = $self->{repo_id};
2592 }
2593 $gs = Git::SVN->init($u, $p, $repo_id, $ref_id, 1);
2594 }
2595 $gs
2596 }
2597
2598 sub check_author {
2599 my ($author) = @_;
2600 if (!defined $author || length $author == 0) {
2601 $author = '(no author)';
2602 } elsif (defined $::_authors && ! defined $::users{$author}) {
2603 die "Author: $author not defined in $::_authors file\n";
2604 }
2605 $author;
2606 }
2607
2608 sub make_log_entry {
2609 my ($self, $rev, $parents, $ed) = @_;
2610 my $untracked = $self->get_untracked($ed);
2611
2612 open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
2613 print $un "r$rev\n" or croak $!;
2614 print $un $_, "\n" foreach @$untracked;
2615 my %log_entry = ( parents => $parents || [], revision => $rev,
2616 log => '');
2617
2618 my $headrev;
2619 my $logged = delete $self->{logged_rev_props};
2620 if (!$logged || $self->{-want_revprops}) {
2621 my $rp = $self->ra->rev_proplist($rev);
2622 foreach (sort keys %$rp) {
2623 my $v = $rp->{$_};
2624 if (/^svn:(author|date|log)$/) {
2625 $log_entry{$1} = $v;
2626 } elsif ($_ eq 'svm:headrev') {
2627 $headrev = $v;
2628 } else {
2629 print $un " rev_prop: ", uri_encode($_), ' ',
2630 uri_encode($v), "\n";
2631 }
2632 }
2633 } else {
2634 map { $log_entry{$_} = $logged->{$_} } keys %$logged;
2635 }
2636 close $un or croak $!;
2637
2638 $log_entry{date} = parse_svn_date($log_entry{date});
2639 $log_entry{log} .= "\n";
2640 my $author = $log_entry{author} = check_author($log_entry{author});
2641 my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
2642 : ($author, undef);
2643
2644 my ($commit_name, $commit_email) = ($name, $email);
2645 if ($_use_log_author) {
2646 my $name_field;
2647 if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
2648 $name_field = $1;
2649 } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
2650 $name_field = $1;
2651 }
2652 if (!defined $name_field) {
2653 if (!defined $email) {
2654 $email = $name;
2655 }
2656 } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
2657 ($name, $email) = ($1, $2);
2658 } elsif ($name_field =~ /(.*)@/) {
2659 ($name, $email) = ($1, $name_field);
2660 } else {
2661 ($name, $email) = ($name_field, $name_field);
2662 }
2663 }
2664 if (defined $headrev && $self->use_svm_props) {
2665 if ($self->rewrite_root) {
2666 die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
2667 "options set!\n";
2668 }
2669 my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$};
2670 # we don't want "SVM: initializing mirror for junk" ...
2671 return undef if $r == 0;
2672 my $svm = $self->svm;
2673 if ($uuid ne $svm->{uuid}) {
2674 die "UUID mismatch on SVM path:\n",
2675 "expected: $svm->{uuid}\n",
2676 " got: $uuid\n";
2677 }
2678 my $full_url = $self->full_url;
2679 $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
2680 die "Failed to replace '$svm->{replace}' with ",
2681 "'$svm->{source}' in $full_url\n";
2682 # throw away username for storing in records
2683 remove_username($full_url);
2684 $log_entry{metadata} = "$full_url\@$r $uuid";
2685 $log_entry{svm_revision} = $r;
2686 $email ||= "$author\@$uuid";
2687 $commit_email ||= "$author\@$uuid";
2688 } elsif ($self->use_svnsync_props) {
2689 my $full_url = $self->svnsync->{url};
2690 $full_url .= "/$self->{path}" if length $self->{path};
2691 remove_username($full_url);
2692 my $uuid = $self->svnsync->{uuid};
2693 $log_entry{metadata} = "$full_url\@$rev $uuid";
2694 $email ||= "$author\@$uuid";
2695 $commit_email ||= "$author\@$uuid";
2696 } else {
2697 my $url = $self->metadata_url;
2698 remove_username($url);
2699 $log_entry{metadata} = "$url\@$rev " .
2700 $self->ra->get_uuid;
2701 $email ||= "$author\@" . $self->ra->get_uuid;
2702 $commit_email ||= "$author\@" . $self->ra->get_uuid;
2703 }
2704 $log_entry{name} = $name;
2705 $log_entry{email} = $email;
2706 $log_entry{commit_name} = $commit_name;
2707 $log_entry{commit_email} = $commit_email;
2708 \%log_entry;
2709 }
2710
2711 sub fetch {
2712 my ($self, $min_rev, $max_rev, @parents) = @_;
2713 my ($last_rev, $last_commit) = $self->last_rev_commit;
2714 my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
2715 $self->ra->gs_fetch_loop_common($base, $head, [$self]);
2716 }
2717
2718 sub set_tree_cb {
2719 my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
2720 $self->{inject_parents} = { $rev => $tree };
2721 $self->fetch(undef, undef);
2722 }
2723
2724 sub set_tree {
2725 my ($self, $tree) = (shift, shift);
2726 my $log_entry = ::get_commit_entry($tree);
2727 unless ($self->{last_rev}) {
2728 ::fatal("Must have an existing revision to commit");
2729 }
2730 my %ed_opts = ( r => $self->{last_rev},
2731 log => $log_entry->{log},
2732 ra => $self->ra,
2733 tree_a => $self->{last_commit},
2734 tree_b => $tree,
2735 editor_cb => sub {
2736 $self->set_tree_cb($log_entry, $tree, @_) },
2737 svn_path => $self->{path} );
2738 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
2739 print "No changes\nr$self->{last_rev} = $tree\n";
2740 }
2741 }
2742
2743 sub rebuild_from_rev_db {
2744 my ($self, $path) = @_;
2745 my $r = -1;
2746 open my $fh, '<', $path or croak "open: $!";
2747 binmode $fh or croak "binmode: $!";
2748 while (<$fh>) {
2749 length($_) == 41 or croak "inconsistent size in ($_) != 41";
2750 chomp($_);
2751 ++$r;
2752 next if $_ eq ('0' x 40);
2753 $self->rev_map_set($r, $_);
2754 print "r$r = $_\n";
2755 }
2756 close $fh or croak "close: $!";
2757 unlink $path or croak "unlink: $!";
2758 }
2759
2760 sub rebuild {
2761 my ($self) = @_;
2762 my $map_path = $self->map_path;
2763 my $partial = (-e $map_path && ! -z $map_path);
2764 return unless ::verify_ref($self->refname.'^0');
2765 if (!$partial && ($self->use_svm_props || $self->no_metadata)) {
2766 my $rev_db = $self->rev_db_path;
2767 $self->rebuild_from_rev_db($rev_db);
2768 if ($self->use_svm_props) {
2769 my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
2770 $self->rebuild_from_rev_db($svm_rev_db);
2771 }
2772 $self->unlink_rev_db_symlink;
2773 return;
2774 }
2775 print "Rebuilding $map_path ...\n" if (!$partial);
2776 my ($base_rev, $head) = ($partial ? $self->rev_map_max_norebuild(1) :
2777 (undef, undef));
2778 my ($log, $ctx) =
2779 command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
2780 ($head ? "$head.." : "") . $self->refname,
2781 '--');
2782 my $metadata_url = $self->metadata_url;
2783 remove_username($metadata_url);
2784 my $svn_uuid = $self->ra_uuid;
2785 my $c;
2786 while (<$log>) {
2787 if ( m{^commit ($::sha1)$} ) {
2788 $c = $1;
2789 next;
2790 }
2791 next unless s{^\s*(git-svn-id:)}{$1};
2792 my ($url, $rev, $uuid) = ::extract_metadata($_);
2793 remove_username($url);
2794
2795 # ignore merges (from set-tree)
2796 next if (!defined $rev || !$uuid);
2797
2798 # if we merged or otherwise started elsewhere, this is
2799 # how we break out of it
2800 if (($uuid ne $svn_uuid) ||
2801 ($metadata_url && $url && ($url ne $metadata_url))) {
2802 next;
2803 }
2804 if ($partial && $head) {
2805 print "Partial-rebuilding $map_path ...\n";
2806 print "Currently at $base_rev = $head\n";
2807 $head = undef;
2808 }
2809
2810 $self->rev_map_set($rev, $c);
2811 print "r$rev = $c\n";
2812 }
2813 command_close_pipe($log, $ctx);
2814 print "Done rebuilding $map_path\n" if (!$partial || !$head);
2815 my $rev_db_path = $self->rev_db_path;
2816 if (-f $self->rev_db_path) {
2817 unlink $self->rev_db_path or croak "unlink: $!";
2818 }
2819 $self->unlink_rev_db_symlink;
2820 }
2821
2822 # rev_map:
2823 # Tie::File seems to be prone to offset errors if revisions get sparse,
2824 # it's not that fast, either. Tie::File is also not in Perl 5.6. So
2825 # one of my favorite modules is out :< Next up would be one of the DBM
2826 # modules, but I'm not sure which is most portable...
2827 #
2828 # This is the replacement for the rev_db format, which was too big
2829 # and inefficient for large repositories with a lot of sparse history
2830 # (mainly tags)
2831 #
2832 # The format is this:
2833 # - 24 bytes for every record,
2834 # * 4 bytes for the integer representing an SVN revision number
2835 # * 20 bytes representing the sha1 of a git commit
2836 # - No empty padding records like the old format
2837 # (except the last record, which can be overwritten)
2838 # - new records are written append-only since SVN revision numbers
2839 # increase monotonically
2840 # - lookups on SVN revision number are done via a binary search
2841 # - Piping the file to xxd -c24 is a good way of dumping it for
2842 # viewing or editing (piped back through xxd -r), should the need
2843 # ever arise.
2844 # - The last record can be padding revision with an all-zero sha1
2845 # This is used to optimize fetch performance when using multiple
2846 # "fetch" directives in .git/config
2847 #
2848 # These files are disposable unless noMetadata or useSvmProps is set
2849
2850 sub _rev_map_set {
2851 my ($fh, $rev, $commit) = @_;
2852
2853 binmode $fh or croak "binmode: $!";
2854 my $size = (stat($fh))[7];
2855 ($size % 24) == 0 or croak "inconsistent size: $size";
2856
2857 my $wr_offset = 0;
2858 if ($size > 0) {
2859 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2860 my $read = sysread($fh, my $buf, 24) or croak "read: $!";
2861 $read == 24 or croak "read only $read bytes (!= 24)";
2862 my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
2863 if ($last_commit eq ('0' x40)) {
2864 if ($size >= 48) {
2865 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2866 $read = sysread($fh, $buf, 24) or
2867 croak "read: $!";
2868 $read == 24 or
2869 croak "read only $read bytes (!= 24)";
2870 ($last_rev, $last_commit) =
2871 unpack(rev_map_fmt, $buf);
2872 if ($last_commit eq ('0' x40)) {
2873 croak "inconsistent .rev_map\n";
2874 }
2875 }
2876 if ($last_rev >= $rev) {
2877 croak "last_rev is higher!: $last_rev >= $rev";
2878 }
2879 $wr_offset = -24;
2880 }
2881 }
2882 sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
2883 syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
2884 croak "write: $!";
2885 }
2886
2887 sub mkfile {
2888 my ($path) = @_;
2889 unless (-e $path) {
2890 my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
2891 mkpath([$dir]) unless -d $dir;
2892 open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
2893 close $fh or die "Couldn't close (create) $path: $!\n";
2894 }
2895 }
2896
2897 sub rev_map_set {
2898 my ($self, $rev, $commit, $update_ref, $uuid) = @_;
2899 length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
2900 my $db = $self->map_path($uuid);
2901 my $db_lock = "$db.lock";
2902 my $sig;
2903 if ($update_ref) {
2904 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2905 $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
2906 }
2907 mkfile($db);
2908
2909 $LOCKFILES{$db_lock} = 1;
2910 my $sync;
2911 # both of these options make our .rev_db file very, very important
2912 # and we can't afford to lose it because rebuild() won't work
2913 if ($self->use_svm_props || $self->no_metadata) {
2914 $sync = 1;
2915 copy($db, $db_lock) or die "rev_map_set(@_): ",
2916 "Failed to copy: ",
2917 "$db => $db_lock ($!)\n";
2918 } else {
2919 rename $db, $db_lock or die "rev_map_set(@_): ",
2920 "Failed to rename: ",
2921 "$db => $db_lock ($!)\n";
2922 }
2923
2924 sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
2925 or croak "Couldn't open $db_lock: $!\n";
2926 _rev_map_set($fh, $rev, $commit);
2927 if ($sync) {
2928 $fh->flush or die "Couldn't flush $db_lock: $!\n";
2929 $fh->sync or die "Couldn't sync $db_lock: $!\n";
2930 }
2931 close $fh or croak $!;
2932 if ($update_ref) {
2933 $_head = $self;
2934 command_noisy('update-ref', '-m', "r$rev",
2935 $self->refname, $commit);
2936 }
2937 rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
2938 "$db_lock => $db ($!)\n";
2939 delete $LOCKFILES{$db_lock};
2940 if ($update_ref) {
2941 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2942 $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
2943 kill $sig, $$ if defined $sig;
2944 }
2945 }
2946
2947 # If want_commit, this will return an array of (rev, commit) where
2948 # commit _must_ be a valid commit in the archive.
2949 # Otherwise, it'll return the max revision (whether or not the
2950 # commit is valid or just a 0x40 placeholder).
2951 sub rev_map_max {
2952 my ($self, $want_commit) = @_;
2953 $self->rebuild;
2954 my ($r, $c) = $self->rev_map_max_norebuild($want_commit);
2955 $want_commit ? ($r, $c) : $r;
2956 }
2957
2958 sub rev_map_max_norebuild {
2959 my ($self, $want_commit) = @_;
2960 my $map_path = $self->map_path;
2961 stat $map_path or return $want_commit ? (0, undef) : 0;
2962 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2963 binmode $fh or croak "binmode: $!";
2964 my $size = (stat($fh))[7];
2965 ($size % 24) == 0 or croak "inconsistent size: $size";
2966
2967 if ($size == 0) {
2968 close $fh or croak "close: $!";
2969 return $want_commit ? (0, undef) : 0;
2970 }
2971
2972 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2973 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2974 my ($r, $c) = unpack(rev_map_fmt, $buf);
2975 if ($want_commit && $c eq ('0' x40)) {
2976 if ($size < 48) {
2977 return $want_commit ? (0, undef) : 0;
2978 }
2979 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2980 sysread($fh, $buf, 24) == 24 or croak "read: $!";
2981 ($r, $c) = unpack(rev_map_fmt, $buf);
2982 if ($c eq ('0'x40)) {
2983 croak "Penultimate record is all-zeroes in $map_path";
2984 }
2985 }
2986 close $fh or croak "close: $!";
2987 $want_commit ? ($r, $c) : $r;
2988 }
2989
2990 sub rev_map_get {
2991 my ($self, $rev, $uuid) = @_;
2992 my $map_path = $self->map_path($uuid);
2993 return undef unless -e $map_path;
2994
2995 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2996 binmode $fh or croak "binmode: $!";
2997 my $size = (stat($fh))[7];
2998 ($size % 24) == 0 or croak "inconsistent size: $size";
2999
3000 if ($size == 0) {
3001 close $fh or croak "close: $fh";
3002 return undef;
3003 }
3004
3005 my ($l, $u) = (0, $size - 24);
3006 my ($r, $c, $buf);
3007
3008 while ($l <= $u) {
3009 my $i = int(($l/24 + $u/24) / 2) * 24;
3010 sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
3011 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
3012 my ($r, $c) = unpack('NH40', $buf);
3013
3014 if ($r < $rev) {
3015 $l = $i + 24;
3016 } elsif ($r > $rev) {
3017 $u = $i - 24;
3018 } else { # $r == $rev
3019 close($fh) or croak "close: $!";
3020 return $c eq ('0' x 40) ? undef : $c;
3021 }
3022 }
3023 close($fh) or croak "close: $!";
3024 undef;
3025 }
3026
3027 # Finds the first svn revision that exists on (if $eq_ok is true) or
3028 # before $rev for the current branch. It will not search any lower
3029 # than $min_rev. Returns the git commit hash and svn revision number
3030 # if found, else (undef, undef).
3031 sub find_rev_before {
3032 my ($self, $rev, $eq_ok, $min_rev) = @_;
3033 --$rev unless $eq_ok;
3034 $min_rev ||= 1;
3035 while ($rev >= $min_rev) {
3036 if (my $c = $self->rev_map_get($rev)) {
3037 return ($rev, $c);
3038 }
3039 --$rev;
3040 }
3041 return (undef, undef);
3042 }
3043
3044 # Finds the first svn revision that exists on (if $eq_ok is true) or
3045 # after $rev for the current branch. It will not search any higher
3046 # than $max_rev. Returns the git commit hash and svn revision number
3047 # if found, else (undef, undef).
3048 sub find_rev_after {
3049 my ($self, $rev, $eq_ok, $max_rev) = @_;
3050 ++$rev unless $eq_ok;
3051 $max_rev ||= $self->rev_map_max;
3052 while ($rev <= $max_rev) {
3053 if (my $c = $self->rev_map_get($rev)) {
3054 return ($rev, $c);
3055 }
3056 ++$rev;
3057 }
3058 return (undef, undef);
3059 }
3060
3061 sub _new {
3062 my ($class, $repo_id, $ref_id, $path) = @_;
3063 unless (defined $repo_id && length $repo_id) {
3064 $repo_id = $Git::SVN::default_repo_id;
3065 }
3066 unless (defined $ref_id && length $ref_id) {
3067 $_[2] = $ref_id = $Git::SVN::default_ref_id;
3068 }
3069 $_[1] = $repo_id;
3070 my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
3071 $_[3] = $path = '' unless (defined $path);
3072 mkpath(["$ENV{GIT_DIR}/svn"]);
3073 bless {
3074 ref_id => $ref_id, dir => $dir, index => "$dir/index",
3075 path => $path, config => "$ENV{GIT_DIR}/svn/config",
3076 map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
3077 }
3078
3079 # for read-only access of old .rev_db formats
3080 sub unlink_rev_db_symlink {
3081 my ($self) = @_;
3082 my $link = $self->rev_db_path;
3083 $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
3084 if (-l $link) {
3085 unlink $link or croak "unlink: $link failed!";
3086 }
3087 }
3088
3089 sub rev_db_path {
3090 my ($self, $uuid) = @_;
3091 my $db_path = $self->map_path($uuid);
3092 $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
3093 or croak "map_path: $db_path does not contain '/.rev_map.' !";
3094 $db_path;
3095 }
3096
3097 # the new replacement for .rev_db
3098 sub map_path {
3099 my ($self, $uuid) = @_;
3100 $uuid ||= $self->ra_uuid;
3101 "$self->{map_root}.$uuid";
3102 }
3103
3104 sub uri_encode {
3105 my ($f) = @_;
3106 $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
3107 $f
3108 }
3109
3110 sub remove_username {
3111 $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
3112 }
3113
3114 package Git::SVN::Prompt;
3115 use strict;
3116 use warnings;
3117 require SVN::Core;
3118 use vars qw/$_no_auth_cache $_username/;
3119
3120 sub simple {
3121 my ($cred, $realm, $default_username, $may_save, $pool) = @_;
3122 $may_save = undef if $_no_auth_cache;
3123 $default_username = $_username if defined $_username;
3124 if (defined $default_username && length $default_username) {
3125 if (defined $realm && length $realm) {
3126 print STDERR "Authentication realm: $realm\n";
3127 STDERR->flush;
3128 }
3129 $cred->username($default_username);
3130 } else {
3131 username($cred, $realm, $may_save, $pool);
3132 }
3133 $cred->password(_read_password("Password for '" .
3134 $cred->username . "': ", $realm));
3135 $cred->may_save($may_save);
3136 $SVN::_Core::SVN_NO_ERROR;
3137 }
3138
3139 sub ssl_server_trust {
3140 my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
3141 $may_save = undef if $_no_auth_cache;
3142 print STDERR "Error validating server certificate for '$realm':\n";
3143 {
3144 no warnings 'once';
3145 # All variables SVN::Auth::SSL::* are used only once,
3146 # so we're shutting up Perl warnings about this.
3147 if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
3148 print STDERR " - The certificate is not issued ",
3149 "by a trusted authority. Use the\n",
3150 " fingerprint to validate ",
3151 "the certificate manually!\n";
3152 }
3153 if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
3154 print STDERR " - The certificate hostname ",
3155 "does not match.\n";
3156 }
3157 if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
3158 print STDERR " - The certificate is not yet valid.\n";
3159 }
3160 if ($failures & $SVN::Auth::SSL::EXPIRED) {
3161 print STDERR " - The certificate has expired.\n";
3162 }
3163 if ($failures & $SVN::Auth::SSL::OTHER) {
3164 print STDERR " - The certificate has ",
3165 "an unknown error.\n";
3166 }
3167 } # no warnings 'once'
3168 printf STDERR
3169 "Certificate information:\n".
3170 " - Hostname: %s\n".
3171 " - Valid: from %s until %s\n".
3172 " - Issuer: %s\n".
3173 " - Fingerprint: %s\n",
3174 map $cert_info->$_, qw(hostname valid_from valid_until
3175 issuer_dname fingerprint);
3176 my $choice;
3177 prompt:
3178 print STDERR $may_save ?
3179 "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
3180 "(R)eject or accept (t)emporarily? ";
3181 STDERR->flush;
3182 $choice = lc(substr(<STDIN> || 'R', 0, 1));
3183 if ($choice =~ /^t$/i) {
3184 $cred->may_save(undef);
3185 } elsif ($choice =~ /^r$/i) {
3186 return -1;
3187 } elsif ($may_save && $choice =~ /^p$/i) {
3188 $cred->may_save($may_save);
3189 } else {
3190 goto prompt;
3191 }
3192 $cred->accepted_failures($failures);
3193 $SVN::_Core::SVN_NO_ERROR;
3194 }
3195
3196 sub ssl_client_cert {
3197 my ($cred, $realm, $may_save, $pool) = @_;
3198 $may_save = undef if $_no_auth_cache;
3199 print STDERR "Client certificate filename: ";
3200 STDERR->flush;
3201 chomp(my $filename = <STDIN>);
3202 $cred->cert_file($filename);
3203 $cred->may_save($may_save);
3204 $SVN::_Core::SVN_NO_ERROR;
3205 }
3206
3207 sub ssl_client_cert_pw {
3208 my ($cred, $realm, $may_save, $pool) = @_;
3209 $may_save = undef if $_no_auth_cache;
3210 $cred->password(_read_password("Password: ", $realm));
3211 $cred->may_save($may_save);
3212 $SVN::_Core::SVN_NO_ERROR;
3213 }
3214
3215 sub username {
3216 my ($cred, $realm, $may_save, $pool) = @_;
3217 $may_save = undef if $_no_auth_cache;
3218 if (defined $realm && length $realm) {
3219 print STDERR "Authentication realm: $realm\n";
3220 }
3221 my $username;
3222 if (defined $_username) {
3223 $username = $_username;
3224 } else {
3225 print STDERR "Username: ";
3226 STDERR->flush;
3227 chomp($username = <STDIN>);
3228 }
3229 $cred->username($username);
3230 $cred->may_save($may_save);
3231 $SVN::_Core::SVN_NO_ERROR;
3232 }
3233
3234 sub _read_password {
3235 my ($prompt, $realm) = @_;
3236 print STDERR $prompt;
3237 STDERR->flush;
3238 require Term::ReadKey;
3239 Term::ReadKey::ReadMode('noecho');
3240 my $password = '';
3241 while (defined(my $key = Term::ReadKey::ReadKey(0))) {
3242 last if $key =~ /[\012\015]/; # \n\r
3243 $password .= $key;
3244 }
3245 Term::ReadKey::ReadMode('restore');
3246 print STDERR "\n";
3247 STDERR->flush;
3248 $password;
3249 }
3250
3251 package SVN::Git::Fetcher;
3252 use vars qw/@ISA/;
3253 use strict;
3254 use warnings;
3255 use Carp qw/croak/;
3256 use File::Temp qw/tempfile/;
3257 use IO::File qw//;
3258 use vars qw/$_ignore_regex/;
3259
3260 # file baton members: path, mode_a, mode_b, pool, fh, blob, base
3261 sub new {
3262 my ($class, $git_svn, $switch_path) = @_;
3263 my $self = SVN::Delta::Editor->new;
3264 bless $self, $class;
3265 if (exists $git_svn->{last_commit}) {
3266 $self->{c} = $git_svn->{last_commit};
3267 $self->{empty_symlinks} =
3268 _mark_empty_symlinks($git_svn, $switch_path);
3269 }
3270 $self->{empty} = {};
3271 $self->{dir_prop} = {};
3272 $self->{file_prop} = {};
3273 $self->{absent_dir} = {};
3274 $self->{absent_file} = {};
3275 $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
3276 $self;
3277 }
3278
3279 # this uses the Ra object, so it must be called before do_{switch,update},
3280 # not inside them (when the Git::SVN::Fetcher object is passed) to
3281 # do_{switch,update}
3282 sub _mark_empty_symlinks {
3283 my ($git_svn, $switch_path) = @_;
3284 my $bool = Git::config_bool('svn.brokenSymlinkWorkaround');
3285 return {} if (defined($bool) && ! $bool);
3286
3287 my %ret;
3288 my ($rev, $cmt) = $git_svn->last_rev_commit;
3289 return {} unless ($rev && $cmt);
3290
3291 # allow the warning to be printed for each revision we fetch to
3292 # ensure the user sees it. The user can also disable the workaround
3293 # on the repository even while git svn is running and the next
3294 # revision fetched will skip this expensive function.
3295 my $printed_warning;
3296 chomp(my $empty_blob = `git hash-object -t blob --stdin < /dev/null`);
3297 my ($ls, $ctx) = command_output_pipe(qw/ls-tree -r -z/, $cmt);
3298 local $/ = "\0";
3299 my $pfx = defined($switch_path) ? $switch_path : $git_svn->{path};
3300 $pfx .= '/' if length($pfx);
3301 while (<$ls>) {
3302 chomp;
3303 s/\A100644 blob $empty_blob\t//o or next;
3304 unless ($printed_warning) {
3305 print STDERR "Scanning for empty symlinks, ",
3306 "this may take a while if you have ",
3307 "many empty files\n",
3308 "You may disable this with `",
3309 "git config svn.brokenSymlinkWorkaround ",
3310 "false'.\n",
3311 "This may be done in a different ",
3312 "terminal without restarting ",
3313 "git svn\n";
3314 $printed_warning = 1;
3315 }
3316 my $path = $_;
3317 my (undef, $props) =
3318 $git_svn->ra->get_file($pfx.$path, $rev, undef);
3319 if ($props->{'svn:special'}) {
3320 $ret{$path} = 1;
3321 }
3322 }
3323 command_close_pipe($ls, $ctx);
3324 \%ret;
3325 }
3326
3327 # returns true if a given path is inside a ".git" directory
3328 sub in_dot_git {
3329 $_[0] =~ m{(?:^|/)\.git(?:/|$)};
3330 }
3331
3332 # return value: 0 -- don't ignore, 1 -- ignore
3333 sub is_path_ignored {
3334 my ($path) = @_;
3335 return 1 if in_dot_git($path);
3336 return 0 unless defined($_ignore_regex);
3337 return 1 if $path =~ m!$_ignore_regex!o;
3338 return 0;
3339 }
3340
3341 sub set_path_strip {
3342 my ($self, $path) = @_;
3343 $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
3344 }
3345
3346 sub open_root {
3347 { path => '' };
3348 }
3349
3350 sub open_directory {
3351 my ($self, $path, $pb, $rev) = @_;
3352 { path => $path };
3353 }
3354
3355 sub git_path {
3356 my ($self, $path) = @_;
3357 if ($self->{path_strip}) {
3358 $path =~ s!$self->{path_strip}!! or
3359 die "Failed to strip path '$path' ($self->{path_strip})\n";
3360 }
3361 $path;
3362 }
3363
3364 sub delete_entry {
3365 my ($self, $path, $rev, $pb) = @_;
3366 return undef if is_path_ignored($path);
3367
3368 my $gpath = $self->git_path($path);
3369 return undef if ($gpath eq '');
3370
3371 # remove entire directories.
3372 if (command('ls-tree', $self->{c}, '--', $gpath) =~ /^040000 tree/) {
3373 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3374 -r --name-only -z/,
3375 $self->{c}, '--', $gpath);
3376 local $/ = "\0";
3377 while (<$ls>) {
3378 chomp;
3379 $self->{gii}->remove($_);
3380 print "\tD\t$_\n" unless $::_q;
3381 }
3382 print "\tD\t$gpath/\n" unless $::_q;
3383 command_close_pipe($ls, $ctx);
3384 $self->{empty}->{$path} = 0
3385 } else {
3386 $self->{gii}->remove($gpath);
3387 print "\tD\t$gpath\n" unless $::_q;
3388 }
3389 undef;
3390 }
3391
3392 sub open_file {
3393 my ($self, $path, $pb, $rev) = @_;
3394 my ($mode, $blob);
3395
3396 goto out if is_path_ignored($path);
3397
3398 my $gpath = $self->git_path($path);
3399 ($mode, $blob) = (command('ls-tree', $self->{c}, '--', $gpath)
3400 =~ /^(\d{6}) blob ([a-f\d]{40})\t/);
3401 unless (defined $mode && defined $blob) {
3402 die "$path was not found in commit $self->{c} (r$rev)\n";
3403 }
3404 if ($mode eq '100644' && $self->{empty_symlinks}->{$path}) {
3405 $mode = '120000';
3406 }
3407 out:
3408 { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3409 pool => SVN::Pool->new, action => 'M' };
3410 }
3411
3412 sub add_file {
3413 my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3414 my $mode;
3415
3416 if (!is_path_ignored($path)) {
3417 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3418 delete $self->{empty}->{$dir};
3419 $mode = '100644';
3420 }
3421 { path => $path, mode_a => $mode, mode_b => $mode,
3422 pool => SVN::Pool->new, action => 'A' };
3423 }
3424
3425 sub add_directory {
3426 my ($self, $path, $cp_path, $cp_rev) = @_;
3427 goto out if is_path_ignored($path);
3428 my $gpath = $self->git_path($path);
3429 if ($gpath eq '') {
3430 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3431 -r --name-only -z/,
3432 $self->{c});
3433 local $/ = "\0";
3434 while (<$ls>) {
3435 chomp;
3436 $self->{gii}->remove($_);
3437 print "\tD\t$_\n" unless $::_q;
3438 }
3439 command_close_pipe($ls, $ctx);
3440 $self->{empty}->{$path} = 0;
3441 }
3442 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3443 delete $self->{empty}->{$dir};
3444 $self->{empty}->{$path} = 1;
3445 out:
3446 { path => $path };
3447 }
3448
3449 sub change_dir_prop {
3450 my ($self, $db, $prop, $value) = @_;
3451 return undef if is_path_ignored($db->{path});
3452 $self->{dir_prop}->{$db->{path}} ||= {};
3453 $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3454 undef;
3455 }
3456
3457 sub absent_directory {
3458 my ($self, $path, $pb) = @_;
3459 return undef if is_path_ignored($path);
3460 $self->{absent_dir}->{$pb->{path}} ||= [];
3461 push @{$self->{absent_dir}->{$pb->{path}}}, $path;
3462 undef;
3463 }
3464
3465 sub absent_file {
3466 my ($self, $path, $pb) = @_;
3467 return undef if is_path_ignored($path);
3468 $self->{absent_file}->{$pb->{path}} ||= [];
3469 push @{$self->{absent_file}->{$pb->{path}}}, $path;
3470 undef;
3471 }
3472
3473 sub change_file_prop {
3474 my ($self, $fb, $prop, $value) = @_;
3475 return undef if is_path_ignored($fb->{path});
3476 if ($prop eq 'svn:executable') {
3477 if ($fb->{mode_b} != 120000) {
3478 $fb->{mode_b} = defined $value ? 100755 : 100644;
3479 }
3480 } elsif ($prop eq 'svn:special') {
3481 $fb->{mode_b} = defined $value ? 120000 : 100644;
3482 } else {
3483 $self->{file_prop}->{$fb->{path}} ||= {};
3484 $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
3485 }
3486 undef;
3487 }
3488
3489 sub apply_textdelta {
3490 my ($self, $fb, $exp) = @_;
3491 return undef if is_path_ignored($fb->{path});
3492 my $fh = $::_repository->temp_acquire('svn_delta');
3493 # $fh gets auto-closed() by SVN::TxDelta::apply(),
3494 # (but $base does not,) so dup() it for reading in close_file
3495 open my $dup, '<&', $fh or croak $!;
3496 my $base = $::_repository->temp_acquire('git_blob');
3497
3498 if ($fb->{blob}) {
3499 my ($base_is_link, $size);
3500
3501 if ($fb->{mode_a} eq '120000' &&
3502 ! $self->{empty_symlinks}->{$fb->{path}}) {
3503 print $base 'link ' or die "print $!\n";
3504 $base_is_link = 1;
3505 }
3506 retry:
3507 $size = $::_repository->cat_blob($fb->{blob}, $base);
3508 die "Failed to read object $fb->{blob}" if ($size < 0);
3509
3510 if (defined $exp) {
3511 seek $base, 0, 0 or croak $!;
3512 my $got = ::md5sum($base);
3513 if ($got ne $exp) {
3514 my $err = "Checksum mismatch: ".
3515 "$fb->{path} $fb->{blob}\n" .
3516 "expected: $exp\n" .
3517 " got: $got\n";
3518 if ($base_is_link) {
3519 warn $err,
3520 "Retrying... (possibly ",
3521 "a bad symlink from SVN)\n";
3522 $::_repository->temp_reset($base);
3523 $base_is_link = 0;
3524 goto retry;
3525 }
3526 die $err;
3527 }
3528 }
3529 }
3530 seek $base, 0, 0 or croak $!;
3531 $fb->{fh} = $fh;
3532 $fb->{base} = $base;
3533 [ SVN::TxDelta::apply($base, $dup, undef, $fb->{path}, $fb->{pool}) ];
3534 }
3535
3536 sub close_file {
3537 my ($self, $fb, $exp) = @_;
3538 return undef if is_path_ignored($fb->{path});
3539
3540 my $hash;
3541 my $path = $self->git_path($fb->{path});
3542 if (my $fh = $fb->{fh}) {
3543 if (defined $exp) {
3544 seek($fh, 0, 0) or croak $!;
3545 my $got = ::md5sum($fh);
3546 if ($got ne $exp) {
3547 die "Checksum mismatch: $path\n",
3548 "expected: $exp\n got: $got\n";
3549 }
3550 }
3551 if ($fb->{mode_b} == 120000) {
3552 sysseek($fh, 0, 0) or croak $!;
3553 my $rd = sysread($fh, my $buf, 5);
3554
3555 if (!defined $rd) {
3556 croak "sysread: $!\n";
3557 } elsif ($rd == 0) {
3558 warn "$path has mode 120000",
3559 " but it points to nothing\n",
3560 "converting to an empty file with mode",
3561 " 100644\n";
3562 $fb->{mode_b} = '100644';
3563 } elsif ($buf ne 'link ') {
3564 warn "$path has mode 120000",
3565 " but is not a link\n";
3566 } else {
3567 my $tmp_fh = $::_repository->temp_acquire(
3568 'svn_hash');
3569 my $res;
3570 while ($res = sysread($fh, my $str, 1024)) {
3571 my $out = syswrite($tmp_fh, $str, $res);
3572 defined($out) && $out == $res
3573 or croak("write ",
3574 Git::temp_path($tmp_fh),
3575 ": $!\n");
3576 }
3577 defined $res or croak $!;
3578
3579 ($fh, $tmp_fh) = ($tmp_fh, $fh);
3580 Git::temp_release($tmp_fh, 1);
3581 }
3582 }
3583
3584 $hash = $::_repository->hash_and_insert_object(
3585 Git::temp_path($fh));
3586 $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
3587
3588 Git::temp_release($fb->{base}, 1);
3589 Git::temp_release($fh, 1);
3590 } else {
3591 $hash = $fb->{blob} or die "no blob information\n";
3592 }
3593 $fb->{pool}->clear;
3594 $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
3595 print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
3596 undef;
3597 }
3598
3599 sub abort_edit {
3600 my $self = shift;
3601 $self->{nr} = $self->{gii}->{nr};
3602 delete $self->{gii};
3603 $self->SUPER::abort_edit(@_);
3604 }
3605
3606 sub close_edit {
3607 my $self = shift;
3608 $self->{git_commit_ok} = 1;
3609 $self->{nr} = $self->{gii}->{nr};
3610 delete $self->{gii};
3611 $self->SUPER::close_edit(@_);
3612 }
3613
3614 package SVN::Git::Editor;
3615 use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
3616 use strict;
3617 use warnings;
3618 use Carp qw/croak/;
3619 use IO::File;
3620
3621 sub new {
3622 my ($class, $opts) = @_;
3623 foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
3624 die "$_ required!\n" unless (defined $opts->{$_});
3625 }
3626
3627 my $pool = SVN::Pool->new;
3628 my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
3629 my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
3630 $opts->{r}, $mods);
3631
3632 # $opts->{ra} functions should not be used after this:
3633 my @ce = $opts->{ra}->get_commit_editor($opts->{log},
3634 $opts->{editor_cb}, $pool);
3635 my $self = SVN::Delta::Editor->new(@ce, $pool);
3636 bless $self, $class;
3637 foreach (qw/svn_path r tree_a tree_b/) {
3638 $self->{$_} = $opts->{$_};
3639 }
3640 $self->{url} = $opts->{ra}->{url};
3641 $self->{mods} = $mods;
3642 $self->{types} = $types;
3643 $self->{pool} = $pool;
3644 $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
3645 $self->{rm} = { };
3646 $self->{path_prefix} = length $self->{svn_path} ?
3647 "$self->{svn_path}/" : '';
3648 $self->{config} = $opts->{config};
3649 return $self;
3650 }
3651
3652 sub generate_diff {
3653 my ($tree_a, $tree_b) = @_;
3654 my @diff_tree = qw(diff-tree -z -r);
3655 if ($_cp_similarity) {
3656 push @diff_tree, "-C$_cp_similarity";
3657 } else {
3658 push @diff_tree, '-C';
3659 }
3660 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
3661 push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
3662 push @diff_tree, $tree_a, $tree_b;
3663 my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
3664 local $/ = "\0";
3665 my $state = 'meta';
3666 my @mods;
3667 while (<$diff_fh>) {
3668 chomp $_; # this gets rid of the trailing "\0"
3669 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
3670 ($::sha1)\s($::sha1)\s
3671 ([MTCRAD])\d*$/xo) {
3672 push @mods, { mode_a => $1, mode_b => $2,
3673 sha1_a => $3, sha1_b => $4,
3674 chg => $5 };
3675 if ($5 =~ /^(?:C|R)$/) {
3676 $state = 'file_a';
3677 } else {
3678 $state = 'file_b';
3679 }
3680 } elsif ($state eq 'file_a') {
3681 my $x = $mods[$#mods] or croak "Empty array\n";
3682 if ($x->{chg} !~ /^(?:C|R)$/) {
3683 croak "Error parsing $_, $x->{chg}\n";
3684 }
3685 $x->{file_a} = $_;
3686 $state = 'file_b';
3687 } elsif ($state eq 'file_b') {
3688 my $x = $mods[$#mods] or croak "Empty array\n";
3689 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
3690 croak "Error parsing $_, $x->{chg}\n";
3691 }
3692 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
3693 croak "Error parsing $_, $x->{chg}\n";
3694 }
3695 $x->{file_b} = $_;
3696 $state = 'meta';
3697 } else {
3698 croak "Error parsing $_\n";
3699 }
3700 }
3701 command_close_pipe($diff_fh, $ctx);
3702 \@mods;
3703 }
3704
3705 sub check_diff_paths {
3706 my ($ra, $pfx, $rev, $mods) = @_;
3707 my %types;
3708 $pfx .= '/' if length $pfx;
3709
3710 sub type_diff_paths {
3711 my ($ra, $types, $path, $rev) = @_;
3712 my @p = split m#/+#, $path;
3713 my $c = shift @p;
3714 unless (defined $types->{$c}) {
3715 $types->{$c} = $ra->check_path($c, $rev);
3716 }
3717 while (@p) {
3718 $c .= '/' . shift @p;
3719 next if defined $types->{$c};
3720 $types->{$c} = $ra->check_path($c, $rev);
3721 }
3722 }
3723
3724 foreach my $m (@$mods) {
3725 foreach my $f (qw/file_a file_b/) {
3726 next unless defined $m->{$f};
3727 my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
3728 if (length $pfx.$dir && ! defined $types{$dir}) {
3729 type_diff_paths($ra, \%types, $pfx.$dir, $rev);
3730 }
3731 }
3732 }
3733 \%types;
3734 }
3735
3736 sub split_path {
3737 return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
3738 }
3739
3740 sub repo_path {
3741 my ($self, $path) = @_;
3742 $self->{path_prefix}.(defined $path ? $path : '');
3743 }
3744
3745 sub url_path {
3746 my ($self, $path) = @_;
3747 if ($self->{url} =~ m#^https?://#) {
3748 $path =~ s/([^~a-zA-Z0-9_.-])/uc sprintf("%%%02x",ord($1))/eg;
3749 }
3750 $self->{url} . '/' . $self->repo_path($path);
3751 }
3752
3753 sub rmdirs {
3754 my ($self) = @_;
3755 my $rm = $self->{rm};
3756 delete $rm->{''}; # we never delete the url we're tracking
3757 return unless %$rm;
3758
3759 foreach (keys %$rm) {
3760 my @d = split m#/#, $_;
3761 my $c = shift @d;
3762 $rm->{$c} = 1;
3763 while (@d) {
3764 $c .= '/' . shift @d;
3765 $rm->{$c} = 1;
3766 }
3767 }
3768 delete $rm->{$self->{svn_path}};
3769 delete $rm->{''}; # we never delete the url we're tracking
3770 return unless %$rm;
3771
3772 my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
3773 $self->{tree_b});
3774 local $/ = "\0";
3775 while (<$fh>) {
3776 chomp;
3777 my @dn = split m#/#, $_;
3778 while (pop @dn) {
3779 delete $rm->{join '/', @dn};
3780 }
3781 unless (%$rm) {
3782 close $fh;
3783 return;
3784 }
3785 }
3786 command_close_pipe($fh, $ctx);
3787
3788 my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
3789 foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
3790 $self->close_directory($bat->{$d}, $p);
3791 my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
3792 print "\tD+\t$d/\n" unless $::_q;
3793 $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
3794 delete $bat->{$d};
3795 }
3796 }
3797
3798 sub open_or_add_dir {
3799 my ($self, $full_path, $baton) = @_;
3800 my $t = $self->{types}->{$full_path};
3801 if (!defined $t) {
3802 die "$full_path not known in r$self->{r} or we have a bug!\n";
3803 }
3804 {
3805 no warnings 'once';
3806 # SVN::Node::none and SVN::Node::file are used only once,
3807 # so we're shutting up Perl's warnings about them.
3808 if ($t == $SVN::Node::none) {
3809 return $self->add_directory($full_path, $baton,
3810 undef, -1, $self->{pool});
3811 } elsif ($t == $SVN::Node::dir) {
3812 return $self->open_directory($full_path, $baton,
3813 $self->{r}, $self->{pool});
3814 } # no warnings 'once'
3815 print STDERR "$full_path already exists in repository at ",
3816 "r$self->{r} and it is not a directory (",
3817 ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
3818 } # no warnings 'once'
3819 exit 1;
3820 }
3821
3822 sub ensure_path {
3823 my ($self, $path) = @_;
3824 my $bat = $self->{bat};
3825 my $repo_path = $self->repo_path($path);
3826 return $bat->{''} unless (length $repo_path);
3827 my @p = split m#/+#, $repo_path;
3828 my $c = shift @p;
3829 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
3830 while (@p) {
3831 my $c0 = $c;
3832 $c .= '/' . shift @p;
3833 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
3834 }
3835 return $bat->{$c};
3836 }
3837
3838 # Subroutine to convert a globbing pattern to a regular expression.
3839 # From perl cookbook.
3840 sub glob2pat {
3841 my $globstr = shift;
3842 my %patmap = ('*' => '.*', '?' => '.', '[' => '[', ']' => ']');
3843 $globstr =~ s{(.)} { $patmap{$1} || "\Q$1" }ge;
3844 return '^' . $globstr . '$';
3845 }
3846
3847 sub check_autoprop {
3848 my ($self, $pattern, $properties, $file, $fbat) = @_;
3849 # Convert the globbing pattern to a regular expression.
3850 my $regex = glob2pat($pattern);
3851 # Check if the pattern matches the file name.
3852 if($file =~ m/($regex)/) {
3853 # Parse the list of properties to set.
3854 my @props = split(/;/, $properties);
3855 foreach my $prop (@props) {
3856 # Parse 'name=value' syntax and set the property.
3857 if ($prop =~ /([^=]+)=(.*)/) {
3858 my ($n,$v) = ($1,$2);
3859 for ($n, $v) {
3860 s/^\s+//; s/\s+$//;
3861 }
3862 $self->change_file_prop($fbat, $n, $v);
3863 }
3864 }
3865 }
3866 }
3867
3868 sub apply_autoprops {
3869 my ($self, $file, $fbat) = @_;
3870 my $conf_t = ${$self->{config}}{'config'};
3871 no warnings 'once';
3872 # Check [miscellany]/enable-auto-props in svn configuration.
3873 if (SVN::_Core::svn_config_get_bool(
3874 $conf_t,
3875 $SVN::_Core::SVN_CONFIG_SECTION_MISCELLANY,
3876 $SVN::_Core::SVN_CONFIG_OPTION_ENABLE_AUTO_PROPS,
3877 0)) {
3878 # Auto-props are enabled. Enumerate them to look for matches.
3879 my $callback = sub {
3880 $self->check_autoprop($_[0], $_[1], $file, $fbat);
3881 };
3882 SVN::_Core::svn_config_enumerate(
3883 $conf_t,
3884 $SVN::_Core::SVN_CONFIG_SECTION_AUTO_PROPS,
3885 $callback);
3886 }
3887 }
3888
3889 sub A {
3890 my ($self, $m) = @_;
3891 my ($dir, $file) = split_path($m->{file_b});
3892 my $pbat = $self->ensure_path($dir);
3893 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3894 undef, -1);
3895 print "\tA\t$m->{file_b}\n" unless $::_q;
3896 $self->apply_autoprops($file, $fbat);
3897 $self->chg_file($fbat, $m);
3898 $self->close_file($fbat,undef,$self->{pool});
3899 }
3900
3901 sub C {
3902 my ($self, $m) = @_;
3903 my ($dir, $file) = split_path($m->{file_b});
3904 my $pbat = $self->ensure_path($dir);
3905 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3906 $self->url_path($m->{file_a}), $self->{r});
3907 print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3908 $self->chg_file($fbat, $m);
3909 $self->close_file($fbat,undef,$self->{pool});
3910 }
3911
3912 sub delete_entry {
3913 my ($self, $path, $pbat) = @_;
3914 my $rpath = $self->repo_path($path);
3915 my ($dir, $file) = split_path($rpath);
3916 $self->{rm}->{$dir} = 1;
3917 $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
3918 }
3919
3920 sub R {
3921 my ($self, $m) = @_;
3922 my ($dir, $file) = split_path($m->{file_b});
3923 my $pbat = $self->ensure_path($dir);
3924 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3925 $self->url_path($m->{file_a}), $self->{r});
3926 print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3927 $self->apply_autoprops($file, $fbat);
3928 $self->chg_file($fbat, $m);
3929 $self->close_file($fbat,undef,$self->{pool});
3930
3931 ($dir, $file) = split_path($m->{file_a});
3932 $pbat = $self->ensure_path($dir);
3933 $self->delete_entry($m->{file_a}, $pbat);
3934 }
3935
3936 sub M {
3937 my ($self, $m) = @_;
3938 my ($dir, $file) = split_path($m->{file_b});
3939 my $pbat = $self->ensure_path($dir);
3940 my $fbat = $self->open_file($self->repo_path($m->{file_b}),
3941 $pbat,$self->{r},$self->{pool});
3942 print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
3943 $self->chg_file($fbat, $m);
3944 $self->close_file($fbat,undef,$self->{pool});
3945 }
3946
3947 sub T { shift->M(@_) }
3948
3949 sub change_file_prop {
3950 my ($self, $fbat, $pname, $pval) = @_;
3951 $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
3952 }
3953
3954 sub _chg_file_get_blob ($$$$) {
3955 my ($self, $fbat, $m, $which) = @_;
3956 my $fh = $::_repository->temp_acquire("git_blob_$which");
3957 if ($m->{"mode_$which"} =~ /^120/) {
3958 print $fh 'link ' or croak $!;
3959 $self->change_file_prop($fbat,'svn:special','*');
3960 } elsif ($m->{mode_a} =~ /^120/ && $m->{"mode_$which"} !~ /^120/) {
3961 $self->change_file_prop($fbat,'svn:special',undef);
3962 }
3963 my $blob = $m->{"sha1_$which"};
3964 return ($fh,) if ($blob =~ /^0{40}$/);
3965 my $size = $::_repository->cat_blob($blob, $fh);
3966 croak "Failed to read object $blob" if ($size < 0);
3967 $fh->flush == 0 or croak $!;
3968 seek $fh, 0, 0 or croak $!;
3969
3970 my $exp = ::md5sum($fh);
3971 seek $fh, 0, 0 or croak $!;
3972 return ($fh, $exp);
3973 }
3974
3975 sub chg_file {
3976 my ($self, $fbat, $m) = @_;
3977 if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
3978 $self->change_file_prop($fbat,'svn:executable','*');
3979 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
3980 $self->change_file_prop($fbat,'svn:executable',undef);
3981 }
3982 my ($fh_a, $exp_a) = _chg_file_get_blob $self, $fbat, $m, 'a';
3983 my ($fh_b, $exp_b) = _chg_file_get_blob $self, $fbat, $m, 'b';
3984 my $pool = SVN::Pool->new;
3985 my $atd = $self->apply_textdelta($fbat, $exp_a, $pool);
3986 if (-s $fh_a) {
3987 my $txstream = SVN::TxDelta::new ($fh_a, $fh_b, $pool);
3988 my $res = SVN::TxDelta::send_txstream($txstream, @$atd, $pool);
3989 if (defined $res) {
3990 die "Unexpected result from send_txstream: $res\n",
3991 "(SVN::Core::VERSION: $SVN::Core::VERSION)\n";
3992 }
3993 } else {
3994 my $got = SVN::TxDelta::send_stream($fh_b, @$atd, $pool);
3995 die "Checksum mismatch\nexpected: $exp_b\ngot: $got\n"
3996 if ($got ne $exp_b);
3997 }
3998 Git::temp_release($fh_b, 1);
3999 Git::temp_release($fh_a, 1);
4000 $pool->clear;
4001 }
4002
4003 sub D {
4004 my ($self, $m) = @_;
4005 my ($dir, $file) = split_path($m->{file_b});
4006 my $pbat = $self->ensure_path($dir);
4007 print "\tD\t$m->{file_b}\n" unless $::_q;
4008 $self->delete_entry($m->{file_b}, $pbat);
4009 }
4010
4011 sub close_edit {
4012 my ($self) = @_;
4013 my ($p,$bat) = ($self->{pool}, $self->{bat});
4014 foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
4015 next if $_ eq '';
4016 $self->close_directory($bat->{$_}, $p);
4017 }
4018 $self->close_directory($bat->{''}, $p);
4019 $self->SUPER::close_edit($p);
4020 $p->clear;
4021 }
4022
4023 sub abort_edit {
4024 my ($self) = @_;
4025 $self->SUPER::abort_edit($self->{pool});
4026 }
4027
4028 sub DESTROY {
4029 my $self = shift;
4030 $self->SUPER::DESTROY(@_);
4031 $self->{pool}->clear;
4032 }
4033
4034 # this drives the editor
4035 sub apply_diff {
4036 my ($self) = @_;
4037 my $mods = $self->{mods};
4038 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
4039 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
4040 my $f = $m->{chg};
4041 if (defined $o{$f}) {
4042 $self->$f($m);
4043 } else {
4044 fatal("Invalid change type: $f");
4045 }
4046 }
4047 $self->rmdirs if $_rmdir;
4048 if (@$mods == 0) {
4049 $self->abort_edit;
4050 } else {
4051 $self->close_edit;
4052 }
4053 return scalar @$mods;
4054 }
4055
4056 package Git::SVN::Ra;
4057 use vars qw/@ISA $config_dir $_log_window_size/;
4058 use strict;
4059 use warnings;
4060 my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
4061
4062 BEGIN {
4063 # enforce temporary pool usage for some simple functions
4064 no strict 'refs';
4065 for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root
4066 get_file/) {
4067 my $SUPER = "SUPER::$f";
4068 *$f = sub {
4069 my $self = shift;
4070 my $pool = SVN::Pool->new;
4071 my @ret = $self->$SUPER(@_,$pool);
4072 $pool->clear;
4073 wantarray ? @ret : $ret[0];
4074 };
4075 }
4076 }
4077
4078 sub _auth_providers () {
4079 [
4080 SVN::Client::get_simple_provider(),
4081 SVN::Client::get_ssl_server_trust_file_provider(),
4082 SVN::Client::get_simple_prompt_provider(
4083 \&Git::SVN::Prompt::simple, 2),
4084 SVN::Client::get_ssl_client_cert_file_provider(),
4085 SVN::Client::get_ssl_client_cert_prompt_provider(
4086 \&Git::SVN::Prompt::ssl_client_cert, 2),
4087 SVN::Client::get_ssl_client_cert_pw_file_provider(),
4088 SVN::Client::get_ssl_client_cert_pw_prompt_provider(
4089 \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
4090 SVN::Client::get_username_provider(),
4091 SVN::Client::get_ssl_server_trust_prompt_provider(
4092 \&Git::SVN::Prompt::ssl_server_trust),
4093 SVN::Client::get_username_prompt_provider(
4094 \&Git::SVN::Prompt::username, 2)
4095 ]
4096 }
4097
4098 sub escape_uri_only {
4099 my ($uri) = @_;
4100 my @tmp;
4101 foreach (split m{/}, $uri) {
4102 s/([^~\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
4103 push @tmp, $_;
4104 }
4105 join('/', @tmp);
4106 }
4107
4108 sub escape_url {
4109 my ($url) = @_;
4110 if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
4111 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
4112 $url = "$scheme://$domain$uri";
4113 }
4114 $url;
4115 }
4116
4117 sub new {
4118 my ($class, $url) = @_;
4119 $url =~ s!/+$!!;
4120 return $RA if ($RA && $RA->{url} eq $url);
4121
4122 SVN::_Core::svn_config_ensure($config_dir, undef);
4123 my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
4124 my $config = SVN::Core::config_get_config($config_dir);
4125 $RA = undef;
4126 my $dont_store_passwords = 1;
4127 my $conf_t = ${$config}{'config'};
4128 {
4129 no warnings 'once';
4130 # The usage of $SVN::_Core::SVN_CONFIG_* variables
4131 # produces warnings that variables are used only once.
4132 # I had not found the better way to shut them up, so
4133 # the warnings of type 'once' are disabled in this block.
4134 if (SVN::_Core::svn_config_get_bool($conf_t,
4135 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
4136 $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
4137 1) == 0) {
4138 SVN::_Core::svn_auth_set_parameter($baton,
4139 $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
4140 bless (\$dont_store_passwords, "_p_void"));
4141 }
4142 if (SVN::_Core::svn_config_get_bool($conf_t,
4143 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
4144 $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
4145 1) == 0) {
4146 $Git::SVN::Prompt::_no_auth_cache = 1;
4147 }
4148 } # no warnings 'once'
4149 my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
4150 config => $config,
4151 pool => SVN::Pool->new,
4152 auth_provider_callbacks => $callbacks);
4153 $self->{url} = $url;
4154 $self->{svn_path} = $url;
4155 $self->{repos_root} = $self->get_repos_root;
4156 $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
4157 $self->{cache} = { check_path => { r => 0, data => {} },
4158 get_dir => { r => 0, data => {} } };
4159 $RA = bless $self, $class;
4160 }
4161
4162 sub check_path {
4163 my ($self, $path, $r) = @_;
4164 my $cache = $self->{cache}->{check_path};
4165 if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
4166 return $cache->{data}->{$path};
4167 }
4168 my $pool = SVN::Pool->new;
4169 my $t = $self->SUPER::check_path($path, $r, $pool);
4170 $pool->clear;
4171 if ($r != $cache->{r}) {
4172 %{$cache->{data}} = ();
4173 $cache->{r} = $r;
4174 }
4175 $cache->{data}->{$path} = $t;
4176 }
4177
4178 sub get_dir {
4179 my ($self, $dir, $r) = @_;
4180 my $cache = $self->{cache}->{get_dir};
4181 if ($r == $cache->{r}) {
4182 if (my $x = $cache->{data}->{$dir}) {
4183 return wantarray ? @$x : $x->[0];
4184 }
4185 }
4186 my $pool = SVN::Pool->new;
4187 my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
4188 my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
4189 $pool->clear;
4190 if ($r != $cache->{r}) {
4191 %{$cache->{data}} = ();
4192 $cache->{r} = $r;
4193 }
4194 $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
4195 wantarray ? (\%dirents, $r, $props) : \%dirents;
4196 }
4197
4198 sub DESTROY {
4199 # do not call the real DESTROY since we store ourselves in $RA
4200 }
4201
4202 # get_log(paths, start, end, limit,
4203 # discover_changed_paths, strict_node_history, receiver)
4204 sub get_log {
4205 my ($self, @args) = @_;
4206 my $pool = SVN::Pool->new;
4207
4208 # the limit parameter was not supported in SVN 1.1.x, so we
4209 # drop it. Therefore, the receiver callback passed to it
4210 # is made aware of this limitation by being wrapped if
4211 # the limit passed to is being wrapped.
4212 if ($SVN::Core::VERSION le '1.2.0') {
4213 my $limit = splice(@args, 3, 1);
4214 if ($limit > 0) {
4215 my $receiver = pop @args;
4216 push(@args, sub { &$receiver(@_) if (--$limit >= 0) });
4217 }
4218 }
4219 my $ret = $self->SUPER::get_log(@args, $pool);
4220 $pool->clear;
4221 $ret;
4222 }
4223
4224 sub trees_match {
4225 my ($self, $url1, $rev1, $url2, $rev2) = @_;
4226 my $ctx = SVN::Client->new(auth => _auth_providers);
4227 my $out = IO::File->new_tmpfile;
4228
4229 # older SVN (1.1.x) doesn't take $pool as the last parameter for
4230 # $ctx->diff(), so we'll create a default one
4231 my $pool = SVN::Pool->new_default_sub;
4232
4233 $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
4234 $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
4235 $out->flush;
4236 my $ret = (($out->stat)[7] == 0);
4237 close $out or croak $!;
4238
4239 $ret;
4240 }
4241
4242 sub get_commit_editor {
4243 my ($self, $log, $cb, $pool) = @_;
4244 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
4245 $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
4246 }
4247
4248 sub gs_do_update {
4249 my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
4250 my $new = ($rev_a == $rev_b);
4251 my $path = $gs->{path};
4252
4253 if ($new && -e $gs->{index}) {
4254 unlink $gs->{index} or die
4255 "Couldn't unlink index: $gs->{index}: $!\n";
4256 }
4257 my $pool = SVN::Pool->new;
4258 $editor->set_path_strip($path);
4259 my (@pc) = split m#/#, $path;
4260 my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
4261 1, $editor, $pool);
4262 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
4263
4264 # Since we can't rely on svn_ra_reparent being available, we'll
4265 # just have to do some magic with set_path to make it so
4266 # we only want a partial path.
4267 my $sp = '';
4268 my $final = join('/', @pc);
4269 while (@pc) {
4270 $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
4271 $sp .= '/' if length $sp;
4272 $sp .= shift @pc;
4273 }
4274 die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
4275
4276 $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
4277
4278 $reporter->finish_report($pool);
4279 $pool->clear;
4280 $editor->{git_commit_ok};
4281 }
4282
4283 # this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
4284 # svn_ra_reparent didn't work before 1.4)
4285 sub gs_do_switch {
4286 my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
4287 my $path = $gs->{path};
4288 my $pool = SVN::Pool->new;
4289
4290 my $full_url = $self->{url};
4291 my $old_url = $full_url;
4292 $full_url .= '/' . escape_uri_only($path) if length $path;
4293 my ($ra, $reparented);
4294
4295 if ($old_url =~ m#^svn(\+ssh)?://#) {
4296 $_[0] = undef;
4297 $self = undef;
4298 $RA = undef;
4299 $ra = Git::SVN::Ra->new($full_url);
4300 $ra_invalid = 1;
4301 } elsif ($old_url ne $full_url) {
4302 SVN::_Ra::svn_ra_reparent($self->{session}, $full_url, $pool);
4303 $self->{url} = $full_url;
4304 $reparented = 1;
4305 }
4306
4307 $ra ||= $self;
4308 $url_b = escape_url($url_b);
4309 my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
4310 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
4311 $reporter->set_path('', $rev_a, 0, @lock, $pool);
4312 $reporter->finish_report($pool);
4313
4314 if ($reparented) {
4315 SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
4316 $self->{url} = $old_url;
4317 }
4318
4319 $pool->clear;
4320 $editor->{git_commit_ok};
4321 }
4322
4323 sub longest_common_path {
4324 my ($gsv, $globs) = @_;
4325 my %common;
4326 my $common_max = scalar @$gsv;
4327
4328 foreach my $gs (@$gsv) {
4329 my @tmp = split m#/#, $gs->{path};
4330 my $p = '';
4331 foreach (@tmp) {
4332 $p .= length($p) ? "/$_" : $_;
4333 $common{$p} ||= 0;
4334 $common{$p}++;
4335 }
4336 }
4337 $globs ||= [];
4338 $common_max += scalar @$globs;
4339 foreach my $glob (@$globs) {
4340 my @tmp = split m#/#, $glob->{path}->{left};
4341 my $p = '';
4342 foreach (@tmp) {
4343 $p .= length($p) ? "/$_" : $_;
4344 $common{$p} ||= 0;
4345 $common{$p}++;
4346 }
4347 }
4348
4349 my $longest_path = '';
4350 foreach (sort {length $b <=> length $a} keys %common) {
4351 if ($common{$_} == $common_max) {
4352 $longest_path = $_;
4353 last;
4354 }
4355 }
4356 $longest_path;
4357 }
4358
4359 sub gs_fetch_loop_common {
4360 my ($self, $base, $head, $gsv, $globs) = @_;
4361 return if ($base > $head);
4362 my $inc = $_log_window_size;
4363 my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
4364 my $longest_path = longest_common_path($gsv, $globs);
4365 my $ra_url = $self->{url};
4366 while (1) {
4367 my %revs;
4368 my $err;
4369 my $err_handler = $SVN::Error::handler;
4370 $SVN::Error::handler = sub {
4371 ($err) = @_;
4372 skip_unknown_revs($err);
4373 };
4374 sub _cb {
4375 my ($paths, $r, $author, $date, $log) = @_;
4376 [ dup_changed_paths($paths),
4377 { author => $author, date => $date, log => $log } ];
4378 }
4379 $self->get_log([$longest_path], $min, $max, 0, 1, 1,
4380 sub { $revs{$_[1]} = _cb(@_) });
4381 if ($err) {
4382 print "Checked through r$max\r";
4383 }
4384 if ($err && $max >= $head) {
4385 print STDERR "Path '$longest_path' ",
4386 "was probably deleted:\n",
4387 $err->expanded_message,
4388 "\nWill attempt to follow ",
4389 "revisions r$min .. r$max ",
4390 "committed before the deletion\n";
4391 my $hi = $max;
4392 while (--$hi >= $min) {
4393 my $ok;
4394 $self->get_log([$longest_path], $min, $hi,
4395 0, 1, 1, sub {
4396 $ok ||= $_[1];
4397 $revs{$_[1]} = _cb(@_) });
4398 if ($ok) {
4399 print STDERR "r$min .. r$ok OK\n";
4400 last;
4401 }
4402 }
4403 }
4404 $SVN::Error::handler = $err_handler;
4405
4406 my %exists = map { $_->{path} => $_ } @$gsv;
4407 foreach my $r (sort {$a <=> $b} keys %revs) {
4408 my ($paths, $logged) = @{$revs{$r}};
4409
4410 foreach my $gs ($self->match_globs(\%exists, $paths,
4411 $globs, $r)) {
4412 if ($gs->rev_map_max >= $r) {
4413 next;
4414 }
4415 next unless $gs->match_paths($paths, $r);
4416 $gs->{logged_rev_props} = $logged;
4417 if (my $last_commit = $gs->last_commit) {
4418 $gs->assert_index_clean($last_commit);
4419 }
4420 my $log_entry = $gs->do_fetch($paths, $r);
4421 if ($log_entry) {
4422 $gs->do_git_commit($log_entry);
4423 }
4424 $INDEX_FILES{$gs->{index}} = 1;
4425 }
4426 foreach my $g (@$globs) {
4427 my $k = "svn-remote.$g->{remote}." .
4428 "$g->{t}-maxRev";
4429 Git::SVN::tmp_config($k, $r);
4430 }
4431 if ($ra_invalid) {
4432 $_[0] = undef;
4433 $self = undef;
4434 $RA = undef;
4435 $self = Git::SVN::Ra->new($ra_url);
4436 $ra_invalid = undef;
4437 }
4438 }
4439 # pre-fill the .rev_db since it'll eventually get filled in
4440 # with '0' x40 if something new gets committed
4441 foreach my $gs (@$gsv) {
4442 next if $gs->rev_map_max >= $max;
4443 next if defined $gs->rev_map_get($max);
4444 $gs->rev_map_set($max, 0 x40);
4445 }
4446 foreach my $g (@$globs) {
4447 my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
4448 Git::SVN::tmp_config($k, $max);
4449 }
4450 last if $max >= $head;
4451 $min = $max + 1;
4452 $max += $inc;
4453 $max = $head if ($max > $head);
4454 }
4455 Git::SVN::gc();
4456 }
4457
4458 sub get_dir_globbed {
4459 my ($self, $left, $depth, $r) = @_;
4460
4461 my @x = eval { $self->get_dir($left, $r) };
4462 return unless scalar @x == 3;
4463 my $dirents = $x[0];
4464 my @finalents;
4465 foreach my $de (keys %$dirents) {
4466 next if $dirents->{$de}->{kind} != $SVN::Node::dir;
4467 if ($depth > 1) {
4468 my @args = ("$left/$de", $depth - 1, $r);
4469 foreach my $dir ($self->get_dir_globbed(@args)) {
4470 push @finalents, "$de/$dir";
4471 }
4472 } else {
4473 push @finalents, $de;
4474 }
4475 }
4476 @finalents;
4477 }
4478
4479 sub match_globs {
4480 my ($self, $exists, $paths, $globs, $r) = @_;
4481
4482 sub get_dir_check {
4483 my ($self, $exists, $g, $r) = @_;
4484
4485 my @dirs = $self->get_dir_globbed($g->{path}->{left},
4486 $g->{path}->{depth},
4487 $r);
4488
4489 foreach my $de (@dirs) {
4490 my $p = $g->{path}->full_path($de);
4491 next if $exists->{$p};
4492 next if (length $g->{path}->{right} &&
4493 ($self->check_path($p, $r) !=
4494 $SVN::Node::dir));
4495 $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
4496 $g->{ref}->full_path($de), 1);
4497 }
4498 }
4499 foreach my $g (@$globs) {
4500 if (my $path = $paths->{"/$g->{path}->{left}"}) {
4501 if ($path->{action} =~ /^[AR]$/) {
4502 get_dir_check($self, $exists, $g, $r);
4503 }
4504 }
4505 foreach (keys %$paths) {
4506 if (/$g->{path}->{left_regex}/ &&
4507 !/$g->{path}->{regex}/) {
4508 next if $paths->{$_}->{action} !~ /^[AR]$/;
4509 get_dir_check($self, $exists, $g, $r);
4510 }
4511 next unless /$g->{path}->{regex}/;
4512 my $p = $1;
4513 my $pathname = $g->{path}->full_path($p);
4514 next if $exists->{$pathname};
4515 next if ($self->check_path($pathname, $r) !=
4516 $SVN::Node::dir);
4517 $exists->{$pathname} = Git::SVN->init(
4518 $self->{url}, $pathname, undef,
4519 $g->{ref}->full_path($p), 1);
4520 }
4521 my $c = '';
4522 foreach (split m#/#, $g->{path}->{left}) {
4523 $c .= "/$_";
4524 next unless ($paths->{$c} &&
4525 ($paths->{$c}->{action} =~ /^[AR]$/));
4526 get_dir_check($self, $exists, $g, $r);
4527 }
4528 }
4529 values %$exists;
4530 }
4531
4532 sub minimize_url {
4533 my ($self) = @_;
4534 return $self->{url} if ($self->{url} eq $self->{repos_root});
4535 my $url = $self->{repos_root};
4536 my @components = split(m!/!, $self->{svn_path});
4537 my $c = '';
4538 do {
4539 $url .= "/$c" if length $c;
4540 eval { (ref $self)->new($url)->get_latest_revnum };
4541 } while ($@ && ($c = shift @components));
4542 $url;
4543 }
4544
4545 sub can_do_switch {
4546 my $self = shift;
4547 unless (defined $can_do_switch) {
4548 my $pool = SVN::Pool->new;
4549 my $rep = eval {
4550 $self->do_switch(1, '', 0, $self->{url},
4551 SVN::Delta::Editor->new, $pool);
4552 };
4553 if ($@) {
4554 $can_do_switch = 0;
4555 } else {
4556 $rep->abort_report($pool);
4557 $can_do_switch = 1;
4558 }
4559 $pool->clear;
4560 }
4561 $can_do_switch;
4562 }
4563
4564 sub skip_unknown_revs {
4565 my ($err) = @_;
4566 my $errno = $err->apr_err();
4567 # Maybe the branch we're tracking didn't
4568 # exist when the repo started, so it's
4569 # not an error if it doesn't, just continue
4570 #
4571 # Wonderfully consistent library, eh?
4572 # 160013 - svn:// and file://
4573 # 175002 - http(s)://
4574 # 175007 - http(s):// (this repo required authorization, too...)
4575 # More codes may be discovered later...
4576 if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
4577 my $err_key = $err->expanded_message;
4578 # revision numbers change every time, filter them out
4579 $err_key =~ s/\d+/\0/g;
4580 $err_key = "$errno\0$err_key";
4581 unless ($ignored_err{$err_key}) {
4582 warn "W: Ignoring error from SVN, path probably ",
4583 "does not exist: ($errno): ",
4584 $err->expanded_message,"\n";
4585 warn "W: Do not be alarmed at the above message ",
4586 "git-svn is just searching aggressively for ",
4587 "old history.\n",
4588 "This may take a while on large repositories\n";
4589 $ignored_err{$err_key} = 1;
4590 }
4591 return;
4592 }
4593 die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
4594 }
4595
4596 # svn_log_changed_path_t objects passed to get_log are likely to be
4597 # overwritten even if only the refs are copied to an external variable,
4598 # so we should dup the structures in their entirety. Using an externally
4599 # passed pool (instead of our temporary and quickly cleared pool in
4600 # Git::SVN::Ra) does not help matters at all...
4601 sub dup_changed_paths {
4602 my ($paths) = @_;
4603 return undef unless $paths;
4604 my %ret;
4605 foreach my $p (keys %$paths) {
4606 my $i = $paths->{$p};
4607 my %s = map { $_ => $i->$_ }
4608 qw/copyfrom_path copyfrom_rev action/;
4609 $ret{$p} = \%s;
4610 }
4611 \%ret;
4612 }
4613
4614 package Git::SVN::Log;
4615 use strict;
4616 use warnings;
4617 use POSIX qw/strftime/;
4618 use constant commit_log_separator => ('-' x 72) . "\n";
4619 use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
4620 %rusers $show_commit $incremental/;
4621 my $l_fmt;
4622
4623 sub cmt_showable {
4624 my ($c) = @_;
4625 return 1 if defined $c->{r};
4626
4627 # big commit message got truncated by the 16k pretty buffer in rev-list
4628 if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
4629 $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
4630 @{$c->{l}} = ();
4631 my @log = command(qw/cat-file commit/, $c->{c});
4632
4633 # shift off the headers
4634 shift @log while ($log[0] ne '');
4635 shift @log;
4636
4637 # TODO: make $c->{l} not have a trailing newline in the future
4638 @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
4639
4640 (undef, $c->{r}, undef) = ::extract_metadata(
4641 (grep(/^git-svn-id: /, @log))[-1]);
4642 }
4643 return defined $c->{r};
4644 }
4645
4646 sub log_use_color {
4647 return $color || Git->repository->get_colorbool('color.diff');
4648 }
4649
4650 sub git_svn_log_cmd {
4651 my ($r_min, $r_max, @args) = @_;
4652 my $head = 'HEAD';
4653 my (@files, @log_opts);
4654 foreach my $x (@args) {
4655 if ($x eq '--' || @files) {
4656 push @files, $x;
4657 } else {
4658 if (::verify_ref("$x^0")) {
4659 $head = $x;
4660 } else {
4661 push @log_opts, $x;
4662 }
4663 }
4664 }
4665
4666 my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
4667 $gs ||= Git::SVN->_new;
4668 my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
4669 $gs->refname);
4670 push @cmd, '-r' unless $non_recursive;
4671 push @cmd, qw/--raw --name-status/ if $verbose;
4672 push @cmd, '--color' if log_use_color();
4673 push @cmd, @log_opts;
4674 if (defined $r_max && $r_max == $r_min) {
4675 push @cmd, '--max-count=1';
4676 if (my $c = $gs->rev_map_get($r_max)) {
4677 push @cmd, $c;
4678 }
4679 } elsif (defined $r_max) {
4680 if ($r_max < $r_min) {
4681 ($r_min, $r_max) = ($r_max, $r_min);
4682 }
4683 my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
4684 my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
4685 # If there are no commits in the range, both $c_max and $c_min
4686 # will be undefined. If there is at least 1 commit in the
4687 # range, both will be defined.
4688 return () if !defined $c_min || !defined $c_max;
4689 if ($c_min eq $c_max) {
4690 push @cmd, '--max-count=1', $c_min;
4691 } else {
4692 push @cmd, '--boundary', "$c_min..$c_max";
4693 }
4694 }
4695 return (@cmd, @files);
4696 }
4697
4698 # adapted from pager.c
4699 sub config_pager {
4700 $pager ||= $ENV{GIT_PAGER} || $ENV{PAGER};
4701 if (!defined $pager) {
4702 $pager = 'less';
4703 } elsif (length $pager == 0 || $pager eq 'cat') {
4704 $pager = undef;
4705 }
4706 $ENV{GIT_PAGER_IN_USE} = defined($pager);
4707 }
4708
4709 sub run_pager {
4710 return unless -t *STDOUT && defined $pager;
4711 pipe my ($rfd, $wfd) or return;
4712 defined(my $pid = fork) or ::fatal "Can't fork: $!";
4713 if (!$pid) {
4714 open STDOUT, '>&', $wfd or
4715 ::fatal "Can't redirect to stdout: $!";
4716 return;
4717 }
4718 open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
4719 $ENV{LESS} ||= 'FRSX';
4720 exec $pager or ::fatal "Can't run pager: $! ($pager)";
4721 }
4722
4723 sub format_svn_date {
4724 return strftime("%Y-%m-%d %H:%M:%S %z (%a, %d %b %Y)", localtime(shift));
4725 }
4726
4727 sub parse_git_date {
4728 my ($t, $tz) = @_;
4729 # Date::Parse isn't in the standard Perl distro :(
4730 if ($tz =~ s/^\+//) {
4731 $t += tz_to_s_offset($tz);
4732 } elsif ($tz =~ s/^\-//) {
4733 $t -= tz_to_s_offset($tz);
4734 }
4735 return $t;
4736 }
4737
4738 sub set_local_timezone {
4739 if (defined $TZ) {
4740 $ENV{TZ} = $TZ;
4741 } else {
4742 delete $ENV{TZ};
4743 }
4744 }
4745
4746 sub tz_to_s_offset {
4747 my ($tz) = @_;
4748 $tz =~ s/(\d\d)$//;
4749 return ($1 * 60) + ($tz * 3600);
4750 }
4751
4752 sub get_author_info {
4753 my ($dest, $author, $t, $tz) = @_;
4754 $author =~ s/(?:^\s*|\s*$)//g;
4755 $dest->{a_raw} = $author;
4756 my $au;
4757 if ($::_authors) {
4758 $au = $rusers{$author} || undef;
4759 }
4760 if (!$au) {
4761 ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
4762 }
4763 $dest->{t} = $t;
4764 $dest->{tz} = $tz;
4765 $dest->{a} = $au;
4766 $dest->{t_utc} = parse_git_date($t, $tz);
4767 }
4768
4769 sub process_commit {
4770 my ($c, $r_min, $r_max, $defer) = @_;
4771 if (defined $r_min && defined $r_max) {
4772 if ($r_min == $c->{r} && $r_min == $r_max) {
4773 show_commit($c);
4774 return 0;
4775 }
4776 return 1 if $r_min == $r_max;
4777 if ($r_min < $r_max) {
4778 # we need to reverse the print order
4779 return 0 if (defined $limit && --$limit < 0);
4780 push @$defer, $c;
4781 return 1;
4782 }
4783 if ($r_min != $r_max) {
4784 return 1 if ($r_min < $c->{r});
4785 return 1 if ($r_max > $c->{r});
4786 }
4787 }
4788 return 0 if (defined $limit && --$limit < 0);
4789 show_commit($c);
4790 return 1;
4791 }
4792
4793 sub show_commit {
4794 my $c = shift;
4795 if ($oneline) {
4796 my $x = "\n";
4797 if (my $l = $c->{l}) {
4798 while ($l->[0] =~ /^\s*$/) { shift @$l }
4799 $x = $l->[0];
4800 }
4801 $l_fmt ||= 'A' . length($c->{r});
4802 print 'r',pack($l_fmt, $c->{r}),' | ';
4803 print "$c->{c} | " if $show_commit;
4804 print $x;
4805 } else {
4806 show_commit_normal($c);
4807 }
4808 }
4809
4810 sub show_commit_changed_paths {
4811 my ($c) = @_;
4812 return unless $c->{changed};
4813 print "Changed paths:\n", @{$c->{changed}};
4814 }
4815
4816 sub show_commit_normal {
4817 my ($c) = @_;
4818 print commit_log_separator, "r$c->{r} | ";
4819 print "$c->{c} | " if $show_commit;
4820 print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
4821 my $nr_line = 0;
4822
4823 if (my $l = $c->{l}) {
4824 while ($l->[$#$l] eq "\n" && $#$l > 0
4825 && $l->[($#$l - 1)] eq "\n") {
4826 pop @$l;
4827 }
4828 $nr_line = scalar @$l;
4829 if (!$nr_line) {
4830 print "1 line\n\n\n";
4831 } else {
4832 if ($nr_line == 1) {
4833 $nr_line = '1 line';
4834 } else {
4835 $nr_line .= ' lines';
4836 }
4837 print $nr_line, "\n";
4838 show_commit_changed_paths($c);
4839 print "\n";
4840 print $_ foreach @$l;
4841 }
4842 } else {
4843 print "1 line\n";
4844 show_commit_changed_paths($c);
4845 print "\n";
4846
4847 }
4848 foreach my $x (qw/raw stat diff/) {
4849 if ($c->{$x}) {
4850 print "\n";
4851 print $_ foreach @{$c->{$x}}
4852 }
4853 }
4854 }
4855
4856 sub cmd_show_log {
4857 my (@args) = @_;
4858 my ($r_min, $r_max);
4859 my $r_last = -1; # prevent dupes
4860 set_local_timezone();
4861 if (defined $::_revision) {
4862 if ($::_revision =~ /^(\d+):(\d+)$/) {
4863 ($r_min, $r_max) = ($1, $2);
4864 } elsif ($::_revision =~ /^\d+$/) {
4865 $r_min = $r_max = $::_revision;
4866 } else {
4867 ::fatal "-r$::_revision is not supported, use ",
4868 "standard 'git log' arguments instead";
4869 }
4870 }
4871
4872 config_pager();
4873 @args = git_svn_log_cmd($r_min, $r_max, @args);
4874 if (!@args) {
4875 print commit_log_separator unless $incremental || $oneline;
4876 return;
4877 }
4878 my $log = command_output_pipe(@args);
4879 run_pager();
4880 my (@k, $c, $d, $stat);
4881 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
4882 while (<$log>) {
4883 if (/^${esc_color}commit -?($::sha1_short)/o) {
4884 my $cmt = $1;
4885 if ($c && cmt_showable($c) && $c->{r} != $r_last) {
4886 $r_last = $c->{r};
4887 process_commit($c, $r_min, $r_max, \@k) or
4888 goto out;
4889 }
4890 $d = undef;
4891 $c = { c => $cmt };
4892 } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
4893 get_author_info($c, $1, $2, $3);
4894 } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
4895 # ignore
4896 } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
4897 push @{$c->{raw}}, $_;
4898 } elsif (/^${esc_color}[ACRMDT]\t/) {
4899 # we could add $SVN->{svn_path} here, but that requires
4900 # remote access at the moment (repo_path_split)...
4901 s#^(${esc_color})([ACRMDT])\t#$1 $2 #o;
4902 push @{$c->{changed}}, $_;
4903 } elsif (/^${esc_color}diff /o) {
4904 $d = 1;
4905 push @{$c->{diff}}, $_;
4906 } elsif ($d) {
4907 push @{$c->{diff}}, $_;
4908 } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
4909 $esc_color*[\+\-]*$esc_color$/x) {
4910 $stat = 1;
4911 push @{$c->{stat}}, $_;
4912 } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
4913 push @{$c->{stat}}, $_;
4914 $stat = undef;
4915 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
4916 ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
4917 } elsif (s/^${esc_color} //o) {
4918 push @{$c->{l}}, $_;
4919 }
4920 }
4921 if ($c && defined $c->{r} && $c->{r} != $r_last) {
4922 $r_last = $c->{r};
4923 process_commit($c, $r_min, $r_max, \@k);
4924 }
4925 if (@k) {
4926 ($r_min, $r_max) = ($r_max, $r_min);
4927 process_commit($_, $r_min, $r_max) foreach reverse @k;
4928 }
4929 out:
4930 close $log;
4931 print commit_log_separator unless $incremental || $oneline;
4932 }
4933
4934 sub cmd_blame {
4935 my $path = pop;
4936
4937 config_pager();
4938 run_pager();
4939
4940 my ($fh, $ctx, $rev);
4941
4942 if ($_git_format) {
4943 ($fh, $ctx) = command_output_pipe('blame', @_, $path);
4944 while (my $line = <$fh>) {
4945 if ($line =~ /^\^?([[:xdigit:]]+)\s/) {
4946 # Uncommitted edits show up as a rev ID of
4947 # all zeros, which we can't look up with
4948 # cmt_metadata
4949 if ($1 !~ /^0+$/) {
4950 (undef, $rev, undef) =
4951 ::cmt_metadata($1);
4952 $rev = '0' if (!$rev);
4953 } else {
4954 $rev = '0';
4955 }
4956 $rev = sprintf('%-10s', $rev);
4957 $line =~ s/^\^?[[:xdigit:]]+(\s)/$rev$1/;
4958 }
4959 print $line;
4960 }
4961 } else {
4962 ($fh, $ctx) = command_output_pipe('blame', '-p', @_, 'HEAD',
4963 '--', $path);
4964 my ($sha1);
4965 my %authors;
4966 while (my $line = <$fh>) {
4967 if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
4968 $sha1 = $1;
4969 (undef, $rev, undef) = ::cmt_metadata($1);
4970 $rev = '0' if (!$rev);
4971 }
4972 elsif ($line =~ /^author (.*)/) {
4973 $authors{$rev} = $1;
4974 $authors{$rev} =~ s/\s/_/g;
4975 }
4976 elsif ($line =~ /^\t(.*)$/) {
4977 printf("%6s %10s %s\n", $rev, $authors{$rev}, $1);
4978 }
4979 }
4980 }
4981 command_close_pipe($fh, $ctx);
4982 }
4983
4984 package Git::SVN::Migration;
4985 # these version numbers do NOT correspond to actual version numbers
4986 # of git nor git-svn. They are just relative.
4987 #
4988 # v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
4989 #
4990 # v1 layout: .git/$id/info/url, refs/remotes/$id
4991 #
4992 # v2 layout: .git/svn/$id/info/url, refs/remotes/$id
4993 #
4994 # v3 layout: .git/svn/$id, refs/remotes/$id
4995 # - info/url may remain for backwards compatibility
4996 # - this is what we migrate up to this layout automatically,
4997 # - this will be used by git svn init on single branches
4998 # v3.1 layout (auto migrated):
4999 # - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
5000 # for backwards compatibility
5001 #
5002 # v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
5003 # - this is only created for newly multi-init-ed
5004 # repositories. Similar in spirit to the
5005 # --use-separate-remotes option in git-clone (now default)
5006 # - we do not automatically migrate to this (following
5007 # the example set by core git)
5008 #
5009 # v5 layout: .rev_db.$UUID => .rev_map.$UUID
5010 # - newer, more-efficient format that uses 24-bytes per record
5011 # with no filler space.
5012 # - use xxd -c24 < .rev_map.$UUID to view and debug
5013 # - This is a one-way migration, repositories updated to the
5014 # new format will not be able to use old git-svn without
5015 # rebuilding the .rev_db. Rebuilding the rev_db is not
5016 # possible if noMetadata or useSvmProps are set; but should
5017 # be no problem for users that use the (sensible) defaults.
5018 use strict;
5019 use warnings;
5020 use Carp qw/croak/;
5021 use File::Path qw/mkpath/;
5022 use File::Basename qw/dirname basename/;
5023 use vars qw/$_minimize/;
5024
5025 sub migrate_from_v0 {
5026 my $git_dir = $ENV{GIT_DIR};
5027 return undef unless -d $git_dir;
5028 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
5029 my $migrated = 0;
5030 while (<$fh>) {
5031 chomp;
5032 my ($id, $orig_ref) = ($_, $_);
5033 next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
5034 next unless -f "$git_dir/$id/info/url";
5035 my $new_ref = "refs/remotes/$id";
5036 if (::verify_ref("$new_ref^0")) {
5037 print STDERR "W: $orig_ref is probably an old ",
5038 "branch used by an ancient version of ",
5039 "git-svn.\n",
5040 "However, $new_ref also exists.\n",
5041 "We will not be able ",
5042 "to use this branch until this ",
5043 "ambiguity is resolved.\n";
5044 next;
5045 }
5046 print STDERR "Migrating from v0 layout...\n" if !$migrated;
5047 print STDERR "Renaming ref: $orig_ref => $new_ref\n";
5048 command_noisy('update-ref', $new_ref, $orig_ref);
5049 command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
5050 $migrated++;
5051 }
5052 command_close_pipe($fh, $ctx);
5053 print STDERR "Done migrating from v0 layout...\n" if $migrated;
5054 $migrated;
5055 }
5056
5057 sub migrate_from_v1 {
5058 my $git_dir = $ENV{GIT_DIR};
5059 my $migrated = 0;
5060 return $migrated unless -d $git_dir;
5061 my $svn_dir = "$git_dir/svn";
5062
5063 # just in case somebody used 'svn' as their $id at some point...
5064 return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
5065
5066 print STDERR "Migrating from a git-svn v1 layout...\n";
5067 mkpath([$svn_dir]);
5068 print STDERR "Data from a previous version of git-svn exists, but\n\t",
5069 "$svn_dir\n\t(required for this version ",
5070 "($::VERSION) of git-svn) does not exist.\n";
5071 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
5072 while (<$fh>) {
5073 my $x = $_;
5074 next unless $x =~ s#^refs/remotes/##;
5075 chomp $x;
5076 next unless -f "$git_dir/$x/info/url";
5077 my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
5078 next unless $u;
5079 my $dn = dirname("$git_dir/svn/$x");
5080 mkpath([$dn]) unless -d $dn;
5081 if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
5082 mkpath(["$git_dir/svn/svn"]);
5083 print STDERR " - $git_dir/$x/info => ",
5084 "$git_dir/svn/$x/info\n";
5085 rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
5086 croak "$!: $x";
5087 # don't worry too much about these, they probably
5088 # don't exist with repos this old (save for index,
5089 # and we can easily regenerate that)
5090 foreach my $f (qw/unhandled.log index .rev_db/) {
5091 rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
5092 }
5093 } else {
5094 print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
5095 rename "$git_dir/$x", "$git_dir/svn/$x" or
5096 croak "$!: $x";
5097 }
5098 $migrated++;
5099 }
5100 command_close_pipe($fh, $ctx);
5101 print STDERR "Done migrating from a git-svn v1 layout\n";
5102 $migrated;
5103 }
5104
5105 sub read_old_urls {
5106 my ($l_map, $pfx, $path) = @_;
5107 my @dir;
5108 foreach (<$path/*>) {
5109 if (-r "$_/info/url") {
5110 $pfx .= '/' if $pfx && $pfx !~ m!/$!;
5111 my $ref_id = $pfx . basename $_;
5112 my $url = ::file_to_s("$_/info/url");
5113 $l_map->{$ref_id} = $url;
5114 } elsif (-d $_) {
5115 push @dir, $_;
5116 }
5117 }
5118 foreach (@dir) {
5119 my $x = $_;
5120 $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
5121 read_old_urls($l_map, $x, $_);
5122 }
5123 }
5124
5125 sub migrate_from_v2 {
5126 my @cfg = command(qw/config -l/);
5127 return if grep /^svn-remote\..+\.url=/, @cfg;
5128 my %l_map;
5129 read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
5130 my $migrated = 0;
5131
5132 foreach my $ref_id (sort keys %l_map) {
5133 eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
5134 if ($@) {
5135 Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
5136 }
5137 $migrated++;
5138 }
5139 $migrated;
5140 }
5141
5142 sub minimize_connections {
5143 my $r = Git::SVN::read_all_remotes();
5144 my $new_urls = {};
5145 my $root_repos = {};
5146 foreach my $repo_id (keys %$r) {
5147 my $url = $r->{$repo_id}->{url} or next;
5148 my $fetch = $r->{$repo_id}->{fetch} or next;
5149 my $ra = Git::SVN::Ra->new($url);
5150
5151 # skip existing cases where we already connect to the root
5152 if (($ra->{url} eq $ra->{repos_root}) ||
5153 ($ra->{repos_root} eq $repo_id)) {
5154 $root_repos->{$ra->{url}} = $repo_id;
5155 next;
5156 }
5157
5158 my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
5159 my $root_path = $ra->{url};
5160 $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
5161 foreach my $path (keys %$fetch) {
5162 my $ref_id = $fetch->{$path};
5163 my $gs = Git::SVN->new($ref_id, $repo_id, $path);
5164
5165 # make sure we can read when connecting to
5166 # a higher level of a repository
5167 my ($last_rev, undef) = $gs->last_rev_commit;
5168 if (!defined $last_rev) {
5169 $last_rev = eval {
5170 $root_ra->get_latest_revnum;
5171 };
5172 next if $@;
5173 }
5174 my $new = $root_path;
5175 $new .= length $path ? "/$path" : '';
5176 eval {
5177 $root_ra->get_log([$new], $last_rev, $last_rev,
5178 0, 0, 1, sub { });
5179 };
5180 next if $@;
5181 $new_urls->{$ra->{repos_root}}->{$new} =
5182 { ref_id => $ref_id,
5183 old_repo_id => $repo_id,
5184 old_path => $path };
5185 }
5186 }
5187
5188 my @emptied;
5189 foreach my $url (keys %$new_urls) {
5190 # see if we can re-use an existing [svn-remote "repo_id"]
5191 # instead of creating a(n ugly) new section:
5192 my $repo_id = $root_repos->{$url} || $url;
5193
5194 my $fetch = $new_urls->{$url};
5195 foreach my $path (keys %$fetch) {
5196 my $x = $fetch->{$path};
5197 Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
5198 my $pfx = "svn-remote.$x->{old_repo_id}";
5199
5200 my $old_fetch = quotemeta("$x->{old_path}:".
5201 "refs/remotes/$x->{ref_id}");
5202 command_noisy(qw/config --unset/,
5203 "$pfx.fetch", '^'. $old_fetch . '$');
5204 delete $r->{$x->{old_repo_id}}->
5205 {fetch}->{$x->{old_path}};
5206 if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
5207 command_noisy(qw/config --unset/,
5208 "$pfx.url");
5209 push @emptied, $x->{old_repo_id}
5210 }
5211 }
5212 }
5213 if (@emptied) {
5214 my $file = $ENV{GIT_CONFIG} || "$ENV{GIT_DIR}/config";
5215 print STDERR <<EOF;
5216 The following [svn-remote] sections in your config file ($file) are empty
5217 and can be safely removed:
5218 EOF
5219 print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
5220 }
5221 }
5222
5223 sub migration_check {
5224 migrate_from_v0();
5225 migrate_from_v1();
5226 migrate_from_v2();
5227 minimize_connections() if $_minimize;
5228 }
5229
5230 package Git::IndexInfo;
5231 use strict;
5232 use warnings;
5233 use Git qw/command_input_pipe command_close_pipe/;
5234
5235 sub new {
5236 my ($class) = @_;
5237 my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
5238 bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
5239 }
5240
5241 sub remove {
5242 my ($self, $path) = @_;
5243 if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
5244 return ++$self->{nr};
5245 }
5246 undef;
5247 }
5248
5249 sub update {
5250 my ($self, $mode, $hash, $path) = @_;
5251 if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
5252 return ++$self->{nr};
5253 }
5254 undef;
5255 }
5256
5257 sub DESTROY {
5258 my ($self) = @_;
5259 command_close_pipe($self->{gui}, $self->{ctx});
5260 }
5261
5262 package Git::SVN::GlobSpec;
5263 use strict;
5264 use warnings;
5265
5266 sub new {
5267 my ($class, $glob) = @_;
5268 my $re = $glob;
5269 $re =~ s!/+$!!g; # no need for trailing slashes
5270 $re =~ m!^([^*]*)(\*(?:/\*)*)([^*]*)$!;
5271 my $temp = $re;
5272 my ($left, $right) = ($1, $3);
5273 $re = $2;
5274 my $depth = $re =~ tr/*/*/;
5275 if ($depth != $temp =~ tr/*/*/) {
5276 die "Only one set of wildcard directories " .
5277 "(e.g. '*' or '*/*/*') is supported: '$glob'\n";
5278 }
5279 if ($depth == 0) {
5280 die "One '*' is needed for glob: '$glob'\n";
5281 }
5282 $re =~ s!\*!\[^/\]*!g;
5283 $re = quotemeta($left) . "($re)" . quotemeta($right);
5284 if (length $left && !($left =~ s!/+$!!g)) {
5285 die "Missing trailing '/' on left side of: '$glob' ($left)\n";
5286 }
5287 if (length $right && !($right =~ s!^/+!!g)) {
5288 die "Missing leading '/' on right side of: '$glob' ($right)\n";
5289 }
5290 my $left_re = qr/^\/\Q$left\E(\/|$)/;
5291 bless { left => $left, right => $right, left_regex => $left_re,
5292 regex => qr/$re/, glob => $glob, depth => $depth }, $class;
5293 }
5294
5295 sub full_path {
5296 my ($self, $path) = @_;
5297 return (length $self->{left} ? "$self->{left}/" : '') .
5298 $path . (length $self->{right} ? "/$self->{right}" : '');
5299 }
5300
5301 __END__
5302
5303 Data structures:
5304
5305
5306 $remotes = { # returned by read_all_remotes()
5307 'svn' => {
5308 # svn-remote.svn.url=https://svn.musicpd.org
5309 url => 'https://svn.musicpd.org',
5310 # svn-remote.svn.fetch=mpd/trunk:trunk
5311 fetch => {
5312 'mpd/trunk' => 'trunk',
5313 },
5314 # svn-remote.svn.tags=mpd/tags/*:tags/*
5315 tags => {
5316 path => {
5317 left => 'mpd/tags',
5318 right => '',
5319 regex => qr!mpd/tags/([^/]+)$!,
5320 glob => 'tags/*',
5321 },
5322 ref => {
5323 left => 'tags',
5324 right => '',
5325 regex => qr!tags/([^/]+)$!,
5326 glob => 'tags/*',
5327 },
5328 }
5329 }
5330 };
5331
5332 $log_entry hashref as returned by libsvn_log_entry()
5333 {
5334 log => 'whitespace-formatted log entry
5335 ', # trailing newline is preserved
5336 revision => '8', # integer
5337 date => '2004-02-24T17:01:44.108345Z', # commit date
5338 author => 'committer name'
5339 };
5340
5341
5342 # this is generated by generate_diff();
5343 @mods = array of diff-index line hashes, each element represents one line
5344 of diff-index output
5345
5346 diff-index line ($m hash)
5347 {
5348 mode_a => first column of diff-index output, no leading ':',
5349 mode_b => second column of diff-index output,
5350 sha1_b => sha1sum of the final blob,
5351 chg => change type [MCRADT],
5352 file_a => original file name of a file (iff chg is 'C' or 'R')
5353 file_b => new/current file name of a file (any chg)
5354 }
5355 ;
5356
5357 # retval of read_url_paths{,_all}();
5358 $l_map = {
5359 # repository root url
5360 'https://svn.musicpd.org' => {
5361 # repository path # GIT_SVN_ID
5362 'mpd/trunk' => 'trunk',
5363 'mpd/tags/0.11.5' => 'tags/0.11.5',
5364 },
5365 }
5366
5367 Notes:
5368 I don't trust the each() function on unless I created %hash myself
5369 because the internal iterator may not have started at base.
Jonathan Nieder’s miscellaneous git branches