![[tokkee]](http://tokkee.org/images/avatar.png){kind=avatar}
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 $_authors_prog %users/;
9 $AUTHOR = 'Eric Wong <normalperson@yhbt.net>';
10 $VERSION = '@@GIT_VERSION@@';
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 } || '';
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 $Git::SVN::_minimize_url = 'unset';
24 if (! exists $ENV{SVN_SSH}) {
25 if (exists $ENV{GIT_SSH}) {
26 $ENV{SVN_SSH} = $ENV{GIT_SSH};
27 if ($^O eq 'msys') {
28 $ENV{SVN_SSH} =~ s/\\/\\\\/g;
29 }
30 }
31 }
33 $Git::SVN::Log::TZ = $ENV{TZ};
34 $ENV{TZ} = 'UTC';
35 $| = 1; # unbuffer STDOUT
37 sub fatal (@) { print STDERR "@_\n"; exit 1 }
38 require SVN::Core; # use()-ing this causes segfaults for me... *shrug*
39 require SVN::Ra;
40 require SVN::Delta;
41 if ($SVN::Core::VERSION lt '1.1.0') {
42 fatal "Need SVN::Core 1.1.0 or better (got $SVN::Core::VERSION)";
43 }
44 my $can_compress = eval { require Compress::Zlib; 1};
45 push @Git::SVN::Ra::ISA, 'SVN::Ra';
46 push @SVN::Git::Editor::ISA, 'SVN::Delta::Editor';
47 push @SVN::Git::Fetcher::ISA, 'SVN::Delta::Editor';
48 use Carp qw/croak/;
49 use Digest::MD5;
50 use IO::File qw//;
51 use File::Basename qw/dirname basename/;
52 use File::Path qw/mkpath/;
53 use File::Spec;
54 use File::Find;
55 use Getopt::Long qw/:config gnu_getopt no_ignore_case auto_abbrev/;
56 use IPC::Open3;
57 use Git;
59 BEGIN {
60 # import functions from Git into our packages, en masse
61 no strict 'refs';
62 foreach (qw/command command_oneline command_noisy command_output_pipe
63 command_input_pipe command_close_pipe
64 command_bidi_pipe command_close_bidi_pipe/) {
65 for my $package ( qw(SVN::Git::Editor SVN::Git::Fetcher
66 Git::SVN::Migration Git::SVN::Log Git::SVN),
67 __PACKAGE__) {
68 *{"${package}::$_"} = \&{"Git::$_"};
69 }
70 }
71 }
73 my ($SVN);
75 $sha1 = qr/[a-f\d]{40}/;
76 $sha1_short = qr/[a-f\d]{4,40}/;
77 my ($_stdin, $_help, $_edit,
78 $_message, $_file, $_branch_dest,
79 $_template, $_shared,
80 $_version, $_fetch_all, $_no_rebase, $_fetch_parent,
81 $_merge, $_strategy, $_dry_run, $_local,
82 $_prefix, $_no_checkout, $_url, $_verbose,
83 $_git_format, $_commit_url, $_tag);
84 $Git::SVN::_follow_parent = 1;
85 $_q ||= 0;
86 my %remote_opts = ( 'username=s' => \$Git::SVN::Prompt::_username,
87 'config-dir=s' => \$Git::SVN::Ra::config_dir,
88 'no-auth-cache' => \$Git::SVN::Prompt::_no_auth_cache,
89 'ignore-paths=s' => \$SVN::Git::Fetcher::_ignore_regex );
90 my %fc_opts = ( 'follow-parent|follow!' => \$Git::SVN::_follow_parent,
91 'authors-file|A=s' => \$_authors,
92 'authors-prog=s' => \$_authors_prog,
93 'repack:i' => \$Git::SVN::_repack,
94 'noMetadata' => \$Git::SVN::_no_metadata,
95 'useSvmProps' => \$Git::SVN::_use_svm_props,
96 'useSvnsyncProps' => \$Git::SVN::_use_svnsync_props,
97 'log-window-size=i' => \$Git::SVN::Ra::_log_window_size,
98 'no-checkout' => \$_no_checkout,
99 'quiet|q+' => \$_q,
100 'repack-flags|repack-args|repack-opts=s' =>
101 \$Git::SVN::_repack_flags,
102 'use-log-author' => \$Git::SVN::_use_log_author,
103 'add-author-from' => \$Git::SVN::_add_author_from,
104 'localtime' => \$Git::SVN::_localtime,
105 %remote_opts );
107 my ($_trunk, @_tags, @_branches, $_stdlayout);
108 my %icv;
109 my %init_opts = ( 'template=s' => \$_template, 'shared:s' => \$_shared,
110 'trunk|T=s' => \$_trunk, 'tags|t=s@' => \@_tags,
111 'branches|b=s@' => \@_branches, 'prefix=s' => \$_prefix,
112 'stdlayout|s' => \$_stdlayout,
113 'minimize-url|m!' => \$Git::SVN::_minimize_url,
114 'no-metadata' => sub { $icv{noMetadata} = 1 },
115 'use-svm-props' => sub { $icv{useSvmProps} = 1 },
116 'use-svnsync-props' => sub { $icv{useSvnsyncProps} = 1 },
117 'rewrite-root=s' => sub { $icv{rewriteRoot} = $_[1] },
118 %remote_opts );
119 my %cmt_opts = ( 'edit|e' => \$_edit,
120 'rmdir' => \$SVN::Git::Editor::_rmdir,
121 'find-copies-harder' => \$SVN::Git::Editor::_find_copies_harder,
122 'l=i' => \$SVN::Git::Editor::_rename_limit,
123 'copy-similarity|C=i'=> \$SVN::Git::Editor::_cp_similarity
124 );
126 my %cmd = (
127 fetch => [ \&cmd_fetch, "Download new revisions from SVN",
128 { 'revision|r=s' => \$_revision,
129 'fetch-all|all' => \$_fetch_all,
130 'parent|p' => \$_fetch_parent,
131 %fc_opts } ],
132 clone => [ \&cmd_clone, "Initialize and fetch revisions",
133 { 'revision|r=s' => \$_revision,
134 %fc_opts, %init_opts } ],
135 init => [ \&cmd_init, "Initialize a repo for tracking" .
136 " (requires URL argument)",
137 \%init_opts ],
138 'multi-init' => [ \&cmd_multi_init,
139 "Deprecated alias for ".
140 "'$0 init -T<trunk> -b<branches> -t<tags>'",
141 \%init_opts ],
142 dcommit => [ \&cmd_dcommit,
143 'Commit several diffs to merge with upstream',
144 { 'merge|m|M' => \$_merge,
145 'strategy|s=s' => \$_strategy,
146 'verbose|v' => \$_verbose,
147 'dry-run|n' => \$_dry_run,
148 'fetch-all|all' => \$_fetch_all,
149 'commit-url=s' => \$_commit_url,
150 'revision|r=i' => \$_revision,
151 'no-rebase' => \$_no_rebase,
152 %cmt_opts, %fc_opts } ],
153 branch => [ \&cmd_branch,
154 'Create a branch in the SVN repository',
155 { 'message|m=s' => \$_message,
156 'destination|d=s' => \$_branch_dest,
157 'dry-run|n' => \$_dry_run,
158 'tag|t' => \$_tag } ],
159 tag => [ sub { $_tag = 1; cmd_branch(@_) },
160 'Create a tag in the SVN repository',
161 { 'message|m=s' => \$_message,
162 'destination|d=s' => \$_branch_dest,
163 'dry-run|n' => \$_dry_run } ],
164 'set-tree' => [ \&cmd_set_tree,
165 "Set an SVN repository to a git tree-ish",
166 { 'stdin' => \$_stdin, %cmt_opts, %fc_opts, } ],
167 'create-ignore' => [ \&cmd_create_ignore,
168 'Create a .gitignore per svn:ignore',
169 { 'revision|r=i' => \$_revision
170 } ],
171 'mkdirs' => [ \&cmd_mkdirs ,
172 "recreate empty directories after a checkout",
173 { 'revision|r=i' => \$_revision } ],
174 'propget' => [ \&cmd_propget,
175 'Print the value of a property on a file or directory',
176 { 'revision|r=i' => \$_revision } ],
177 'proplist' => [ \&cmd_proplist,
178 'List all properties of a file or directory',
179 { 'revision|r=i' => \$_revision } ],
180 'show-ignore' => [ \&cmd_show_ignore, "Show svn:ignore listings",
181 { 'revision|r=i' => \$_revision
182 } ],
183 'show-externals' => [ \&cmd_show_externals, "Show svn:externals listings",
184 { 'revision|r=i' => \$_revision
185 } ],
186 'multi-fetch' => [ \&cmd_multi_fetch,
187 "Deprecated alias for $0 fetch --all",
188 { 'revision|r=s' => \$_revision, %fc_opts } ],
189 'migrate' => [ sub { },
190 # no-op, we automatically run this anyways,
191 'Migrate configuration/metadata/layout from
192 previous versions of git-svn',
193 { 'minimize' => \$Git::SVN::Migration::_minimize,
194 %remote_opts } ],
195 'log' => [ \&Git::SVN::Log::cmd_show_log, 'Show commit logs',
196 { 'limit=i' => \$Git::SVN::Log::limit,
197 'revision|r=s' => \$_revision,
198 'verbose|v' => \$Git::SVN::Log::verbose,
199 'incremental' => \$Git::SVN::Log::incremental,
200 'oneline' => \$Git::SVN::Log::oneline,
201 'show-commit' => \$Git::SVN::Log::show_commit,
202 'non-recursive' => \$Git::SVN::Log::non_recursive,
203 'authors-file|A=s' => \$_authors,
204 'color' => \$Git::SVN::Log::color,
205 'pager=s' => \$Git::SVN::Log::pager
206 } ],
207 'find-rev' => [ \&cmd_find_rev,
208 "Translate between SVN revision numbers and tree-ish",
209 {} ],
210 'rebase' => [ \&cmd_rebase, "Fetch and rebase your working directory",
211 { 'merge|m|M' => \$_merge,
212 'verbose|v' => \$_verbose,
213 'strategy|s=s' => \$_strategy,
214 'local|l' => \$_local,
215 'fetch-all|all' => \$_fetch_all,
216 'dry-run|n' => \$_dry_run,
217 %fc_opts } ],
218 'commit-diff' => [ \&cmd_commit_diff,
219 'Commit a diff between two trees',
220 { 'message|m=s' => \$_message,
221 'file|F=s' => \$_file,
222 'revision|r=s' => \$_revision,
223 %cmt_opts } ],
224 'info' => [ \&cmd_info,
225 "Show info about the latest SVN revision
226 on the current branch",
227 { 'url' => \$_url, } ],
228 'blame' => [ \&Git::SVN::Log::cmd_blame,
229 "Show what revision and author last modified each line of a file",
230 { 'git-format' => \$_git_format } ],
231 'reset' => [ \&cmd_reset,
232 "Undo fetches back to the specified SVN revision",
233 { 'revision|r=s' => \$_revision,
234 'parent|p' => \$_fetch_parent } ],
235 'gc' => [ \&cmd_gc,
236 "Compress unhandled.log files in .git/svn and remove " .
237 "index files in .git/svn",
238 {} ],
239 );
241 my $cmd;
242 for (my $i = 0; $i < @ARGV; $i++) {
243 if (defined $cmd{$ARGV[$i]}) {
244 $cmd = $ARGV[$i];
245 splice @ARGV, $i, 1;
246 last;
247 } elsif ($ARGV[$i] eq 'help') {
248 $cmd = $ARGV[$i+1];
249 usage(0);
250 }
251 };
253 # make sure we're always running at the top-level working directory
254 unless ($cmd && $cmd =~ /(?:clone|init|multi-init)$/) {
255 unless (-d $ENV{GIT_DIR}) {
256 if ($git_dir_user_set) {
257 die "GIT_DIR=$ENV{GIT_DIR} explicitly set, ",
258 "but it is not a directory\n";
259 }
260 my $git_dir = delete $ENV{GIT_DIR};
261 my $cdup = undef;
262 git_cmd_try {
263 $cdup = command_oneline(qw/rev-parse --show-cdup/);
264 $git_dir = '.' unless ($cdup);
265 chomp $cdup if ($cdup);
266 $cdup = "." unless ($cdup && length $cdup);
267 } "Already at toplevel, but $git_dir not found\n";
268 chdir $cdup or die "Unable to chdir up to '$cdup'\n";
269 unless (-d $git_dir) {
270 die "$git_dir still not found after going to ",
271 "'$cdup'\n";
272 }
273 $ENV{GIT_DIR} = $git_dir;
274 }
275 $_repository = Git->repository(Repository => $ENV{GIT_DIR});
276 }
278 my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
280 read_git_config(\%opts);
281 if ($cmd && ($cmd eq 'log' || $cmd eq 'blame')) {
282 Getopt::Long::Configure('pass_through');
283 }
284 my $rv = GetOptions(%opts, 'help|H|h' => \$_help, 'version|V' => \$_version,
285 'minimize-connections' => \$Git::SVN::Migration::_minimize,
286 'id|i=s' => \$Git::SVN::default_ref_id,
287 'svn-remote|remote|R=s' => sub {
288 $Git::SVN::no_reuse_existing = 1;
289 $Git::SVN::default_repo_id = $_[1] });
290 exit 1 if (!$rv && $cmd && $cmd ne 'log');
292 usage(0) if $_help;
293 version() if $_version;
294 usage(1) unless defined $cmd;
295 load_authors() if $_authors;
296 if (defined $_authors_prog) {
297 $_authors_prog = "'" . File::Spec->rel2abs($_authors_prog) . "'";
298 }
300 unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
301 Git::SVN::Migration::migration_check();
302 }
303 Git::SVN::init_vars();
304 eval {
305 Git::SVN::verify_remotes_sanity();
306 $cmd{$cmd}->[0]->(@ARGV);
307 };
308 fatal $@ if $@;
309 post_fetch_checkout();
310 exit 0;
312 ####################### primary functions ######################
313 sub usage {
314 my $exit = shift || 0;
315 my $fd = $exit ? \*STDERR : \*STDOUT;
316 print $fd <<"";
317 git-svn - bidirectional operations between a single Subversion tree and git
318 Usage: git svn <command> [options] [arguments]\n
320 print $fd "Available commands:\n" unless $cmd;
322 foreach (sort keys %cmd) {
323 next if $cmd && $cmd ne $_;
324 next if /^multi-/; # don't show deprecated commands
325 print $fd ' ',pack('A17',$_),$cmd{$_}->[1],"\n";
326 foreach (sort keys %{$cmd{$_}->[2]}) {
327 # mixed-case options are for .git/config only
328 next if /[A-Z]/ && /^[a-z]+$/i;
329 # prints out arguments as they should be passed:
330 my $x = s#[:=]s$## ? '<arg>' : s#[:=]i$## ? '<num>' : '';
331 print $fd ' ' x 21, join(', ', map { length $_ > 1 ?
332 "--$_" : "-$_" }
333 split /\|/,$_)," $x\n";
334 }
335 }
336 print $fd <<"";
337 \nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
338 arbitrary identifier if you're tracking multiple SVN branches/repositories in
339 one git repository and want to keep them separate. See git-svn(1) for more
340 information.
342 exit $exit;
343 }
345 sub version {
346 print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
347 exit 0;
348 }
350 sub do_git_init_db {
351 unless (-d $ENV{GIT_DIR}) {
352 my @init_db = ('init');
353 push @init_db, "--template=$_template" if defined $_template;
354 if (defined $_shared) {
355 if ($_shared =~ /[a-z]/) {
356 push @init_db, "--shared=$_shared";
357 } else {
358 push @init_db, "--shared";
359 }
360 }
361 command_noisy(@init_db);
362 $_repository = Git->repository(Repository => ".git");
363 }
364 command_noisy('config', 'core.autocrlf', 'false');
365 my $set;
366 my $pfx = "svn-remote.$Git::SVN::default_repo_id";
367 foreach my $i (keys %icv) {
368 die "'$set' and '$i' cannot both be set\n" if $set;
369 next unless defined $icv{$i};
370 command_noisy('config', "$pfx.$i", $icv{$i});
371 $set = $i;
372 }
373 my $ignore_regex = \$SVN::Git::Fetcher::_ignore_regex;
374 command_noisy('config', "$pfx.ignore-paths", $$ignore_regex)
375 if defined $$ignore_regex;
376 }
378 sub init_subdir {
379 my $repo_path = shift or return;
380 mkpath([$repo_path]) unless -d $repo_path;
381 chdir $repo_path or die "Couldn't chdir to $repo_path: $!\n";
382 $ENV{GIT_DIR} = '.git';
383 $_repository = Git->repository(Repository => $ENV{GIT_DIR});
384 }
386 sub cmd_clone {
387 my ($url, $path) = @_;
388 if (!defined $path &&
389 (defined $_trunk || @_branches || @_tags ||
390 defined $_stdlayout) &&
391 $url !~ m#^[a-z\+]+://#) {
392 $path = $url;
393 }
394 $path = basename($url) if !defined $path || !length $path;
395 my $authors_absolute = $_authors ? File::Spec->rel2abs($_authors) : "";
396 cmd_init($url, $path);
397 command_oneline('config', 'svn.authorsfile', $authors_absolute)
398 if $_authors;
399 Git::SVN::fetch_all($Git::SVN::default_repo_id);
400 }
402 sub cmd_init {
403 if (defined $_stdlayout) {
404 $_trunk = 'trunk' if (!defined $_trunk);
405 @_tags = 'tags' if (! @_tags);
406 @_branches = 'branches' if (! @_branches);
407 }
408 if (defined $_trunk || @_branches || @_tags) {
409 return cmd_multi_init(@_);
410 }
411 my $url = shift or die "SVN repository location required ",
412 "as a command-line argument\n";
413 $url = canonicalize_url($url);
414 init_subdir(@_);
415 do_git_init_db();
417 if ($Git::SVN::_minimize_url eq 'unset') {
418 $Git::SVN::_minimize_url = 0;
419 }
421 Git::SVN->init($url);
422 }
424 sub cmd_fetch {
425 if (grep /^\d+=./, @_) {
426 die "'<rev>=<commit>' fetch arguments are ",
427 "no longer supported.\n";
428 }
429 my ($remote) = @_;
430 if (@_ > 1) {
431 die "Usage: $0 fetch [--all] [--parent] [svn-remote]\n";
432 }
433 $Git::SVN::no_reuse_existing = undef;
434 if ($_fetch_parent) {
435 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
436 unless ($gs) {
437 die "Unable to determine upstream SVN information from ",
438 "working tree history\n";
439 }
440 # just fetch, don't checkout.
441 $_no_checkout = 'true';
442 $_fetch_all ? $gs->fetch_all : $gs->fetch;
443 } elsif ($_fetch_all) {
444 cmd_multi_fetch();
445 } else {
446 $remote ||= $Git::SVN::default_repo_id;
447 Git::SVN::fetch_all($remote, Git::SVN::read_all_remotes());
448 }
449 }
451 sub cmd_set_tree {
452 my (@commits) = @_;
453 if ($_stdin || !@commits) {
454 print "Reading from stdin...\n";
455 @commits = ();
456 while (<STDIN>) {
457 if (/\b($sha1_short)\b/o) {
458 unshift @commits, $1;
459 }
460 }
461 }
462 my @revs;
463 foreach my $c (@commits) {
464 my @tmp = command('rev-parse',$c);
465 if (scalar @tmp == 1) {
466 push @revs, $tmp[0];
467 } elsif (scalar @tmp > 1) {
468 push @revs, reverse(command('rev-list',@tmp));
469 } else {
470 fatal "Failed to rev-parse $c";
471 }
472 }
473 my $gs = Git::SVN->new;
474 my ($r_last, $cmt_last) = $gs->last_rev_commit;
475 $gs->fetch;
476 if (defined $gs->{last_rev} && $r_last != $gs->{last_rev}) {
477 fatal "There are new revisions that were fetched ",
478 "and need to be merged (or acknowledged) ",
479 "before committing.\nlast rev: $r_last\n",
480 " current: $gs->{last_rev}";
481 }
482 $gs->set_tree($_) foreach @revs;
483 print "Done committing ",scalar @revs," revisions to SVN\n";
484 unlink $gs->{index};
485 }
487 sub cmd_dcommit {
488 my $head = shift;
489 git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
490 'Cannot dcommit with a dirty index. Commit your changes first, '
491 . "or stash them with `git stash'.\n";
492 $head ||= 'HEAD';
494 my $old_head;
495 if ($head ne 'HEAD') {
496 $old_head = eval {
497 command_oneline([qw/symbolic-ref -q HEAD/])
498 };
499 if ($old_head) {
500 $old_head =~ s{^refs/heads/}{};
501 } else {
502 $old_head = eval { command_oneline(qw/rev-parse HEAD/) };
503 }
504 command(['checkout', $head], STDERR => 0);
505 }
507 my @refs;
508 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD', \@refs);
509 unless ($gs) {
510 die "Unable to determine upstream SVN information from ",
511 "$head history.\nPerhaps the repository is empty.";
512 }
514 if (defined $_commit_url) {
515 $url = $_commit_url;
516 } else {
517 $url = eval { command_oneline('config', '--get',
518 "svn-remote.$gs->{repo_id}.commiturl") };
519 if (!$url) {
520 $url = $gs->full_url
521 }
522 }
524 my $last_rev = $_revision if defined $_revision;
525 if ($url) {
526 print "Committing to $url ...\n";
527 }
528 my ($linear_refs, $parents) = linearize_history($gs, \@refs);
529 if ($_no_rebase && scalar(@$linear_refs) > 1) {
530 warn "Attempting to commit more than one change while ",
531 "--no-rebase is enabled.\n",
532 "If these changes depend on each other, re-running ",
533 "without --no-rebase may be required."
534 }
535 my $expect_url = $url;
536 Git::SVN::remove_username($expect_url);
537 while (1) {
538 my $d = shift @$linear_refs or last;
539 unless (defined $last_rev) {
540 (undef, $last_rev, undef) = cmt_metadata("$d~1");
541 unless (defined $last_rev) {
542 fatal "Unable to extract revision information ",
543 "from commit $d~1";
544 }
545 }
546 if ($_dry_run) {
547 print "diff-tree $d~1 $d\n";
548 } else {
549 my $cmt_rev;
550 my %ed_opts = ( r => $last_rev,
551 log => get_commit_entry($d)->{log},
552 ra => Git::SVN::Ra->new($url),
553 config => SVN::Core::config_get_config(
554 $Git::SVN::Ra::config_dir
555 ),
556 tree_a => "$d~1",
557 tree_b => $d,
558 editor_cb => sub {
559 print "Committed r$_[0]\n";
560 $cmt_rev = $_[0];
561 },
562 svn_path => '');
563 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
564 print "No changes\n$d~1 == $d\n";
565 } elsif ($parents->{$d} && @{$parents->{$d}}) {
566 $gs->{inject_parents_dcommit}->{$cmt_rev} =
567 $parents->{$d};
568 }
569 $_fetch_all ? $gs->fetch_all : $gs->fetch;
570 $last_rev = $cmt_rev;
571 next if $_no_rebase;
573 # we always want to rebase against the current HEAD,
574 # not any head that was passed to us
575 my @diff = command('diff-tree', $d,
576 $gs->refname, '--');
577 my @finish;
578 if (@diff) {
579 @finish = rebase_cmd();
580 print STDERR "W: $d and ", $gs->refname,
581 " differ, using @finish:\n",
582 join("\n", @diff), "\n";
583 } else {
584 print "No changes between current HEAD and ",
585 $gs->refname,
586 "\nResetting to the latest ",
587 $gs->refname, "\n";
588 @finish = qw/reset --mixed/;
589 }
590 command_noisy(@finish, $gs->refname);
591 if (@diff) {
592 @refs = ();
593 my ($url_, $rev_, $uuid_, $gs_) =
594 working_head_info('HEAD', \@refs);
595 my ($linear_refs_, $parents_) =
596 linearize_history($gs_, \@refs);
597 if (scalar(@$linear_refs) !=
598 scalar(@$linear_refs_)) {
599 fatal "# of revisions changed ",
600 "\nbefore:\n",
601 join("\n", @$linear_refs),
602 "\n\nafter:\n",
603 join("\n", @$linear_refs_), "\n",
604 'If you are attempting to commit ',
605 "merges, try running:\n\t",
606 'git rebase --interactive',
607 '--preserve-merges ',
608 $gs->refname,
609 "\nBefore dcommitting";
610 }
611 if ($url_ ne $expect_url) {
612 if ($url_ eq $gs->metadata_url) {
613 print
614 "Accepting rewritten URL:",
615 " $url_\n";
616 } else {
617 fatal
618 "URL mismatch after rebase:",
619 " $url_ != $expect_url";
620 }
621 }
622 if ($uuid_ ne $uuid) {
623 fatal "uuid mismatch after rebase: ",
624 "$uuid_ != $uuid";
625 }
626 # remap parents
627 my (%p, @l, $i);
628 for ($i = 0; $i < scalar @$linear_refs; $i++) {
629 my $new = $linear_refs_->[$i] or next;
630 $p{$new} =
631 $parents->{$linear_refs->[$i]};
632 push @l, $new;
633 }
634 $parents = \%p;
635 $linear_refs = \@l;
636 }
637 }
638 }
640 if ($old_head) {
641 my $new_head = command_oneline(qw/rev-parse HEAD/);
642 my $new_is_symbolic = eval {
643 command_oneline(qw/symbolic-ref -q HEAD/);
644 };
645 if ($new_is_symbolic) {
646 print "dcommitted the branch ", $head, "\n";
647 } else {
648 print "dcommitted on a detached HEAD because you gave ",
649 "a revision argument.\n",
650 "The rewritten commit is: ", $new_head, "\n";
651 }
652 command(['checkout', $old_head], STDERR => 0);
653 }
655 unlink $gs->{index};
656 }
658 sub cmd_branch {
659 my ($branch_name, $head) = @_;
661 unless (defined $branch_name && length $branch_name) {
662 die(($_tag ? "tag" : "branch") . " name required\n");
663 }
664 $head ||= 'HEAD';
666 my ($src, $rev, undef, $gs) = working_head_info($head);
668 my $remote = Git::SVN::read_all_remotes()->{$gs->{repo_id}};
669 my $allglobs = $remote->{ $_tag ? 'tags' : 'branches' };
670 my $glob;
671 if ($#{$allglobs} == 0) {
672 $glob = $allglobs->[0];
673 } else {
674 unless(defined $_branch_dest) {
675 die "Multiple ",
676 $_tag ? "tag" : "branch",
677 " paths defined for Subversion repository.\n",
678 "You must specify where you want to create the ",
679 $_tag ? "tag" : "branch",
680 " with the --destination argument.\n";
681 }
682 foreach my $g (@{$allglobs}) {
683 # SVN::Git::Editor could probably be moved to Git.pm..
684 my $re = SVN::Git::Editor::glob2pat($g->{path}->{left});
685 if ($_branch_dest =~ /$re/) {
686 $glob = $g;
687 last;
688 }
689 }
690 unless (defined $glob) {
691 my $dest_re = qr/\b\Q$_branch_dest\E\b/;
692 foreach my $g (@{$allglobs}) {
693 $g->{path}->{left} =~ /$dest_re/ or next;
694 if (defined $glob) {
695 die "Ambiguous destination: ",
696 $_branch_dest, "\nmatches both '",
697 $glob->{path}->{left}, "' and '",
698 $g->{path}->{left}, "'\n";
699 }
700 $glob = $g;
701 }
702 unless (defined $glob) {
703 die "Unknown ",
704 $_tag ? "tag" : "branch",
705 " destination $_branch_dest\n";
706 }
707 }
708 }
709 my ($lft, $rgt) = @{ $glob->{path} }{qw/left right/};
710 my $dst = join '/', $remote->{url}, $lft, $branch_name, ($rgt || ());
712 my $ctx = SVN::Client->new(
713 auth => Git::SVN::Ra::_auth_providers(),
714 log_msg => sub {
715 ${ $_[0] } = defined $_message
716 ? $_message
717 : 'Create ' . ($_tag ? 'tag ' : 'branch ' )
718 . $branch_name;
719 },
720 );
722 eval {
723 $ctx->ls($dst, 'HEAD', 0);
724 } and die "branch ${branch_name} already exists\n";
726 print "Copying ${src} at r${rev} to ${dst}...\n";
727 $ctx->copy($src, $rev, $dst)
728 unless $_dry_run;
730 $gs->fetch_all;
731 }
733 sub cmd_find_rev {
734 my $revision_or_hash = shift or die "SVN or git revision required ",
735 "as a command-line argument\n";
736 my $result;
737 if ($revision_or_hash =~ /^r\d+$/) {
738 my $head = shift;
739 $head ||= 'HEAD';
740 my @refs;
741 my (undef, undef, $uuid, $gs) = working_head_info($head, \@refs);
742 unless ($gs) {
743 die "Unable to determine upstream SVN information from ",
744 "$head history\n";
745 }
746 my $desired_revision = substr($revision_or_hash, 1);
747 $result = $gs->rev_map_get($desired_revision, $uuid);
748 } else {
749 my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
750 $result = $rev;
751 }
752 print "$result\n" if $result;
753 }
755 sub cmd_rebase {
756 command_noisy(qw/update-index --refresh/);
757 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
758 unless ($gs) {
759 die "Unable to determine upstream SVN information from ",
760 "working tree history\n";
761 }
762 if ($_dry_run) {
763 print "Remote Branch: " . $gs->refname . "\n";
764 print "SVN URL: " . $url . "\n";
765 return;
766 }
767 if (command(qw/diff-index HEAD --/)) {
768 print STDERR "Cannot rebase with uncommited changes:\n";
769 command_noisy('status');
770 exit 1;
771 }
772 unless ($_local) {
773 # rebase will checkout for us, so no need to do it explicitly
774 $_no_checkout = 'true';
775 $_fetch_all ? $gs->fetch_all : $gs->fetch;
776 }
777 command_noisy(rebase_cmd(), $gs->refname);
778 $gs->mkemptydirs;
779 }
781 sub cmd_show_ignore {
782 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
783 $gs ||= Git::SVN->new;
784 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
785 $gs->prop_walk($gs->{path}, $r, sub {
786 my ($gs, $path, $props) = @_;
787 print STDOUT "\n# $path\n";
788 my $s = $props->{'svn:ignore'} or return;
789 $s =~ s/[\r\n]+/\n/g;
790 $s =~ s/^\n+//;
791 chomp $s;
792 $s =~ s#^#$path#gm;
793 print STDOUT "$s\n";
794 });
795 }
797 sub cmd_show_externals {
798 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
799 $gs ||= Git::SVN->new;
800 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
801 $gs->prop_walk($gs->{path}, $r, sub {
802 my ($gs, $path, $props) = @_;
803 print STDOUT "\n# $path\n";
804 my $s = $props->{'svn:externals'} or return;
805 $s =~ s/[\r\n]+/\n/g;
806 chomp $s;
807 $s =~ s#^#$path#gm;
808 print STDOUT "$s\n";
809 });
810 }
812 sub cmd_create_ignore {
813 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
814 $gs ||= Git::SVN->new;
815 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
816 $gs->prop_walk($gs->{path}, $r, sub {
817 my ($gs, $path, $props) = @_;
818 # $path is of the form /path/to/dir/
819 $path = '.' . $path;
820 # SVN can have attributes on empty directories,
821 # which git won't track
822 mkpath([$path]) unless -d $path;
823 my $ignore = $path . '.gitignore';
824 my $s = $props->{'svn:ignore'} or return;
825 open(GITIGNORE, '>', $ignore)
826 or fatal("Failed to open `$ignore' for writing: $!");
827 $s =~ s/[\r\n]+/\n/g;
828 $s =~ s/^\n+//;
829 chomp $s;
830 # Prefix all patterns so that the ignore doesn't apply
831 # to sub-directories.
832 $s =~ s#^#/#gm;
833 print GITIGNORE "$s\n";
834 close(GITIGNORE)
835 or fatal("Failed to close `$ignore': $!");
836 command_noisy('add', '-f', $ignore);
837 });
838 }
840 sub cmd_mkdirs {
841 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
842 $gs ||= Git::SVN->new;
843 $gs->mkemptydirs($_revision);
844 }
846 sub canonicalize_path {
847 my ($path) = @_;
848 my $dot_slash_added = 0;
849 if (substr($path, 0, 1) ne "/") {
850 $path = "./" . $path;
851 $dot_slash_added = 1;
852 }
853 # File::Spec->canonpath doesn't collapse x/../y into y (for a
854 # good reason), so let's do this manually.
855 $path =~ s#/+#/#g;
856 $path =~ s#/\.(?:/|$)#/#g;
857 $path =~ s#/[^/]+/\.\.##g;
858 $path =~ s#/$##g;
859 $path =~ s#^\./## if $dot_slash_added;
860 $path =~ s#^/##;
861 $path =~ s#^\.$##;
862 return $path;
863 }
865 sub canonicalize_url {
866 my ($url) = @_;
867 $url =~ s#^([^:]+://[^/]*/)(.*)$#$1 . canonicalize_path($2)#e;
868 return $url;
869 }
871 # get_svnprops(PATH)
872 # ------------------
873 # Helper for cmd_propget and cmd_proplist below.
874 sub get_svnprops {
875 my $path = shift;
876 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
877 $gs ||= Git::SVN->new;
879 # prefix THE PATH by the sub-directory from which the user
880 # invoked us.
881 $path = $cmd_dir_prefix . $path;
882 fatal("No such file or directory: $path") unless -e $path;
883 my $is_dir = -d $path ? 1 : 0;
884 $path = $gs->{path} . '/' . $path;
886 # canonicalize the path (otherwise libsvn will abort or fail to
887 # find the file)
888 $path = canonicalize_path($path);
890 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
891 my $props;
892 if ($is_dir) {
893 (undef, undef, $props) = $gs->ra->get_dir($path, $r);
894 }
895 else {
896 (undef, $props) = $gs->ra->get_file($path, $r, undef);
897 }
898 return $props;
899 }
901 # cmd_propget (PROP, PATH)
902 # ------------------------
903 # Print the SVN property PROP for PATH.
904 sub cmd_propget {
905 my ($prop, $path) = @_;
906 $path = '.' if not defined $path;
907 usage(1) if not defined $prop;
908 my $props = get_svnprops($path);
909 if (not defined $props->{$prop}) {
910 fatal("`$path' does not have a `$prop' SVN property.");
911 }
912 print $props->{$prop} . "\n";
913 }
915 # cmd_proplist (PATH)
916 # -------------------
917 # Print the list of SVN properties for PATH.
918 sub cmd_proplist {
919 my $path = shift;
920 $path = '.' if not defined $path;
921 my $props = get_svnprops($path);
922 print "Properties on '$path':\n";
923 foreach (sort keys %{$props}) {
924 print " $_\n";
925 }
926 }
928 sub cmd_multi_init {
929 my $url = shift;
930 unless (defined $_trunk || @_branches || @_tags) {
931 usage(1);
932 }
934 $_prefix = '' unless defined $_prefix;
935 if (defined $url) {
936 $url = canonicalize_url($url);
937 init_subdir(@_);
938 }
939 do_git_init_db();
940 if (defined $_trunk) {
941 my $trunk_ref = 'refs/remotes/' . $_prefix . 'trunk';
942 # try both old-style and new-style lookups:
943 my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
944 unless ($gs_trunk) {
945 my ($trunk_url, $trunk_path) =
946 complete_svn_url($url, $_trunk);
947 $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
948 undef, $trunk_ref);
949 }
950 }
951 return unless @_branches || @_tags;
952 my $ra = $url ? Git::SVN::Ra->new($url) : undef;
953 foreach my $path (@_branches) {
954 complete_url_ls_init($ra, $path, '--branches/-b', $_prefix);
955 }
956 foreach my $path (@_tags) {
957 complete_url_ls_init($ra, $path, '--tags/-t', $_prefix.'tags/');
958 }
959 }
961 sub cmd_multi_fetch {
962 $Git::SVN::no_reuse_existing = undef;
963 my $remotes = Git::SVN::read_all_remotes();
964 foreach my $repo_id (sort keys %$remotes) {
965 if ($remotes->{$repo_id}->{url}) {
966 Git::SVN::fetch_all($repo_id, $remotes);
967 }
968 }
969 }
971 # this command is special because it requires no metadata
972 sub cmd_commit_diff {
973 my ($ta, $tb, $url) = @_;
974 my $usage = "Usage: $0 commit-diff -r<revision> ".
975 "<tree-ish> <tree-ish> [<URL>]";
976 fatal($usage) if (!defined $ta || !defined $tb);
977 my $svn_path = '';
978 if (!defined $url) {
979 my $gs = eval { Git::SVN->new };
980 if (!$gs) {
981 fatal("Needed URL or usable git-svn --id in ",
982 "the command-line\n", $usage);
983 }
984 $url = $gs->{url};
985 $svn_path = $gs->{path};
986 }
987 unless (defined $_revision) {
988 fatal("-r|--revision is a required argument\n", $usage);
989 }
990 if (defined $_message && defined $_file) {
991 fatal("Both --message/-m and --file/-F specified ",
992 "for the commit message.\n",
993 "I have no idea what you mean");
994 }
995 if (defined $_file) {
996 $_message = file_to_s($_file);
997 } else {
998 $_message ||= get_commit_entry($tb)->{log};
999 }
1000 my $ra ||= Git::SVN::Ra->new($url);
1001 my $r = $_revision;
1002 if ($r eq 'HEAD') {
1003 $r = $ra->get_latest_revnum;
1004 } elsif ($r !~ /^\d+$/) {
1005 die "revision argument: $r not understood by git-svn\n";
1006 }
1007 my %ed_opts = ( r => $r,
1008 log => $_message,
1009 ra => $ra,
1010 tree_a => $ta,
1011 tree_b => $tb,
1012 editor_cb => sub { print "Committed r$_[0]\n" },
1013 svn_path => $svn_path );
1014 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
1015 print "No changes\n$ta == $tb\n";
1016 }
1017 }
1019 sub escape_uri_only {
1020 my ($uri) = @_;
1021 my @tmp;
1022 foreach (split m{/}, $uri) {
1023 s/([^~\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
1024 push @tmp, $_;
1025 }
1026 join('/', @tmp);
1027 }
1029 sub escape_url {
1030 my ($url) = @_;
1031 if ($url =~ m#^([^:]+)://([^/]*)(.*)$#) {
1032 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
1033 $url = "$scheme://$domain$uri";
1034 }
1035 $url;
1036 }
1038 sub cmd_info {
1039 my $path = canonicalize_path(defined($_[0]) ? $_[0] : ".");
1040 my $fullpath = canonicalize_path($cmd_dir_prefix . $path);
1041 if (exists $_[1]) {
1042 die "Too many arguments specified\n";
1043 }
1045 my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
1047 if (!$file_type && !$diff_status) {
1048 print STDERR "svn: '$path' is not under version control\n";
1049 exit 1;
1050 }
1052 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
1053 unless ($gs) {
1054 die "Unable to determine upstream SVN information from ",
1055 "working tree history\n";
1056 }
1058 # canonicalize_path() will return "" to make libsvn 1.5.x happy,
1059 $path = "." if $path eq "";
1061 my $full_url = $url . ($fullpath eq "" ? "" : "/$fullpath");
1063 if ($_url) {
1064 print escape_url($full_url), "\n";
1065 return;
1066 }
1068 my $result = "Path: $path\n";
1069 $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
1070 $result .= "URL: " . escape_url($full_url) . "\n";
1072 eval {
1073 my $repos_root = $gs->repos_root;
1074 Git::SVN::remove_username($repos_root);
1075 $result .= "Repository Root: " . escape_url($repos_root) . "\n";
1076 };
1077 if ($@) {
1078 $result .= "Repository Root: (offline)\n";
1079 }
1080 $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A" &&
1081 ($SVN::Core::VERSION le '1.5.4' || $file_type ne "dir");
1082 $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
1084 $result .= "Node Kind: " .
1085 ($file_type eq "dir" ? "directory" : "file") . "\n";
1087 my $schedule = $diff_status eq "A"
1088 ? "add"
1089 : ($diff_status eq "D" ? "delete" : "normal");
1090 $result .= "Schedule: $schedule\n";
1092 if ($diff_status eq "A") {
1093 print $result, "\n";
1094 return;
1095 }
1097 my ($lc_author, $lc_rev, $lc_date_utc);
1098 my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $fullpath);
1099 my $log = command_output_pipe(@args);
1100 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
1101 while (<$log>) {
1102 if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
1103 $lc_author = $1;
1104 $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
1105 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
1106 (undef, $lc_rev, undef) = ::extract_metadata($1);
1107 }
1108 }
1109 close $log;
1111 Git::SVN::Log::set_local_timezone();
1113 $result .= "Last Changed Author: $lc_author\n";
1114 $result .= "Last Changed Rev: $lc_rev\n";
1115 $result .= "Last Changed Date: " .
1116 Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
1118 if ($file_type ne "dir") {
1119 my $text_last_updated_date =
1120 ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
1121 $result .=
1122 "Text Last Updated: " .
1123 Git::SVN::Log::format_svn_date($text_last_updated_date) .
1124 "\n";
1125 my $checksum;
1126 if ($diff_status eq "D") {
1127 my ($fh, $ctx) =
1128 command_output_pipe(qw(cat-file blob), "HEAD:$path");
1129 if ($file_type eq "link") {
1130 my $file_name = <$fh>;
1131 $checksum = md5sum("link $file_name");
1132 } else {
1133 $checksum = md5sum($fh);
1134 }
1135 command_close_pipe($fh, $ctx);
1136 } elsif ($file_type eq "link") {
1137 my $file_name =
1138 command(qw(cat-file blob), "HEAD:$path");
1139 $checksum =
1140 md5sum("link " . $file_name);
1141 } else {
1142 open FILE, "<", $path or die $!;
1143 $checksum = md5sum(\*FILE);
1144 close FILE or die $!;
1145 }
1146 $result .= "Checksum: " . $checksum . "\n";
1147 }
1149 print $result, "\n";
1150 }
1152 sub cmd_reset {
1153 my $target = shift || $_revision or die "SVN revision required\n";
1154 $target = $1 if $target =~ /^r(\d+)$/;
1155 $target =~ /^\d+$/ or die "Numeric SVN revision expected\n";
1156 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
1157 unless ($gs) {
1158 die "Unable to determine upstream SVN information from ".
1159 "history\n";
1160 }
1161 my ($r, $c) = $gs->find_rev_before($target, not $_fetch_parent);
1162 $gs->rev_map_set($r, $c, 'reset', $uuid);
1163 print "r$r = $c ($gs->{ref_id})\n";
1164 }
1166 sub cmd_gc {
1167 if (!$can_compress) {
1168 warn "Compress::Zlib could not be found; unhandled.log " .
1169 "files will not be compressed.\n";
1170 }
1171 find({ wanted => \&gc_directory, no_chdir => 1}, "$ENV{GIT_DIR}/svn");
1172 }
1174 ########################### utility functions #########################
1176 sub rebase_cmd {
1177 my @cmd = qw/rebase/;
1178 push @cmd, '-v' if $_verbose;
1179 push @cmd, qw/--merge/ if $_merge;
1180 push @cmd, "--strategy=$_strategy" if $_strategy;
1181 @cmd;
1182 }
1184 sub post_fetch_checkout {
1185 return if $_no_checkout;
1186 my $gs = $Git::SVN::_head or return;
1187 return if verify_ref('refs/heads/master^0');
1189 # look for "trunk" ref if it exists
1190 my $remote = Git::SVN::read_all_remotes()->{$gs->{repo_id}};
1191 my $fetch = $remote->{fetch};
1192 if ($fetch) {
1193 foreach my $p (keys %$fetch) {
1194 basename($fetch->{$p}) eq 'trunk' or next;
1195 $gs = Git::SVN->new($fetch->{$p}, $gs->{repo_id}, $p);
1196 last;
1197 }
1198 }
1200 my $valid_head = verify_ref('HEAD^0');
1201 command_noisy(qw(update-ref refs/heads/master), $gs->refname);
1202 return if ($valid_head || !verify_ref('HEAD^0'));
1204 return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
1205 my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
1206 return if -f $index;
1208 return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
1209 return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
1210 command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
1211 print STDERR "Checked out HEAD:\n ",
1212 $gs->full_url, " r", $gs->last_rev, "\n";
1213 $gs->mkemptydirs($gs->last_rev);
1214 }
1216 sub complete_svn_url {
1217 my ($url, $path) = @_;
1218 $path =~ s#/+$##;
1219 if ($path !~ m#^[a-z\+]+://#) {
1220 if (!defined $url || $url !~ m#^[a-z\+]+://#) {
1221 fatal("E: '$path' is not a complete URL ",
1222 "and a separate URL is not specified");
1223 }
1224 return ($url, $path);
1225 }
1226 return ($path, '');
1227 }
1229 sub complete_url_ls_init {
1230 my ($ra, $repo_path, $switch, $pfx) = @_;
1231 unless ($repo_path) {
1232 print STDERR "W: $switch not specified\n";
1233 return;
1234 }
1235 $repo_path =~ s#/+$##;
1236 if ($repo_path =~ m#^[a-z\+]+://#) {
1237 $ra = Git::SVN::Ra->new($repo_path);
1238 $repo_path = '';
1239 } else {
1240 $repo_path =~ s#^/+##;
1241 unless ($ra) {
1242 fatal("E: '$repo_path' is not a complete URL ",
1243 "and a separate URL is not specified");
1244 }
1245 }
1246 my $url = $ra->{url};
1247 my $gs = Git::SVN->init($url, undef, undef, undef, 1);
1248 my $k = "svn-remote.$gs->{repo_id}.url";
1249 my $orig_url = eval { command_oneline(qw/config --get/, $k) };
1250 if ($orig_url && ($orig_url ne $gs->{url})) {
1251 die "$k already set: $orig_url\n",
1252 "wanted to set to: $gs->{url}\n";
1253 }
1254 command_oneline('config', $k, $gs->{url}) unless $orig_url;
1255 my $remote_path = "$gs->{path}/$repo_path";
1256 $remote_path =~ s{%([0-9A-F]{2})}{chr hex($1)}ieg;
1257 $remote_path =~ s#/+#/#g;
1258 $remote_path =~ s#^/##g;
1259 $remote_path .= "/*" if $remote_path !~ /\*/;
1260 my ($n) = ($switch =~ /^--(\w+)/);
1261 if (length $pfx && $pfx !~ m#/$#) {
1262 die "--prefix='$pfx' must have a trailing slash '/'\n";
1263 }
1264 command_noisy('config',
1265 '--add',
1266 "svn-remote.$gs->{repo_id}.$n",
1267 "$remote_path:refs/remotes/$pfx*" .
1268 ('/*' x (($remote_path =~ tr/*/*/) - 1)) );
1269 }
1271 sub verify_ref {
1272 my ($ref) = @_;
1273 eval { command_oneline([ 'rev-parse', '--verify', $ref ],
1274 { STDERR => 0 }); };
1275 }
1277 sub get_tree_from_treeish {
1278 my ($treeish) = @_;
1279 # $treeish can be a symbolic ref, too:
1280 my $type = command_oneline(qw/cat-file -t/, $treeish);
1281 my $expected;
1282 while ($type eq 'tag') {
1283 ($treeish, $type) = command(qw/cat-file tag/, $treeish);
1284 }
1285 if ($type eq 'commit') {
1286 $expected = (grep /^tree /, command(qw/cat-file commit/,
1287 $treeish))[0];
1288 ($expected) = ($expected =~ /^tree ($sha1)$/o);
1289 die "Unable to get tree from $treeish\n" unless $expected;
1290 } elsif ($type eq 'tree') {
1291 $expected = $treeish;
1292 } else {
1293 die "$treeish is a $type, expected tree, tag or commit\n";
1294 }
1295 return $expected;
1296 }
1298 sub get_commit_entry {
1299 my ($treeish) = shift;
1300 my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
1301 my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
1302 my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
1303 open my $log_fh, '>', $commit_editmsg or croak $!;
1305 my $type = command_oneline(qw/cat-file -t/, $treeish);
1306 if ($type eq 'commit' || $type eq 'tag') {
1307 my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1308 $type, $treeish);
1309 my $in_msg = 0;
1310 my $author;
1311 my $saw_from = 0;
1312 my $msgbuf = "";
1313 while (<$msg_fh>) {
1314 if (!$in_msg) {
1315 $in_msg = 1 if (/^\s*$/);
1316 $author = $1 if (/^author (.*>)/);
1317 } elsif (/^git-svn-id: /) {
1318 # skip this for now, we regenerate the
1319 # correct one on re-fetch anyways
1320 # TODO: set *:merge properties or like...
1321 } else {
1322 if (/^From:/ || /^Signed-off-by:/) {
1323 $saw_from = 1;
1324 }
1325 $msgbuf .= $_;
1326 }
1327 }
1328 $msgbuf =~ s/\s+$//s;
1329 if ($Git::SVN::_add_author_from && defined($author)
1330 && !$saw_from) {
1331 $msgbuf .= "\n\nFrom: $author";
1332 }
1333 print $log_fh $msgbuf or croak $!;
1334 command_close_pipe($msg_fh, $ctx);
1335 }
1336 close $log_fh or croak $!;
1338 if ($_edit || ($type eq 'tree')) {
1339 chomp(my $editor = command_oneline(qw(var GIT_EDITOR)));
1340 system('sh', '-c', $editor.' "$@"', $editor, $commit_editmsg);
1341 }
1342 rename $commit_editmsg, $commit_msg or croak $!;
1343 {
1344 require Encode;
1345 # SVN requires messages to be UTF-8 when entering the repo
1346 local $/;
1347 open $log_fh, '<', $commit_msg or croak $!;
1348 binmode $log_fh;
1349 chomp($log_entry{log} = <$log_fh>);
1351 my $enc = Git::config('i18n.commitencoding') || 'UTF-8';
1352 my $msg = $log_entry{log};
1354 eval { $msg = Encode::decode($enc, $msg, 1) };
1355 if ($@) {
1356 die "Could not decode as $enc:\n", $msg,
1357 "\nPerhaps you need to set i18n.commitencoding\n";
1358 }
1360 eval { $msg = Encode::encode('UTF-8', $msg, 1) };
1361 die "Could not encode as UTF-8:\n$msg\n" if $@;
1363 $log_entry{log} = $msg;
1365 close $log_fh or croak $!;
1366 }
1367 unlink $commit_msg;
1368 \%log_entry;
1369 }
1371 sub s_to_file {
1372 my ($str, $file, $mode) = @_;
1373 open my $fd,'>',$file or croak $!;
1374 print $fd $str,"\n" or croak $!;
1375 close $fd or croak $!;
1376 chmod ($mode &~ umask, $file) if (defined $mode);
1377 }
1379 sub file_to_s {
1380 my $file = shift;
1381 open my $fd,'<',$file or croak "$!: file: $file\n";
1382 local $/;
1383 my $ret = <$fd>;
1384 close $fd or croak $!;
1385 $ret =~ s/\s*$//s;
1386 return $ret;
1387 }
1389 # '<svn username> = real-name <email address>' mapping based on git-svnimport:
1390 sub load_authors {
1391 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1392 my $log = $cmd eq 'log';
1393 while (<$authors>) {
1394 chomp;
1395 next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1396 my ($user, $name, $email) = ($1, $2, $3);
1397 if ($log) {
1398 $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1399 } else {
1400 $users{$user} = [$name, $email];
1401 }
1402 }
1403 close $authors or croak $!;
1404 }
1406 # convert GetOpt::Long specs for use by git-config
1407 sub read_git_config {
1408 my $opts = shift;
1409 my @config_only;
1410 foreach my $o (keys %$opts) {
1411 # if we have mixedCase and a long option-only, then
1412 # it's a config-only variable that we don't need for
1413 # the command-line.
1414 push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1415 my $v = $opts->{$o};
1416 my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1417 $key =~ s/-//g;
1418 my $arg = 'git config';
1419 $arg .= ' --int' if ($o =~ /[:=]i$/);
1420 $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1421 if (ref $v eq 'ARRAY') {
1422 chomp(my @tmp = `$arg --get-all svn.$key`);
1423 @$v = @tmp if @tmp;
1424 } else {
1425 chomp(my $tmp = `$arg --get svn.$key`);
1426 if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1427 $$v = $tmp;
1428 }
1429 }
1430 }
1431 delete @$opts{@config_only} if @config_only;
1432 }
1434 sub extract_metadata {
1435 my $id = shift or return (undef, undef, undef);
1436 my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1437 \s([a-f\d\-]+)$/ix);
1438 if (!defined $rev || !$uuid || !$url) {
1439 # some of the original repositories I made had
1440 # identifiers like this:
1441 ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/i);
1442 }
1443 return ($url, $rev, $uuid);
1444 }
1446 sub cmt_metadata {
1447 return extract_metadata((grep(/^git-svn-id: /,
1448 command(qw/cat-file commit/, shift)))[-1]);
1449 }
1451 sub cmt_sha2rev_batch {
1452 my %s2r;
1453 my ($pid, $in, $out, $ctx) = command_bidi_pipe(qw/cat-file --batch/);
1454 my $list = shift;
1456 foreach my $sha (@{$list}) {
1457 my $first = 1;
1458 my $size = 0;
1459 print $out $sha, "\n";
1461 while (my $line = <$in>) {
1462 if ($first && $line =~ /^[[:xdigit:]]{40}\smissing$/) {
1463 last;
1464 } elsif ($first &&
1465 $line =~ /^[[:xdigit:]]{40}\scommit\s(\d+)$/) {
1466 $first = 0;
1467 $size = $1;
1468 next;
1469 } elsif ($line =~ /^(git-svn-id: )/) {
1470 my (undef, $rev, undef) =
1471 extract_metadata($line);
1472 $s2r{$sha} = $rev;
1473 }
1475 $size -= length($line);
1476 last if ($size == 0);
1477 }
1478 }
1480 command_close_bidi_pipe($pid, $in, $out, $ctx);
1482 return \%s2r;
1483 }
1485 sub working_head_info {
1486 my ($head, $refs) = @_;
1487 my @args = ('log', '--no-color', '--first-parent', '--pretty=medium');
1488 my ($fh, $ctx) = command_output_pipe(@args, $head);
1489 my $hash;
1490 my %max;
1491 while (<$fh>) {
1492 if ( m{^commit ($::sha1)$} ) {
1493 unshift @$refs, $hash if $hash and $refs;
1494 $hash = $1;
1495 next;
1496 }
1497 next unless s{^\s*(git-svn-id:)}{$1};
1498 my ($url, $rev, $uuid) = extract_metadata($_);
1499 if (defined $url && defined $rev) {
1500 next if $max{$url} and $max{$url} < $rev;
1501 if (my $gs = Git::SVN->find_by_url($url)) {
1502 my $c = $gs->rev_map_get($rev, $uuid);
1503 if ($c && $c eq $hash) {
1504 close $fh; # break the pipe
1505 return ($url, $rev, $uuid, $gs);
1506 } else {
1507 $max{$url} ||= $gs->rev_map_max;
1508 }
1509 }
1510 }
1511 }
1512 command_close_pipe($fh, $ctx);
1513 (undef, undef, undef, undef);
1514 }
1516 sub read_commit_parents {
1517 my ($parents, $c) = @_;
1518 chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1519 $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1520 @{$parents->{$c}} = split(/ /, $p);
1521 }
1523 sub linearize_history {
1524 my ($gs, $refs) = @_;
1525 my %parents;
1526 foreach my $c (@$refs) {
1527 read_commit_parents(\%parents, $c);
1528 }
1530 my @linear_refs;
1531 my %skip = ();
1532 my $last_svn_commit = $gs->last_commit;
1533 foreach my $c (reverse @$refs) {
1534 next if $c eq $last_svn_commit;
1535 last if $skip{$c};
1537 unshift @linear_refs, $c;
1538 $skip{$c} = 1;
1540 # we only want the first parent to diff against for linear
1541 # history, we save the rest to inject when we finalize the
1542 # svn commit
1543 my $fp_a = verify_ref("$c~1");
1544 my $fp_b = shift @{$parents{$c}} if $parents{$c};
1545 if (!$fp_a || !$fp_b) {
1546 die "Commit $c\n",
1547 "has no parent commit, and therefore ",
1548 "nothing to diff against.\n",
1549 "You should be working from a repository ",
1550 "originally created by git-svn\n";
1551 }
1552 if ($fp_a ne $fp_b) {
1553 die "$c~1 = $fp_a, however parsing commit $c ",
1554 "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1555 }
1557 foreach my $p (@{$parents{$c}}) {
1558 $skip{$p} = 1;
1559 }
1560 }
1561 (\@linear_refs, \%parents);
1562 }
1564 sub find_file_type_and_diff_status {
1565 my ($path) = @_;
1566 return ('dir', '') if $path eq '';
1568 my $diff_output =
1569 command_oneline(qw(diff --cached --name-status --), $path) || "";
1570 my $diff_status = (split(' ', $diff_output))[0] || "";
1572 my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1574 return (undef, undef) if !$diff_status && !$ls_tree;
1576 if ($diff_status eq "A") {
1577 return ("link", $diff_status) if -l $path;
1578 return ("dir", $diff_status) if -d $path;
1579 return ("file", $diff_status);
1580 }
1582 my $mode = (split(' ', $ls_tree))[0] || "";
1584 return ("link", $diff_status) if $mode eq "120000";
1585 return ("dir", $diff_status) if $mode eq "040000";
1586 return ("file", $diff_status);
1587 }
1589 sub md5sum {
1590 my $arg = shift;
1591 my $ref = ref $arg;
1592 my $md5 = Digest::MD5->new();
1593 if ($ref eq 'GLOB' || $ref eq 'IO::File' || $ref eq 'File::Temp') {
1594 $md5->addfile($arg) or croak $!;
1595 } elsif ($ref eq 'SCALAR') {
1596 $md5->add($$arg) or croak $!;
1597 } elsif (!$ref) {
1598 $md5->add($arg) or croak $!;
1599 } else {
1600 ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
1601 }
1602 return $md5->hexdigest();
1603 }
1605 sub gc_directory {
1606 if ($can_compress && -f $_ && basename($_) eq "unhandled.log") {
1607 my $out_filename = $_ . ".gz";
1608 open my $in_fh, "<", $_ or die "Unable to open $_: $!\n";
1609 binmode $in_fh;
1610 my $gz = Compress::Zlib::gzopen($out_filename, "ab") or
1611 die "Unable to open $out_filename: $!\n";
1613 my $res;
1614 while ($res = sysread($in_fh, my $str, 1024)) {
1615 $gz->gzwrite($str) or
1616 die "Unable to write: ".$gz->gzerror()."!\n";
1617 }
1618 unlink $_ or die "unlink $File::Find::name: $!\n";
1619 } elsif (-f $_ && basename($_) eq "index") {
1620 unlink $_ or die "unlink $_: $!\n";
1621 }
1622 }
1624 package Git::SVN;
1625 use strict;
1626 use warnings;
1627 use Fcntl qw/:DEFAULT :seek/;
1628 use constant rev_map_fmt => 'NH40';
1629 use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1630 $_repack $_repack_flags $_use_svm_props $_head
1631 $_use_svnsync_props $no_reuse_existing $_minimize_url
1632 $_use_log_author $_add_author_from $_localtime/;
1633 use Carp qw/croak/;
1634 use File::Path qw/mkpath/;
1635 use File::Copy qw/copy/;
1636 use IPC::Open3;
1638 my ($_gc_nr, $_gc_period);
1640 # properties that we do not log:
1641 my %SKIP_PROP;
1642 BEGIN {
1643 %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1644 svn:special svn:executable
1645 svn:entry:committed-rev
1646 svn:entry:last-author
1647 svn:entry:uuid
1648 svn:entry:committed-date/;
1650 # some options are read globally, but can be overridden locally
1651 # per [svn-remote "..."] section. Command-line options will *NOT*
1652 # override options set in an [svn-remote "..."] section
1653 no strict 'refs';
1654 for my $option (qw/follow_parent no_metadata use_svm_props
1655 use_svnsync_props/) {
1656 my $key = $option;
1657 $key =~ tr/_//d;
1658 my $prop = "-$option";
1659 *$option = sub {
1660 my ($self) = @_;
1661 return $self->{$prop} if exists $self->{$prop};
1662 my $k = "svn-remote.$self->{repo_id}.$key";
1663 eval { command_oneline(qw/config --get/, $k) };
1664 if ($@) {
1665 $self->{$prop} = ${"Git::SVN::_$option"};
1666 } else {
1667 my $v = command_oneline(qw/config --bool/,$k);
1668 $self->{$prop} = $v eq 'false' ? 0 : 1;
1669 }
1670 return $self->{$prop};
1671 }
1672 }
1673 }
1676 my (%LOCKFILES, %INDEX_FILES);
1677 END {
1678 unlink keys %LOCKFILES if %LOCKFILES;
1679 unlink keys %INDEX_FILES if %INDEX_FILES;
1680 }
1682 sub resolve_local_globs {
1683 my ($url, $fetch, $glob_spec) = @_;
1684 return unless defined $glob_spec;
1685 my $ref = $glob_spec->{ref};
1686 my $path = $glob_spec->{path};
1687 foreach (command(qw#for-each-ref --format=%(refname) refs/#)) {
1688 next unless m#^$ref->{regex}$#;
1689 my $p = $1;
1690 my $pathname = desanitize_refname($path->full_path($p));
1691 my $refname = desanitize_refname($ref->full_path($p));
1692 if (my $existing = $fetch->{$pathname}) {
1693 if ($existing ne $refname) {
1694 die "Refspec conflict:\n",
1695 "existing: $existing\n",
1696 " globbed: $refname\n";
1697 }
1698 my $u = (::cmt_metadata("$refname"))[0];
1699 $u =~ s!^\Q$url\E(/|$)!! or die
1700 "$refname: '$url' not found in '$u'\n";
1701 if ($pathname ne $u) {
1702 warn "W: Refspec glob conflict ",
1703 "(ref: $refname):\n",
1704 "expected path: $pathname\n",
1705 " real path: $u\n",
1706 "Continuing ahead with $u\n";
1707 next;
1708 }
1709 } else {
1710 $fetch->{$pathname} = $refname;
1711 }
1712 }
1713 }
1715 sub parse_revision_argument {
1716 my ($base, $head) = @_;
1717 if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1718 return ($base, $head);
1719 }
1720 return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1721 return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1722 return ($head, $head) if ($::_revision eq 'HEAD');
1723 return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1724 return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1725 die "revision argument: $::_revision not understood by git-svn\n";
1726 }
1728 sub fetch_all {
1729 my ($repo_id, $remotes) = @_;
1730 if (ref $repo_id) {
1731 my $gs = $repo_id;
1732 $repo_id = undef;
1733 $repo_id = $gs->{repo_id};
1734 }
1735 $remotes ||= read_all_remotes();
1736 my $remote = $remotes->{$repo_id} or
1737 die "[svn-remote \"$repo_id\"] unknown\n";
1738 my $fetch = $remote->{fetch};
1739 my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1740 my (@gs, @globs);
1741 my $ra = Git::SVN::Ra->new($url);
1742 my $uuid = $ra->get_uuid;
1743 my $head = $ra->get_latest_revnum;
1744 $ra->get_log("", $head, 0, 1, 0, 1, sub { $head = $_[1] });
1745 my $base = defined $fetch ? $head : 0;
1747 # read the max revs for wildcard expansion (branches/*, tags/*)
1748 foreach my $t (qw/branches tags/) {
1749 defined $remote->{$t} or next;
1750 push @globs, @{$remote->{$t}};
1752 my $max_rev = eval { tmp_config(qw/--int --get/,
1753 "svn-remote.$repo_id.${t}-maxRev") };
1754 if (defined $max_rev && ($max_rev < $base)) {
1755 $base = $max_rev;
1756 } elsif (!defined $max_rev) {
1757 $base = 0;
1758 }
1759 }
1761 if ($fetch) {
1762 foreach my $p (sort keys %$fetch) {
1763 my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1764 my $lr = $gs->rev_map_max;
1765 if (defined $lr) {
1766 $base = $lr if ($lr < $base);
1767 }
1768 push @gs, $gs;
1769 }
1770 }
1772 ($base, $head) = parse_revision_argument($base, $head);
1773 $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1774 }
1776 sub read_all_remotes {
1777 my $r = {};
1778 my $use_svm_props = eval { command_oneline(qw/config --bool
1779 svn.useSvmProps/) };
1780 $use_svm_props = $use_svm_props eq 'true' if $use_svm_props;
1781 my $svn_refspec = qr{\s*(.*?)\s*:\s*(.+?)\s*};
1782 foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1783 if (m!^(.+)\.fetch=$svn_refspec$!) {
1784 my ($remote, $local_ref, $remote_ref) = ($1, $2, $3);
1785 die("svn-remote.$remote: remote ref '$remote_ref' "
1786 . "must start with 'refs/'\n")
1787 unless $remote_ref =~ m{^refs/};
1788 $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1789 $r->{$remote}->{svm} = {} if $use_svm_props;
1790 } elsif (m!^(.+)\.usesvmprops=\s*(.*)\s*$!) {
1791 $r->{$1}->{svm} = {};
1792 } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1793 $r->{$1}->{url} = $2;
1794 } elsif (m!^(.+)\.(branches|tags)=$svn_refspec$!) {
1795 my ($remote, $t, $local_ref, $remote_ref) =
1796 ($1, $2, $3, $4);
1797 die("svn-remote.$remote: remote ref '$remote_ref' ($t) "
1798 . "must start with 'refs/'\n")
1799 unless $remote_ref =~ m{^refs/};
1800 my $rs = {
1801 t => $t,
1802 remote => $remote,
1803 path => Git::SVN::GlobSpec->new($local_ref),
1804 ref => Git::SVN::GlobSpec->new($remote_ref) };
1805 if (length($rs->{ref}->{right}) != 0) {
1806 die "The '*' glob character must be the last ",
1807 "character of '$remote_ref'\n";
1808 }
1809 push @{ $r->{$remote}->{$t} }, $rs;
1810 }
1811 }
1813 map {
1814 if (defined $r->{$_}->{svm}) {
1815 my $svm;
1816 eval {
1817 my $section = "svn-remote.$_";
1818 $svm = {
1819 source => tmp_config('--get',
1820 "$section.svm-source"),
1821 replace => tmp_config('--get',
1822 "$section.svm-replace"),
1823 }
1824 };
1825 $r->{$_}->{svm} = $svm;
1826 }
1827 } keys %$r;
1829 $r;
1830 }
1832 sub init_vars {
1833 $_gc_nr = $_gc_period = 1000;
1834 if (defined $_repack || defined $_repack_flags) {
1835 warn "Repack options are obsolete; they have no effect.\n";
1836 }
1837 }
1839 sub verify_remotes_sanity {
1840 return unless -d $ENV{GIT_DIR};
1841 my %seen;
1842 foreach (command(qw/config -l/)) {
1843 if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1844 if ($seen{$1}) {
1845 die "Remote ref refs/remote/$1 is tracked by",
1846 "\n \"$_\"\nand\n \"$seen{$1}\"\n",
1847 "Please resolve this ambiguity in ",
1848 "your git configuration file before ",
1849 "continuing\n";
1850 }
1851 $seen{$1} = $_;
1852 }
1853 }
1854 }
1856 sub find_existing_remote {
1857 my ($url, $remotes) = @_;
1858 return undef if $no_reuse_existing;
1859 my $existing;
1860 foreach my $repo_id (keys %$remotes) {
1861 my $u = $remotes->{$repo_id}->{url} or next;
1862 next if $u ne $url;
1863 $existing = $repo_id;
1864 last;
1865 }
1866 $existing;
1867 }
1869 sub init_remote_config {
1870 my ($self, $url, $no_write) = @_;
1871 $url =~ s!/+$!!; # strip trailing slash
1872 my $r = read_all_remotes();
1873 my $existing = find_existing_remote($url, $r);
1874 if ($existing) {
1875 unless ($no_write) {
1876 print STDERR "Using existing ",
1877 "[svn-remote \"$existing\"]\n";
1878 }
1879 $self->{repo_id} = $existing;
1880 } elsif ($_minimize_url) {
1881 my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1882 $existing = find_existing_remote($min_url, $r);
1883 if ($existing) {
1884 unless ($no_write) {
1885 print STDERR "Using existing ",
1886 "[svn-remote \"$existing\"]\n";
1887 }
1888 $self->{repo_id} = $existing;
1889 }
1890 if ($min_url ne $url) {
1891 unless ($no_write) {
1892 print STDERR "Using higher level of URL: ",
1893 "$url => $min_url\n";
1894 }
1895 my $old_path = $self->{path};
1896 $self->{path} = $url;
1897 $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1898 if (length $old_path) {
1899 $self->{path} .= "/$old_path";
1900 }
1901 $url = $min_url;
1902 }
1903 }
1904 my $orig_url;
1905 if (!$existing) {
1906 # verify that we aren't overwriting anything:
1907 $orig_url = eval {
1908 command_oneline('config', '--get',
1909 "svn-remote.$self->{repo_id}.url")
1910 };
1911 if ($orig_url && ($orig_url ne $url)) {
1912 die "svn-remote.$self->{repo_id}.url already set: ",
1913 "$orig_url\nwanted to set to: $url\n";
1914 }
1915 }
1916 my ($xrepo_id, $xpath) = find_ref($self->refname);
1917 if (!$no_write && defined $xpath) {
1918 die "svn-remote.$xrepo_id.fetch already set to track ",
1919 "$xpath:", $self->refname, "\n";
1920 }
1921 unless ($no_write) {
1922 command_noisy('config',
1923 "svn-remote.$self->{repo_id}.url", $url);
1924 $self->{path} =~ s{^/}{};
1925 $self->{path} =~ s{%([0-9A-F]{2})}{chr hex($1)}ieg;
1926 command_noisy('config', '--add',
1927 "svn-remote.$self->{repo_id}.fetch",
1928 "$self->{path}:".$self->refname);
1929 }
1930 $self->{url} = $url;
1931 }
1933 sub find_by_url { # repos_root and, path are optional
1934 my ($class, $full_url, $repos_root, $path) = @_;
1936 return undef unless defined $full_url;
1937 remove_username($full_url);
1938 remove_username($repos_root) if defined $repos_root;
1939 my $remotes = read_all_remotes();
1940 if (defined $full_url && defined $repos_root && !defined $path) {
1941 $path = $full_url;
1942 $path =~ s#^\Q$repos_root\E(?:/|$)##;
1943 }
1944 foreach my $repo_id (keys %$remotes) {
1945 my $u = $remotes->{$repo_id}->{url} or next;
1946 remove_username($u);
1947 next if defined $repos_root && $repos_root ne $u;
1949 my $fetch = $remotes->{$repo_id}->{fetch} || {};
1950 foreach my $t (qw/branches tags/) {
1951 foreach my $globspec (@{$remotes->{$repo_id}->{$t}}) {
1952 resolve_local_globs($u, $fetch, $globspec);
1953 }
1954 }
1955 my $p = $path;
1956 my $rwr = rewrite_root({repo_id => $repo_id});
1957 my $svm = $remotes->{$repo_id}->{svm}
1958 if defined $remotes->{$repo_id}->{svm};
1959 unless (defined $p) {
1960 $p = $full_url;
1961 my $z = $u;
1962 my $prefix = '';
1963 if ($rwr) {
1964 $z = $rwr;
1965 remove_username($z);
1966 } elsif (defined $svm) {
1967 $z = $svm->{source};
1968 $prefix = $svm->{replace};
1969 $prefix =~ s#^\Q$u\E(?:/|$)##;
1970 $prefix =~ s#/$##;
1971 }
1972 $p =~ s#^\Q$z\E(?:/|$)#$prefix# or next;
1973 }
1974 foreach my $f (keys %$fetch) {
1975 next if $f ne $p;
1976 return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1977 }
1978 }
1979 undef;
1980 }
1982 sub init {
1983 my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1984 my $self = _new($class, $repo_id, $ref_id, $path);
1985 if (defined $url) {
1986 $self->init_remote_config($url, $no_write);
1987 }
1988 $self;
1989 }
1991 sub find_ref {
1992 my ($ref_id) = @_;
1993 foreach (command(qw/config -l/)) {
1994 next unless m!^svn-remote\.(.+)\.fetch=
1995 \s*(.*?)\s*:\s*(.+?)\s*$!x;
1996 my ($repo_id, $path, $ref) = ($1, $2, $3);
1997 if ($ref eq $ref_id) {
1998 $path = '' if ($path =~ m#^\./?#);
1999 return ($repo_id, $path);
2000 }
2001 }
2002 (undef, undef, undef);
2003 }
2005 sub new {
2006 my ($class, $ref_id, $repo_id, $path) = @_;
2007 if (defined $ref_id && !defined $repo_id && !defined $path) {
2008 ($repo_id, $path) = find_ref($ref_id);
2009 if (!defined $repo_id) {
2010 die "Could not find a \"svn-remote.*.fetch\" key ",
2011 "in the repository configuration matching: ",
2012 "$ref_id\n";
2013 }
2014 }
2015 my $self = _new($class, $repo_id, $ref_id, $path);
2016 if (!defined $self->{path} || !length $self->{path}) {
2017 my $fetch = command_oneline('config', '--get',
2018 "svn-remote.$repo_id.fetch",
2019 ":$ref_id\$") or
2020 die "Failed to read \"svn-remote.$repo_id.fetch\" ",
2021 "\":$ref_id\$\" in config\n";
2022 ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
2023 }
2024 $self->{url} = command_oneline('config', '--get',
2025 "svn-remote.$repo_id.url") or
2026 die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
2027 $self->rebuild;
2028 $self;
2029 }
2031 sub refname {
2032 my ($refname) = $_[0]->{ref_id} ;
2034 # It cannot end with a slash /, we'll throw up on this because
2035 # SVN can't have directories with a slash in their name, either:
2036 if ($refname =~ m{/$}) {
2037 die "ref: '$refname' ends with a trailing slash, this is ",
2038 "not permitted by git nor Subversion\n";
2039 }
2041 # It cannot have ASCII control character space, tilde ~, caret ^,
2042 # colon :, question-mark ?, asterisk *, space, or open bracket [
2043 # anywhere.
2044 #
2045 # Additionally, % must be escaped because it is used for escaping
2046 # and we want our escaped refname to be reversible
2047 $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
2049 # no slash-separated component can begin with a dot .
2050 # /.* becomes /%2E*
2051 $refname =~ s{/\.}{/%2E}g;
2053 # It cannot have two consecutive dots .. anywhere
2054 # .. becomes %2E%2E
2055 $refname =~ s{\.\.}{%2E%2E}g;
2057 return $refname;
2058 }
2060 sub desanitize_refname {
2061 my ($refname) = @_;
2062 $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
2063 return $refname;
2064 }
2066 sub svm_uuid {
2067 my ($self) = @_;
2068 return $self->{svm}->{uuid} if $self->svm;
2069 $self->ra;
2070 unless ($self->{svm}) {
2071 die "SVM UUID not cached, and reading remotely failed\n";
2072 }
2073 $self->{svm}->{uuid};
2074 }
2076 sub svm {
2077 my ($self) = @_;
2078 return $self->{svm} if $self->{svm};
2079 my $svm;
2080 # see if we have it in our config, first:
2081 eval {
2082 my $section = "svn-remote.$self->{repo_id}";
2083 $svm = {
2084 source => tmp_config('--get', "$section.svm-source"),
2085 uuid => tmp_config('--get', "$section.svm-uuid"),
2086 replace => tmp_config('--get', "$section.svm-replace"),
2087 }
2088 };
2089 if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
2090 $self->{svm} = $svm;
2091 }
2092 $self->{svm};
2093 }
2095 sub _set_svm_vars {
2096 my ($self, $ra) = @_;
2097 return $ra if $self->svm;
2099 my @err = ( "useSvmProps set, but failed to read SVM properties\n",
2100 "(svm:source, svm:uuid) ",
2101 "from the following URLs:\n" );
2102 sub read_svm_props {
2103 my ($self, $ra, $path, $r) = @_;
2104 my $props = ($ra->get_dir($path, $r))[2];
2105 my $src = $props->{'svm:source'};
2106 my $uuid = $props->{'svm:uuid'};
2107 return undef if (!$src || !$uuid);
2109 chomp($src, $uuid);
2111 $uuid =~ m{^[0-9a-f\-]{30,}$}i
2112 or die "doesn't look right - svm:uuid is '$uuid'\n";
2114 # the '!' is used to mark the repos_root!/relative/path
2115 $src =~ s{/?!/?}{/};
2116 $src =~ s{/+$}{}; # no trailing slashes please
2117 # username is of no interest
2118 $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
2120 my $replace = $ra->{url};
2121 $replace .= "/$path" if length $path;
2123 my $section = "svn-remote.$self->{repo_id}";
2124 tmp_config("$section.svm-source", $src);
2125 tmp_config("$section.svm-replace", $replace);
2126 tmp_config("$section.svm-uuid", $uuid);
2127 $self->{svm} = {
2128 source => $src,
2129 uuid => $uuid,
2130 replace => $replace
2131 };
2132 }
2134 my $r = $ra->get_latest_revnum;
2135 my $path = $self->{path};
2136 my %tried;
2137 while (length $path) {
2138 unless ($tried{"$self->{url}/$path"}) {
2139 return $ra if $self->read_svm_props($ra, $path, $r);
2140 $tried{"$self->{url}/$path"} = 1;
2141 }
2142 $path =~ s#/?[^/]+$##;
2143 }
2144 die "Path: '$path' should be ''\n" if $path ne '';
2145 return $ra if $self->read_svm_props($ra, $path, $r);
2146 $tried{"$self->{url}/$path"} = 1;
2148 if ($ra->{repos_root} eq $self->{url}) {
2149 die @err, (map { " $_\n" } keys %tried), "\n";
2150 }
2152 # nope, make sure we're connected to the repository root:
2153 my $ok;
2154 my @tried_b;
2155 $path = $ra->{svn_path};
2156 $ra = Git::SVN::Ra->new($ra->{repos_root});
2157 while (length $path) {
2158 unless ($tried{"$ra->{url}/$path"}) {
2159 $ok = $self->read_svm_props($ra, $path, $r);
2160 last if $ok;
2161 $tried{"$ra->{url}/$path"} = 1;
2162 }
2163 $path =~ s#/?[^/]+$##;
2164 }
2165 die "Path: '$path' should be ''\n" if $path ne '';
2166 $ok ||= $self->read_svm_props($ra, $path, $r);
2167 $tried{"$ra->{url}/$path"} = 1;
2168 if (!$ok) {
2169 die @err, (map { " $_\n" } keys %tried), "\n";
2170 }
2171 Git::SVN::Ra->new($self->{url});
2172 }
2174 sub svnsync {
2175 my ($self) = @_;
2176 return $self->{svnsync} if $self->{svnsync};
2178 if ($self->no_metadata) {
2179 die "Can't have both 'noMetadata' and ",
2180 "'useSvnsyncProps' options set!\n";
2181 }
2182 if ($self->rewrite_root) {
2183 die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
2184 "options set!\n";
2185 }
2187 my $svnsync;
2188 # see if we have it in our config, first:
2189 eval {
2190 my $section = "svn-remote.$self->{repo_id}";
2192 my $url = tmp_config('--get', "$section.svnsync-url");
2193 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
2194 die "doesn't look right - svn:sync-from-url is '$url'\n";
2196 my $uuid = tmp_config('--get', "$section.svnsync-uuid");
2197 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}i) or
2198 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
2200 $svnsync = { url => $url, uuid => $uuid }
2201 };
2202 if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
2203 return $self->{svnsync} = $svnsync;
2204 }
2206 my $err = "useSvnsyncProps set, but failed to read " .
2207 "svnsync property: svn:sync-from-";
2208 my $rp = $self->ra->rev_proplist(0);
2210 my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
2211 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
2212 die "doesn't look right - svn:sync-from-url is '$url'\n";
2214 my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
2215 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}i) or
2216 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
2218 my $section = "svn-remote.$self->{repo_id}";
2219 tmp_config('--add', "$section.svnsync-uuid", $uuid);
2220 tmp_config('--add', "$section.svnsync-url", $url);
2221 return $self->{svnsync} = { url => $url, uuid => $uuid };
2222 }
2224 # this allows us to memoize our SVN::Ra UUID locally and avoid a
2225 # remote lookup (useful for 'git svn log').
2226 sub ra_uuid {
2227 my ($self) = @_;
2228 unless ($self->{ra_uuid}) {
2229 my $key = "svn-remote.$self->{repo_id}.uuid";
2230 my $uuid = eval { tmp_config('--get', $key) };
2231 if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/i) {
2232 $self->{ra_uuid} = $uuid;
2233 } else {
2234 die "ra_uuid called without URL\n" unless $self->{url};
2235 $self->{ra_uuid} = $self->ra->get_uuid;
2236 tmp_config('--add', $key, $self->{ra_uuid});
2237 }
2238 }
2239 $self->{ra_uuid};
2240 }
2242 sub _set_repos_root {
2243 my ($self, $repos_root) = @_;
2244 my $k = "svn-remote.$self->{repo_id}.reposRoot";
2245 $repos_root ||= $self->ra->{repos_root};
2246 tmp_config($k, $repos_root);
2247 $repos_root;
2248 }
2250 sub repos_root {
2251 my ($self) = @_;
2252 my $k = "svn-remote.$self->{repo_id}.reposRoot";
2253 eval { tmp_config('--get', $k) } || $self->_set_repos_root;
2254 }
2256 sub ra {
2257 my ($self) = shift;
2258 my $ra = Git::SVN::Ra->new($self->{url});
2259 $self->_set_repos_root($ra->{repos_root});
2260 if ($self->use_svm_props && !$self->{svm}) {
2261 if ($self->no_metadata) {
2262 die "Can't have both 'noMetadata' and ",
2263 "'useSvmProps' options set!\n";
2264 } elsif ($self->use_svnsync_props) {
2265 die "Can't have both 'useSvnsyncProps' and ",
2266 "'useSvmProps' options set!\n";
2267 }
2268 $ra = $self->_set_svm_vars($ra);
2269 $self->{-want_revprops} = 1;
2270 }
2271 $ra;
2272 }
2274 # prop_walk(PATH, REV, SUB)
2275 # -------------------------
2276 # Recursively traverse PATH at revision REV and invoke SUB for each
2277 # directory that contains a SVN property. SUB will be invoked as
2278 # follows: &SUB(gs, path, props); where `gs' is this instance of
2279 # Git::SVN, `path' the path to the directory where the properties
2280 # `props' were found. The `path' will be relative to point of checkout,
2281 # that is, if url://repo/trunk is the current Git branch, and that
2282 # directory contains a sub-directory `d', SUB will be invoked with `/d/'
2283 # as `path' (note the trailing `/').
2284 sub prop_walk {
2285 my ($self, $path, $rev, $sub) = @_;
2287 $path =~ s#^/##;
2288 my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
2289 $path =~ s#^/*#/#g;
2290 my $p = $path;
2291 # Strip the irrelevant part of the path.
2292 $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
2293 # Ensure the path is terminated by a `/'.
2294 $p =~ s#/*$#/#;
2296 # The properties contain all the internal SVN stuff nobody
2297 # (usually) cares about.
2298 my $interesting_props = 0;
2299 foreach (keys %{$props}) {
2300 # If it doesn't start with `svn:', it must be a
2301 # user-defined property.
2302 ++$interesting_props and next if $_ !~ /^svn:/;
2303 # FIXME: Fragile, if SVN adds new public properties,
2304 # this needs to be updated.
2305 ++$interesting_props if /^svn:(?:ignore|keywords|executable
2306 |eol-style|mime-type
2307 |externals|needs-lock)$/x;
2308 }
2309 &$sub($self, $p, $props) if $interesting_props;
2311 foreach (sort keys %$dirent) {
2312 next if $dirent->{$_}->{kind} != $SVN::Node::dir;
2313 $self->prop_walk($self->{path} . $p . $_, $rev, $sub);
2314 }
2315 }
2317 sub last_rev { ($_[0]->last_rev_commit)[0] }
2318 sub last_commit { ($_[0]->last_rev_commit)[1] }
2320 # returns the newest SVN revision number and newest commit SHA1
2321 sub last_rev_commit {
2322 my ($self) = @_;
2323 if (defined $self->{last_rev} && defined $self->{last_commit}) {
2324 return ($self->{last_rev}, $self->{last_commit});
2325 }
2326 my $c = ::verify_ref($self->refname.'^0');
2327 if ($c && !$self->use_svm_props && !$self->no_metadata) {
2328 my $rev = (::cmt_metadata($c))[1];
2329 if (defined $rev) {
2330 ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
2331 return ($rev, $c);
2332 }
2333 }
2334 my $map_path = $self->map_path;
2335 unless (-e $map_path) {
2336 ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
2337 return (undef, undef);
2338 }
2339 my ($rev, $commit) = $self->rev_map_max(1);
2340 ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
2341 return ($rev, $commit);
2342 }
2344 sub get_fetch_range {
2345 my ($self, $min, $max) = @_;
2346 $max ||= $self->ra->get_latest_revnum;
2347 $min ||= $self->rev_map_max;
2348 (++$min, $max);
2349 }
2351 sub tmp_config {
2352 my (@args) = @_;
2353 my $old_def_config = "$ENV{GIT_DIR}/svn/config";
2354 my $config = "$ENV{GIT_DIR}/svn/.metadata";
2355 if (! -f $config && -f $old_def_config) {
2356 rename $old_def_config, $config or
2357 die "Failed rename $old_def_config => $config: $!\n";
2358 }
2359 my $old_config = $ENV{GIT_CONFIG};
2360 $ENV{GIT_CONFIG} = $config;
2361 $@ = undef;
2362 my @ret = eval {
2363 unless (-f $config) {
2364 mkfile($config);
2365 open my $fh, '>', $config or
2366 die "Can't open $config: $!\n";
2367 print $fh "; This file is used internally by ",
2368 "git-svn\n" or die
2369 "Couldn't write to $config: $!\n";
2370 print $fh "; You should not have to edit it\n" or
2371 die "Couldn't write to $config: $!\n";
2372 close $fh or die "Couldn't close $config: $!\n";
2373 }
2374 command('config', @args);
2375 };
2376 my $err = $@;
2377 if (defined $old_config) {
2378 $ENV{GIT_CONFIG} = $old_config;
2379 } else {
2380 delete $ENV{GIT_CONFIG};
2381 }
2382 die $err if $err;
2383 wantarray ? @ret : $ret[0];
2384 }
2386 sub tmp_index_do {
2387 my ($self, $sub) = @_;
2388 my $old_index = $ENV{GIT_INDEX_FILE};
2389 $ENV{GIT_INDEX_FILE} = $self->{index};
2390 $@ = undef;
2391 my @ret = eval {
2392 my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
2393 mkpath([$dir]) unless -d $dir;
2394 &$sub;
2395 };
2396 my $err = $@;
2397 if (defined $old_index) {
2398 $ENV{GIT_INDEX_FILE} = $old_index;
2399 } else {
2400 delete $ENV{GIT_INDEX_FILE};
2401 }
2402 die $err if $err;
2403 wantarray ? @ret : $ret[0];
2404 }
2406 sub assert_index_clean {
2407 my ($self, $treeish) = @_;
2409 $self->tmp_index_do(sub {
2410 command_noisy('read-tree', $treeish) unless -e $self->{index};
2411 my $x = command_oneline('write-tree');
2412 my ($y) = (command(qw/cat-file commit/, $treeish) =~
2413 /^tree ($::sha1)/mo);
2414 return if $y eq $x;
2416 warn "Index mismatch: $y != $x\nrereading $treeish\n";
2417 unlink $self->{index} or die "unlink $self->{index}: $!\n";
2418 command_noisy('read-tree', $treeish);
2419 $x = command_oneline('write-tree');
2420 if ($y ne $x) {
2421 ::fatal "trees ($treeish) $y != $x\n",
2422 "Something is seriously wrong...";
2423 }
2424 });
2425 }
2427 sub get_commit_parents {
2428 my ($self, $log_entry) = @_;
2429 my (%seen, @ret, @tmp);
2430 # legacy support for 'set-tree'; this is only used by set_tree_cb:
2431 if (my $ip = $self->{inject_parents}) {
2432 if (my $commit = delete $ip->{$log_entry->{revision}}) {
2433 push @tmp, $commit;
2434 }
2435 }
2436 if (my $cur = ::verify_ref($self->refname.'^0')) {
2437 push @tmp, $cur;
2438 }
2439 if (my $ipd = $self->{inject_parents_dcommit}) {
2440 if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2441 push @tmp, @$commit;
2442 }
2443 }
2444 push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2445 while (my $p = shift @tmp) {
2446 next if $seen{$p};
2447 $seen{$p} = 1;
2448 push @ret, $p;
2449 # MAXPARENT is defined to 16 in commit-tree.c:
2450 last if @ret >= 16;
2451 }
2452 if (@tmp) {
2453 die "r$log_entry->{revision}: No room for parents:\n\t",
2454 join("\n\t", @tmp), "\n";
2455 }
2456 @ret;
2457 }
2459 sub rewrite_root {
2460 my ($self) = @_;
2461 return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2462 my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2463 my $rwr = eval { command_oneline(qw/config --get/, $k) };
2464 if ($rwr) {
2465 $rwr =~ s#/+$##;
2466 if ($rwr !~ m#^[a-z\+]+://#) {
2467 die "$rwr is not a valid URL (key: $k)\n";
2468 }
2469 }
2470 $self->{-rewrite_root} = $rwr;
2471 }
2473 sub metadata_url {
2474 my ($self) = @_;
2475 ($self->rewrite_root || $self->{url}) .
2476 (length $self->{path} ? '/' . $self->{path} : '');
2477 }
2479 sub full_url {
2480 my ($self) = @_;
2481 $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2482 }
2485 sub set_commit_header_env {
2486 my ($log_entry) = @_;
2487 my %env;
2488 foreach my $ned (qw/NAME EMAIL DATE/) {
2489 foreach my $ac (qw/AUTHOR COMMITTER/) {
2490 $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2491 }
2492 }
2494 $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2495 $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2496 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2498 $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2499 ? $log_entry->{commit_name}
2500 : $log_entry->{name};
2501 $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2502 ? $log_entry->{commit_email}
2503 : $log_entry->{email};
2504 \%env;
2505 }
2507 sub restore_commit_header_env {
2508 my ($env) = @_;
2509 foreach my $ned (qw/NAME EMAIL DATE/) {
2510 foreach my $ac (qw/AUTHOR COMMITTER/) {
2511 my $k = "GIT_${ac}_${ned}";
2512 if (defined $env->{$k}) {
2513 $ENV{$k} = $env->{$k};
2514 } else {
2515 delete $ENV{$k};
2516 }
2517 }
2518 }
2519 }
2521 sub gc {
2522 command_noisy('gc', '--auto');
2523 };
2525 sub do_git_commit {
2526 my ($self, $log_entry) = @_;
2527 my $lr = $self->last_rev;
2528 if (defined $lr && $lr >= $log_entry->{revision}) {
2529 die "Last fetched revision of ", $self->refname,
2530 " was r$lr, but we are about to fetch: ",
2531 "r$log_entry->{revision}!\n";
2532 }
2533 if (my $c = $self->rev_map_get($log_entry->{revision})) {
2534 croak "$log_entry->{revision} = $c already exists! ",
2535 "Why are we refetching it?\n";
2536 }
2537 my $old_env = set_commit_header_env($log_entry);
2538 my $tree = $log_entry->{tree};
2539 if (!defined $tree) {
2540 $tree = $self->tmp_index_do(sub {
2541 command_oneline('write-tree') });
2542 }
2543 die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2545 my @exec = ('git', 'commit-tree', $tree);
2546 foreach ($self->get_commit_parents($log_entry)) {
2547 push @exec, '-p', $_;
2548 }
2549 defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2550 or croak $!;
2551 binmode $msg_fh;
2553 # we always get UTF-8 from SVN, but we may want our commits in
2554 # a different encoding.
2555 if (my $enc = Git::config('i18n.commitencoding')) {
2556 require Encode;
2557 Encode::from_to($log_entry->{log}, 'UTF-8', $enc);
2558 }
2559 print $msg_fh $log_entry->{log} or croak $!;
2560 restore_commit_header_env($old_env);
2561 unless ($self->no_metadata) {
2562 print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2563 or croak $!;
2564 }
2565 $msg_fh->flush == 0 or croak $!;
2566 close $msg_fh or croak $!;
2567 chomp(my $commit = do { local $/; <$out_fh> });
2568 close $out_fh or croak $!;
2569 waitpid $pid, 0;
2570 croak $? if $?;
2571 if ($commit !~ /^$::sha1$/o) {
2572 die "Failed to commit, invalid sha1: $commit\n";
2573 }
2575 $self->rev_map_set($log_entry->{revision}, $commit, 1);
2577 $self->{last_rev} = $log_entry->{revision};
2578 $self->{last_commit} = $commit;
2579 print "r$log_entry->{revision}" unless $::_q > 1;
2580 if (defined $log_entry->{svm_revision}) {
2581 print " (\@$log_entry->{svm_revision})" unless $::_q > 1;
2582 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2583 0, $self->svm_uuid);
2584 }
2585 print " = $commit ($self->{ref_id})\n" unless $::_q > 1;
2586 if (--$_gc_nr == 0) {
2587 $_gc_nr = $_gc_period;
2588 gc();
2589 }
2590 return $commit;
2591 }
2593 sub match_paths {
2594 my ($self, $paths, $r) = @_;
2595 return 1 if $self->{path} eq '';
2596 if (my $path = $paths->{"/$self->{path}"}) {
2597 return ($path->{action} eq 'D') ? 0 : 1;
2598 }
2599 $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
2600 if (grep /$self->{path_regex}/, keys %$paths) {
2601 return 1;
2602 }
2603 my $c = '';
2604 foreach (split m#/#, $self->{path}) {
2605 $c .= "/$_";
2606 next unless ($paths->{$c} &&
2607 ($paths->{$c}->{action} =~ /^[AR]$/));
2608 if ($self->ra->check_path($self->{path}, $r) ==
2609 $SVN::Node::dir) {
2610 return 1;
2611 }
2612 }
2613 return 0;
2614 }
2616 sub find_parent_branch {
2617 my ($self, $paths, $rev) = @_;
2618 return undef unless $self->follow_parent;
2619 unless (defined $paths) {
2620 my $err_handler = $SVN::Error::handler;
2621 $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2622 $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1,
2623 sub { $paths = $_[0] });
2624 $SVN::Error::handler = $err_handler;
2625 }
2626 return undef unless defined $paths;
2628 # look for a parent from another branch:
2629 my @b_path_components = split m#/#, $self->{path};
2630 my @a_path_components;
2631 my $i;
2632 while (@b_path_components) {
2633 $i = $paths->{'/'.join('/', @b_path_components)};
2634 last if $i && defined $i->{copyfrom_path};
2635 unshift(@a_path_components, pop(@b_path_components));
2636 }
2637 return undef unless defined $i && defined $i->{copyfrom_path};
2638 my $branch_from = $i->{copyfrom_path};
2639 if (@a_path_components) {
2640 print STDERR "branch_from: $branch_from => ";
2641 $branch_from .= '/'.join('/', @a_path_components);
2642 print STDERR $branch_from, "\n";
2643 }
2644 my $r = $i->{copyfrom_rev};
2645 my $repos_root = $self->ra->{repos_root};
2646 my $url = $self->ra->{url};
2647 my $new_url = $url . $branch_from;
2648 print STDERR "Found possible branch point: ",
2649 "$new_url => ", $self->full_url, ", $r\n"
2650 unless $::_q > 1;
2651 $branch_from =~ s#^/##;
2652 my $gs = $self->other_gs($new_url, $url,
2653 $branch_from, $r, $self->{ref_id});
2654 my ($r0, $parent) = $gs->find_rev_before($r, 1);
2655 {
2656 my ($base, $head);
2657 if (!defined $r0 || !defined $parent) {
2658 ($base, $head) = parse_revision_argument(0, $r);
2659 } else {
2660 if ($r0 < $r) {
2661 $gs->ra->get_log([$gs->{path}], $r0 + 1, $r, 1,
2662 0, 1, sub { $base = $_[1] - 1 });
2663 }
2664 }
2665 if (defined $base && $base <= $r) {
2666 $gs->fetch($base, $r);
2667 }
2668 ($r0, $parent) = $gs->find_rev_before($r, 1);
2669 }
2670 if (defined $r0 && defined $parent) {
2671 print STDERR "Found branch parent: ($self->{ref_id}) $parent\n"
2672 unless $::_q > 1;
2673 my $ed;
2674 if ($self->ra->can_do_switch) {
2675 $self->assert_index_clean($parent);
2676 print STDERR "Following parent with do_switch\n"
2677 unless $::_q > 1;
2678 # do_switch works with svn/trunk >= r22312, but that
2679 # is not included with SVN 1.4.3 (the latest version
2680 # at the moment), so we can't rely on it
2681 $self->{last_rev} = $r0;
2682 $self->{last_commit} = $parent;
2683 $ed = SVN::Git::Fetcher->new($self, $gs->{path});
2684 $gs->ra->gs_do_switch($r0, $rev, $gs,
2685 $self->full_url, $ed)
2686 or die "SVN connection failed somewhere...\n";
2687 } elsif ($self->ra->trees_match($new_url, $r0,
2688 $self->full_url, $rev)) {
2689 print STDERR "Trees match:\n",
2690 " $new_url\@$r0\n",
2691 " ${\$self->full_url}\@$rev\n",
2692 "Following parent with no changes\n"
2693 unless $::_q > 1;
2694 $self->tmp_index_do(sub {
2695 command_noisy('read-tree', $parent);
2696 });
2697 $self->{last_commit} = $parent;
2698 } else {
2699 print STDERR "Following parent with do_update\n"
2700 unless $::_q > 1;
2701 $ed = SVN::Git::Fetcher->new($self);
2702 $self->ra->gs_do_update($rev, $rev, $self, $ed)
2703 or die "SVN connection failed somewhere...\n";
2704 }
2705 print STDERR "Successfully followed parent\n" unless $::_q > 1;
2706 return $self->make_log_entry($rev, [$parent], $ed);
2707 }
2708 return undef;
2709 }
2711 sub do_fetch {
2712 my ($self, $paths, $rev) = @_;
2713 my $ed;
2714 my ($last_rev, @parents);
2715 if (my $lc = $self->last_commit) {
2716 # we can have a branch that was deleted, then re-added
2717 # under the same name but copied from another path, in
2718 # which case we'll have multiple parents (we don't
2719 # want to break the original ref, nor lose copypath info):
2720 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2721 push @{$log_entry->{parents}}, $lc;
2722 return $log_entry;
2723 }
2724 $ed = SVN::Git::Fetcher->new($self);
2725 $last_rev = $self->{last_rev};
2726 $ed->{c} = $lc;
2727 @parents = ($lc);
2728 } else {
2729 $last_rev = $rev;
2730 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2731 return $log_entry;
2732 }
2733 $ed = SVN::Git::Fetcher->new($self);
2734 }
2735 unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2736 die "SVN connection failed somewhere...\n";
2737 }
2738 $self->make_log_entry($rev, \@parents, $ed);
2739 }
2741 sub mkemptydirs {
2742 my ($self, $r) = @_;
2743 my %empty_dirs = ();
2745 open my $fh, '<', "$self->{dir}/unhandled.log" or return;
2746 binmode $fh or croak "binmode: $!";
2747 while (<$fh>) {
2748 if (defined $r && /^r(\d+)$/) {
2749 last if $1 > $r;
2750 } elsif (/^ \+empty_dir: (.+)$/) {
2751 $empty_dirs{$1} = 1;
2752 } elsif (/^ \-empty_dir: (.+)$/) {
2753 my @d = grep {m[^\Q$1\E(/|$)]} (keys %empty_dirs);
2754 delete @empty_dirs{@d};
2755 }
2756 }
2757 close $fh;
2759 my $strip = qr/\A\Q$self->{path}\E(?:\/|$)/;
2760 foreach my $d (sort keys %empty_dirs) {
2761 $d = uri_decode($d);
2762 $d =~ s/$strip//;
2763 next if -d $d;
2764 if (-e _) {
2765 warn "$d exists but is not a directory\n";
2766 } else {
2767 print "creating empty directory: $d\n";
2768 mkpath([$d]);
2769 }
2770 }
2771 }
2773 sub get_untracked {
2774 my ($self, $ed) = @_;
2775 my @out;
2776 my $h = $ed->{empty};
2777 foreach (sort keys %$h) {
2778 my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2779 push @out, " $act: " . uri_encode($_);
2780 warn "W: $act: $_\n";
2781 }
2782 foreach my $t (qw/dir_prop file_prop/) {
2783 $h = $ed->{$t} or next;
2784 foreach my $path (sort keys %$h) {
2785 my $ppath = $path eq '' ? '.' : $path;
2786 foreach my $prop (sort keys %{$h->{$path}}) {
2787 next if $SKIP_PROP{$prop};
2788 my $v = $h->{$path}->{$prop};
2789 my $t_ppath_prop = "$t: " .
2790 uri_encode($ppath) . ' ' .
2791 uri_encode($prop);
2792 if (defined $v) {
2793 push @out, " +$t_ppath_prop " .
2794 uri_encode($v);
2795 } else {
2796 push @out, " -$t_ppath_prop";
2797 }
2798 }
2799 }
2800 }
2801 foreach my $t (qw/absent_file absent_directory/) {
2802 $h = $ed->{$t} or next;
2803 foreach my $parent (sort keys %$h) {
2804 foreach my $path (sort @{$h->{$parent}}) {
2805 push @out, " $t: " .
2806 uri_encode("$parent/$path");
2807 warn "W: $t: $parent/$path ",
2808 "Insufficient permissions?\n";
2809 }
2810 }
2811 }
2812 \@out;
2813 }
2815 # parse_svn_date(DATE)
2816 # --------------------
2817 # Given a date (in UTC) from Subversion, return a string in the format
2818 # "<TZ Offset> <local date/time>" that Git will use.
2819 #
2820 # By default the parsed date will be in UTC; if $Git::SVN::_localtime
2821 # is true we'll convert it to the local timezone instead.
2822 sub parse_svn_date {
2823 my $date = shift || return '+0000 1970-01-01 00:00:00';
2824 my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2825 (\d\d)\:(\d\d)\:(\d\d)\.\d*Z$/x) or
2826 croak "Unable to parse date: $date\n";
2827 my $parsed_date; # Set next.
2829 if ($Git::SVN::_localtime) {
2830 # Translate the Subversion datetime to an epoch time.
2831 # Begin by switching ourselves to $date's timezone, UTC.
2832 my $old_env_TZ = $ENV{TZ};
2833 $ENV{TZ} = 'UTC';
2835 my $epoch_in_UTC =
2836 POSIX::strftime('%s', $S, $M, $H, $d, $m - 1, $Y - 1900);
2838 # Determine our local timezone (including DST) at the
2839 # time of $epoch_in_UTC. $Git::SVN::Log::TZ stored the
2840 # value of TZ, if any, at the time we were run.
2841 if (defined $Git::SVN::Log::TZ) {
2842 $ENV{TZ} = $Git::SVN::Log::TZ;
2843 } else {
2844 delete $ENV{TZ};
2845 }
2847 my $our_TZ =
2848 POSIX::strftime('%Z', $S, $M, $H, $d, $m - 1, $Y - 1900);
2850 # This converts $epoch_in_UTC into our local timezone.
2851 my ($sec, $min, $hour, $mday, $mon, $year,
2852 $wday, $yday, $isdst) = localtime($epoch_in_UTC);
2854 $parsed_date = sprintf('%s %04d-%02d-%02d %02d:%02d:%02d',
2855 $our_TZ, $year + 1900, $mon + 1,
2856 $mday, $hour, $min, $sec);
2858 # Reset us to the timezone in effect when we entered
2859 # this routine.
2860 if (defined $old_env_TZ) {
2861 $ENV{TZ} = $old_env_TZ;
2862 } else {
2863 delete $ENV{TZ};
2864 }
2865 } else {
2866 $parsed_date = "+0000 $Y-$m-$d $H:$M:$S";
2867 }
2869 return $parsed_date;
2870 }
2872 sub other_gs {
2873 my ($self, $new_url, $url,
2874 $branch_from, $r, $old_ref_id) = @_;
2875 my $gs = Git::SVN->find_by_url($new_url, $url, $branch_from);
2876 unless ($gs) {
2877 my $ref_id = $old_ref_id;
2878 $ref_id =~ s/\@\d+$//;
2879 $ref_id .= "\@$r";
2880 # just grow a tail if we're not unique enough :x
2881 $ref_id .= '-' while find_ref($ref_id);
2882 print STDERR "Initializing parent: $ref_id\n" unless $::_q > 1;
2883 my ($u, $p, $repo_id) = ($new_url, '', $ref_id);
2884 if ($u =~ s#^\Q$url\E(/|$)##) {
2885 $p = $u;
2886 $u = $url;
2887 $repo_id = $self->{repo_id};
2888 }
2889 $gs = Git::SVN->init($u, $p, $repo_id, $ref_id, 1);
2890 }
2891 $gs
2892 }
2894 sub call_authors_prog {
2895 my ($orig_author) = @_;
2896 $orig_author = command_oneline('rev-parse', '--sq-quote', $orig_author);
2897 my $author = `$::_authors_prog $orig_author`;
2898 if ($? != 0) {
2899 die "$::_authors_prog failed with exit code $?\n"
2900 }
2901 if ($author =~ /^\s*(.+?)\s*<(.*)>\s*$/) {
2902 my ($name, $email) = ($1, $2);
2903 $email = undef if length $2 == 0;
2904 return [$name, $email];
2905 } else {
2906 die "Author: $orig_author: $::_authors_prog returned "
2907 . "invalid author format: $author\n";
2908 }
2909 }
2911 sub check_author {
2912 my ($author) = @_;
2913 if (!defined $author || length $author == 0) {
2914 $author = '(no author)';
2915 }
2916 if (!defined $::users{$author}) {
2917 if (defined $::_authors_prog) {
2918 $::users{$author} = call_authors_prog($author);
2919 } elsif (defined $::_authors) {
2920 die "Author: $author not defined in $::_authors file\n";
2921 }
2922 }
2923 $author;
2924 }
2926 sub find_extra_svk_parents {
2927 my ($self, $ed, $tickets, $parents) = @_;
2928 # aha! svk:merge property changed...
2929 my @tickets = split "\n", $tickets;
2930 my @known_parents;
2931 for my $ticket ( @tickets ) {
2932 my ($uuid, $path, $rev) = split /:/, $ticket;
2933 if ( $uuid eq $self->ra_uuid ) {
2934 my $url = $self->rewrite_root || $self->{url};
2935 my $repos_root = $url;
2936 my $branch_from = $path;
2937 $branch_from =~ s{^/}{};
2938 my $gs = $self->other_gs($repos_root."/".$branch_from,
2939 $url,
2940 $branch_from,
2941 $rev,
2942 $self->{ref_id});
2943 if ( my $commit = $gs->rev_map_get($rev, $uuid) ) {
2944 # wahey! we found it, but it might be
2945 # an old one (!)
2946 push @known_parents, [ $rev, $commit ];
2947 }
2948 }
2949 }
2950 # Ordering matters; highest-numbered commit merge tickets
2951 # first, as they may account for later merge ticket additions
2952 # or changes.
2953 @known_parents = map {$_->[1]} sort {$b->[0] <=> $a->[0]} @known_parents;
2954 for my $parent ( @known_parents ) {
2955 my @cmd = ('rev-list', $parent, map { "^$_" } @$parents );
2956 my ($msg_fh, $ctx) = command_output_pipe(@cmd);
2957 my $new;
2958 while ( <$msg_fh> ) {
2959 $new=1;last;
2960 }
2961 command_close_pipe($msg_fh, $ctx);
2962 if ( $new ) {
2963 print STDERR
2964 "Found merge parent (svk:merge ticket): $parent\n";
2965 push @$parents, $parent;
2966 }
2967 }
2968 }
2970 # note: this function should only be called if the various dirprops
2971 # have actually changed
2972 sub find_extra_svn_parents {
2973 my ($self, $ed, $mergeinfo, $parents) = @_;
2974 # aha! svk:merge property changed...
2976 # We first search for merged tips which are not in our
2977 # history. Then, we figure out which git revisions are in
2978 # that tip, but not this revision. If all of those revisions
2979 # are now marked as merge, we can add the tip as a parent.
2980 my @merges = split "\n", $mergeinfo;
2981 my @merge_tips;
2982 my @merged_commit_ranges;
2983 my $url = $self->rewrite_root || $self->{url};
2984 for my $merge ( @merges ) {
2985 my ($source, $revs) = split ":", $merge;
2986 my $path = $source;
2987 $path =~ s{^/}{};
2988 my $gs = Git::SVN->find_by_url($url.$source, $url, $path);
2989 if ( !$gs ) {
2990 warn "Couldn't find revmap for $url$source\n";
2991 next;
2992 }
2993 my @ranges = split ",", $revs;
2994 my ($tip, $tip_commit);
2995 # find the tip
2996 for my $range ( @ranges ) {
2997 my ($bottom, $top) = split "-", $range;
2998 $top ||= $bottom;
2999 my $bottom_commit =
3000 $gs->rev_map_get($bottom, $self->ra_uuid) ||
3001 $gs->rev_map_get($bottom+1, $self->ra_uuid);
3002 my $top_commit;
3003 for (; !$top_commit && $top >= $bottom; --$top) {
3004 $top_commit =
3005 $gs->rev_map_get($top, $self->ra_uuid);
3006 }
3008 unless ($top_commit and $bottom_commit) {
3009 warn "W:unknown path/rev in svn:mergeinfo "
3010 ."dirprop: $source:$range\n";
3011 next;
3012 }
3014 push @merged_commit_ranges,
3015 "$bottom_commit..$top_commit";
3017 if ( !defined $tip or $top > $tip ) {
3018 $tip = $top;
3019 $tip_commit = $top_commit;
3020 }
3021 }
3022 unless (!$tip_commit or
3023 grep { $_ eq $tip_commit } @$parents ) {
3024 push @merge_tips, $tip_commit;
3025 } else {
3026 push @merge_tips, undef;
3027 }
3028 }
3029 for my $merge_tip ( @merge_tips ) {
3030 my $spec = shift @merges;
3031 next unless $merge_tip;
3032 my @cmd = ('rev-list', "-1", $merge_tip,
3033 "--not", @$parents );
3034 my ($msg_fh, $ctx) = command_output_pipe(@cmd);
3035 my $new;
3036 while ( <$msg_fh> ) {
3037 $new=1;last;
3038 }
3039 command_close_pipe($msg_fh, $ctx);
3040 if ( $new ) {
3041 push @cmd, @merged_commit_ranges;
3042 my ($msg_fh, $ctx) = command_output_pipe(@cmd);
3043 my $unmerged;
3044 while ( <$msg_fh> ) {
3045 $unmerged=1;last;
3046 }
3047 command_close_pipe($msg_fh, $ctx);
3048 if ( $unmerged ) {
3049 warn "W:svn cherry-pick ignored ($spec)\n";
3050 } else {
3051 warn
3052 "Found merge parent (svn:mergeinfo prop): ",
3053 $merge_tip, "\n";
3054 push @$parents, $merge_tip;
3055 }
3056 }
3057 }
3058 }
3060 sub make_log_entry {
3061 my ($self, $rev, $parents, $ed) = @_;
3062 my $untracked = $self->get_untracked($ed);
3064 my @parents = @$parents;
3065 my $ps = $ed->{path_strip} || "";
3066 for my $path ( grep { m/$ps/ } %{$ed->{dir_prop}} ) {
3067 my $props = $ed->{dir_prop}{$path};
3068 if ( $props->{"svk:merge"} ) {
3069 $self->find_extra_svk_parents
3070 ($ed, $props->{"svk:merge"}, \@parents);
3071 }
3072 if ( $props->{"svn:mergeinfo"} ) {
3073 $self->find_extra_svn_parents
3074 ($ed,
3075 $props->{"svn:mergeinfo"},
3076 \@parents);
3077 }
3078 }
3080 open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
3081 print $un "r$rev\n" or croak $!;
3082 print $un $_, "\n" foreach @$untracked;
3083 my %log_entry = ( parents => \@parents, revision => $rev,
3084 log => '');
3086 my $headrev;
3087 my $logged = delete $self->{logged_rev_props};
3088 if (!$logged || $self->{-want_revprops}) {
3089 my $rp = $self->ra->rev_proplist($rev);
3090 foreach (sort keys %$rp) {
3091 my $v = $rp->{$_};
3092 if (/^svn:(author|date|log)$/) {
3093 $log_entry{$1} = $v;
3094 } elsif ($_ eq 'svm:headrev') {
3095 $headrev = $v;
3096 } else {
3097 print $un " rev_prop: ", uri_encode($_), ' ',
3098 uri_encode($v), "\n";
3099 }
3100 }
3101 } else {
3102 map { $log_entry{$_} = $logged->{$_} } keys %$logged;
3103 }
3104 close $un or croak $!;
3106 $log_entry{date} = parse_svn_date($log_entry{date});
3107 $log_entry{log} .= "\n";
3108 my $author = $log_entry{author} = check_author($log_entry{author});
3109 my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
3110 : ($author, undef);
3112 my ($commit_name, $commit_email) = ($name, $email);
3113 if ($_use_log_author) {
3114 my $name_field;
3115 if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
3116 $name_field = $1;
3117 } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
3118 $name_field = $1;
3119 }
3120 if (!defined $name_field) {
3121 if (!defined $email) {
3122 $email = $name;
3123 }
3124 } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
3125 ($name, $email) = ($1, $2);
3126 } elsif ($name_field =~ /(.*)@/) {
3127 ($name, $email) = ($1, $name_field);
3128 } else {
3129 ($name, $email) = ($name_field, $name_field);
3130 }
3131 }
3132 if (defined $headrev && $self->use_svm_props) {
3133 if ($self->rewrite_root) {
3134 die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
3135 "options set!\n";
3136 }
3137 my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$}i;
3138 # we don't want "SVM: initializing mirror for junk" ...
3139 return undef if $r == 0;
3140 my $svm = $self->svm;
3141 if ($uuid ne $svm->{uuid}) {
3142 die "UUID mismatch on SVM path:\n",
3143 "expected: $svm->{uuid}\n",
3144 " got: $uuid\n";
3145 }
3146 my $full_url = $self->full_url;
3147 $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
3148 die "Failed to replace '$svm->{replace}' with ",
3149 "'$svm->{source}' in $full_url\n";
3150 # throw away username for storing in records
3151 remove_username($full_url);
3152 $log_entry{metadata} = "$full_url\@$r $uuid";
3153 $log_entry{svm_revision} = $r;
3154 $email ||= "$author\@$uuid";
3155 $commit_email ||= "$author\@$uuid";
3156 } elsif ($self->use_svnsync_props) {
3157 my $full_url = $self->svnsync->{url};
3158 $full_url .= "/$self->{path}" if length $self->{path};
3159 remove_username($full_url);
3160 my $uuid = $self->svnsync->{uuid};
3161 $log_entry{metadata} = "$full_url\@$rev $uuid";
3162 $email ||= "$author\@$uuid";
3163 $commit_email ||= "$author\@$uuid";
3164 } else {
3165 my $url = $self->metadata_url;
3166 remove_username($url);
3167 $log_entry{metadata} = "$url\@$rev " .
3168 $self->ra->get_uuid;
3169 $email ||= "$author\@" . $self->ra->get_uuid;
3170 $commit_email ||= "$author\@" . $self->ra->get_uuid;
3171 }
3172 $log_entry{name} = $name;
3173 $log_entry{email} = $email;
3174 $log_entry{commit_name} = $commit_name;
3175 $log_entry{commit_email} = $commit_email;
3176 \%log_entry;
3177 }
3179 sub fetch {
3180 my ($self, $min_rev, $max_rev, @parents) = @_;
3181 my ($last_rev, $last_commit) = $self->last_rev_commit;
3182 my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
3183 $self->ra->gs_fetch_loop_common($base, $head, [$self]);
3184 }
3186 sub set_tree_cb {
3187 my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
3188 $self->{inject_parents} = { $rev => $tree };
3189 $self->fetch(undef, undef);
3190 }
3192 sub set_tree {
3193 my ($self, $tree) = (shift, shift);
3194 my $log_entry = ::get_commit_entry($tree);
3195 unless ($self->{last_rev}) {
3196 ::fatal("Must have an existing revision to commit");
3197 }
3198 my %ed_opts = ( r => $self->{last_rev},
3199 log => $log_entry->{log},
3200 ra => $self->ra,
3201 tree_a => $self->{last_commit},
3202 tree_b => $tree,
3203 editor_cb => sub {
3204 $self->set_tree_cb($log_entry, $tree, @_) },
3205 svn_path => $self->{path} );
3206 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
3207 print "No changes\nr$self->{last_rev} = $tree\n";
3208 }
3209 }
3211 sub rebuild_from_rev_db {
3212 my ($self, $path) = @_;
3213 my $r = -1;
3214 open my $fh, '<', $path or croak "open: $!";
3215 binmode $fh or croak "binmode: $!";
3216 while (<$fh>) {
3217 length($_) == 41 or croak "inconsistent size in ($_) != 41";
3218 chomp($_);
3219 ++$r;
3220 next if $_ eq ('0' x 40);
3221 $self->rev_map_set($r, $_);
3222 print "r$r = $_\n";
3223 }
3224 close $fh or croak "close: $!";
3225 unlink $path or croak "unlink: $!";
3226 }
3228 sub rebuild {
3229 my ($self) = @_;
3230 my $map_path = $self->map_path;
3231 my $partial = (-e $map_path && ! -z $map_path);
3232 return unless ::verify_ref($self->refname.'^0');
3233 if (!$partial && ($self->use_svm_props || $self->no_metadata)) {
3234 my $rev_db = $self->rev_db_path;
3235 $self->rebuild_from_rev_db($rev_db);
3236 if ($self->use_svm_props) {
3237 my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
3238 $self->rebuild_from_rev_db($svm_rev_db);
3239 }
3240 $self->unlink_rev_db_symlink;
3241 return;
3242 }
3243 print "Rebuilding $map_path ...\n" if (!$partial);
3244 my ($base_rev, $head) = ($partial ? $self->rev_map_max_norebuild(1) :
3245 (undef, undef));
3246 my ($log, $ctx) =
3247 command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
3248 ($head ? "$head.." : "") . $self->refname,
3249 '--');
3250 my $metadata_url = $self->metadata_url;
3251 remove_username($metadata_url);
3252 my $svn_uuid = $self->ra_uuid;
3253 my $c;
3254 while (<$log>) {
3255 if ( m{^commit ($::sha1)$} ) {
3256 $c = $1;
3257 next;
3258 }
3259 next unless s{^\s*(git-svn-id:)}{$1};
3260 my ($url, $rev, $uuid) = ::extract_metadata($_);
3261 remove_username($url);
3263 # ignore merges (from set-tree)
3264 next if (!defined $rev || !$uuid);
3266 # if we merged or otherwise started elsewhere, this is
3267 # how we break out of it
3268 if (($uuid ne $svn_uuid) ||
3269 ($metadata_url && $url && ($url ne $metadata_url))) {
3270 next;
3271 }
3272 if ($partial && $head) {
3273 print "Partial-rebuilding $map_path ...\n";
3274 print "Currently at $base_rev = $head\n";
3275 $head = undef;
3276 }
3278 $self->rev_map_set($rev, $c);
3279 print "r$rev = $c\n";
3280 }
3281 command_close_pipe($log, $ctx);
3282 print "Done rebuilding $map_path\n" if (!$partial || !$head);
3283 my $rev_db_path = $self->rev_db_path;
3284 if (-f $self->rev_db_path) {
3285 unlink $self->rev_db_path or croak "unlink: $!";
3286 }
3287 $self->unlink_rev_db_symlink;
3288 }
3290 # rev_map:
3291 # Tie::File seems to be prone to offset errors if revisions get sparse,
3292 # it's not that fast, either. Tie::File is also not in Perl 5.6. So
3293 # one of my favorite modules is out :< Next up would be one of the DBM
3294 # modules, but I'm not sure which is most portable...
3295 #
3296 # This is the replacement for the rev_db format, which was too big
3297 # and inefficient for large repositories with a lot of sparse history
3298 # (mainly tags)
3299 #
3300 # The format is this:
3301 # - 24 bytes for every record,
3302 # * 4 bytes for the integer representing an SVN revision number
3303 # * 20 bytes representing the sha1 of a git commit
3304 # - No empty padding records like the old format
3305 # (except the last record, which can be overwritten)
3306 # - new records are written append-only since SVN revision numbers
3307 # increase monotonically
3308 # - lookups on SVN revision number are done via a binary search
3309 # - Piping the file to xxd -c24 is a good way of dumping it for
3310 # viewing or editing (piped back through xxd -r), should the need
3311 # ever arise.
3312 # - The last record can be padding revision with an all-zero sha1
3313 # This is used to optimize fetch performance when using multiple
3314 # "fetch" directives in .git/config
3315 #
3316 # These files are disposable unless noMetadata or useSvmProps is set
3318 sub _rev_map_set {
3319 my ($fh, $rev, $commit) = @_;
3321 binmode $fh or croak "binmode: $!";
3322 my $size = (stat($fh))[7];
3323 ($size % 24) == 0 or croak "inconsistent size: $size";
3325 my $wr_offset = 0;
3326 if ($size > 0) {
3327 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
3328 my $read = sysread($fh, my $buf, 24) or croak "read: $!";
3329 $read == 24 or croak "read only $read bytes (!= 24)";
3330 my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
3331 if ($last_commit eq ('0' x40)) {
3332 if ($size >= 48) {
3333 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
3334 $read = sysread($fh, $buf, 24) or
3335 croak "read: $!";
3336 $read == 24 or
3337 croak "read only $read bytes (!= 24)";
3338 ($last_rev, $last_commit) =
3339 unpack(rev_map_fmt, $buf);
3340 if ($last_commit eq ('0' x40)) {
3341 croak "inconsistent .rev_map\n";
3342 }
3343 }
3344 if ($last_rev >= $rev) {
3345 croak "last_rev is higher!: $last_rev >= $rev";
3346 }
3347 $wr_offset = -24;
3348 }
3349 }
3350 sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
3351 syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
3352 croak "write: $!";
3353 }
3355 sub _rev_map_reset {
3356 my ($fh, $rev, $commit) = @_;
3357 my $c = _rev_map_get($fh, $rev);
3358 $c eq $commit or die "_rev_map_reset(@_) commit $c does not match!\n";
3359 my $offset = sysseek($fh, 0, SEEK_CUR) or croak "seek: $!";
3360 truncate $fh, $offset or croak "truncate: $!";
3361 }
3363 sub mkfile {
3364 my ($path) = @_;
3365 unless (-e $path) {
3366 my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
3367 mkpath([$dir]) unless -d $dir;
3368 open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
3369 close $fh or die "Couldn't close (create) $path: $!\n";
3370 }
3371 }
3373 sub rev_map_set {
3374 my ($self, $rev, $commit, $update_ref, $uuid) = @_;
3375 length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
3376 my $db = $self->map_path($uuid);
3377 my $db_lock = "$db.lock";
3378 my $sig;
3379 $update_ref ||= 0;
3380 if ($update_ref) {
3381 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
3382 $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
3383 }
3384 mkfile($db);
3386 $LOCKFILES{$db_lock} = 1;
3387 my $sync;
3388 # both of these options make our .rev_db file very, very important
3389 # and we can't afford to lose it because rebuild() won't work
3390 if ($self->use_svm_props || $self->no_metadata) {
3391 $sync = 1;
3392 copy($db, $db_lock) or die "rev_map_set(@_): ",
3393 "Failed to copy: ",
3394 "$db => $db_lock ($!)\n";
3395 } else {
3396 rename $db, $db_lock or die "rev_map_set(@_): ",
3397 "Failed to rename: ",
3398 "$db => $db_lock ($!)\n";
3399 }
3401 sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
3402 or croak "Couldn't open $db_lock: $!\n";
3403 $update_ref eq 'reset' ? _rev_map_reset($fh, $rev, $commit) :
3404 _rev_map_set($fh, $rev, $commit);
3405 if ($sync) {
3406 $fh->flush or die "Couldn't flush $db_lock: $!\n";
3407 $fh->sync or die "Couldn't sync $db_lock: $!\n";
3408 }
3409 close $fh or croak $!;
3410 if ($update_ref) {
3411 $_head = $self;
3412 my $note = "";
3413 $note = " ($update_ref)" if ($update_ref !~ /^\d*$/);
3414 command_noisy('update-ref', '-m', "r$rev$note",
3415 $self->refname, $commit);
3416 }
3417 rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
3418 "$db_lock => $db ($!)\n";
3419 delete $LOCKFILES{$db_lock};
3420 if ($update_ref) {
3421 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
3422 $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
3423 kill $sig, $$ if defined $sig;
3424 }
3425 }
3427 # If want_commit, this will return an array of (rev, commit) where
3428 # commit _must_ be a valid commit in the archive.
3429 # Otherwise, it'll return the max revision (whether or not the
3430 # commit is valid or just a 0x40 placeholder).
3431 sub rev_map_max {
3432 my ($self, $want_commit) = @_;
3433 $self->rebuild;
3434 my ($r, $c) = $self->rev_map_max_norebuild($want_commit);
3435 $want_commit ? ($r, $c) : $r;
3436 }
3438 sub rev_map_max_norebuild {
3439 my ($self, $want_commit) = @_;
3440 my $map_path = $self->map_path;
3441 stat $map_path or return $want_commit ? (0, undef) : 0;
3442 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
3443 binmode $fh or croak "binmode: $!";
3444 my $size = (stat($fh))[7];
3445 ($size % 24) == 0 or croak "inconsistent size: $size";
3447 if ($size == 0) {
3448 close $fh or croak "close: $!";
3449 return $want_commit ? (0, undef) : 0;
3450 }
3452 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
3453 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
3454 my ($r, $c) = unpack(rev_map_fmt, $buf);
3455 if ($want_commit && $c eq ('0' x40)) {
3456 if ($size < 48) {
3457 return $want_commit ? (0, undef) : 0;
3458 }
3459 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
3460 sysread($fh, $buf, 24) == 24 or croak "read: $!";
3461 ($r, $c) = unpack(rev_map_fmt, $buf);
3462 if ($c eq ('0'x40)) {
3463 croak "Penultimate record is all-zeroes in $map_path";
3464 }
3465 }
3466 close $fh or croak "close: $!";
3467 $want_commit ? ($r, $c) : $r;
3468 }
3470 sub rev_map_get {
3471 my ($self, $rev, $uuid) = @_;
3472 my $map_path = $self->map_path($uuid);
3473 return undef unless -e $map_path;
3475 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
3476 my $c = _rev_map_get($fh, $rev);
3477 close($fh) or croak "close: $!";
3478 $c
3479 }
3481 sub _rev_map_get {
3482 my ($fh, $rev) = @_;
3484 binmode $fh or croak "binmode: $!";
3485 my $size = (stat($fh))[7];
3486 ($size % 24) == 0 or croak "inconsistent size: $size";
3488 if ($size == 0) {
3489 return undef;
3490 }
3492 my ($l, $u) = (0, $size - 24);
3493 my ($r, $c, $buf);
3495 while ($l <= $u) {
3496 my $i = int(($l/24 + $u/24) / 2) * 24;
3497 sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
3498 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
3499 my ($r, $c) = unpack(rev_map_fmt, $buf);
3501 if ($r < $rev) {
3502 $l = $i + 24;
3503 } elsif ($r > $rev) {
3504 $u = $i - 24;
3505 } else { # $r == $rev
3506 return $c eq ('0' x 40) ? undef : $c;
3507 }
3508 }
3509 undef;
3510 }
3512 # Finds the first svn revision that exists on (if $eq_ok is true) or
3513 # before $rev for the current branch. It will not search any lower
3514 # than $min_rev. Returns the git commit hash and svn revision number
3515 # if found, else (undef, undef).
3516 sub find_rev_before {
3517 my ($self, $rev, $eq_ok, $min_rev) = @_;
3518 --$rev unless $eq_ok;
3519 $min_rev ||= 1;
3520 my $max_rev = $self->rev_map_max;
3521 $rev = $max_rev if ($rev > $max_rev);
3522 while ($rev >= $min_rev) {
3523 if (my $c = $self->rev_map_get($rev)) {
3524 return ($rev, $c);
3525 }
3526 --$rev;
3527 }
3528 return (undef, undef);
3529 }
3531 # Finds the first svn revision that exists on (if $eq_ok is true) or
3532 # after $rev for the current branch. It will not search any higher
3533 # than $max_rev. Returns the git commit hash and svn revision number
3534 # if found, else (undef, undef).
3535 sub find_rev_after {
3536 my ($self, $rev, $eq_ok, $max_rev) = @_;
3537 ++$rev unless $eq_ok;
3538 $max_rev ||= $self->rev_map_max;
3539 while ($rev <= $max_rev) {
3540 if (my $c = $self->rev_map_get($rev)) {
3541 return ($rev, $c);
3542 }
3543 ++$rev;
3544 }
3545 return (undef, undef);
3546 }
3548 sub _new {
3549 my ($class, $repo_id, $ref_id, $path) = @_;
3550 unless (defined $repo_id && length $repo_id) {
3551 $repo_id = $Git::SVN::default_repo_id;
3552 }
3553 unless (defined $ref_id && length $ref_id) {
3554 $_prefix = '' unless defined($_prefix);
3555 $_[2] = $ref_id =
3556 "refs/remotes/$_prefix$Git::SVN::default_ref_id";
3557 }
3558 $_[1] = $repo_id;
3559 my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
3561 # Older repos imported by us used $GIT_DIR/svn/foo instead of
3562 # $GIT_DIR/svn/refs/remotes/foo when tracking refs/remotes/foo
3563 if ($ref_id =~ m{^refs/remotes/(.*)}) {
3564 my $old_dir = "$ENV{GIT_DIR}/svn/$1";
3565 if (-d $old_dir && ! -d $dir) {
3566 $dir = $old_dir;
3567 }
3568 }
3570 $_[3] = $path = '' unless (defined $path);
3571 mkpath([$dir]);
3572 bless {
3573 ref_id => $ref_id, dir => $dir, index => "$dir/index",
3574 path => $path, config => "$ENV{GIT_DIR}/svn/config",
3575 map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
3576 }
3578 # for read-only access of old .rev_db formats
3579 sub unlink_rev_db_symlink {
3580 my ($self) = @_;
3581 my $link = $self->rev_db_path;
3582 $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
3583 if (-l $link) {
3584 unlink $link or croak "unlink: $link failed!";
3585 }
3586 }
3588 sub rev_db_path {
3589 my ($self, $uuid) = @_;
3590 my $db_path = $self->map_path($uuid);
3591 $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
3592 or croak "map_path: $db_path does not contain '/.rev_map.' !";
3593 $db_path;
3594 }
3596 # the new replacement for .rev_db
3597 sub map_path {
3598 my ($self, $uuid) = @_;
3599 $uuid ||= $self->ra_uuid;
3600 "$self->{map_root}.$uuid";
3601 }
3603 sub uri_encode {
3604 my ($f) = @_;
3605 $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
3606 $f
3607 }
3609 sub uri_decode {
3610 my ($f) = @_;
3611 $f =~ s#%([0-9a-fA-F]{2})#chr(hex($1))#eg;
3612 $f
3613 }
3615 sub remove_username {
3616 $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
3617 }
3619 package Git::SVN::Prompt;
3620 use strict;
3621 use warnings;
3622 require SVN::Core;
3623 use vars qw/$_no_auth_cache $_username/;
3625 sub simple {
3626 my ($cred, $realm, $default_username, $may_save, $pool) = @_;
3627 $may_save = undef if $_no_auth_cache;
3628 $default_username = $_username if defined $_username;
3629 if (defined $default_username && length $default_username) {
3630 if (defined $realm && length $realm) {
3631 print STDERR "Authentication realm: $realm\n";
3632 STDERR->flush;
3633 }
3634 $cred->username($default_username);
3635 } else {
3636 username($cred, $realm, $may_save, $pool);
3637 }
3638 $cred->password(_read_password("Password for '" .
3639 $cred->username . "': ", $realm));
3640 $cred->may_save($may_save);
3641 $SVN::_Core::SVN_NO_ERROR;
3642 }
3644 sub ssl_server_trust {
3645 my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
3646 $may_save = undef if $_no_auth_cache;
3647 print STDERR "Error validating server certificate for '$realm':\n";
3648 {
3649 no warnings 'once';
3650 # All variables SVN::Auth::SSL::* are used only once,
3651 # so we're shutting up Perl warnings about this.
3652 if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
3653 print STDERR " - The certificate is not issued ",
3654 "by a trusted authority. Use the\n",
3655 " fingerprint to validate ",
3656 "the certificate manually!\n";
3657 }
3658 if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
3659 print STDERR " - The certificate hostname ",
3660 "does not match.\n";
3661 }
3662 if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
3663 print STDERR " - The certificate is not yet valid.\n";
3664 }
3665 if ($failures & $SVN::Auth::SSL::EXPIRED) {
3666 print STDERR " - The certificate has expired.\n";
3667 }
3668 if ($failures & $SVN::Auth::SSL::OTHER) {
3669 print STDERR " - The certificate has ",
3670 "an unknown error.\n";
3671 }
3672 } # no warnings 'once'
3673 printf STDERR
3674 "Certificate information:\n".
3675 " - Hostname: %s\n".
3676 " - Valid: from %s until %s\n".
3677 " - Issuer: %s\n".
3678 " - Fingerprint: %s\n",
3679 map $cert_info->$_, qw(hostname valid_from valid_until
3680 issuer_dname fingerprint);
3681 my $choice;
3682 prompt:
3683 print STDERR $may_save ?
3684 "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
3685 "(R)eject or accept (t)emporarily? ";
3686 STDERR->flush;
3687 $choice = lc(substr(<STDIN> || 'R', 0, 1));
3688 if ($choice =~ /^t$/i) {
3689 $cred->may_save(undef);
3690 } elsif ($choice =~ /^r$/i) {
3691 return -1;
3692 } elsif ($may_save && $choice =~ /^p$/i) {
3693 $cred->may_save($may_save);
3694 } else {
3695 goto prompt;
3696 }
3697 $cred->accepted_failures($failures);
3698 $SVN::_Core::SVN_NO_ERROR;
3699 }
3701 sub ssl_client_cert {
3702 my ($cred, $realm, $may_save, $pool) = @_;
3703 $may_save = undef if $_no_auth_cache;
3704 print STDERR "Client certificate filename: ";
3705 STDERR->flush;
3706 chomp(my $filename = <STDIN>);
3707 $cred->cert_file($filename);
3708 $cred->may_save($may_save);
3709 $SVN::_Core::SVN_NO_ERROR;
3710 }
3712 sub ssl_client_cert_pw {
3713 my ($cred, $realm, $may_save, $pool) = @_;
3714 $may_save = undef if $_no_auth_cache;
3715 $cred->password(_read_password("Password: ", $realm));
3716 $cred->may_save($may_save);
3717 $SVN::_Core::SVN_NO_ERROR;
3718 }
3720 sub username {
3721 my ($cred, $realm, $may_save, $pool) = @_;
3722 $may_save = undef if $_no_auth_cache;
3723 if (defined $realm && length $realm) {
3724 print STDERR "Authentication realm: $realm\n";
3725 }
3726 my $username;
3727 if (defined $_username) {
3728 $username = $_username;
3729 } else {
3730 print STDERR "Username: ";
3731 STDERR->flush;
3732 chomp($username = <STDIN>);
3733 }
3734 $cred->username($username);
3735 $cred->may_save($may_save);
3736 $SVN::_Core::SVN_NO_ERROR;
3737 }
3739 sub _read_password {
3740 my ($prompt, $realm) = @_;
3741 print STDERR $prompt;
3742 STDERR->flush;
3743 require Term::ReadKey;
3744 Term::ReadKey::ReadMode('noecho');
3745 my $password = '';
3746 while (defined(my $key = Term::ReadKey::ReadKey(0))) {
3747 last if $key =~ /[\012\015]/; # \n\r
3748 $password .= $key;
3749 }
3750 Term::ReadKey::ReadMode('restore');
3751 print STDERR "\n";
3752 STDERR->flush;
3753 $password;
3754 }
3756 package SVN::Git::Fetcher;
3757 use vars qw/@ISA/;
3758 use strict;
3759 use warnings;
3760 use Carp qw/croak/;
3761 use File::Temp qw/tempfile/;
3762 use IO::File qw//;
3763 use vars qw/$_ignore_regex/;
3765 # file baton members: path, mode_a, mode_b, pool, fh, blob, base
3766 sub new {
3767 my ($class, $git_svn, $switch_path) = @_;
3768 my $self = SVN::Delta::Editor->new;
3769 bless $self, $class;
3770 if (exists $git_svn->{last_commit}) {
3771 $self->{c} = $git_svn->{last_commit};
3772 $self->{empty_symlinks} =
3773 _mark_empty_symlinks($git_svn, $switch_path);
3774 }
3775 $self->{ignore_regex} = eval { command_oneline('config', '--get',
3776 "svn-remote.$git_svn->{repo_id}.ignore-paths") };
3777 $self->{empty} = {};
3778 $self->{dir_prop} = {};
3779 $self->{file_prop} = {};
3780 $self->{absent_dir} = {};
3781 $self->{absent_file} = {};
3782 $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
3783 $self;
3784 }
3786 # this uses the Ra object, so it must be called before do_{switch,update},
3787 # not inside them (when the Git::SVN::Fetcher object is passed) to
3788 # do_{switch,update}
3789 sub _mark_empty_symlinks {
3790 my ($git_svn, $switch_path) = @_;
3791 my $bool = Git::config_bool('svn.brokenSymlinkWorkaround');
3792 return {} if (!defined($bool)) || (defined($bool) && ! $bool);
3794 my %ret;
3795 my ($rev, $cmt) = $git_svn->last_rev_commit;
3796 return {} unless ($rev && $cmt);
3798 # allow the warning to be printed for each revision we fetch to
3799 # ensure the user sees it. The user can also disable the workaround
3800 # on the repository even while git svn is running and the next
3801 # revision fetched will skip this expensive function.
3802 my $printed_warning;
3803 chomp(my $empty_blob = `git hash-object -t blob --stdin < /dev/null`);
3804 my ($ls, $ctx) = command_output_pipe(qw/ls-tree -r -z/, $cmt);
3805 local $/ = "\0";
3806 my $pfx = defined($switch_path) ? $switch_path : $git_svn->{path};
3807 $pfx .= '/' if length($pfx);
3808 while (<$ls>) {
3809 chomp;
3810 s/\A100644 blob $empty_blob\t//o or next;
3811 unless ($printed_warning) {
3812 print STDERR "Scanning for empty symlinks, ",
3813 "this may take a while if you have ",
3814 "many empty files\n",
3815 "You may disable this with `",
3816 "git config svn.brokenSymlinkWorkaround ",
3817 "false'.\n",
3818 "This may be done in a different ",
3819 "terminal without restarting ",
3820 "git svn\n";
3821 $printed_warning = 1;
3822 }
3823 my $path = $_;
3824 my (undef, $props) =
3825 $git_svn->ra->get_file($pfx.$path, $rev, undef);
3826 if ($props->{'svn:special'}) {
3827 $ret{$path} = 1;
3828 }
3829 }
3830 command_close_pipe($ls, $ctx);
3831 \%ret;
3832 }
3834 # returns true if a given path is inside a ".git" directory
3835 sub in_dot_git {
3836 $_[0] =~ m{(?:^|/)\.git(?:/|$)};
3837 }
3839 # return value: 0 -- don't ignore, 1 -- ignore
3840 sub is_path_ignored {
3841 my ($self, $path) = @_;
3842 return 1 if in_dot_git($path);
3843 return 1 if defined($self->{ignore_regex}) &&
3844 $path =~ m!$self->{ignore_regex}!;
3845 return 0 unless defined($_ignore_regex);
3846 return 1 if $path =~ m!$_ignore_regex!o;
3847 return 0;
3848 }
3850 sub set_path_strip {
3851 my ($self, $path) = @_;
3852 $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
3853 }
3855 sub open_root {
3856 { path => '' };
3857 }
3859 sub open_directory {
3860 my ($self, $path, $pb, $rev) = @_;
3861 { path => $path };
3862 }
3864 sub git_path {
3865 my ($self, $path) = @_;
3866 if ($self->{path_strip}) {
3867 $path =~ s!$self->{path_strip}!! or
3868 die "Failed to strip path '$path' ($self->{path_strip})\n";
3869 }
3870 $path;
3871 }
3873 sub delete_entry {
3874 my ($self, $path, $rev, $pb) = @_;
3875 return undef if $self->is_path_ignored($path);
3877 my $gpath = $self->git_path($path);
3878 return undef if ($gpath eq '');
3880 # remove entire directories.
3881 my ($tree) = (command('ls-tree', '-z', $self->{c}, "./$gpath")
3882 =~ /\A040000 tree ([a-f\d]{40})\t\Q$gpath\E\0/);
3883 if ($tree) {
3884 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3885 -r --name-only -z/,
3886 $tree);
3887 local $/ = "\0";
3888 while (<$ls>) {
3889 chomp;
3890 my $rmpath = "$gpath/$_";
3891 $self->{gii}->remove($rmpath);
3892 print "\tD\t$rmpath\n" unless $::_q;
3893 }
3894 print "\tD\t$gpath/\n" unless $::_q;
3895 command_close_pipe($ls, $ctx);
3896 } else {
3897 $self->{gii}->remove($gpath);
3898 print "\tD\t$gpath\n" unless $::_q;
3899 }
3900 $self->{empty}->{$path} = 0;
3901 undef;
3902 }
3904 sub open_file {
3905 my ($self, $path, $pb, $rev) = @_;
3906 my ($mode, $blob);
3908 goto out if $self->is_path_ignored($path);
3910 my $gpath = $self->git_path($path);
3911 ($mode, $blob) = (command('ls-tree', '-z', $self->{c}, "./$gpath")
3912 =~ /\A(\d{6}) blob ([a-f\d]{40})\t\Q$gpath\E\0/);
3913 unless (defined $mode && defined $blob) {
3914 die "$path was not found in commit $self->{c} (r$rev)\n";
3915 }
3916 if ($mode eq '100644' && $self->{empty_symlinks}->{$path}) {
3917 $mode = '120000';
3918 }
3919 out:
3920 { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3921 pool => SVN::Pool->new, action => 'M' };
3922 }
3924 sub add_file {
3925 my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3926 my $mode;
3928 if (!$self->is_path_ignored($path)) {
3929 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3930 delete $self->{empty}->{$dir};
3931 $mode = '100644';
3932 }
3933 { path => $path, mode_a => $mode, mode_b => $mode,
3934 pool => SVN::Pool->new, action => 'A' };
3935 }
3937 sub add_directory {
3938 my ($self, $path, $cp_path, $cp_rev) = @_;
3939 goto out if $self->is_path_ignored($path);
3940 my $gpath = $self->git_path($path);
3941 if ($gpath eq '') {
3942 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3943 -r --name-only -z/,
3944 $self->{c});
3945 local $/ = "\0";
3946 while (<$ls>) {
3947 chomp;
3948 $self->{gii}->remove($_);
3949 print "\tD\t$_\n" unless $::_q;
3950 }
3951 command_close_pipe($ls, $ctx);
3952 $self->{empty}->{$path} = 0;
3953 }
3954 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3955 delete $self->{empty}->{$dir};
3956 $self->{empty}->{$path} = 1;
3957 out:
3958 { path => $path };
3959 }
3961 sub change_dir_prop {
3962 my ($self, $db, $prop, $value) = @_;
3963 return undef if $self->is_path_ignored($db->{path});
3964 $self->{dir_prop}->{$db->{path}} ||= {};
3965 $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3966 undef;
3967 }
3969 sub absent_directory {
3970 my ($self, $path, $pb) = @_;
3971 return undef if $self->is_path_ignored($path);
3972 $self->{absent_dir}->{$pb->{path}} ||= [];
3973 push @{$self->{absent_dir}->{$pb->{path}}}, $path;
3974 undef;
3975 }
3977 sub absent_file {
3978 my ($self, $path, $pb) = @_;
3979 return undef if $self->is_path_ignored($path);
3980 $self->{absent_file}->{$pb->{path}} ||= [];
3981 push @{$self->{absent_file}->{$pb->{path}}}, $path;
3982 undef;
3983 }
3985 sub change_file_prop {
3986 my ($self, $fb, $prop, $value) = @_;
3987 return undef if $self->is_path_ignored($fb->{path});
3988 if ($prop eq 'svn:executable') {
3989 if ($fb->{mode_b} != 120000) {
3990 $fb->{mode_b} = defined $value ? 100755 : 100644;
3991 }
3992 } elsif ($prop eq 'svn:special') {
3993 $fb->{mode_b} = defined $value ? 120000 : 100644;
3994 } else {
3995 $self->{file_prop}->{$fb->{path}} ||= {};
3996 $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
3997 }
3998 undef;
3999 }
4001 sub apply_textdelta {
4002 my ($self, $fb, $exp) = @_;
4003 return undef if $self->is_path_ignored($fb->{path});
4004 my $fh = $::_repository->temp_acquire('svn_delta');
4005 # $fh gets auto-closed() by SVN::TxDelta::apply(),
4006 # (but $base does not,) so dup() it for reading in close_file
4007 open my $dup, '<&', $fh or croak $!;
4008 my $base = $::_repository->temp_acquire('git_blob');
4010 if ($fb->{blob}) {
4011 my ($base_is_link, $size);
4013 if ($fb->{mode_a} eq '120000' &&
4014 ! $self->{empty_symlinks}->{$fb->{path}}) {
4015 print $base 'link ' or die "print $!\n";
4016 $base_is_link = 1;
4017 }
4018 retry:
4019 $size = $::_repository->cat_blob($fb->{blob}, $base);
4020 die "Failed to read object $fb->{blob}" if ($size < 0);
4022 if (defined $exp) {
4023 seek $base, 0, 0 or croak $!;
4024 my $got = ::md5sum($base);
4025 if ($got ne $exp) {
4026 my $err = "Checksum mismatch: ".
4027 "$fb->{path} $fb->{blob}\n" .
4028 "expected: $exp\n" .
4029 " got: $got\n";
4030 if ($base_is_link) {
4031 warn $err,
4032 "Retrying... (possibly ",
4033 "a bad symlink from SVN)\n";
4034 $::_repository->temp_reset($base);
4035 $base_is_link = 0;
4036 goto retry;
4037 }
4038 die $err;
4039 }
4040 }
4041 }
4042 seek $base, 0, 0 or croak $!;
4043 $fb->{fh} = $fh;
4044 $fb->{base} = $base;
4045 [ SVN::TxDelta::apply($base, $dup, undef, $fb->{path}, $fb->{pool}) ];
4046 }
4048 sub close_file {
4049 my ($self, $fb, $exp) = @_;
4050 return undef if $self->is_path_ignored($fb->{path});
4052 my $hash;
4053 my $path = $self->git_path($fb->{path});
4054 if (my $fh = $fb->{fh}) {
4055 if (defined $exp) {
4056 seek($fh, 0, 0) or croak $!;
4057 my $got = ::md5sum($fh);
4058 if ($got ne $exp) {
4059 die "Checksum mismatch: $path\n",
4060 "expected: $exp\n got: $got\n";
4061 }
4062 }
4063 if ($fb->{mode_b} == 120000) {
4064 sysseek($fh, 0, 0) or croak $!;
4065 my $rd = sysread($fh, my $buf, 5);
4067 if (!defined $rd) {
4068 croak "sysread: $!\n";
4069 } elsif ($rd == 0) {
4070 warn "$path has mode 120000",
4071 " but it points to nothing\n",
4072 "converting to an empty file with mode",
4073 " 100644\n";
4074 $fb->{mode_b} = '100644';
4075 } elsif ($buf ne 'link ') {
4076 warn "$path has mode 120000",
4077 " but is not a link\n";
4078 } else {
4079 my $tmp_fh = $::_repository->temp_acquire(
4080 'svn_hash');
4081 my $res;
4082 while ($res = sysread($fh, my $str, 1024)) {
4083 my $out = syswrite($tmp_fh, $str, $res);
4084 defined($out) && $out == $res
4085 or croak("write ",
4086 Git::temp_path($tmp_fh),
4087 ": $!\n");
4088 }
4089 defined $res or croak $!;
4091 ($fh, $tmp_fh) = ($tmp_fh, $fh);
4092 Git::temp_release($tmp_fh, 1);
4093 }
4094 }
4096 $hash = $::_repository->hash_and_insert_object(
4097 Git::temp_path($fh));
4098 $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
4100 Git::temp_release($fb->{base}, 1);
4101 Git::temp_release($fh, 1);
4102 } else {
4103 $hash = $fb->{blob} or die "no blob information\n";
4104 }
4105 $fb->{pool}->clear;
4106 $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
4107 print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
4108 undef;
4109 }
4111 sub abort_edit {
4112 my $self = shift;
4113 $self->{nr} = $self->{gii}->{nr};
4114 delete $self->{gii};
4115 $self->SUPER::abort_edit(@_);
4116 }
4118 sub close_edit {
4119 my $self = shift;
4120 $self->{git_commit_ok} = 1;
4121 $self->{nr} = $self->{gii}->{nr};
4122 delete $self->{gii};
4123 $self->SUPER::close_edit(@_);
4124 }
4126 package SVN::Git::Editor;
4127 use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
4128 use strict;
4129 use warnings;
4130 use Carp qw/croak/;
4131 use IO::File;
4133 sub new {
4134 my ($class, $opts) = @_;
4135 foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
4136 die "$_ required!\n" unless (defined $opts->{$_});
4137 }
4139 my $pool = SVN::Pool->new;
4140 my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
4141 my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
4142 $opts->{r}, $mods);
4144 # $opts->{ra} functions should not be used after this:
4145 my @ce = $opts->{ra}->get_commit_editor($opts->{log},
4146 $opts->{editor_cb}, $pool);
4147 my $self = SVN::Delta::Editor->new(@ce, $pool);
4148 bless $self, $class;
4149 foreach (qw/svn_path r tree_a tree_b/) {
4150 $self->{$_} = $opts->{$_};
4151 }
4152 $self->{url} = $opts->{ra}->{url};
4153 $self->{mods} = $mods;
4154 $self->{types} = $types;
4155 $self->{pool} = $pool;
4156 $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
4157 $self->{rm} = { };
4158 $self->{path_prefix} = length $self->{svn_path} ?
4159 "$self->{svn_path}/" : '';
4160 $self->{config} = $opts->{config};
4161 return $self;
4162 }
4164 sub generate_diff {
4165 my ($tree_a, $tree_b) = @_;
4166 my @diff_tree = qw(diff-tree -z -r);
4167 if ($_cp_similarity) {
4168 push @diff_tree, "-C$_cp_similarity";
4169 } else {
4170 push @diff_tree, '-C';
4171 }
4172 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
4173 push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
4174 push @diff_tree, $tree_a, $tree_b;
4175 my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
4176 local $/ = "\0";
4177 my $state = 'meta';
4178 my @mods;
4179 while (<$diff_fh>) {
4180 chomp $_; # this gets rid of the trailing "\0"
4181 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
4182 ($::sha1)\s($::sha1)\s
4183 ([MTCRAD])\d*$/xo) {
4184 push @mods, { mode_a => $1, mode_b => $2,
4185 sha1_a => $3, sha1_b => $4,
4186 chg => $5 };
4187 if ($5 =~ /^(?:C|R)$/) {
4188 $state = 'file_a';
4189 } else {
4190 $state = 'file_b';
4191 }
4192 } elsif ($state eq 'file_a') {
4193 my $x = $mods[$#mods] or croak "Empty array\n";
4194 if ($x->{chg} !~ /^(?:C|R)$/) {
4195 croak "Error parsing $_, $x->{chg}\n";
4196 }
4197 $x->{file_a} = $_;
4198 $state = 'file_b';
4199 } elsif ($state eq 'file_b') {
4200 my $x = $mods[$#mods] or croak "Empty array\n";
4201 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
4202 croak "Error parsing $_, $x->{chg}\n";
4203 }
4204 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
4205 croak "Error parsing $_, $x->{chg}\n";
4206 }
4207 $x->{file_b} = $_;
4208 $state = 'meta';
4209 } else {
4210 croak "Error parsing $_\n";
4211 }
4212 }
4213 command_close_pipe($diff_fh, $ctx);
4214 \@mods;
4215 }
4217 sub check_diff_paths {
4218 my ($ra, $pfx, $rev, $mods) = @_;
4219 my %types;
4220 $pfx .= '/' if length $pfx;
4222 sub type_diff_paths {
4223 my ($ra, $types, $path, $rev) = @_;
4224 my @p = split m#/+#, $path;
4225 my $c = shift @p;
4226 unless (defined $types->{$c}) {
4227 $types->{$c} = $ra->check_path($c, $rev);
4228 }
4229 while (@p) {
4230 $c .= '/' . shift @p;
4231 next if defined $types->{$c};
4232 $types->{$c} = $ra->check_path($c, $rev);
4233 }
4234 }
4236 foreach my $m (@$mods) {
4237 foreach my $f (qw/file_a file_b/) {
4238 next unless defined $m->{$f};
4239 my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
4240 if (length $pfx.$dir && ! defined $types{$dir}) {
4241 type_diff_paths($ra, \%types, $pfx.$dir, $rev);
4242 }
4243 }
4244 }
4245 \%types;
4246 }
4248 sub split_path {
4249 return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
4250 }
4252 sub repo_path {
4253 my ($self, $path) = @_;
4254 $self->{path_prefix}.(defined $path ? $path : '');
4255 }
4257 sub url_path {
4258 my ($self, $path) = @_;
4259 if ($self->{url} =~ m#^https?://#) {
4260 $path =~ s!([^~a-zA-Z0-9_./-])!uc sprintf("%%%02x",ord($1))!eg;
4261 }
4262 $self->{url} . '/' . $self->repo_path($path);
4263 }
4265 sub rmdirs {
4266 my ($self) = @_;
4267 my $rm = $self->{rm};
4268 delete $rm->{''}; # we never delete the url we're tracking
4269 return unless %$rm;
4271 foreach (keys %$rm) {
4272 my @d = split m#/#, $_;
4273 my $c = shift @d;
4274 $rm->{$c} = 1;
4275 while (@d) {
4276 $c .= '/' . shift @d;
4277 $rm->{$c} = 1;
4278 }
4279 }
4280 delete $rm->{$self->{svn_path}};
4281 delete $rm->{''}; # we never delete the url we're tracking
4282 return unless %$rm;
4284 my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
4285 $self->{tree_b});
4286 local $/ = "\0";
4287 while (<$fh>) {
4288 chomp;
4289 my @dn = split m#/#, $_;
4290 while (pop @dn) {
4291 delete $rm->{join '/', @dn};
4292 }
4293 unless (%$rm) {
4294 close $fh;
4295 return;
4296 }
4297 }
4298 command_close_pipe($fh, $ctx);
4300 my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
4301 foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
4302 $self->close_directory($bat->{$d}, $p);
4303 my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
4304 print "\tD+\t$d/\n" unless $::_q;
4305 $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
4306 delete $bat->{$d};
4307 }
4308 }
4310 sub open_or_add_dir {
4311 my ($self, $full_path, $baton) = @_;
4312 my $t = $self->{types}->{$full_path};
4313 if (!defined $t) {
4314 die "$full_path not known in r$self->{r} or we have a bug!\n";
4315 }
4316 {
4317 no warnings 'once';
4318 # SVN::Node::none and SVN::Node::file are used only once,
4319 # so we're shutting up Perl's warnings about them.
4320 if ($t == $SVN::Node::none) {
4321 return $self->add_directory($full_path, $baton,
4322 undef, -1, $self->{pool});
4323 } elsif ($t == $SVN::Node::dir) {
4324 return $self->open_directory($full_path, $baton,
4325 $self->{r}, $self->{pool});
4326 } # no warnings 'once'
4327 print STDERR "$full_path already exists in repository at ",
4328 "r$self->{r} and it is not a directory (",
4329 ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
4330 } # no warnings 'once'
4331 exit 1;
4332 }
4334 sub ensure_path {
4335 my ($self, $path) = @_;
4336 my $bat = $self->{bat};
4337 my $repo_path = $self->repo_path($path);
4338 return $bat->{''} unless (length $repo_path);
4339 my @p = split m#/+#, $repo_path;
4340 my $c = shift @p;
4341 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
4342 while (@p) {
4343 my $c0 = $c;
4344 $c .= '/' . shift @p;
4345 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
4346 }
4347 return $bat->{$c};
4348 }
4350 # Subroutine to convert a globbing pattern to a regular expression.
4351 # From perl cookbook.
4352 sub glob2pat {
4353 my $globstr = shift;
4354 my %patmap = ('*' => '.*', '?' => '.', '[' => '[', ']' => ']');
4355 $globstr =~ s{(.)} { $patmap{$1} || "\Q$1" }ge;
4356 return '^' . $globstr . '$';
4357 }
4359 sub check_autoprop {
4360 my ($self, $pattern, $properties, $file, $fbat) = @_;
4361 # Convert the globbing pattern to a regular expression.
4362 my $regex = glob2pat($pattern);
4363 # Check if the pattern matches the file name.
4364 if($file =~ m/($regex)/) {
4365 # Parse the list of properties to set.
4366 my @props = split(/;/, $properties);
4367 foreach my $prop (@props) {
4368 # Parse 'name=value' syntax and set the property.
4369 if ($prop =~ /([^=]+)=(.*)/) {
4370 my ($n,$v) = ($1,$2);
4371 for ($n, $v) {
4372 s/^\s+//; s/\s+$//;
4373 }
4374 $self->change_file_prop($fbat, $n, $v);
4375 }
4376 }
4377 }
4378 }
4380 sub apply_autoprops {
4381 my ($self, $file, $fbat) = @_;
4382 my $conf_t = ${$self->{config}}{'config'};
4383 no warnings 'once';
4384 # Check [miscellany]/enable-auto-props in svn configuration.
4385 if (SVN::_Core::svn_config_get_bool(
4386 $conf_t,
4387 $SVN::_Core::SVN_CONFIG_SECTION_MISCELLANY,
4388 $SVN::_Core::SVN_CONFIG_OPTION_ENABLE_AUTO_PROPS,
4389 0)) {
4390 # Auto-props are enabled. Enumerate them to look for matches.
4391 my $callback = sub {
4392 $self->check_autoprop($_[0], $_[1], $file, $fbat);
4393 };
4394 SVN::_Core::svn_config_enumerate(
4395 $conf_t,
4396 $SVN::_Core::SVN_CONFIG_SECTION_AUTO_PROPS,
4397 $callback);
4398 }
4399 }
4401 sub A {
4402 my ($self, $m) = @_;
4403 my ($dir, $file) = split_path($m->{file_b});
4404 my $pbat = $self->ensure_path($dir);
4405 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
4406 undef, -1);
4407 print "\tA\t$m->{file_b}\n" unless $::_q;
4408 $self->apply_autoprops($file, $fbat);
4409 $self->chg_file($fbat, $m);
4410 $self->close_file($fbat,undef,$self->{pool});
4411 }
4413 sub C {
4414 my ($self, $m) = @_;
4415 my ($dir, $file) = split_path($m->{file_b});
4416 my $pbat = $self->ensure_path($dir);
4417 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
4418 $self->url_path($m->{file_a}), $self->{r});
4419 print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
4420 $self->chg_file($fbat, $m);
4421 $self->close_file($fbat,undef,$self->{pool});
4422 }
4424 sub delete_entry {
4425 my ($self, $path, $pbat) = @_;
4426 my $rpath = $self->repo_path($path);
4427 my ($dir, $file) = split_path($rpath);
4428 $self->{rm}->{$dir} = 1;
4429 $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
4430 }
4432 sub R {
4433 my ($self, $m) = @_;
4434 my ($dir, $file) = split_path($m->{file_b});
4435 my $pbat = $self->ensure_path($dir);
4436 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
4437 $self->url_path($m->{file_a}), $self->{r});
4438 print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
4439 $self->apply_autoprops($file, $fbat);
4440 $self->chg_file($fbat, $m);
4441 $self->close_file($fbat,undef,$self->{pool});
4443 ($dir, $file) = split_path($m->{file_a});
4444 $pbat = $self->ensure_path($dir);
4445 $self->delete_entry($m->{file_a}, $pbat);
4446 }
4448 sub M {
4449 my ($self, $m) = @_;
4450 my ($dir, $file) = split_path($m->{file_b});
4451 my $pbat = $self->ensure_path($dir);
4452 my $fbat = $self->open_file($self->repo_path($m->{file_b}),
4453 $pbat,$self->{r},$self->{pool});
4454 print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
4455 $self->chg_file($fbat, $m);
4456 $self->close_file($fbat,undef,$self->{pool});
4457 }
4459 sub T { shift->M(@_) }
4461 sub change_file_prop {
4462 my ($self, $fbat, $pname, $pval) = @_;
4463 $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
4464 }
4466 sub _chg_file_get_blob ($$$$) {
4467 my ($self, $fbat, $m, $which) = @_;
4468 my $fh = $::_repository->temp_acquire("git_blob_$which");
4469 if ($m->{"mode_$which"} =~ /^120/) {
4470 print $fh 'link ' or croak $!;
4471 $self->change_file_prop($fbat,'svn:special','*');
4472 } elsif ($m->{mode_a} =~ /^120/ && $m->{"mode_$which"} !~ /^120/) {
4473 $self->change_file_prop($fbat,'svn:special',undef);
4474 }
4475 my $blob = $m->{"sha1_$which"};
4476 return ($fh,) if ($blob =~ /^0{40}$/);
4477 my $size = $::_repository->cat_blob($blob, $fh);
4478 croak "Failed to read object $blob" if ($size < 0);
4479 $fh->flush == 0 or croak $!;
4480 seek $fh, 0, 0 or croak $!;
4482 my $exp = ::md5sum($fh);
4483 seek $fh, 0, 0 or croak $!;
4484 return ($fh, $exp);
4485 }
4487 sub chg_file {
4488 my ($self, $fbat, $m) = @_;
4489 if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
4490 $self->change_file_prop($fbat,'svn:executable','*');
4491 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
4492 $self->change_file_prop($fbat,'svn:executable',undef);
4493 }
4494 my ($fh_a, $exp_a) = _chg_file_get_blob $self, $fbat, $m, 'a';
4495 my ($fh_b, $exp_b) = _chg_file_get_blob $self, $fbat, $m, 'b';
4496 my $pool = SVN::Pool->new;
4497 my $atd = $self->apply_textdelta($fbat, $exp_a, $pool);
4498 if (-s $fh_a) {
4499 my $txstream = SVN::TxDelta::new ($fh_a, $fh_b, $pool);
4500 my $res = SVN::TxDelta::send_txstream($txstream, @$atd, $pool);
4501 if (defined $res) {
4502 die "Unexpected result from send_txstream: $res\n",
4503 "(SVN::Core::VERSION: $SVN::Core::VERSION)\n";
4504 }
4505 } else {
4506 my $got = SVN::TxDelta::send_stream($fh_b, @$atd, $pool);
4507 die "Checksum mismatch\nexpected: $exp_b\ngot: $got\n"
4508 if ($got ne $exp_b);
4509 }
4510 Git::temp_release($fh_b, 1);
4511 Git::temp_release($fh_a, 1);
4512 $pool->clear;
4513 }
4515 sub D {
4516 my ($self, $m) = @_;
4517 my ($dir, $file) = split_path($m->{file_b});
4518 my $pbat = $self->ensure_path($dir);
4519 print "\tD\t$m->{file_b}\n" unless $::_q;
4520 $self->delete_entry($m->{file_b}, $pbat);
4521 }
4523 sub close_edit {
4524 my ($self) = @_;
4525 my ($p,$bat) = ($self->{pool}, $self->{bat});
4526 foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
4527 next if $_ eq '';
4528 $self->close_directory($bat->{$_}, $p);
4529 }
4530 $self->close_directory($bat->{''}, $p);
4531 $self->SUPER::close_edit($p);
4532 $p->clear;
4533 }
4535 sub abort_edit {
4536 my ($self) = @_;
4537 $self->SUPER::abort_edit($self->{pool});
4538 }
4540 sub DESTROY {
4541 my $self = shift;
4542 $self->SUPER::DESTROY(@_);
4543 $self->{pool}->clear;
4544 }
4546 # this drives the editor
4547 sub apply_diff {
4548 my ($self) = @_;
4549 my $mods = $self->{mods};
4550 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
4551 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
4552 my $f = $m->{chg};
4553 if (defined $o{$f}) {
4554 $self->$f($m);
4555 } else {
4556 fatal("Invalid change type: $f");
4557 }
4558 }
4559 $self->rmdirs if $_rmdir;
4560 if (@$mods == 0) {
4561 $self->abort_edit;
4562 } else {
4563 $self->close_edit;
4564 }
4565 return scalar @$mods;
4566 }
4568 package Git::SVN::Ra;
4569 use vars qw/@ISA $config_dir $_log_window_size/;
4570 use strict;
4571 use warnings;
4572 my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
4574 BEGIN {
4575 # enforce temporary pool usage for some simple functions
4576 no strict 'refs';
4577 for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root
4578 get_file/) {
4579 my $SUPER = "SUPER::$f";
4580 *$f = sub {
4581 my $self = shift;
4582 my $pool = SVN::Pool->new;
4583 my @ret = $self->$SUPER(@_,$pool);
4584 $pool->clear;
4585 wantarray ? @ret : $ret[0];
4586 };
4587 }
4588 }
4590 sub _auth_providers () {
4591 [
4592 SVN::Client::get_simple_provider(),
4593 SVN::Client::get_ssl_server_trust_file_provider(),
4594 SVN::Client::get_simple_prompt_provider(
4595 \&Git::SVN::Prompt::simple, 2),
4596 SVN::Client::get_ssl_client_cert_file_provider(),
4597 SVN::Client::get_ssl_client_cert_prompt_provider(
4598 \&Git::SVN::Prompt::ssl_client_cert, 2),
4599 SVN::Client::get_ssl_client_cert_pw_file_provider(),
4600 SVN::Client::get_ssl_client_cert_pw_prompt_provider(
4601 \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
4602 SVN::Client::get_username_provider(),
4603 SVN::Client::get_ssl_server_trust_prompt_provider(
4604 \&Git::SVN::Prompt::ssl_server_trust),
4605 SVN::Client::get_username_prompt_provider(
4606 \&Git::SVN::Prompt::username, 2)
4607 ]
4608 }
4610 sub escape_uri_only {
4611 my ($uri) = @_;
4612 my @tmp;
4613 foreach (split m{/}, $uri) {
4614 s/([^~\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
4615 push @tmp, $_;
4616 }
4617 join('/', @tmp);
4618 }
4620 sub escape_url {
4621 my ($url) = @_;
4622 if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
4623 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
4624 $url = "$scheme://$domain$uri";
4625 }
4626 $url;
4627 }
4629 sub new {
4630 my ($class, $url) = @_;
4631 $url =~ s!/+$!!;
4632 return $RA if ($RA && $RA->{url} eq $url);
4634 SVN::_Core::svn_config_ensure($config_dir, undef);
4635 my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
4636 my $config = SVN::Core::config_get_config($config_dir);
4637 $RA = undef;
4638 my $dont_store_passwords = 1;
4639 my $conf_t = ${$config}{'config'};
4640 {
4641 no warnings 'once';
4642 # The usage of $SVN::_Core::SVN_CONFIG_* variables
4643 # produces warnings that variables are used only once.
4644 # I had not found the better way to shut them up, so
4645 # the warnings of type 'once' are disabled in this block.
4646 if (SVN::_Core::svn_config_get_bool($conf_t,
4647 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
4648 $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
4649 1) == 0) {
4650 SVN::_Core::svn_auth_set_parameter($baton,
4651 $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
4652 bless (\$dont_store_passwords, "_p_void"));
4653 }
4654 if (SVN::_Core::svn_config_get_bool($conf_t,
4655 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
4656 $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
4657 1) == 0) {
4658 $Git::SVN::Prompt::_no_auth_cache = 1;
4659 }
4660 } # no warnings 'once'
4661 my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
4662 config => $config,
4663 pool => SVN::Pool->new,
4664 auth_provider_callbacks => $callbacks);
4665 $self->{url} = $url;
4666 $self->{svn_path} = $url;
4667 $self->{repos_root} = $self->get_repos_root;
4668 $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
4669 $self->{cache} = { check_path => { r => 0, data => {} },
4670 get_dir => { r => 0, data => {} } };
4671 $RA = bless $self, $class;
4672 }
4674 sub check_path {
4675 my ($self, $path, $r) = @_;
4676 my $cache = $self->{cache}->{check_path};
4677 if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
4678 return $cache->{data}->{$path};
4679 }
4680 my $pool = SVN::Pool->new;
4681 my $t = $self->SUPER::check_path($path, $r, $pool);
4682 $pool->clear;
4683 if ($r != $cache->{r}) {
4684 %{$cache->{data}} = ();
4685 $cache->{r} = $r;
4686 }
4687 $cache->{data}->{$path} = $t;
4688 }
4690 sub get_dir {
4691 my ($self, $dir, $r) = @_;
4692 my $cache = $self->{cache}->{get_dir};
4693 if ($r == $cache->{r}) {
4694 if (my $x = $cache->{data}->{$dir}) {
4695 return wantarray ? @$x : $x->[0];
4696 }
4697 }
4698 my $pool = SVN::Pool->new;
4699 my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
4700 my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
4701 $pool->clear;
4702 if ($r != $cache->{r}) {
4703 %{$cache->{data}} = ();
4704 $cache->{r} = $r;
4705 }
4706 $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
4707 wantarray ? (\%dirents, $r, $props) : \%dirents;
4708 }
4710 sub DESTROY {
4711 # do not call the real DESTROY since we store ourselves in $RA
4712 }
4714 # get_log(paths, start, end, limit,
4715 # discover_changed_paths, strict_node_history, receiver)
4716 sub get_log {
4717 my ($self, @args) = @_;
4718 my $pool = SVN::Pool->new;
4720 # svn_log_changed_path_t objects passed to get_log are likely to be
4721 # overwritten even if only the refs are copied to an external variable,
4722 # so we should dup the structures in their entirety. Using an
4723 # externally passed pool (instead of our temporary and quickly cleared
4724 # pool in Git::SVN::Ra) does not help matters at all...
4725 my $receiver = pop @args;
4726 my $prefix = "/".$self->{svn_path};
4727 $prefix =~ s#/+($)##;
4728 my $prefix_regex = qr#^\Q$prefix\E#;
4729 push(@args, sub {
4730 my ($paths) = $_[0];
4731 return &$receiver(@_) unless $paths;
4732 $_[0] = ();
4733 foreach my $p (keys %$paths) {
4734 my $i = $paths->{$p};
4735 # Make path relative to our url, not repos_root
4736 $p =~ s/$prefix_regex//;
4737 my %s = map { $_ => $i->$_; }
4738 qw/copyfrom_path copyfrom_rev action/;
4739 if ($s{'copyfrom_path'}) {
4740 $s{'copyfrom_path'} =~ s/$prefix_regex//;
4741 }
4742 $_[0]{$p} = \%s;
4743 }
4744 &$receiver(@_);
4745 });
4748 # the limit parameter was not supported in SVN 1.1.x, so we
4749 # drop it. Therefore, the receiver callback passed to it
4750 # is made aware of this limitation by being wrapped if
4751 # the limit passed to is being wrapped.
4752 if ($SVN::Core::VERSION le '1.2.0') {
4753 my $limit = splice(@args, 3, 1);
4754 if ($limit > 0) {
4755 my $receiver = pop @args;
4756 push(@args, sub { &$receiver(@_) if (--$limit >= 0) });
4757 }
4758 }
4759 my $ret = $self->SUPER::get_log(@args, $pool);
4760 $pool->clear;
4761 $ret;
4762 }
4764 sub trees_match {
4765 my ($self, $url1, $rev1, $url2, $rev2) = @_;
4766 my $ctx = SVN::Client->new(auth => _auth_providers);
4767 my $out = IO::File->new_tmpfile;
4769 # older SVN (1.1.x) doesn't take $pool as the last parameter for
4770 # $ctx->diff(), so we'll create a default one
4771 my $pool = SVN::Pool->new_default_sub;
4773 $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
4774 $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
4775 $out->flush;
4776 my $ret = (($out->stat)[7] == 0);
4777 close $out or croak $!;
4779 $ret;
4780 }
4782 sub get_commit_editor {
4783 my ($self, $log, $cb, $pool) = @_;
4784 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
4785 $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
4786 }
4788 sub gs_do_update {
4789 my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
4790 my $new = ($rev_a == $rev_b);
4791 my $path = $gs->{path};
4793 if ($new && -e $gs->{index}) {
4794 unlink $gs->{index} or die
4795 "Couldn't unlink index: $gs->{index}: $!\n";
4796 }
4797 my $pool = SVN::Pool->new;
4798 $editor->set_path_strip($path);
4799 my (@pc) = split m#/#, $path;
4800 my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
4801 1, $editor, $pool);
4802 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
4804 # Since we can't rely on svn_ra_reparent being available, we'll
4805 # just have to do some magic with set_path to make it so
4806 # we only want a partial path.
4807 my $sp = '';
4808 my $final = join('/', @pc);
4809 while (@pc) {
4810 $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
4811 $sp .= '/' if length $sp;
4812 $sp .= shift @pc;
4813 }
4814 die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
4816 $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
4818 $reporter->finish_report($pool);
4819 $pool->clear;
4820 $editor->{git_commit_ok};
4821 }
4823 # this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
4824 # svn_ra_reparent didn't work before 1.4)
4825 sub gs_do_switch {
4826 my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
4827 my $path = $gs->{path};
4828 my $pool = SVN::Pool->new;
4830 my $full_url = $self->{url};
4831 my $old_url = $full_url;
4832 $full_url .= '/' . $path if length $path;
4833 my ($ra, $reparented);
4835 if ($old_url =~ m#^svn(\+ssh)?://# ||
4836 ($full_url =~ m#^https?://# &&
4837 escape_url($full_url) ne $full_url)) {
4838 $_[0] = undef;
4839 $self = undef;
4840 $RA = undef;
4841 $ra = Git::SVN::Ra->new($full_url);
4842 $ra_invalid = 1;
4843 } elsif ($old_url ne $full_url) {
4844 SVN::_Ra::svn_ra_reparent($self->{session}, $full_url, $pool);
4845 $self->{url} = $full_url;
4846 $reparented = 1;
4847 }
4849 $ra ||= $self;
4850 $url_b = escape_url($url_b);
4851 my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
4852 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
4853 $reporter->set_path('', $rev_a, 0, @lock, $pool);
4854 $reporter->finish_report($pool);
4856 if ($reparented) {
4857 SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
4858 $self->{url} = $old_url;
4859 }
4861 $pool->clear;
4862 $editor->{git_commit_ok};
4863 }
4865 sub longest_common_path {
4866 my ($gsv, $globs) = @_;
4867 my %common;
4868 my $common_max = scalar @$gsv;
4870 foreach my $gs (@$gsv) {
4871 my @tmp = split m#/#, $gs->{path};
4872 my $p = '';
4873 foreach (@tmp) {
4874 $p .= length($p) ? "/$_" : $_;
4875 $common{$p} ||= 0;
4876 $common{$p}++;
4877 }
4878 }
4879 $globs ||= [];
4880 $common_max += scalar @$globs;
4881 foreach my $glob (@$globs) {
4882 my @tmp = split m#/#, $glob->{path}->{left};
4883 my $p = '';
4884 foreach (@tmp) {
4885 $p .= length($p) ? "/$_" : $_;
4886 $common{$p} ||= 0;
4887 $common{$p}++;
4888 }
4889 }
4891 my $longest_path = '';
4892 foreach (sort {length $b <=> length $a} keys %common) {
4893 if ($common{$_} == $common_max) {
4894 $longest_path = $_;
4895 last;
4896 }
4897 }
4898 $longest_path;
4899 }
4901 sub gs_fetch_loop_common {
4902 my ($self, $base, $head, $gsv, $globs) = @_;
4903 return if ($base > $head);
4904 my $inc = $_log_window_size;
4905 my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
4906 my $longest_path = longest_common_path($gsv, $globs);
4907 my $ra_url = $self->{url};
4908 my $find_trailing_edge;
4909 while (1) {
4910 my %revs;
4911 my $err;
4912 my $err_handler = $SVN::Error::handler;
4913 $SVN::Error::handler = sub {
4914 ($err) = @_;
4915 skip_unknown_revs($err);
4916 };
4917 sub _cb {
4918 my ($paths, $r, $author, $date, $log) = @_;
4919 [ $paths,
4920 { author => $author, date => $date, log => $log } ];
4921 }
4922 $self->get_log([$longest_path], $min, $max, 0, 1, 1,
4923 sub { $revs{$_[1]} = _cb(@_) });
4924 if ($err) {
4925 print "Checked through r$max\r";
4926 } else {
4927 $find_trailing_edge = 1;
4928 }
4929 if ($err and $find_trailing_edge) {
4930 print STDERR "Path '$longest_path' ",
4931 "was probably deleted:\n",
4932 $err->expanded_message,
4933 "\nWill attempt to follow ",
4934 "revisions r$min .. r$max ",
4935 "committed before the deletion\n";
4936 my $hi = $max;
4937 while (--$hi >= $min) {
4938 my $ok;
4939 $self->get_log([$longest_path], $min, $hi,
4940 0, 1, 1, sub {
4941 $ok = $_[1];
4942 $revs{$_[1]} = _cb(@_) });
4943 if ($ok) {
4944 print STDERR "r$min .. r$ok OK\n";
4945 last;
4946 }
4947 }
4948 $find_trailing_edge = 0;
4949 }
4950 $SVN::Error::handler = $err_handler;
4952 my %exists = map { $_->{path} => $_ } @$gsv;
4953 foreach my $r (sort {$a <=> $b} keys %revs) {
4954 my ($paths, $logged) = @{$revs{$r}};
4956 foreach my $gs ($self->match_globs(\%exists, $paths,
4957 $globs, $r)) {
4958 if ($gs->rev_map_max >= $r) {
4959 next;
4960 }
4961 next unless $gs->match_paths($paths, $r);
4962 $gs->{logged_rev_props} = $logged;
4963 if (my $last_commit = $gs->last_commit) {
4964 $gs->assert_index_clean($last_commit);
4965 }
4966 my $log_entry = $gs->do_fetch($paths, $r);
4967 if ($log_entry) {
4968 $gs->do_git_commit($log_entry);
4969 }
4970 $INDEX_FILES{$gs->{index}} = 1;
4971 }
4972 foreach my $g (@$globs) {
4973 my $k = "svn-remote.$g->{remote}." .
4974 "$g->{t}-maxRev";
4975 Git::SVN::tmp_config($k, $r);
4976 }
4977 if ($ra_invalid) {
4978 $_[0] = undef;
4979 $self = undef;
4980 $RA = undef;
4981 $self = Git::SVN::Ra->new($ra_url);
4982 $ra_invalid = undef;
4983 }
4984 }
4985 # pre-fill the .rev_db since it'll eventually get filled in
4986 # with '0' x40 if something new gets committed
4987 foreach my $gs (@$gsv) {
4988 next if $gs->rev_map_max >= $max;
4989 next if defined $gs->rev_map_get($max);
4990 $gs->rev_map_set($max, 0 x40);
4991 }
4992 foreach my $g (@$globs) {
4993 my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
4994 Git::SVN::tmp_config($k, $max);
4995 }
4996 last if $max >= $head;
4997 $min = $max + 1;
4998 $max += $inc;
4999 $max = $head if ($max > $head);
5000 }
5001 Git::SVN::gc();
5002 }
5004 sub get_dir_globbed {
5005 my ($self, $left, $depth, $r) = @_;
5007 my @x = eval { $self->get_dir($left, $r) };
5008 return unless scalar @x == 3;
5009 my $dirents = $x[0];
5010 my @finalents;
5011 foreach my $de (keys %$dirents) {
5012 next if $dirents->{$de}->{kind} != $SVN::Node::dir;
5013 if ($depth > 1) {
5014 my @args = ("$left/$de", $depth - 1, $r);
5015 foreach my $dir ($self->get_dir_globbed(@args)) {
5016 push @finalents, "$de/$dir";
5017 }
5018 } else {
5019 push @finalents, $de;
5020 }
5021 }
5022 @finalents;
5023 }
5025 sub match_globs {
5026 my ($self, $exists, $paths, $globs, $r) = @_;
5028 sub get_dir_check {
5029 my ($self, $exists, $g, $r) = @_;
5031 my @dirs = $self->get_dir_globbed($g->{path}->{left},
5032 $g->{path}->{depth},
5033 $r);
5035 foreach my $de (@dirs) {
5036 my $p = $g->{path}->full_path($de);
5037 next if $exists->{$p};
5038 next if (length $g->{path}->{right} &&
5039 ($self->check_path($p, $r) !=
5040 $SVN::Node::dir));
5041 $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
5042 $g->{ref}->full_path($de), 1);
5043 }
5044 }
5045 foreach my $g (@$globs) {
5046 if (my $path = $paths->{"/$g->{path}->{left}"}) {
5047 if ($path->{action} =~ /^[AR]$/) {
5048 get_dir_check($self, $exists, $g, $r);
5049 }
5050 }
5051 foreach (keys %$paths) {
5052 if (/$g->{path}->{left_regex}/ &&
5053 !/$g->{path}->{regex}/) {
5054 next if $paths->{$_}->{action} !~ /^[AR]$/;
5055 get_dir_check($self, $exists, $g, $r);
5056 }
5057 next unless /$g->{path}->{regex}/;
5058 my $p = $1;
5059 my $pathname = $g->{path}->full_path($p);
5060 next if $exists->{$pathname};
5061 next if ($self->check_path($pathname, $r) !=
5062 $SVN::Node::dir);
5063 $exists->{$pathname} = Git::SVN->init(
5064 $self->{url}, $pathname, undef,
5065 $g->{ref}->full_path($p), 1);
5066 }
5067 my $c = '';
5068 foreach (split m#/#, $g->{path}->{left}) {
5069 $c .= "/$_";
5070 next unless ($paths->{$c} &&
5071 ($paths->{$c}->{action} =~ /^[AR]$/));
5072 get_dir_check($self, $exists, $g, $r);
5073 }
5074 }
5075 values %$exists;
5076 }
5078 sub minimize_url {
5079 my ($self) = @_;
5080 return $self->{url} if ($self->{url} eq $self->{repos_root});
5081 my $url = $self->{repos_root};
5082 my @components = split(m!/!, $self->{svn_path});
5083 my $c = '';
5084 do {
5085 $url .= "/$c" if length $c;
5086 eval {
5087 my $ra = (ref $self)->new($url);
5088 my $latest = $ra->get_latest_revnum;
5089 $ra->get_log("", $latest, 0, 1, 0, 1, sub {});
5090 };
5091 } while ($@ && ($c = shift @components));
5092 $url;
5093 }
5095 sub can_do_switch {
5096 my $self = shift;
5097 unless (defined $can_do_switch) {
5098 my $pool = SVN::Pool->new;
5099 my $rep = eval {
5100 $self->do_switch(1, '', 0, $self->{url},
5101 SVN::Delta::Editor->new, $pool);
5102 };
5103 if ($@) {
5104 $can_do_switch = 0;
5105 } else {
5106 $rep->abort_report($pool);
5107 $can_do_switch = 1;
5108 }
5109 $pool->clear;
5110 }
5111 $can_do_switch;
5112 }
5114 sub skip_unknown_revs {
5115 my ($err) = @_;
5116 my $errno = $err->apr_err();
5117 # Maybe the branch we're tracking didn't
5118 # exist when the repo started, so it's
5119 # not an error if it doesn't, just continue
5120 #
5121 # Wonderfully consistent library, eh?
5122 # 160013 - svn:// and file://
5123 # 175002 - http(s)://
5124 # 175007 - http(s):// (this repo required authorization, too...)
5125 # More codes may be discovered later...
5126 if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
5127 my $err_key = $err->expanded_message;
5128 # revision numbers change every time, filter them out
5129 $err_key =~ s/\d+/\0/g;
5130 $err_key = "$errno\0$err_key";
5131 unless ($ignored_err{$err_key}) {
5132 warn "W: Ignoring error from SVN, path probably ",
5133 "does not exist: ($errno): ",
5134 $err->expanded_message,"\n";
5135 warn "W: Do not be alarmed at the above message ",
5136 "git-svn is just searching aggressively for ",
5137 "old history.\n",
5138 "This may take a while on large repositories\n";
5139 $ignored_err{$err_key} = 1;
5140 }
5141 return;
5142 }
5143 die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
5144 }
5146 package Git::SVN::Log;
5147 use strict;
5148 use warnings;
5149 use POSIX qw/strftime/;
5150 use Time::Local;
5151 use constant commit_log_separator => ('-' x 72) . "\n";
5152 use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
5153 %rusers $show_commit $incremental/;
5154 my $l_fmt;
5156 sub cmt_showable {
5157 my ($c) = @_;
5158 return 1 if defined $c->{r};
5160 # big commit message got truncated by the 16k pretty buffer in rev-list
5161 if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
5162 $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
5163 @{$c->{l}} = ();
5164 my @log = command(qw/cat-file commit/, $c->{c});
5166 # shift off the headers
5167 shift @log while ($log[0] ne '');
5168 shift @log;
5170 # TODO: make $c->{l} not have a trailing newline in the future
5171 @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
5173 (undef, $c->{r}, undef) = ::extract_metadata(
5174 (grep(/^git-svn-id: /, @log))[-1]);
5175 }
5176 return defined $c->{r};
5177 }
5179 sub log_use_color {
5180 return $color || Git->repository->get_colorbool('color.diff');
5181 }
5183 sub git_svn_log_cmd {
5184 my ($r_min, $r_max, @args) = @_;
5185 my $head = 'HEAD';
5186 my (@files, @log_opts);
5187 foreach my $x (@args) {
5188 if ($x eq '--' || @files) {
5189 push @files, $x;
5190 } else {
5191 if (::verify_ref("$x^0")) {
5192 $head = $x;
5193 } else {
5194 push @log_opts, $x;
5195 }
5196 }
5197 }
5199 my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
5200 $gs ||= Git::SVN->_new;
5201 my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
5202 $gs->refname);
5203 push @cmd, '-r' unless $non_recursive;
5204 push @cmd, qw/--raw --name-status/ if $verbose;
5205 push @cmd, '--color' if log_use_color();
5206 push @cmd, @log_opts;
5207 if (defined $r_max && $r_max == $r_min) {
5208 push @cmd, '--max-count=1';
5209 if (my $c = $gs->rev_map_get($r_max)) {
5210 push @cmd, $c;
5211 }
5212 } elsif (defined $r_max) {
5213 if ($r_max < $r_min) {
5214 ($r_min, $r_max) = ($r_max, $r_min);
5215 }
5216 my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
5217 my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
5218 # If there are no commits in the range, both $c_max and $c_min
5219 # will be undefined. If there is at least 1 commit in the
5220 # range, both will be defined.
5221 return () if !defined $c_min || !defined $c_max;
5222 if ($c_min eq $c_max) {
5223 push @cmd, '--max-count=1', $c_min;
5224 } else {
5225 push @cmd, '--boundary', "$c_min..$c_max";
5226 }
5227 }
5228 return (@cmd, @files);
5229 }
5231 # adapted from pager.c
5232 sub config_pager {
5233 chomp(my $pager = command_oneline(qw(var GIT_PAGER)));
5234 if ($pager eq 'cat') {
5235 $pager = undef;
5236 }
5237 $ENV{GIT_PAGER_IN_USE} = defined($pager);
5238 }
5240 sub run_pager {
5241 return unless -t *STDOUT && defined $pager;
5242 pipe my ($rfd, $wfd) or return;
5243 defined(my $pid = fork) or ::fatal "Can't fork: $!";
5244 if (!$pid) {
5245 open STDOUT, '>&', $wfd or
5246 ::fatal "Can't redirect to stdout: $!";
5247 return;
5248 }
5249 open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
5250 $ENV{LESS} ||= 'FRSX';
5251 exec $pager or ::fatal "Can't run pager: $! ($pager)";
5252 }
5254 sub format_svn_date {
5255 # some systmes don't handle or mishandle %z, so be creative.
5256 my $t = shift || time;
5257 my $gm = timelocal(gmtime($t));
5258 my $sign = qw( + + - )[ $t <=> $gm ];
5259 my $gmoff = sprintf("%s%02d%02d", $sign, (gmtime(abs($t - $gm)))[2,1]);
5260 return strftime("%Y-%m-%d %H:%M:%S $gmoff (%a, %d %b %Y)", localtime($t));
5261 }
5263 sub parse_git_date {
5264 my ($t, $tz) = @_;
5265 # Date::Parse isn't in the standard Perl distro :(
5266 if ($tz =~ s/^\+//) {
5267 $t += tz_to_s_offset($tz);
5268 } elsif ($tz =~ s/^\-//) {
5269 $t -= tz_to_s_offset($tz);
5270 }
5271 return $t;
5272 }
5274 sub set_local_timezone {
5275 if (defined $TZ) {
5276 $ENV{TZ} = $TZ;
5277 } else {
5278 delete $ENV{TZ};
5279 }
5280 }
5282 sub tz_to_s_offset {
5283 my ($tz) = @_;
5284 $tz =~ s/(\d\d)$//;
5285 return ($1 * 60) + ($tz * 3600);
5286 }
5288 sub get_author_info {
5289 my ($dest, $author, $t, $tz) = @_;
5290 $author =~ s/(?:^\s*|\s*$)//g;
5291 $dest->{a_raw} = $author;
5292 my $au;
5293 if ($::_authors) {
5294 $au = $rusers{$author} || undef;
5295 }
5296 if (!$au) {
5297 ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
5298 }
5299 $dest->{t} = $t;
5300 $dest->{tz} = $tz;
5301 $dest->{a} = $au;
5302 $dest->{t_utc} = parse_git_date($t, $tz);
5303 }
5305 sub process_commit {
5306 my ($c, $r_min, $r_max, $defer) = @_;
5307 if (defined $r_min && defined $r_max) {
5308 if ($r_min == $c->{r} && $r_min == $r_max) {
5309 show_commit($c);
5310 return 0;
5311 }
5312 return 1 if $r_min == $r_max;
5313 if ($r_min < $r_max) {
5314 # we need to reverse the print order
5315 return 0 if (defined $limit && --$limit < 0);
5316 push @$defer, $c;
5317 return 1;
5318 }
5319 if ($r_min != $r_max) {
5320 return 1 if ($r_min < $c->{r});
5321 return 1 if ($r_max > $c->{r});
5322 }
5323 }
5324 return 0 if (defined $limit && --$limit < 0);
5325 show_commit($c);
5326 return 1;
5327 }
5329 sub show_commit {
5330 my $c = shift;
5331 if ($oneline) {
5332 my $x = "\n";
5333 if (my $l = $c->{l}) {
5334 while ($l->[0] =~ /^\s*$/) { shift @$l }
5335 $x = $l->[0];
5336 }
5337 $l_fmt ||= 'A' . length($c->{r});
5338 print 'r',pack($l_fmt, $c->{r}),' | ';
5339 print "$c->{c} | " if $show_commit;
5340 print $x;
5341 } else {
5342 show_commit_normal($c);
5343 }
5344 }
5346 sub show_commit_changed_paths {
5347 my ($c) = @_;
5348 return unless $c->{changed};
5349 print "Changed paths:\n", @{$c->{changed}};
5350 }
5352 sub show_commit_normal {
5353 my ($c) = @_;
5354 print commit_log_separator, "r$c->{r} | ";
5355 print "$c->{c} | " if $show_commit;
5356 print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
5357 my $nr_line = 0;
5359 if (my $l = $c->{l}) {
5360 while ($l->[$#$l] eq "\n" && $#$l > 0
5361 && $l->[($#$l - 1)] eq "\n") {
5362 pop @$l;
5363 }
5364 $nr_line = scalar @$l;
5365 if (!$nr_line) {
5366 print "1 line\n\n\n";
5367 } else {
5368 if ($nr_line == 1) {
5369 $nr_line = '1 line';
5370 } else {
5371 $nr_line .= ' lines';
5372 }
5373 print $nr_line, "\n";
5374 show_commit_changed_paths($c);
5375 print "\n";
5376 print $_ foreach @$l;
5377 }
5378 } else {
5379 print "1 line\n";
5380 show_commit_changed_paths($c);
5381 print "\n";
5383 }
5384 foreach my $x (qw/raw stat diff/) {
5385 if ($c->{$x}) {
5386 print "\n";
5387 print $_ foreach @{$c->{$x}}
5388 }
5389 }
5390 }
5392 sub cmd_show_log {
5393 my (@args) = @_;
5394 my ($r_min, $r_max);
5395 my $r_last = -1; # prevent dupes
5396 set_local_timezone();
5397 if (defined $::_revision) {
5398 if ($::_revision =~ /^(\d+):(\d+)$/) {
5399 ($r_min, $r_max) = ($1, $2);
5400 } elsif ($::_revision =~ /^\d+$/) {
5401 $r_min = $r_max = $::_revision;
5402 } else {
5403 ::fatal "-r$::_revision is not supported, use ",
5404 "standard 'git log' arguments instead";
5405 }
5406 }
5408 config_pager();
5409 @args = git_svn_log_cmd($r_min, $r_max, @args);
5410 if (!@args) {
5411 print commit_log_separator unless $incremental || $oneline;
5412 return;
5413 }
5414 my $log = command_output_pipe(@args);
5415 run_pager();
5416 my (@k, $c, $d, $stat);
5417 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
5418 while (<$log>) {
5419 if (/^${esc_color}commit -?($::sha1_short)/o) {
5420 my $cmt = $1;
5421 if ($c && cmt_showable($c) && $c->{r} != $r_last) {
5422 $r_last = $c->{r};
5423 process_commit($c, $r_min, $r_max, \@k) or
5424 goto out;
5425 }
5426 $d = undef;
5427 $c = { c => $cmt };
5428 } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
5429 get_author_info($c, $1, $2, $3);
5430 } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
5431 # ignore
5432 } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
5433 push @{$c->{raw}}, $_;
5434 } elsif (/^${esc_color}[ACRMDT]\t/) {
5435 # we could add $SVN->{svn_path} here, but that requires
5436 # remote access at the moment (repo_path_split)...
5437 s#^(${esc_color})([ACRMDT])\t#$1 $2 #o;
5438 push @{$c->{changed}}, $_;
5439 } elsif (/^${esc_color}diff /o) {
5440 $d = 1;
5441 push @{$c->{diff}}, $_;
5442 } elsif ($d) {
5443 push @{$c->{diff}}, $_;
5444 } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
5445 $esc_color*[\+\-]*$esc_color$/x) {
5446 $stat = 1;
5447 push @{$c->{stat}}, $_;
5448 } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
5449 push @{$c->{stat}}, $_;
5450 $stat = undef;
5451 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
5452 ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
5453 } elsif (s/^${esc_color} //o) {
5454 push @{$c->{l}}, $_;
5455 }
5456 }
5457 if ($c && defined $c->{r} && $c->{r} != $r_last) {
5458 $r_last = $c->{r};
5459 process_commit($c, $r_min, $r_max, \@k);
5460 }
5461 if (@k) {
5462 ($r_min, $r_max) = ($r_max, $r_min);
5463 process_commit($_, $r_min, $r_max) foreach reverse @k;
5464 }
5465 out:
5466 close $log;
5467 print commit_log_separator unless $incremental || $oneline;
5468 }
5470 sub cmd_blame {
5471 my $path = pop;
5473 config_pager();
5474 run_pager();
5476 my ($fh, $ctx, $rev);
5478 if ($_git_format) {
5479 ($fh, $ctx) = command_output_pipe('blame', @_, $path);
5480 while (my $line = <$fh>) {
5481 if ($line =~ /^\^?([[:xdigit:]]+)\s/) {
5482 # Uncommitted edits show up as a rev ID of
5483 # all zeros, which we can't look up with
5484 # cmt_metadata
5485 if ($1 !~ /^0+$/) {
5486 (undef, $rev, undef) =
5487 ::cmt_metadata($1);
5488 $rev = '0' if (!$rev);
5489 } else {
5490 $rev = '0';
5491 }
5492 $rev = sprintf('%-10s', $rev);
5493 $line =~ s/^\^?[[:xdigit:]]+(\s)/$rev$1/;
5494 }
5495 print $line;
5496 }
5497 } else {
5498 ($fh, $ctx) = command_output_pipe('blame', '-p', @_, 'HEAD',
5499 '--', $path);
5500 my ($sha1);
5501 my %authors;
5502 my @buffer;
5503 my %dsha; #distinct sha keys
5505 while (my $line = <$fh>) {
5506 push @buffer, $line;
5507 if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
5508 $dsha{$1} = 1;
5509 }
5510 }
5512 my $s2r = ::cmt_sha2rev_batch([keys %dsha]);
5514 foreach my $line (@buffer) {
5515 if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
5516 $rev = $s2r->{$1};
5517 $rev = '0' if (!$rev)
5518 }
5519 elsif ($line =~ /^author (.*)/) {
5520 $authors{$rev} = $1;
5521 $authors{$rev} =~ s/\s/_/g;
5522 }
5523 elsif ($line =~ /^\t(.*)$/) {
5524 printf("%6s %10s %s\n", $rev, $authors{$rev}, $1);
5525 }
5526 }
5527 }
5528 command_close_pipe($fh, $ctx);
5529 }
5531 package Git::SVN::Migration;
5532 # these version numbers do NOT correspond to actual version numbers
5533 # of git nor git-svn. They are just relative.
5534 #
5535 # v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
5536 #
5537 # v1 layout: .git/$id/info/url, refs/remotes/$id
5538 #
5539 # v2 layout: .git/svn/$id/info/url, refs/remotes/$id
5540 #
5541 # v3 layout: .git/svn/$id, refs/remotes/$id
5542 # - info/url may remain for backwards compatibility
5543 # - this is what we migrate up to this layout automatically,
5544 # - this will be used by git svn init on single branches
5545 # v3.1 layout (auto migrated):
5546 # - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
5547 # for backwards compatibility
5548 #
5549 # v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
5550 # - this is only created for newly multi-init-ed
5551 # repositories. Similar in spirit to the
5552 # --use-separate-remotes option in git-clone (now default)
5553 # - we do not automatically migrate to this (following
5554 # the example set by core git)
5555 #
5556 # v5 layout: .rev_db.$UUID => .rev_map.$UUID
5557 # - newer, more-efficient format that uses 24-bytes per record
5558 # with no filler space.
5559 # - use xxd -c24 < .rev_map.$UUID to view and debug
5560 # - This is a one-way migration, repositories updated to the
5561 # new format will not be able to use old git-svn without
5562 # rebuilding the .rev_db. Rebuilding the rev_db is not
5563 # possible if noMetadata or useSvmProps are set; but should
5564 # be no problem for users that use the (sensible) defaults.
5565 use strict;
5566 use warnings;
5567 use Carp qw/croak/;
5568 use File::Path qw/mkpath/;
5569 use File::Basename qw/dirname basename/;
5570 use vars qw/$_minimize/;
5572 sub migrate_from_v0 {
5573 my $git_dir = $ENV{GIT_DIR};
5574 return undef unless -d $git_dir;
5575 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
5576 my $migrated = 0;
5577 while (<$fh>) {
5578 chomp;
5579 my ($id, $orig_ref) = ($_, $_);
5580 next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
5581 next unless -f "$git_dir/$id/info/url";
5582 my $new_ref = "refs/remotes/$id";
5583 if (::verify_ref("$new_ref^0")) {
5584 print STDERR "W: $orig_ref is probably an old ",
5585 "branch used by an ancient version of ",
5586 "git-svn.\n",
5587 "However, $new_ref also exists.\n",
5588 "We will not be able ",
5589 "to use this branch until this ",
5590 "ambiguity is resolved.\n";
5591 next;
5592 }
5593 print STDERR "Migrating from v0 layout...\n" if !$migrated;
5594 print STDERR "Renaming ref: $orig_ref => $new_ref\n";
5595 command_noisy('update-ref', $new_ref, $orig_ref);
5596 command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
5597 $migrated++;
5598 }
5599 command_close_pipe($fh, $ctx);
5600 print STDERR "Done migrating from v0 layout...\n" if $migrated;
5601 $migrated;
5602 }
5604 sub migrate_from_v1 {
5605 my $git_dir = $ENV{GIT_DIR};
5606 my $migrated = 0;
5607 return $migrated unless -d $git_dir;
5608 my $svn_dir = "$git_dir/svn";
5610 # just in case somebody used 'svn' as their $id at some point...
5611 return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
5613 print STDERR "Migrating from a git-svn v1 layout...\n";
5614 mkpath([$svn_dir]);
5615 print STDERR "Data from a previous version of git-svn exists, but\n\t",
5616 "$svn_dir\n\t(required for this version ",
5617 "($::VERSION) of git-svn) does not exist.\n";
5618 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
5619 while (<$fh>) {
5620 my $x = $_;
5621 next unless $x =~ s#^refs/remotes/##;
5622 chomp $x;
5623 next unless -f "$git_dir/$x/info/url";
5624 my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
5625 next unless $u;
5626 my $dn = dirname("$git_dir/svn/$x");
5627 mkpath([$dn]) unless -d $dn;
5628 if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
5629 mkpath(["$git_dir/svn/svn"]);
5630 print STDERR " - $git_dir/$x/info => ",
5631 "$git_dir/svn/$x/info\n";
5632 rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
5633 croak "$!: $x";
5634 # don't worry too much about these, they probably
5635 # don't exist with repos this old (save for index,
5636 # and we can easily regenerate that)
5637 foreach my $f (qw/unhandled.log index .rev_db/) {
5638 rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
5639 }
5640 } else {
5641 print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
5642 rename "$git_dir/$x", "$git_dir/svn/$x" or
5643 croak "$!: $x";
5644 }
5645 $migrated++;
5646 }
5647 command_close_pipe($fh, $ctx);
5648 print STDERR "Done migrating from a git-svn v1 layout\n";
5649 $migrated;
5650 }
5652 sub read_old_urls {
5653 my ($l_map, $pfx, $path) = @_;
5654 my @dir;
5655 foreach (<$path/*>) {
5656 if (-r "$_/info/url") {
5657 $pfx .= '/' if $pfx && $pfx !~ m!/$!;
5658 my $ref_id = $pfx . basename $_;
5659 my $url = ::file_to_s("$_/info/url");
5660 $l_map->{$ref_id} = $url;
5661 } elsif (-d $_) {
5662 push @dir, $_;
5663 }
5664 }
5665 foreach (@dir) {
5666 my $x = $_;
5667 $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
5668 read_old_urls($l_map, $x, $_);
5669 }
5670 }
5672 sub migrate_from_v2 {
5673 my @cfg = command(qw/config -l/);
5674 return if grep /^svn-remote\..+\.url=/, @cfg;
5675 my %l_map;
5676 read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
5677 my $migrated = 0;
5679 foreach my $ref_id (sort keys %l_map) {
5680 eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
5681 if ($@) {
5682 Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
5683 }
5684 $migrated++;
5685 }
5686 $migrated;
5687 }
5689 sub minimize_connections {
5690 my $r = Git::SVN::read_all_remotes();
5691 my $new_urls = {};
5692 my $root_repos = {};
5693 foreach my $repo_id (keys %$r) {
5694 my $url = $r->{$repo_id}->{url} or next;
5695 my $fetch = $r->{$repo_id}->{fetch} or next;
5696 my $ra = Git::SVN::Ra->new($url);
5698 # skip existing cases where we already connect to the root
5699 if (($ra->{url} eq $ra->{repos_root}) ||
5700 ($ra->{repos_root} eq $repo_id)) {
5701 $root_repos->{$ra->{url}} = $repo_id;
5702 next;
5703 }
5705 my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
5706 my $root_path = $ra->{url};
5707 $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
5708 foreach my $path (keys %$fetch) {
5709 my $ref_id = $fetch->{$path};
5710 my $gs = Git::SVN->new($ref_id, $repo_id, $path);
5712 # make sure we can read when connecting to
5713 # a higher level of a repository
5714 my ($last_rev, undef) = $gs->last_rev_commit;
5715 if (!defined $last_rev) {
5716 $last_rev = eval {
5717 $root_ra->get_latest_revnum;
5718 };
5719 next if $@;
5720 }
5721 my $new = $root_path;
5722 $new .= length $path ? "/$path" : '';
5723 eval {
5724 $root_ra->get_log([$new], $last_rev, $last_rev,
5725 0, 0, 1, sub { });
5726 };
5727 next if $@;
5728 $new_urls->{$ra->{repos_root}}->{$new} =
5729 { ref_id => $ref_id,
5730 old_repo_id => $repo_id,
5731 old_path => $path };
5732 }
5733 }
5735 my @emptied;
5736 foreach my $url (keys %$new_urls) {
5737 # see if we can re-use an existing [svn-remote "repo_id"]
5738 # instead of creating a(n ugly) new section:
5739 my $repo_id = $root_repos->{$url} || $url;
5741 my $fetch = $new_urls->{$url};
5742 foreach my $path (keys %$fetch) {
5743 my $x = $fetch->{$path};
5744 Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
5745 my $pfx = "svn-remote.$x->{old_repo_id}";
5747 my $old_fetch = quotemeta("$x->{old_path}:".
5748 "$x->{ref_id}");
5749 command_noisy(qw/config --unset/,
5750 "$pfx.fetch", '^'. $old_fetch . '$');
5751 delete $r->{$x->{old_repo_id}}->
5752 {fetch}->{$x->{old_path}};
5753 if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
5754 command_noisy(qw/config --unset/,
5755 "$pfx.url");
5756 push @emptied, $x->{old_repo_id}
5757 }
5758 }
5759 }
5760 if (@emptied) {
5761 my $file = $ENV{GIT_CONFIG} || "$ENV{GIT_DIR}/config";
5762 print STDERR <<EOF;
5763 The following [svn-remote] sections in your config file ($file) are empty
5764 and can be safely removed:
5765 EOF
5766 print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
5767 }
5768 }
5770 sub migration_check {
5771 migrate_from_v0();
5772 migrate_from_v1();
5773 migrate_from_v2();
5774 minimize_connections() if $_minimize;
5775 }
5777 package Git::IndexInfo;
5778 use strict;
5779 use warnings;
5780 use Git qw/command_input_pipe command_close_pipe/;
5782 sub new {
5783 my ($class) = @_;
5784 my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
5785 bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
5786 }
5788 sub remove {
5789 my ($self, $path) = @_;
5790 if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
5791 return ++$self->{nr};
5792 }
5793 undef;
5794 }
5796 sub update {
5797 my ($self, $mode, $hash, $path) = @_;
5798 if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
5799 return ++$self->{nr};
5800 }
5801 undef;
5802 }
5804 sub DESTROY {
5805 my ($self) = @_;
5806 command_close_pipe($self->{gui}, $self->{ctx});
5807 }
5809 package Git::SVN::GlobSpec;
5810 use strict;
5811 use warnings;
5813 sub new {
5814 my ($class, $glob) = @_;
5815 my $re = $glob;
5816 $re =~ s!/+$!!g; # no need for trailing slashes
5817 $re =~ m!^([^*]*)(\*(?:/\*)*)(.*)$!;
5818 my $temp = $re;
5819 my ($left, $right) = ($1, $3);
5820 $re = $2;
5821 my $depth = $re =~ tr/*/*/;
5822 if ($depth != $temp =~ tr/*/*/) {
5823 die "Only one set of wildcard directories " .
5824 "(e.g. '*' or '*/*/*') is supported: '$glob'\n";
5825 }
5826 if ($depth == 0) {
5827 die "One '*' is needed for glob: '$glob'\n";
5828 }
5829 $re =~ s!\*!\[^/\]*!g;
5830 $re = quotemeta($left) . "($re)" . quotemeta($right);
5831 if (length $left && !($left =~ s!/+$!!g)) {
5832 die "Missing trailing '/' on left side of: '$glob' ($left)\n";
5833 }
5834 if (length $right && !($right =~ s!^/+!!g)) {
5835 die "Missing leading '/' on right side of: '$glob' ($right)\n";
5836 }
5837 my $left_re = qr/^\/\Q$left\E(\/|$)/;
5838 bless { left => $left, right => $right, left_regex => $left_re,
5839 regex => qr/$re/, glob => $glob, depth => $depth }, $class;
5840 }
5842 sub full_path {
5843 my ($self, $path) = @_;
5844 return (length $self->{left} ? "$self->{left}/" : '') .
5845 $path . (length $self->{right} ? "/$self->{right}" : '');
5846 }
5848 __END__
5850 Data structures:
5853 $remotes = { # returned by read_all_remotes()
5854 'svn' => {
5855 # svn-remote.svn.url=https://svn.musicpd.org
5856 url => 'https://svn.musicpd.org',
5857 # svn-remote.svn.fetch=mpd/trunk:trunk
5858 fetch => {
5859 'mpd/trunk' => 'trunk',
5860 },
5861 # svn-remote.svn.tags=mpd/tags/*:tags/*
5862 tags => {
5863 path => {
5864 left => 'mpd/tags',
5865 right => '',
5866 regex => qr!mpd/tags/([^/]+)$!,
5867 glob => 'tags/*',
5868 },
5869 ref => {
5870 left => 'tags',
5871 right => '',
5872 regex => qr!tags/([^/]+)$!,
5873 glob => 'tags/*',
5874 },
5875 }
5876 }
5877 };
5879 $log_entry hashref as returned by libsvn_log_entry()
5880 {
5881 log => 'whitespace-formatted log entry
5882 ', # trailing newline is preserved
5883 revision => '8', # integer
5884 date => '2004-02-24T17:01:44.108345Z', # commit date
5885 author => 'committer name'
5886 };
5889 # this is generated by generate_diff();
5890 @mods = array of diff-index line hashes, each element represents one line
5891 of diff-index output
5893 diff-index line ($m hash)
5894 {
5895 mode_a => first column of diff-index output, no leading ':',
5896 mode_b => second column of diff-index output,
5897 sha1_b => sha1sum of the final blob,
5898 chg => change type [MCRADT],
5899 file_a => original file name of a file (iff chg is 'C' or 'R')
5900 file_b => new/current file name of a file (any chg)
5901 }
5902 ;
5904 # retval of read_url_paths{,_all}();
5905 $l_map = {
5906 # repository root url
5907 'https://svn.musicpd.org' => {
5908 # repository path # GIT_SVN_ID
5909 'mpd/trunk' => 'trunk',
5910 'mpd/tags/0.11.5' => 'tags/0.11.5',
5911 },
5912 }
5914 Notes:
5915 I don't trust the each() function on unless I created %hash myself
5916 because the internal iterator may not have started at base.