a6f5061c3c356076af797f6d3373ef5f9ca870a3
1#!/usr/bin/env perl
2# Copyright (C) 2006, Eric Wong <normalperson@yhbt.net>
3# License: GPL v2 or later
4use warnings;
5use strict;
6use 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@@';
11# From which subdir have we been invoked?
13my $cmd_dir_prefix = eval {
14 command_oneline([qw/rev-parse --show-prefix/], STDERR => 0)
15} || '';
16my $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';
23if (! exists $ENV{SVN_SSH}) {
25 if (exists $ENV{GIT_SSH}) {
26 $ENV{SVN_SSH} = $ENV{GIT_SSH};
27 if ($^O eq 'msys') {
28 $ENV{SVN_SSH} =~ s/\\/\\\\/g;
29 }
30 }
31}
32$Git::SVN::Log::TZ = $ENV{TZ};
34$ENV{TZ} = 'UTC';
35$| = 1; # unbuffer STDOUT
36sub fatal (@) { print STDERR "@_\n"; exit 1 }
38require SVN::Core; # use()-ing this causes segfaults for me... *shrug*
39require SVN::Ra;
40require SVN::Delta;
41if ($SVN::Core::VERSION lt '1.1.0') {
42 fatal "Need SVN::Core 1.1.0 or better (got $SVN::Core::VERSION)";
43}
44my $can_compress = eval { require Compress::Zlib; 1};
45push @Git::SVN::Ra::ISA, 'SVN::Ra';
46push @SVN::Git::Editor::ISA, 'SVN::Delta::Editor';
47push @SVN::Git::Fetcher::ISA, 'SVN::Delta::Editor';
48use Carp qw/croak/;
49use Digest::MD5;
50use IO::File qw//;
51use File::Basename qw/dirname basename/;
52use File::Path qw/mkpath/;
53use File::Spec;
54use File::Find;
55use Getopt::Long qw/:config gnu_getopt no_ignore_case auto_abbrev/;
56use IPC::Open3;
57use Git;
58BEGIN {
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}
72my ($SVN);
74$sha1 = qr/[a-f\d]{40}/;
76$sha1_short = qr/[a-f\d]{4,40}/;
77my ($_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$_q ||= 0;
86my %remote_opts = ( 'username=s' => \$Git::SVN::Prompt::_username,
87 'config-dir=s' => \$Git::SVN::Ra::config_dir,
88 'no-auth-cache' => \$Git::SVN::Prompt::_no_auth_cache,
89 'ignore-paths=s' => \$SVN::Git::Fetcher::_ignore_regex );
90my %fc_opts = ( 'follow-parent|follow!' => \$Git::SVN::_follow_parent,
91 'authors-file|A=s' => \$_authors,
92 'authors-prog=s' => \$_authors_prog,
93 'repack:i' => \$Git::SVN::_repack,
94 'noMetadata' => \$Git::SVN::_no_metadata,
95 'useSvmProps' => \$Git::SVN::_use_svm_props,
96 'useSvnsyncProps' => \$Git::SVN::_use_svnsync_props,
97 'log-window-size=i' => \$Git::SVN::Ra::_log_window_size,
98 'no-checkout' => \$_no_checkout,
99 'quiet|q+' => \$_q,
100 'repack-flags|repack-args|repack-opts=s' =>
101 \$Git::SVN::_repack_flags,
102 'use-log-author' => \$Git::SVN::_use_log_author,
103 'add-author-from' => \$Git::SVN::_add_author_from,
104 'localtime' => \$Git::SVN::_localtime,
105 %remote_opts );
106my ($_trunk, @_tags, @_branches, $_stdlayout);
108my %icv;
109my %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 );
119my %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);
125my %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);
240my $cmd;
242for (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};
252# make sure we're always running at the top-level working directory
254unless ($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 => $ENV{GIT_DIR});
276}
277my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
279read_git_config(\%opts);
281if ($cmd && ($cmd eq 'log' || $cmd eq 'blame')) {
282 Getopt::Long::Configure('pass_through');
283}
284my $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 = $_[1] });
290exit 1 if (!$rv && $cmd && $cmd ne 'log');
291usage(0) if $_help;
293version() if $_version;
294usage(1) unless defined $cmd;
295load_authors() if $_authors;
296if (defined $_authors_prog) {
297 $_authors_prog = "'" . File::Spec->rel2abs($_authors_prog) . "'";
298}
299unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
301 Git::SVN::Migration::migration_check();
302}
303Git::SVN::init_vars();
304eval {
305 Git::SVN::verify_remotes_sanity();
306 $cmd{$cmd}->[0]->(@ARGV);
307};
308fatal $@ if $@;
309post_fetch_checkout();
310exit 0;
311####################### primary functions ######################
313sub usage {
314 my $exit = shift || 0;
315 my $fd = $exit ? \*STDERR : \*STDOUT;
316 print $fd <<"";
317git-svn - bidirectional operations between a single Subversion tree and git
318Usage: git svn <command> [options] [arguments]\n
319 print $fd "Available commands:\n" unless $cmd;
321 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
338arbitrary identifier if you're tracking multiple SVN branches/repositories in
339one git repository and want to keep them separate. See git-svn(1) for more
340information.
341 exit $exit;
343}
344sub version {
346 print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
347 exit 0;
348}
349sub 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}
377sub 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 => $ENV{GIT_DIR});
384}
385sub cmd_clone {
387 my ($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}
401sub 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();
416 if ($Git::SVN::_minimize_url eq 'unset') {
418 $Git::SVN::_minimize_url = 0;
419 }
420 Git::SVN->init($url);
422}
423sub 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}
450sub 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}
486sub 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';
493 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', $head], STDERR => 0);
505 }
506 my @refs;
508 my ($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 }
513 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 }
523 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;
572 # 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 $linear_refs = \@l;
636 }
637 }
638 }
639 if ($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 }
654 unlink $gs->{index};
656}
657sub cmd_branch {
659 my ($branch_name, $head) = @_;
660 unless (defined $branch_name && length $branch_name) {
662 die(($_tag ? "tag" : "branch") . " name required\n");
663 }
664 $head ||= 'HEAD';
665 my ($src, $rev, undef, $gs) = working_head_info($head);
667 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 || ());
711 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 );
721 eval {
723 $ctx->ls($dst, 'HEAD', 0);
724 } and die "branch ${branch_name} already exists\n";
725 print "Copying ${src} at r${rev} to ${dst}...\n";
727 $ctx->copy($src, $rev, $dst)
728 unless $_dry_run;
729 $gs->fetch_all;
731}
732sub 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 $result;
753}
754sub cmd_rebase {
756 command_noisy(qw/update-index --refresh/);
757 my ($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 $gs->mkemptydirs;
779}
780sub cmd_show_ignore {
782 my ($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 "$s\n";
794 });
795}
796sub cmd_show_externals {
798 my ($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 "$s\n";
809 });
810}
811sub cmd_create_ignore {
813 my ($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', $ignore);
837 });
838}
839sub cmd_mkdirs {
841 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
842 $gs ||= Git::SVN->new;
843 $gs->mkemptydirs($_revision);
844}
845sub 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 $path;
863}
864sub canonicalize_url {
866 my ($url) = @_;
867 $url =~ s#^([^:]+://[^/]*/)(.*)$#$1 . canonicalize_path($2)#e;
868 return $url;
869}
870# get_svnprops(PATH)
872# ------------------
873# Helper for cmd_propget and cmd_proplist below.
874sub get_svnprops {
875 my $path = shift;
876 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
877 $gs ||= Git::SVN->new;
878 # 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;
885 # canonicalize the path (otherwise libsvn will abort or fail to
887 # find the file)
888 $path = canonicalize_path($path);
889 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 $props;
899}
900# cmd_propget (PROP, PATH)
902# ------------------------
903# Print the SVN property PROP for PATH.
904sub cmd_propget {
905 my ($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}
914# cmd_proplist (PATH)
916# -------------------
917# Print the list of SVN properties for PATH.
918sub 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}
927sub cmd_multi_init {
929 my $url = shift;
930 unless (defined $_trunk || @_branches || @_tags) {
931 usage(1);
932 }
933 $_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}
960sub 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, $remotes);
967 }
968 }
969}
970# this command is special because it requires no metadata
972sub cmd_commit_diff {
973 my ($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 == $tb\n";
1016 }
1017}
1018sub escape_uri_only {
1020 my ($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, $_;
1025 }
1026 join('/', @tmp);
1027}
1028sub escape_url {
1030 my ($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}
1037sub cmd_info {
1039 my $path = canonicalize_path(defined($_[0]) ? $_[0] : ".");
1040 my $fullpath = canonicalize_path($cmd_dir_prefix . $path);
1041 if (exists $_[1]) {
1042 die "Too many arguments specified\n";
1043 }
1044 my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
1046 if (!$file_type && !$diff_status) {
1048 print STDERR "svn: '$path' is not under version control\n";
1049 exit 1;
1050 }
1051 my ($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 }
1057 # canonicalize_path() will return "" to make libsvn 1.5.x happy,
1059 $path = "." if $path eq "";
1060 my $full_url = $url . ($fullpath eq "" ? "" : "/$fullpath");
1062 if ($_url) {
1064 print escape_url($full_url), "\n";
1065 return;
1066 }
1067 my $result = "Path: $path\n";
1069 $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
1070 $result .= "URL: " . escape_url($full_url) . "\n";
1071 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";
1083 $result .= "Node Kind: " .
1085 ($file_type eq "dir" ? "directory" : "file") . "\n";
1086 my $schedule = $diff_status eq "A"
1088 ? "add"
1089 : ($diff_status eq "D" ? "delete" : "normal");
1090 $result .= "Schedule: $schedule\n";
1091 if ($diff_status eq "A") {
1093 print $result, "\n";
1094 return;
1095 }
1096 my ($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;
1110 Git::SVN::Log::set_local_timezone();
1112 $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";
1117 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 }
1148 print $result, "\n";
1150}
1151sub 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}
1165sub 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}
1173########################### utility functions #########################
1175sub 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}
1183sub 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');
1188 # 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 }
1199 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'));
1203 return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
1205 my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
1206 return if -f $index;
1207 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}
1215sub 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}
1228sub 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}
1270sub verify_ref {
1272 my ($ref) = @_;
1273 eval { command_oneline([ 'rev-parse', '--verify', $ref ],
1274 { STDERR => 0 }); };
1275}
1276sub get_tree_from_treeish {
1278 my ($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 $expected;
1296}
1297sub get_commit_entry {
1299 my ($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 $!;
1304 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 $!;
1337 if ($_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>);
1350 my $enc = Git::config('i18n.commitencoding') || 'UTF-8';
1352 my $msg = $log_entry{log};
1353 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 }
1359 eval { $msg = Encode::encode('UTF-8', $msg, 1) };
1361 die "Could not encode as UTF-8:\n$msg\n" if $@;
1362 $log_entry{log} = $msg;
1364 close $log_fh or croak $!;
1366 }
1367 unlink $commit_msg;
1368 \%log_entry;
1369}
1370sub s_to_file {
1372 my ($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}
1378sub 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}
1388# '<svn username> = real-name <email address>' mapping based on git-svnimport:
1390sub 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}
1405# convert GetOpt::Long specs for use by git-config
1407sub 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}
1433sub 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}
1445sub cmt_metadata {
1447 return extract_metadata((grep(/^git-svn-id: /,
1448 command(qw/cat-file commit/, shift)))[-1]);
1449}
1450sub cmt_sha2rev_batch {
1452 my %s2r;
1453 my ($pid, $in, $out, $ctx) = command_bidi_pipe(qw/cat-file --batch/);
1454 my $list = shift;
1455 foreach my $sha (@{$list}) {
1457 my $first = 1;
1458 my $size = 0;
1459 print $out $sha, "\n";
1460 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 }
1474 $size -= length($line);
1476 last if ($size == 0);
1477 }
1478 }
1479 command_close_bidi_pipe($pid, $in, $out, $ctx);
1481 return \%s2r;
1483}
1484sub 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}
1515sub 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}
1522sub linearize_history {
1524 my ($gs, $refs) = @_;
1525 my %parents;
1526 foreach my $c (@$refs) {
1527 read_commit_parents(\%parents, $c);
1528 }
1529 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};
1536 unshift @linear_refs, $c;
1538 $skip{$c} = 1;
1539 # 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 }
1556 foreach my $p (@{$parents{$c}}) {
1558 $skip{$p} = 1;
1559 }
1560 }
1561 (\@linear_refs, \%parents);
1562}
1563sub find_file_type_and_diff_status {
1565 my ($path) = @_;
1566 return ('dir', '') if $path eq '';
1567 my $diff_output =
1569 command_oneline(qw(diff --cached --name-status --), $path) || "";
1570 my $diff_status = (split(' ', $diff_output))[0] || "";
1571 my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1573 return (undef, undef) if !$diff_status && !$ls_tree;
1575 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 }
1581 my $mode = (split(' ', $ls_tree))[0] || "";
1583 return ("link", $diff_status) if $mode eq "120000";
1585 return ("dir", $diff_status) if $mode eq "040000";
1586 return ("file", $diff_status);
1587}
1588sub 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 $md5->hexdigest();
1603}
1604sub gc_directory {
1606 if ($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";
1612 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 $_: $!\n";
1621 }
1622}
1623package Git::SVN;
1625use strict;
1626use warnings;
1627use Fcntl qw/:DEFAULT :seek/;
1628use constant rev_map_fmt => 'NH40';
1629use vars qw/$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 $_localtime/;
1633use Carp qw/croak/;
1634use File::Path qw/mkpath/;
1635use File::Copy qw/copy/;
1636use IPC::Open3;
1637my ($_gc_nr, $_gc_period);
1639# properties that we do not log:
1641my %SKIP_PROP;
1642BEGIN {
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/;
1649 # 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}
1674my (%LOCKFILES, %INDEX_FILES);
1677END {
1678 unlink keys %LOCKFILES if %LOCKFILES;
1679 unlink keys %INDEX_FILES if %INDEX_FILES;
1680}
1681sub 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} = $refname;
1711 }
1712 }
1713}
1714sub parse_revision_argument {
1716 my ($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: $::_revision not understood by git-svn\n";
1726}
1727sub fetch_all {
1729 my ($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 # ignore errors, $head revision may not even exist anymore
1746 eval { $ra->get_log("", $head, 0, 1, 0, 1, sub { $head = $_[1] }) };
1747 warn "W: $@\n" if $@;
1748 my $base = defined $fetch ? $head : 0;
1750 # read the max revs for wildcard expansion (branches/*, tags/*)
1752 foreach my $t (qw/branches tags/) {
1753 defined $remote->{$t} or next;
1754 push @globs, @{$remote->{$t}};
1755 my $max_rev = eval { tmp_config(qw/--int --get/,
1757 "svn-remote.$repo_id.${t}-maxRev") };
1758 if (defined $max_rev && ($max_rev < $base)) {
1759 $base = $max_rev;
1760 } elsif (!defined $max_rev) {
1761 $base = 0;
1762 }
1763 }
1764 if ($fetch) {
1766 foreach my $p (sort keys %$fetch) {
1767 my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1768 my $lr = $gs->rev_map_max;
1769 if (defined $lr) {
1770 $base = $lr if ($lr < $base);
1771 }
1772 push @gs, $gs;
1773 }
1774 }
1775 ($base, $head) = parse_revision_argument($base, $head);
1777 $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1778}
1779sub read_all_remotes {
1781 my $r = {};
1782 my $use_svm_props = eval { command_oneline(qw/config --bool
1783 svn.useSvmProps/) };
1784 $use_svm_props = $use_svm_props eq 'true' if $use_svm_props;
1785 my $svn_refspec = qr{\s*(.*?)\s*:\s*(.+?)\s*};
1786 foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1787 if (m!^(.+)\.fetch=$svn_refspec$!) {
1788 my ($remote, $local_ref, $remote_ref) = ($1, $2, $3);
1789 die("svn-remote.$remote: remote ref '$remote_ref' "
1790 . "must start with 'refs/'\n")
1791 unless $remote_ref =~ m{^refs/};
1792 $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1793 $r->{$remote}->{svm} = {} if $use_svm_props;
1794 } elsif (m!^(.+)\.usesvmprops=\s*(.*)\s*$!) {
1795 $r->{$1}->{svm} = {};
1796 } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1797 $r->{$1}->{url} = $2;
1798 } elsif (m!^(.+)\.(branches|tags)=$svn_refspec$!) {
1799 my ($remote, $t, $local_ref, $remote_ref) =
1800 ($1, $2, $3, $4);
1801 die("svn-remote.$remote: remote ref '$remote_ref' ($t) "
1802 . "must start with 'refs/'\n")
1803 unless $remote_ref =~ m{^refs/};
1804 my $rs = {
1805 t => $t,
1806 remote => $remote,
1807 path => Git::SVN::GlobSpec->new($local_ref),
1808 ref => Git::SVN::GlobSpec->new($remote_ref) };
1809 if (length($rs->{ref}->{right}) != 0) {
1810 die "The '*' glob character must be the last ",
1811 "character of '$remote_ref'\n";
1812 }
1813 push @{ $r->{$remote}->{$t} }, $rs;
1814 }
1815 }
1816 map {
1818 if (defined $r->{$_}->{svm}) {
1819 my $svm;
1820 eval {
1821 my $section = "svn-remote.$_";
1822 $svm = {
1823 source => tmp_config('--get',
1824 "$section.svm-source"),
1825 replace => tmp_config('--get',
1826 "$section.svm-replace"),
1827 }
1828 };
1829 $r->{$_}->{svm} = $svm;
1830 }
1831 } keys %$r;
1832 $r;
1834}
1835sub init_vars {
1837 $_gc_nr = $_gc_period = 1000;
1838 if (defined $_repack || defined $_repack_flags) {
1839 warn "Repack options are obsolete; they have no effect.\n";
1840 }
1841}
1842sub verify_remotes_sanity {
1844 return unless -d $ENV{GIT_DIR};
1845 my %seen;
1846 foreach (command(qw/config -l/)) {
1847 if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1848 if ($seen{$1}) {
1849 die "Remote ref refs/remote/$1 is tracked by",
1850 "\n \"$_\"\nand\n \"$seen{$1}\"\n",
1851 "Please resolve this ambiguity in ",
1852 "your git configuration file before ",
1853 "continuing\n";
1854 }
1855 $seen{$1} = $_;
1856 }
1857 }
1858}
1859sub find_existing_remote {
1861 my ($url, $remotes) = @_;
1862 return undef if $no_reuse_existing;
1863 my $existing;
1864 foreach my $repo_id (keys %$remotes) {
1865 my $u = $remotes->{$repo_id}->{url} or next;
1866 next if $u ne $url;
1867 $existing = $repo_id;
1868 last;
1869 }
1870 $existing;
1871}
1872sub init_remote_config {
1874 my ($self, $url, $no_write) = @_;
1875 $url =~ s!/+$!!; # strip trailing slash
1876 my $r = read_all_remotes();
1877 my $existing = find_existing_remote($url, $r);
1878 if ($existing) {
1879 unless ($no_write) {
1880 print STDERR "Using existing ",
1881 "[svn-remote \"$existing\"]\n";
1882 }
1883 $self->{repo_id} = $existing;
1884 } elsif ($_minimize_url) {
1885 my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1886 $existing = find_existing_remote($min_url, $r);
1887 if ($existing) {
1888 unless ($no_write) {
1889 print STDERR "Using existing ",
1890 "[svn-remote \"$existing\"]\n";
1891 }
1892 $self->{repo_id} = $existing;
1893 }
1894 if ($min_url ne $url) {
1895 unless ($no_write) {
1896 print STDERR "Using higher level of URL: ",
1897 "$url => $min_url\n";
1898 }
1899 my $old_path = $self->{path};
1900 $self->{path} = $url;
1901 $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1902 if (length $old_path) {
1903 $self->{path} .= "/$old_path";
1904 }
1905 $url = $min_url;
1906 }
1907 }
1908 my $orig_url;
1909 if (!$existing) {
1910 # verify that we aren't overwriting anything:
1911 $orig_url = eval {
1912 command_oneline('config', '--get',
1913 "svn-remote.$self->{repo_id}.url")
1914 };
1915 if ($orig_url && ($orig_url ne $url)) {
1916 die "svn-remote.$self->{repo_id}.url already set: ",
1917 "$orig_url\nwanted to set to: $url\n";
1918 }
1919 }
1920 my ($xrepo_id, $xpath) = find_ref($self->refname);
1921 if (!$no_write && defined $xpath) {
1922 die "svn-remote.$xrepo_id.fetch already set to track ",
1923 "$xpath:", $self->refname, "\n";
1924 }
1925 unless ($no_write) {
1926 command_noisy('config',
1927 "svn-remote.$self->{repo_id}.url", $url);
1928 $self->{path} =~ s{^/}{};
1929 $self->{path} =~ s{%([0-9A-F]{2})}{chr hex($1)}ieg;
1930 command_noisy('config', '--add',
1931 "svn-remote.$self->{repo_id}.fetch",
1932 "$self->{path}:".$self->refname);
1933 }
1934 $self->{url} = $url;
1935}
1936sub find_by_url { # repos_root and, path are optional
1938 my ($class, $full_url, $repos_root, $path) = @_;
1939 return undef unless defined $full_url;
1941 remove_username($full_url);
1942 remove_username($repos_root) if defined $repos_root;
1943 my $remotes = read_all_remotes();
1944 if (defined $full_url && defined $repos_root && !defined $path) {
1945 $path = $full_url;
1946 $path =~ s#^\Q$repos_root\E(?:/|$)##;
1947 }
1948 foreach my $repo_id (keys %$remotes) {
1949 my $u = $remotes->{$repo_id}->{url} or next;
1950 remove_username($u);
1951 next if defined $repos_root && $repos_root ne $u;
1952 my $fetch = $remotes->{$repo_id}->{fetch} || {};
1954 foreach my $t (qw/branches tags/) {
1955 foreach my $globspec (@{$remotes->{$repo_id}->{$t}}) {
1956 resolve_local_globs($u, $fetch, $globspec);
1957 }
1958 }
1959 my $p = $path;
1960 my $rwr = rewrite_root({repo_id => $repo_id});
1961 my $svm = $remotes->{$repo_id}->{svm}
1962 if defined $remotes->{$repo_id}->{svm};
1963 unless (defined $p) {
1964 $p = $full_url;
1965 my $z = $u;
1966 my $prefix = '';
1967 if ($rwr) {
1968 $z = $rwr;
1969 remove_username($z);
1970 } elsif (defined $svm) {
1971 $z = $svm->{source};
1972 $prefix = $svm->{replace};
1973 $prefix =~ s#^\Q$u\E(?:/|$)##;
1974 $prefix =~ s#/$##;
1975 }
1976 $p =~ s#^\Q$z\E(?:/|$)#$prefix# or next;
1977 }
1978 foreach my $f (keys %$fetch) {
1979 next if $f ne $p;
1980 return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1981 }
1982 }
1983 undef;
1984}
1985sub init {
1987 my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1988 my $self = _new($class, $repo_id, $ref_id, $path);
1989 if (defined $url) {
1990 $self->init_remote_config($url, $no_write);
1991 }
1992 $self;
1993}
1994sub find_ref {
1996 my ($ref_id) = @_;
1997 foreach (command(qw/config -l/)) {
1998 next unless m!^svn-remote\.(.+)\.fetch=
1999 \s*(.*?)\s*:\s*(.+?)\s*$!x;
2000 my ($repo_id, $path, $ref) = ($1, $2, $3);
2001 if ($ref eq $ref_id) {
2002 $path = '' if ($path =~ m#^\./?#);
2003 return ($repo_id, $path);
2004 }
2005 }
2006 (undef, undef, undef);
2007}
2008sub new {
2010 my ($class, $ref_id, $repo_id, $path) = @_;
2011 if (defined $ref_id && !defined $repo_id && !defined $path) {
2012 ($repo_id, $path) = find_ref($ref_id);
2013 if (!defined $repo_id) {
2014 die "Could not find a \"svn-remote.*.fetch\" key ",
2015 "in the repository configuration matching: ",
2016 "$ref_id\n";
2017 }
2018 }
2019 my $self = _new($class, $repo_id, $ref_id, $path);
2020 if (!defined $self->{path} || !length $self->{path}) {
2021 my $fetch = command_oneline('config', '--get',
2022 "svn-remote.$repo_id.fetch",
2023 ":$ref_id\$") or
2024 die "Failed to read \"svn-remote.$repo_id.fetch\" ",
2025 "\":$ref_id\$\" in config\n";
2026 ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
2027 }
2028 $self->{url} = command_oneline('config', '--get',
2029 "svn-remote.$repo_id.url") or
2030 die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
2031 $self->rebuild;
2032 $self;
2033}
2034sub refname {
2036 my ($refname) = $_[0]->{ref_id} ;
2037 # It cannot end with a slash /, we'll throw up on this because
2039 # SVN can't have directories with a slash in their name, either:
2040 if ($refname =~ m{/$}) {
2041 die "ref: '$refname' ends with a trailing slash, this is ",
2042 "not permitted by git nor Subversion\n";
2043 }
2044 # It cannot have ASCII control character space, tilde ~, caret ^,
2046 # colon :, question-mark ?, asterisk *, space, or open bracket [
2047 # anywhere.
2048 #
2049 # Additionally, % must be escaped because it is used for escaping
2050 # and we want our escaped refname to be reversible
2051 $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
2052 # no slash-separated component can begin with a dot .
2054 # /.* becomes /%2E*
2055 $refname =~ s{/\.}{/%2E}g;
2056 # It cannot have two consecutive dots .. anywhere
2058 # .. becomes %2E%2E
2059 $refname =~ s{\.\.}{%2E%2E}g;
2060 return $refname;
2062}
2063sub desanitize_refname {
2065 my ($refname) = @_;
2066 $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
2067 return $refname;
2068}
2069sub svm_uuid {
2071 my ($self) = @_;
2072 return $self->{svm}->{uuid} if $self->svm;
2073 $self->ra;
2074 unless ($self->{svm}) {
2075 die "SVM UUID not cached, and reading remotely failed\n";
2076 }
2077 $self->{svm}->{uuid};
2078}
2079sub svm {
2081 my ($self) = @_;
2082 return $self->{svm} if $self->{svm};
2083 my $svm;
2084 # see if we have it in our config, first:
2085 eval {
2086 my $section = "svn-remote.$self->{repo_id}";
2087 $svm = {
2088 source => tmp_config('--get', "$section.svm-source"),
2089 uuid => tmp_config('--get', "$section.svm-uuid"),
2090 replace => tmp_config('--get', "$section.svm-replace"),
2091 }
2092 };
2093 if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
2094 $self->{svm} = $svm;
2095 }
2096 $self->{svm};
2097}
2098sub _set_svm_vars {
2100 my ($self, $ra) = @_;
2101 return $ra if $self->svm;
2102 my @err = ( "useSvmProps set, but failed to read SVM properties\n",
2104 "(svm:source, svm:uuid) ",
2105 "from the following URLs:\n" );
2106 sub read_svm_props {
2107 my ($self, $ra, $path, $r) = @_;
2108 my $props = ($ra->get_dir($path, $r))[2];
2109 my $src = $props->{'svm:source'};
2110 my $uuid = $props->{'svm:uuid'};
2111 return undef if (!$src || !$uuid);
2112 chomp($src, $uuid);
2114 $uuid =~ m{^[0-9a-f\-]{30,}$}i
2116 or die "doesn't look right - svm:uuid is '$uuid'\n";
2117 # the '!' is used to mark the repos_root!/relative/path
2119 $src =~ s{/?!/?}{/};
2120 $src =~ s{/+$}{}; # no trailing slashes please
2121 # username is of no interest
2122 $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
2123 my $replace = $ra->{url};
2125 $replace .= "/$path" if length $path;
2126 my $section = "svn-remote.$self->{repo_id}";
2128 tmp_config("$section.svm-source", $src);
2129 tmp_config("$section.svm-replace", $replace);
2130 tmp_config("$section.svm-uuid", $uuid);
2131 $self->{svm} = {
2132 source => $src,
2133 uuid => $uuid,
2134 replace => $replace
2135 };
2136 }
2137 my $r = $ra->get_latest_revnum;
2139 my $path = $self->{path};
2140 my %tried;
2141 while (length $path) {
2142 unless ($tried{"$self->{url}/$path"}) {
2143 return $ra if $self->read_svm_props($ra, $path, $r);
2144 $tried{"$self->{url}/$path"} = 1;
2145 }
2146 $path =~ s#/?[^/]+$##;
2147 }
2148 die "Path: '$path' should be ''\n" if $path ne '';
2149 return $ra if $self->read_svm_props($ra, $path, $r);
2150 $tried{"$self->{url}/$path"} = 1;
2151 if ($ra->{repos_root} eq $self->{url}) {
2153 die @err, (map { " $_\n" } keys %tried), "\n";
2154 }
2155 # nope, make sure we're connected to the repository root:
2157 my $ok;
2158 my @tried_b;
2159 $path = $ra->{svn_path};
2160 $ra = Git::SVN::Ra->new($ra->{repos_root});
2161 while (length $path) {
2162 unless ($tried{"$ra->{url}/$path"}) {
2163 $ok = $self->read_svm_props($ra, $path, $r);
2164 last if $ok;
2165 $tried{"$ra->{url}/$path"} = 1;
2166 }
2167 $path =~ s#/?[^/]+$##;
2168 }
2169 die "Path: '$path' should be ''\n" if $path ne '';
2170 $ok ||= $self->read_svm_props($ra, $path, $r);
2171 $tried{"$ra->{url}/$path"} = 1;
2172 if (!$ok) {
2173 die @err, (map { " $_\n" } keys %tried), "\n";
2174 }
2175 Git::SVN::Ra->new($self->{url});
2176}
2177sub svnsync {
2179 my ($self) = @_;
2180 return $self->{svnsync} if $self->{svnsync};
2181 if ($self->no_metadata) {
2183 die "Can't have both 'noMetadata' and ",
2184 "'useSvnsyncProps' options set!\n";
2185 }
2186 if ($self->rewrite_root) {
2187 die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
2188 "options set!\n";
2189 }
2190 my $svnsync;
2192 # see if we have it in our config, first:
2193 eval {
2194 my $section = "svn-remote.$self->{repo_id}";
2195 my $url = tmp_config('--get', "$section.svnsync-url");
2197 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
2198 die "doesn't look right - svn:sync-from-url is '$url'\n";
2199 my $uuid = tmp_config('--get', "$section.svnsync-uuid");
2201 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}i) or
2202 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
2203 $svnsync = { url => $url, uuid => $uuid }
2205 };
2206 if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
2207 return $self->{svnsync} = $svnsync;
2208 }
2209 my $err = "useSvnsyncProps set, but failed to read " .
2211 "svnsync property: svn:sync-from-";
2212 my $rp = $self->ra->rev_proplist(0);
2213 my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
2215 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
2216 die "doesn't look right - svn:sync-from-url is '$url'\n";
2217 my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
2219 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}i) or
2220 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
2221 my $section = "svn-remote.$self->{repo_id}";
2223 tmp_config('--add', "$section.svnsync-uuid", $uuid);
2224 tmp_config('--add', "$section.svnsync-url", $url);
2225 return $self->{svnsync} = { url => $url, uuid => $uuid };
2226}
2227# this allows us to memoize our SVN::Ra UUID locally and avoid a
2229# remote lookup (useful for 'git svn log').
2230sub ra_uuid {
2231 my ($self) = @_;
2232 unless ($self->{ra_uuid}) {
2233 my $key = "svn-remote.$self->{repo_id}.uuid";
2234 my $uuid = eval { tmp_config('--get', $key) };
2235 if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/i) {
2236 $self->{ra_uuid} = $uuid;
2237 } else {
2238 die "ra_uuid called without URL\n" unless $self->{url};
2239 $self->{ra_uuid} = $self->ra->get_uuid;
2240 tmp_config('--add', $key, $self->{ra_uuid});
2241 }
2242 }
2243 $self->{ra_uuid};
2244}
2245sub _set_repos_root {
2247 my ($self, $repos_root) = @_;
2248 my $k = "svn-remote.$self->{repo_id}.reposRoot";
2249 $repos_root ||= $self->ra->{repos_root};
2250 tmp_config($k, $repos_root);
2251 $repos_root;
2252}
2253sub repos_root {
2255 my ($self) = @_;
2256 my $k = "svn-remote.$self->{repo_id}.reposRoot";
2257 eval { tmp_config('--get', $k) } || $self->_set_repos_root;
2258}
2259sub ra {
2261 my ($self) = shift;
2262 my $ra = Git::SVN::Ra->new($self->{url});
2263 $self->_set_repos_root($ra->{repos_root});
2264 if ($self->use_svm_props && !$self->{svm}) {
2265 if ($self->no_metadata) {
2266 die "Can't have both 'noMetadata' and ",
2267 "'useSvmProps' options set!\n";
2268 } elsif ($self->use_svnsync_props) {
2269 die "Can't have both 'useSvnsyncProps' and ",
2270 "'useSvmProps' options set!\n";
2271 }
2272 $ra = $self->_set_svm_vars($ra);
2273 $self->{-want_revprops} = 1;
2274 }
2275 $ra;
2276}
2277# prop_walk(PATH, REV, SUB)
2279# -------------------------
2280# Recursively traverse PATH at revision REV and invoke SUB for each
2281# directory that contains a SVN property. SUB will be invoked as
2282# follows: &SUB(gs, path, props); where `gs' is this instance of
2283# Git::SVN, `path' the path to the directory where the properties
2284# `props' were found. The `path' will be relative to point of checkout,
2285# that is, if url://repo/trunk is the current Git branch, and that
2286# directory contains a sub-directory `d', SUB will be invoked with `/d/'
2287# as `path' (note the trailing `/').
2288sub prop_walk {
2289 my ($self, $path, $rev, $sub) = @_;
2290 $path =~ s#^/##;
2292 my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
2293 $path =~ s#^/*#/#g;
2294 my $p = $path;
2295 # Strip the irrelevant part of the path.
2296 $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
2297 # Ensure the path is terminated by a `/'.
2298 $p =~ s#/*$#/#;
2299 # The properties contain all the internal SVN stuff nobody
2301 # (usually) cares about.
2302 my $interesting_props = 0;
2303 foreach (keys %{$props}) {
2304 # If it doesn't start with `svn:', it must be a
2305 # user-defined property.
2306 ++$interesting_props and next if $_ !~ /^svn:/;
2307 # FIXME: Fragile, if SVN adds new public properties,
2308 # this needs to be updated.
2309 ++$interesting_props if /^svn:(?:ignore|keywords|executable
2310 |eol-style|mime-type
2311 |externals|needs-lock)$/x;
2312 }
2313 &$sub($self, $p, $props) if $interesting_props;
2314 foreach (sort keys %$dirent) {
2316 next if $dirent->{$_}->{kind} != $SVN::Node::dir;
2317 $self->prop_walk($self->{path} . $p . $_, $rev, $sub);
2318 }
2319}
2320sub last_rev { ($_[0]->last_rev_commit)[0] }
2322sub last_commit { ($_[0]->last_rev_commit)[1] }
2323# returns the newest SVN revision number and newest commit SHA1
2325sub last_rev_commit {
2326 my ($self) = @_;
2327 if (defined $self->{last_rev} && defined $self->{last_commit}) {
2328 return ($self->{last_rev}, $self->{last_commit});
2329 }
2330 my $c = ::verify_ref($self->refname.'^0');
2331 if ($c && !$self->use_svm_props && !$self->no_metadata) {
2332 my $rev = (::cmt_metadata($c))[1];
2333 if (defined $rev) {
2334 ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
2335 return ($rev, $c);
2336 }
2337 }
2338 my $map_path = $self->map_path;
2339 unless (-e $map_path) {
2340 ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
2341 return (undef, undef);
2342 }
2343 my ($rev, $commit) = $self->rev_map_max(1);
2344 ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
2345 return ($rev, $commit);
2346}
2347sub get_fetch_range {
2349 my ($self, $min, $max) = @_;
2350 $max ||= $self->ra->get_latest_revnum;
2351 $min ||= $self->rev_map_max;
2352 (++$min, $max);
2353}
2354sub tmp_config {
2356 my (@args) = @_;
2357 my $old_def_config = "$ENV{GIT_DIR}/svn/config";
2358 my $config = "$ENV{GIT_DIR}/svn/.metadata";
2359 if (! -f $config && -f $old_def_config) {
2360 rename $old_def_config, $config or
2361 die "Failed rename $old_def_config => $config: $!\n";
2362 }
2363 my $old_config = $ENV{GIT_CONFIG};
2364 $ENV{GIT_CONFIG} = $config;
2365 $@ = undef;
2366 my @ret = eval {
2367 unless (-f $config) {
2368 mkfile($config);
2369 open my $fh, '>', $config or
2370 die "Can't open $config: $!\n";
2371 print $fh "; This file is used internally by ",
2372 "git-svn\n" or die
2373 "Couldn't write to $config: $!\n";
2374 print $fh "; You should not have to edit it\n" or
2375 die "Couldn't write to $config: $!\n";
2376 close $fh or die "Couldn't close $config: $!\n";
2377 }
2378 command('config', @args);
2379 };
2380 my $err = $@;
2381 if (defined $old_config) {
2382 $ENV{GIT_CONFIG} = $old_config;
2383 } else {
2384 delete $ENV{GIT_CONFIG};
2385 }
2386 die $err if $err;
2387 wantarray ? @ret : $ret[0];
2388}
2389sub tmp_index_do {
2391 my ($self, $sub) = @_;
2392 my $old_index = $ENV{GIT_INDEX_FILE};
2393 $ENV{GIT_INDEX_FILE} = $self->{index};
2394 $@ = undef;
2395 my @ret = eval {
2396 my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
2397 mkpath([$dir]) unless -d $dir;
2398 &$sub;
2399 };
2400 my $err = $@;
2401 if (defined $old_index) {
2402 $ENV{GIT_INDEX_FILE} = $old_index;
2403 } else {
2404 delete $ENV{GIT_INDEX_FILE};
2405 }
2406 die $err if $err;
2407 wantarray ? @ret : $ret[0];
2408}
2409sub assert_index_clean {
2411 my ($self, $treeish) = @_;
2412 $self->tmp_index_do(sub {
2414 command_noisy('read-tree', $treeish) unless -e $self->{index};
2415 my $x = command_oneline('write-tree');
2416 my ($y) = (command(qw/cat-file commit/, $treeish) =~
2417 /^tree ($::sha1)/mo);
2418 return if $y eq $x;
2419 warn "Index mismatch: $y != $x\nrereading $treeish\n";
2421 unlink $self->{index} or die "unlink $self->{index}: $!\n";
2422 command_noisy('read-tree', $treeish);
2423 $x = command_oneline('write-tree');
2424 if ($y ne $x) {
2425 ::fatal "trees ($treeish) $y != $x\n",
2426 "Something is seriously wrong...";
2427 }
2428 });
2429}
2430sub get_commit_parents {
2432 my ($self, $log_entry) = @_;
2433 my (%seen, @ret, @tmp);
2434 # legacy support for 'set-tree'; this is only used by set_tree_cb:
2435 if (my $ip = $self->{inject_parents}) {
2436 if (my $commit = delete $ip->{$log_entry->{revision}}) {
2437 push @tmp, $commit;
2438 }
2439 }
2440 if (my $cur = ::verify_ref($self->refname.'^0')) {
2441 push @tmp, $cur;
2442 }
2443 if (my $ipd = $self->{inject_parents_dcommit}) {
2444 if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2445 push @tmp, @$commit;
2446 }
2447 }
2448 push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2449 while (my $p = shift @tmp) {
2450 next if $seen{$p};
2451 $seen{$p} = 1;
2452 push @ret, $p;
2453 # MAXPARENT is defined to 16 in commit-tree.c:
2454 last if @ret >= 16;
2455 }
2456 if (@tmp) {
2457 die "r$log_entry->{revision}: No room for parents:\n\t",
2458 join("\n\t", @tmp), "\n";
2459 }
2460 @ret;
2461}
2462sub rewrite_root {
2464 my ($self) = @_;
2465 return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2466 my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2467 my $rwr = eval { command_oneline(qw/config --get/, $k) };
2468 if ($rwr) {
2469 $rwr =~ s#/+$##;
2470 if ($rwr !~ m#^[a-z\+]+://#) {
2471 die "$rwr is not a valid URL (key: $k)\n";
2472 }
2473 }
2474 $self->{-rewrite_root} = $rwr;
2475}
2476sub metadata_url {
2478 my ($self) = @_;
2479 ($self->rewrite_root || $self->{url}) .
2480 (length $self->{path} ? '/' . $self->{path} : '');
2481}
2482sub full_url {
2484 my ($self) = @_;
2485 $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2486}
2487sub set_commit_header_env {
2490 my ($log_entry) = @_;
2491 my %env;
2492 foreach my $ned (qw/NAME EMAIL DATE/) {
2493 foreach my $ac (qw/AUTHOR COMMITTER/) {
2494 $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2495 }
2496 }
2497 $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2499 $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2500 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2501 $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2503 ? $log_entry->{commit_name}
2504 : $log_entry->{name};
2505 $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2506 ? $log_entry->{commit_email}
2507 : $log_entry->{email};
2508 \%env;
2509}
2510sub restore_commit_header_env {
2512 my ($env) = @_;
2513 foreach my $ned (qw/NAME EMAIL DATE/) {
2514 foreach my $ac (qw/AUTHOR COMMITTER/) {
2515 my $k = "GIT_${ac}_${ned}";
2516 if (defined $env->{$k}) {
2517 $ENV{$k} = $env->{$k};
2518 } else {
2519 delete $ENV{$k};
2520 }
2521 }
2522 }
2523}
2524sub gc {
2526 command_noisy('gc', '--auto');
2527};
2528sub do_git_commit {
2530 my ($self, $log_entry) = @_;
2531 my $lr = $self->last_rev;
2532 if (defined $lr && $lr >= $log_entry->{revision}) {
2533 die "Last fetched revision of ", $self->refname,
2534 " was r$lr, but we are about to fetch: ",
2535 "r$log_entry->{revision}!\n";
2536 }
2537 if (my $c = $self->rev_map_get($log_entry->{revision})) {
2538 croak "$log_entry->{revision} = $c already exists! ",
2539 "Why are we refetching it?\n";
2540 }
2541 my $old_env = set_commit_header_env($log_entry);
2542 my $tree = $log_entry->{tree};
2543 if (!defined $tree) {
2544 $tree = $self->tmp_index_do(sub {
2545 command_oneline('write-tree') });
2546 }
2547 die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2548 my @exec = ('git', 'commit-tree', $tree);
2550 foreach ($self->get_commit_parents($log_entry)) {
2551 push @exec, '-p', $_;
2552 }
2553 defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2554 or croak $!;
2555 binmode $msg_fh;
2556 # we always get UTF-8 from SVN, but we may want our commits in
2558 # a different encoding.
2559 if (my $enc = Git::config('i18n.commitencoding')) {
2560 require Encode;
2561 Encode::from_to($log_entry->{log}, 'UTF-8', $enc);
2562 }
2563 print $msg_fh $log_entry->{log} or croak $!;
2564 restore_commit_header_env($old_env);
2565 unless ($self->no_metadata) {
2566 print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2567 or croak $!;
2568 }
2569 $msg_fh->flush == 0 or croak $!;
2570 close $msg_fh or croak $!;
2571 chomp(my $commit = do { local $/; <$out_fh> });
2572 close $out_fh or croak $!;
2573 waitpid $pid, 0;
2574 croak $? if $?;
2575 if ($commit !~ /^$::sha1$/o) {
2576 die "Failed to commit, invalid sha1: $commit\n";
2577 }
2578 $self->rev_map_set($log_entry->{revision}, $commit, 1);
2580 $self->{last_rev} = $log_entry->{revision};
2582 $self->{last_commit} = $commit;
2583 print "r$log_entry->{revision}" unless $::_q > 1;
2584 if (defined $log_entry->{svm_revision}) {
2585 print " (\@$log_entry->{svm_revision})" unless $::_q > 1;
2586 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2587 0, $self->svm_uuid);
2588 }
2589 print " = $commit ($self->{ref_id})\n" unless $::_q > 1;
2590 if (--$_gc_nr == 0) {
2591 $_gc_nr = $_gc_period;
2592 gc();
2593 }
2594 return $commit;
2595}
2596sub match_paths {
2598 my ($self, $paths, $r) = @_;
2599 return 1 if $self->{path} eq '';
2600 if (my $path = $paths->{"/$self->{path}"}) {
2601 return ($path->{action} eq 'D') ? 0 : 1;
2602 }
2603 $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
2604 if (grep /$self->{path_regex}/, keys %$paths) {
2605 return 1;
2606 }
2607 my $c = '';
2608 foreach (split m#/#, $self->{path}) {
2609 $c .= "/$_";
2610 next unless ($paths->{$c} &&
2611 ($paths->{$c}->{action} =~ /^[AR]$/));
2612 if ($self->ra->check_path($self->{path}, $r) ==
2613 $SVN::Node::dir) {
2614 return 1;
2615 }
2616 }
2617 return 0;
2618}
2619sub find_parent_branch {
2621 my ($self, $paths, $rev) = @_;
2622 return undef unless $self->follow_parent;
2623 unless (defined $paths) {
2624 my $err_handler = $SVN::Error::handler;
2625 $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2626 $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1,
2627 sub { $paths = $_[0] });
2628 $SVN::Error::handler = $err_handler;
2629 }
2630 return undef unless defined $paths;
2631 # look for a parent from another branch:
2633 my @b_path_components = split m#/#, $self->{path};
2634 my @a_path_components;
2635 my $i;
2636 while (@b_path_components) {
2637 $i = $paths->{'/'.join('/', @b_path_components)};
2638 last if $i && defined $i->{copyfrom_path};
2639 unshift(@a_path_components, pop(@b_path_components));
2640 }
2641 return undef unless defined $i && defined $i->{copyfrom_path};
2642 my $branch_from = $i->{copyfrom_path};
2643 if (@a_path_components) {
2644 print STDERR "branch_from: $branch_from => ";
2645 $branch_from .= '/'.join('/', @a_path_components);
2646 print STDERR $branch_from, "\n";
2647 }
2648 my $r = $i->{copyfrom_rev};
2649 my $repos_root = $self->ra->{repos_root};
2650 my $url = $self->ra->{url};
2651 my $new_url = $url . $branch_from;
2652 print STDERR "Found possible branch point: ",
2653 "$new_url => ", $self->full_url, ", $r\n"
2654 unless $::_q > 1;
2655 $branch_from =~ s#^/##;
2656 my $gs = $self->other_gs($new_url, $url,
2657 $branch_from, $r, $self->{ref_id});
2658 my ($r0, $parent) = $gs->find_rev_before($r, 1);
2659 {
2660 my ($base, $head);
2661 if (!defined $r0 || !defined $parent) {
2662 ($base, $head) = parse_revision_argument(0, $r);
2663 } else {
2664 if ($r0 < $r) {
2665 $gs->ra->get_log([$gs->{path}], $r0 + 1, $r, 1,
2666 0, 1, sub { $base = $_[1] - 1 });
2667 }
2668 }
2669 if (defined $base && $base <= $r) {
2670 $gs->fetch($base, $r);
2671 }
2672 ($r0, $parent) = $gs->find_rev_before($r, 1);
2673 }
2674 if (defined $r0 && defined $parent) {
2675 print STDERR "Found branch parent: ($self->{ref_id}) $parent\n"
2676 unless $::_q > 1;
2677 my $ed;
2678 if ($self->ra->can_do_switch) {
2679 $self->assert_index_clean($parent);
2680 print STDERR "Following parent with do_switch\n"
2681 unless $::_q > 1;
2682 # do_switch works with svn/trunk >= r22312, but that
2683 # is not included with SVN 1.4.3 (the latest version
2684 # at the moment), so we can't rely on it
2685 $self->{last_rev} = $r0;
2686 $self->{last_commit} = $parent;
2687 $ed = SVN::Git::Fetcher->new($self, $gs->{path});
2688 $gs->ra->gs_do_switch($r0, $rev, $gs,
2689 $self->full_url, $ed)
2690 or die "SVN connection failed somewhere...\n";
2691 } elsif ($self->ra->trees_match($new_url, $r0,
2692 $self->full_url, $rev)) {
2693 print STDERR "Trees match:\n",
2694 " $new_url\@$r0\n",
2695 " ${\$self->full_url}\@$rev\n",
2696 "Following parent with no changes\n"
2697 unless $::_q > 1;
2698 $self->tmp_index_do(sub {
2699 command_noisy('read-tree', $parent);
2700 });
2701 $self->{last_commit} = $parent;
2702 } else {
2703 print STDERR "Following parent with do_update\n"
2704 unless $::_q > 1;
2705 $ed = SVN::Git::Fetcher->new($self);
2706 $self->ra->gs_do_update($rev, $rev, $self, $ed)
2707 or die "SVN connection failed somewhere...\n";
2708 }
2709 print STDERR "Successfully followed parent\n" unless $::_q > 1;
2710 return $self->make_log_entry($rev, [$parent], $ed);
2711 }
2712 return undef;
2713}
2714sub do_fetch {
2716 my ($self, $paths, $rev) = @_;
2717 my $ed;
2718 my ($last_rev, @parents);
2719 if (my $lc = $self->last_commit) {
2720 # we can have a branch that was deleted, then re-added
2721 # under the same name but copied from another path, in
2722 # which case we'll have multiple parents (we don't
2723 # want to break the original ref, nor lose copypath info):
2724 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2725 push @{$log_entry->{parents}}, $lc;
2726 return $log_entry;
2727 }
2728 $ed = SVN::Git::Fetcher->new($self);
2729 $last_rev = $self->{last_rev};
2730 $ed->{c} = $lc;
2731 @parents = ($lc);
2732 } else {
2733 $last_rev = $rev;
2734 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2735 return $log_entry;
2736 }
2737 $ed = SVN::Git::Fetcher->new($self);
2738 }
2739 unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2740 die "SVN connection failed somewhere...\n";
2741 }
2742 $self->make_log_entry($rev, \@parents, $ed);
2743}
2744sub mkemptydirs {
2746 my ($self, $r) = @_;
2747 sub scan {
2749 my ($r, $empty_dirs, $line) = @_;
2750 if (defined $r && $line =~ /^r(\d+)$/) {
2751 return 0 if $1 > $r;
2752 } elsif ($line =~ /^ \+empty_dir: (.+)$/) {
2753 $empty_dirs->{$1} = 1;
2754 } elsif ($line =~ /^ \-empty_dir: (.+)$/) {
2755 my @d = grep {m[^\Q$1\E(/|$)]} (keys %$empty_dirs);
2756 delete @$empty_dirs{@d};
2757 }
2758 1; # continue
2759 };
2760 my %empty_dirs = ();
2762 my $gz_file = "$self->{dir}/unhandled.log.gz";
2763 if (-f $gz_file) {
2764 if (!$can_compress) {
2765 warn "Compress::Zlib could not be found; ",
2766 "empty directories in $gz_file will not be read\n";
2767 } else {
2768 my $gz = Compress::Zlib::gzopen($gz_file, "rb") or
2769 die "Unable to open $gz_file: $!\n";
2770 my $line;
2771 while ($gz->gzreadline($line) > 0) {
2772 scan($r, \%empty_dirs, $line) or last;
2773 }
2774 $gz->gzclose;
2775 }
2776 }
2777 if (open my $fh, '<', "$self->{dir}/unhandled.log") {
2779 binmode $fh or croak "binmode: $!";
2780 while (<$fh>) {
2781 scan($r, \%empty_dirs, $_) or last;
2782 }
2783 close $fh;
2784 }
2785 my $strip = qr/\A\Q$self->{path}\E(?:\/|$)/;
2787 foreach my $d (sort keys %empty_dirs) {
2788 $d = uri_decode($d);
2789 $d =~ s/$strip//;
2790 next if -d $d;
2791 if (-e _) {
2792 warn "$d exists but is not a directory\n";
2793 } else {
2794 print "creating empty directory: $d\n";
2795 mkpath([$d]);
2796 }
2797 }
2798}
2799sub get_untracked {
2801 my ($self, $ed) = @_;
2802 my @out;
2803 my $h = $ed->{empty};
2804 foreach (sort keys %$h) {
2805 my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2806 push @out, " $act: " . uri_encode($_);
2807 warn "W: $act: $_\n";
2808 }
2809 foreach my $t (qw/dir_prop file_prop/) {
2810 $h = $ed->{$t} or next;
2811 foreach my $path (sort keys %$h) {
2812 my $ppath = $path eq '' ? '.' : $path;
2813 foreach my $prop (sort keys %{$h->{$path}}) {
2814 next if $SKIP_PROP{$prop};
2815 my $v = $h->{$path}->{$prop};
2816 my $t_ppath_prop = "$t: " .
2817 uri_encode($ppath) . ' ' .
2818 uri_encode($prop);
2819 if (defined $v) {
2820 push @out, " +$t_ppath_prop " .
2821 uri_encode($v);
2822 } else {
2823 push @out, " -$t_ppath_prop";
2824 }
2825 }
2826 }
2827 }
2828 foreach my $t (qw/absent_file absent_directory/) {
2829 $h = $ed->{$t} or next;
2830 foreach my $parent (sort keys %$h) {
2831 foreach my $path (sort @{$h->{$parent}}) {
2832 push @out, " $t: " .
2833 uri_encode("$parent/$path");
2834 warn "W: $t: $parent/$path ",
2835 "Insufficient permissions?\n";
2836 }
2837 }
2838 }
2839 \@out;
2840}
2841# parse_svn_date(DATE)
2843# --------------------
2844# Given a date (in UTC) from Subversion, return a string in the format
2845# "<TZ Offset> <local date/time>" that Git will use.
2846#
2847# By default the parsed date will be in UTC; if $Git::SVN::_localtime
2848# is true we'll convert it to the local timezone instead.
2849sub parse_svn_date {
2850 my $date = shift || return '+0000 1970-01-01 00:00:00';
2851 my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2852 (\d\d)\:(\d\d)\:(\d\d)\.\d*Z$/x) or
2853 croak "Unable to parse date: $date\n";
2854 my $parsed_date; # Set next.
2855 if ($Git::SVN::_localtime) {
2857 # Translate the Subversion datetime to an epoch time.
2858 # Begin by switching ourselves to $date's timezone, UTC.
2859 my $old_env_TZ = $ENV{TZ};
2860 $ENV{TZ} = 'UTC';
2861 my $epoch_in_UTC =
2863 POSIX::strftime('%s', $S, $M, $H, $d, $m - 1, $Y - 1900);
2864 # Determine our local timezone (including DST) at the
2866 # time of $epoch_in_UTC. $Git::SVN::Log::TZ stored the
2867 # value of TZ, if any, at the time we were run.
2868 if (defined $Git::SVN::Log::TZ) {
2869 $ENV{TZ} = $Git::SVN::Log::TZ;
2870 } else {
2871 delete $ENV{TZ};
2872 }
2873 my $our_TZ =
2875 POSIX::strftime('%Z', $S, $M, $H, $d, $m - 1, $Y - 1900);
2876 # This converts $epoch_in_UTC into our local timezone.
2878 my ($sec, $min, $hour, $mday, $mon, $year,
2879 $wday, $yday, $isdst) = localtime($epoch_in_UTC);
2880 $parsed_date = sprintf('%s %04d-%02d-%02d %02d:%02d:%02d',
2882 $our_TZ, $year + 1900, $mon + 1,
2883 $mday, $hour, $min, $sec);
2884 # Reset us to the timezone in effect when we entered
2886 # this routine.
2887 if (defined $old_env_TZ) {
2888 $ENV{TZ} = $old_env_TZ;
2889 } else {
2890 delete $ENV{TZ};
2891 }
2892 } else {
2893 $parsed_date = "+0000 $Y-$m-$d $H:$M:$S";
2894 }
2895 return $parsed_date;
2897}
2898sub other_gs {
2900 my ($self, $new_url, $url,
2901 $branch_from, $r, $old_ref_id) = @_;
2902 my $gs = Git::SVN->find_by_url($new_url, $url, $branch_from);
2903 unless ($gs) {
2904 my $ref_id = $old_ref_id;
2905 $ref_id =~ s/\@\d+$//;
2906 $ref_id .= "\@$r";
2907 # just grow a tail if we're not unique enough :x
2908 $ref_id .= '-' while find_ref($ref_id);
2909 print STDERR "Initializing parent: $ref_id\n" unless $::_q > 1;
2910 my ($u, $p, $repo_id) = ($new_url, '', $ref_id);
2911 if ($u =~ s#^\Q$url\E(/|$)##) {
2912 $p = $u;
2913 $u = $url;
2914 $repo_id = $self->{repo_id};
2915 }
2916 $gs = Git::SVN->init($u, $p, $repo_id, $ref_id, 1);
2917 }
2918 $gs
2919}
2920sub call_authors_prog {
2922 my ($orig_author) = @_;
2923 $orig_author = command_oneline('rev-parse', '--sq-quote', $orig_author);
2924 my $author = `$::_authors_prog $orig_author`;
2925 if ($? != 0) {
2926 die "$::_authors_prog failed with exit code $?\n"
2927 }
2928 if ($author =~ /^\s*(.+?)\s*<(.*)>\s*$/) {
2929 my ($name, $email) = ($1, $2);
2930 $email = undef if length $2 == 0;
2931 return [$name, $email];
2932 } else {
2933 die "Author: $orig_author: $::_authors_prog returned "
2934 . "invalid author format: $author\n";
2935 }
2936}
2937sub check_author {
2939 my ($author) = @_;
2940 if (!defined $author || length $author == 0) {
2941 $author = '(no author)';
2942 }
2943 if (!defined $::users{$author}) {
2944 if (defined $::_authors_prog) {
2945 $::users{$author} = call_authors_prog($author);
2946 } elsif (defined $::_authors) {
2947 die "Author: $author not defined in $::_authors file\n";
2948 }
2949 }
2950 $author;
2951}
2952sub find_extra_svk_parents {
2954 my ($self, $ed, $tickets, $parents) = @_;
2955 # aha! svk:merge property changed...
2956 my @tickets = split "\n", $tickets;
2957 my @known_parents;
2958 for my $ticket ( @tickets ) {
2959 my ($uuid, $path, $rev) = split /:/, $ticket;
2960 if ( $uuid eq $self->ra_uuid ) {
2961 my $url = $self->rewrite_root || $self->{url};
2962 my $repos_root = $url;
2963 my $branch_from = $path;
2964 $branch_from =~ s{^/}{};
2965 my $gs = $self->other_gs($repos_root."/".$branch_from,
2966 $url,
2967 $branch_from,
2968 $rev,
2969 $self->{ref_id});
2970 if ( my $commit = $gs->rev_map_get($rev, $uuid) ) {
2971 # wahey! we found it, but it might be
2972 # an old one (!)
2973 push @known_parents, [ $rev, $commit ];
2974 }
2975 }
2976 }
2977 # Ordering matters; highest-numbered commit merge tickets
2978 # first, as they may account for later merge ticket additions
2979 # or changes.
2980 @known_parents = map {$_->[1]} sort {$b->[0] <=> $a->[0]} @known_parents;
2981 for my $parent ( @known_parents ) {
2982 my @cmd = ('rev-list', $parent, map { "^$_" } @$parents );
2983 my ($msg_fh, $ctx) = command_output_pipe(@cmd);
2984 my $new;
2985 while ( <$msg_fh> ) {
2986 $new=1;last;
2987 }
2988 command_close_pipe($msg_fh, $ctx);
2989 if ( $new ) {
2990 print STDERR
2991 "Found merge parent (svk:merge ticket): $parent\n";
2992 push @$parents, $parent;
2993 }
2994 }
2995}
2996# note: this function should only be called if the various dirprops
2998# have actually changed
2999sub find_extra_svn_parents {
3000 my ($self, $ed, $mergeinfo, $parents) = @_;
3001 # aha! svk:merge property changed...
3002 # We first search for merged tips which are not in our
3004 # history. Then, we figure out which git revisions are in
3005 # that tip, but not this revision. If all of those revisions
3006 # are now marked as merge, we can add the tip as a parent.
3007 my @merges = split "\n", $mergeinfo;
3008 my @merge_tips;
3009 my @merged_commit_ranges;
3010 my $url = $self->rewrite_root || $self->{url};
3011 for my $merge ( @merges ) {
3012 my ($source, $revs) = split ":", $merge;
3013 my $path = $source;
3014 $path =~ s{^/}{};
3015 my $gs = Git::SVN->find_by_url($url.$source, $url, $path);
3016 if ( !$gs ) {
3017 warn "Couldn't find revmap for $url$source\n";
3018 next;
3019 }
3020 my @ranges = split ",", $revs;
3021 my ($tip, $tip_commit);
3022 # find the tip
3023 for my $range ( @ranges ) {
3024 my ($bottom, $top) = split "-", $range;
3025 $top ||= $bottom;
3026 my $bottom_commit =
3027 $gs->rev_map_get($bottom, $self->ra_uuid) ||
3028 $gs->rev_map_get($bottom+1, $self->ra_uuid);
3029 my $top_commit;
3030 for (; !$top_commit && $top >= $bottom; --$top) {
3031 $top_commit =
3032 $gs->rev_map_get($top, $self->ra_uuid);
3033 }
3034 unless ($top_commit and $bottom_commit) {
3036 warn "W:unknown path/rev in svn:mergeinfo "
3037 ."dirprop: $source:$range\n";
3038 next;
3039 }
3040 push @merged_commit_ranges,
3042 "$bottom_commit..$top_commit";
3043 if ( !defined $tip or $top > $tip ) {
3045 $tip = $top;
3046 $tip_commit = $top_commit;
3047 }
3048 }
3049 unless (!$tip_commit or
3050 grep { $_ eq $tip_commit } @$parents ) {
3051 push @merge_tips, $tip_commit;
3052 } else {
3053 push @merge_tips, undef;
3054 }
3055 }
3056 for my $merge_tip ( @merge_tips ) {
3057 my $spec = shift @merges;
3058 next unless $merge_tip;
3059 my @cmd = ('rev-list', "-1", $merge_tip,
3060 "--not", @$parents );
3061 my ($msg_fh, $ctx) = command_output_pipe(@cmd);
3062 my $new;
3063 while ( <$msg_fh> ) {
3064 $new=1;last;
3065 }
3066 command_close_pipe($msg_fh, $ctx);
3067 if ( $new ) {
3068 push @cmd, @merged_commit_ranges;
3069 my ($msg_fh, $ctx) = command_output_pipe(@cmd);
3070 my $unmerged;
3071 while ( <$msg_fh> ) {
3072 $unmerged=1;last;
3073 }
3074 command_close_pipe($msg_fh, $ctx);
3075 if ( $unmerged ) {
3076 warn "W:svn cherry-pick ignored ($spec)\n";
3077 } else {
3078 warn
3079 "Found merge parent (svn:mergeinfo prop): ",
3080 $merge_tip, "\n";
3081 push @$parents, $merge_tip;
3082 }
3083 }
3084 }
3085}
3086sub make_log_entry {
3088 my ($self, $rev, $parents, $ed) = @_;
3089 my $untracked = $self->get_untracked($ed);
3090 my @parents = @$parents;
3092 my $ps = $ed->{path_strip} || "";
3093 for my $path ( grep { m/$ps/ } %{$ed->{dir_prop}} ) {
3094 my $props = $ed->{dir_prop}{$path};
3095 if ( $props->{"svk:merge"} ) {
3096 $self->find_extra_svk_parents
3097 ($ed, $props->{"svk:merge"}, \@parents);
3098 }
3099 if ( $props->{"svn:mergeinfo"} ) {
3100 $self->find_extra_svn_parents
3101 ($ed,
3102 $props->{"svn:mergeinfo"},
3103 \@parents);
3104 }
3105 }
3106 open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
3108 print $un "r$rev\n" or croak $!;
3109 print $un $_, "\n" foreach @$untracked;
3110 my %log_entry = ( parents => \@parents, revision => $rev,
3111 log => '');
3112 my $headrev;
3114 my $logged = delete $self->{logged_rev_props};
3115 if (!$logged || $self->{-want_revprops}) {
3116 my $rp = $self->ra->rev_proplist($rev);
3117 foreach (sort keys %$rp) {
3118 my $v = $rp->{$_};
3119 if (/^svn:(author|date|log)$/) {
3120 $log_entry{$1} = $v;
3121 } elsif ($_ eq 'svm:headrev') {
3122 $headrev = $v;
3123 } else {
3124 print $un " rev_prop: ", uri_encode($_), ' ',
3125 uri_encode($v), "\n";
3126 }
3127 }
3128 } else {
3129 map { $log_entry{$_} = $logged->{$_} } keys %$logged;
3130 }
3131 close $un or croak $!;
3132 $log_entry{date} = parse_svn_date($log_entry{date});
3134 $log_entry{log} .= "\n";
3135 my $author = $log_entry{author} = check_author($log_entry{author});
3136 my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
3137 : ($author, undef);
3138 my ($commit_name, $commit_email) = ($name, $email);
3140 if ($_use_log_author) {
3141 my $name_field;
3142 if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
3143 $name_field = $1;
3144 } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
3145 $name_field = $1;
3146 }
3147 if (!defined $name_field) {
3148 if (!defined $email) {
3149 $email = $name;
3150 }
3151 } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
3152 ($name, $email) = ($1, $2);
3153 } elsif ($name_field =~ /(.*)@/) {
3154 ($name, $email) = ($1, $name_field);
3155 } else {
3156 ($name, $email) = ($name_field, $name_field);
3157 }
3158 }
3159 if (defined $headrev && $self->use_svm_props) {
3160 if ($self->rewrite_root) {
3161 die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
3162 "options set!\n";
3163 }
3164 my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$}i;
3165 # we don't want "SVM: initializing mirror for junk" ...
3166 return undef if $r == 0;
3167 my $svm = $self->svm;
3168 if ($uuid ne $svm->{uuid}) {
3169 die "UUID mismatch on SVM path:\n",
3170 "expected: $svm->{uuid}\n",
3171 " got: $uuid\n";
3172 }
3173 my $full_url = $self->full_url;
3174 $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
3175 die "Failed to replace '$svm->{replace}' with ",
3176 "'$svm->{source}' in $full_url\n";
3177 # throw away username for storing in records
3178 remove_username($full_url);
3179 $log_entry{metadata} = "$full_url\@$r $uuid";
3180 $log_entry{svm_revision} = $r;
3181 $email ||= "$author\@$uuid";
3182 $commit_email ||= "$author\@$uuid";
3183 } elsif ($self->use_svnsync_props) {
3184 my $full_url = $self->svnsync->{url};
3185 $full_url .= "/$self->{path}" if length $self->{path};
3186 remove_username($full_url);
3187 my $uuid = $self->svnsync->{uuid};
3188 $log_entry{metadata} = "$full_url\@$rev $uuid";
3189 $email ||= "$author\@$uuid";
3190 $commit_email ||= "$author\@$uuid";
3191 } else {
3192 my $url = $self->metadata_url;
3193 remove_username($url);
3194 $log_entry{metadata} = "$url\@$rev " .
3195 $self->ra->get_uuid;
3196 $email ||= "$author\@" . $self->ra->get_uuid;
3197 $commit_email ||= "$author\@" . $self->ra->get_uuid;
3198 }
3199 $log_entry{name} = $name;
3200 $log_entry{email} = $email;
3201 $log_entry{commit_name} = $commit_name;
3202 $log_entry{commit_email} = $commit_email;
3203 \%log_entry;
3204}
3205sub fetch {
3207 my ($self, $min_rev, $max_rev, @parents) = @_;
3208 my ($last_rev, $last_commit) = $self->last_rev_commit;
3209 my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
3210 $self->ra->gs_fetch_loop_common($base, $head, [$self]);
3211}
3212sub set_tree_cb {
3214 my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
3215 $self->{inject_parents} = { $rev => $tree };
3216 $self->fetch(undef, undef);
3217}
3218sub set_tree {
3220 my ($self, $tree) = (shift, shift);
3221 my $log_entry = ::get_commit_entry($tree);
3222 unless ($self->{last_rev}) {
3223 ::fatal("Must have an existing revision to commit");
3224 }
3225 my %ed_opts = ( r => $self->{last_rev},
3226 log => $log_entry->{log},
3227 ra => $self->ra,
3228 tree_a => $self->{last_commit},
3229 tree_b => $tree,
3230 editor_cb => sub {
3231 $self->set_tree_cb($log_entry, $tree, @_) },
3232 svn_path => $self->{path} );
3233 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
3234 print "No changes\nr$self->{last_rev} = $tree\n";
3235 }
3236}
3237sub rebuild_from_rev_db {
3239 my ($self, $path) = @_;
3240 my $r = -1;
3241 open my $fh, '<', $path or croak "open: $!";
3242 binmode $fh or croak "binmode: $!";
3243 while (<$fh>) {
3244 length($_) == 41 or croak "inconsistent size in ($_) != 41";
3245 chomp($_);
3246 ++$r;
3247 next if $_ eq ('0' x 40);
3248 $self->rev_map_set($r, $_);
3249 print "r$r = $_\n";
3250 }
3251 close $fh or croak "close: $!";
3252 unlink $path or croak "unlink: $!";
3253}
3254sub rebuild {
3256 my ($self) = @_;
3257 my $map_path = $self->map_path;
3258 my $partial = (-e $map_path && ! -z $map_path);
3259 return unless ::verify_ref($self->refname.'^0');
3260 if (!$partial && ($self->use_svm_props || $self->no_metadata)) {
3261 my $rev_db = $self->rev_db_path;
3262 $self->rebuild_from_rev_db($rev_db);
3263 if ($self->use_svm_props) {
3264 my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
3265 $self->rebuild_from_rev_db($svm_rev_db);
3266 }
3267 $self->unlink_rev_db_symlink;
3268 return;
3269 }
3270 print "Rebuilding $map_path ...\n" if (!$partial);
3271 my ($base_rev, $head) = ($partial ? $self->rev_map_max_norebuild(1) :
3272 (undef, undef));
3273 my ($log, $ctx) =
3274 command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
3275 ($head ? "$head.." : "") . $self->refname,
3276 '--');
3277 my $metadata_url = $self->metadata_url;
3278 remove_username($metadata_url);
3279 my $svn_uuid = $self->ra_uuid;
3280 my $c;
3281 while (<$log>) {
3282 if ( m{^commit ($::sha1)$} ) {
3283 $c = $1;
3284 next;
3285 }
3286 next unless s{^\s*(git-svn-id:)}{$1};
3287 my ($url, $rev, $uuid) = ::extract_metadata($_);
3288 remove_username($url);
3289 # ignore merges (from set-tree)
3291 next if (!defined $rev || !$uuid);
3292 # if we merged or otherwise started elsewhere, this is
3294 # how we break out of it
3295 if (($uuid ne $svn_uuid) ||
3296 ($metadata_url && $url && ($url ne $metadata_url))) {
3297 next;
3298 }
3299 if ($partial && $head) {
3300 print "Partial-rebuilding $map_path ...\n";
3301 print "Currently at $base_rev = $head\n";
3302 $head = undef;
3303 }
3304 $self->rev_map_set($rev, $c);
3306 print "r$rev = $c\n";
3307 }
3308 command_close_pipe($log, $ctx);
3309 print "Done rebuilding $map_path\n" if (!$partial || !$head);
3310 my $rev_db_path = $self->rev_db_path;
3311 if (-f $self->rev_db_path) {
3312 unlink $self->rev_db_path or croak "unlink: $!";
3313 }
3314 $self->unlink_rev_db_symlink;
3315}
3316# rev_map:
3318# Tie::File seems to be prone to offset errors if revisions get sparse,
3319# it's not that fast, either. Tie::File is also not in Perl 5.6. So
3320# one of my favorite modules is out :< Next up would be one of the DBM
3321# modules, but I'm not sure which is most portable...
3322#
3323# This is the replacement for the rev_db format, which was too big
3324# and inefficient for large repositories with a lot of sparse history
3325# (mainly tags)
3326#
3327# The format is this:
3328# - 24 bytes for every record,
3329# * 4 bytes for the integer representing an SVN revision number
3330# * 20 bytes representing the sha1 of a git commit
3331# - No empty padding records like the old format
3332# (except the last record, which can be overwritten)
3333# - new records are written append-only since SVN revision numbers
3334# increase monotonically
3335# - lookups on SVN revision number are done via a binary search
3336# - Piping the file to xxd -c24 is a good way of dumping it for
3337# viewing or editing (piped back through xxd -r), should the need
3338# ever arise.
3339# - The last record can be padding revision with an all-zero sha1
3340# This is used to optimize fetch performance when using multiple
3341# "fetch" directives in .git/config
3342#
3343# These files are disposable unless noMetadata or useSvmProps is set
3344sub _rev_map_set {
3346 my ($fh, $rev, $commit) = @_;
3347 binmode $fh or croak "binmode: $!";
3349 my $size = (stat($fh))[7];
3350 ($size % 24) == 0 or croak "inconsistent size: $size";
3351 my $wr_offset = 0;
3353 if ($size > 0) {
3354 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
3355 my $read = sysread($fh, my $buf, 24) or croak "read: $!";
3356 $read == 24 or croak "read only $read bytes (!= 24)";
3357 my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
3358 if ($last_commit eq ('0' x40)) {
3359 if ($size >= 48) {
3360 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
3361 $read = sysread($fh, $buf, 24) or
3362 croak "read: $!";
3363 $read == 24 or
3364 croak "read only $read bytes (!= 24)";
3365 ($last_rev, $last_commit) =
3366 unpack(rev_map_fmt, $buf);
3367 if ($last_commit eq ('0' x40)) {
3368 croak "inconsistent .rev_map\n";
3369 }
3370 }
3371 if ($last_rev >= $rev) {
3372 croak "last_rev is higher!: $last_rev >= $rev";
3373 }
3374 $wr_offset = -24;
3375 }
3376 }
3377 sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
3378 syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
3379 croak "write: $!";
3380}
3381sub _rev_map_reset {
3383 my ($fh, $rev, $commit) = @_;
3384 my $c = _rev_map_get($fh, $rev);
3385 $c eq $commit or die "_rev_map_reset(@_) commit $c does not match!\n";
3386 my $offset = sysseek($fh, 0, SEEK_CUR) or croak "seek: $!";
3387 truncate $fh, $offset or croak "truncate: $!";
3388}
3389sub mkfile {
3391 my ($path) = @_;
3392 unless (-e $path) {
3393 my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
3394 mkpath([$dir]) unless -d $dir;
3395 open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
3396 close $fh or die "Couldn't close (create) $path: $!\n";
3397 }
3398}
3399sub rev_map_set {
3401 my ($self, $rev, $commit, $update_ref, $uuid) = @_;
3402 length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
3403 my $db = $self->map_path($uuid);
3404 my $db_lock = "$db.lock";
3405 my $sig;
3406 $update_ref ||= 0;
3407 if ($update_ref) {
3408 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
3409 $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
3410 }
3411 mkfile($db);
3412 $LOCKFILES{$db_lock} = 1;
3414 my $sync;
3415 # both of these options make our .rev_db file very, very important
3416 # and we can't afford to lose it because rebuild() won't work
3417 if ($self->use_svm_props || $self->no_metadata) {
3418 $sync = 1;
3419 copy($db, $db_lock) or die "rev_map_set(@_): ",
3420 "Failed to copy: ",
3421 "$db => $db_lock ($!)\n";
3422 } else {
3423 rename $db, $db_lock or die "rev_map_set(@_): ",
3424 "Failed to rename: ",
3425 "$db => $db_lock ($!)\n";
3426 }
3427 sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
3429 or croak "Couldn't open $db_lock: $!\n";
3430 $update_ref eq 'reset' ? _rev_map_reset($fh, $rev, $commit) :
3431 _rev_map_set($fh, $rev, $commit);
3432 if ($sync) {
3433 $fh->flush or die "Couldn't flush $db_lock: $!\n";
3434 $fh->sync or die "Couldn't sync $db_lock: $!\n";
3435 }
3436 close $fh or croak $!;
3437 if ($update_ref) {
3438 $_head = $self;
3439 my $note = "";
3440 $note = " ($update_ref)" if ($update_ref !~ /^\d*$/);
3441 command_noisy('update-ref', '-m', "r$rev$note",
3442 $self->refname, $commit);
3443 }
3444 rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
3445 "$db_lock => $db ($!)\n";
3446 delete $LOCKFILES{$db_lock};
3447 if ($update_ref) {
3448 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
3449 $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
3450 kill $sig, $$ if defined $sig;
3451 }
3452}
3453# If want_commit, this will return an array of (rev, commit) where
3455# commit _must_ be a valid commit in the archive.
3456# Otherwise, it'll return the max revision (whether or not the
3457# commit is valid or just a 0x40 placeholder).
3458sub rev_map_max {
3459 my ($self, $want_commit) = @_;
3460 $self->rebuild;
3461 my ($r, $c) = $self->rev_map_max_norebuild($want_commit);
3462 $want_commit ? ($r, $c) : $r;
3463}
3464sub rev_map_max_norebuild {
3466 my ($self, $want_commit) = @_;
3467 my $map_path = $self->map_path;
3468 stat $map_path or return $want_commit ? (0, undef) : 0;
3469 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
3470 binmode $fh or croak "binmode: $!";
3471 my $size = (stat($fh))[7];
3472 ($size % 24) == 0 or croak "inconsistent size: $size";
3473 if ($size == 0) {
3475 close $fh or croak "close: $!";
3476 return $want_commit ? (0, undef) : 0;
3477 }
3478 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
3480 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
3481 my ($r, $c) = unpack(rev_map_fmt, $buf);
3482 if ($want_commit && $c eq ('0' x40)) {
3483 if ($size < 48) {
3484 return $want_commit ? (0, undef) : 0;
3485 }
3486 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
3487 sysread($fh, $buf, 24) == 24 or croak "read: $!";
3488 ($r, $c) = unpack(rev_map_fmt, $buf);
3489 if ($c eq ('0'x40)) {
3490 croak "Penultimate record is all-zeroes in $map_path";
3491 }
3492 }
3493 close $fh or croak "close: $!";
3494 $want_commit ? ($r, $c) : $r;
3495}
3496sub rev_map_get {
3498 my ($self, $rev, $uuid) = @_;
3499 my $map_path = $self->map_path($uuid);
3500 return undef unless -e $map_path;
3501 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
3503 my $c = _rev_map_get($fh, $rev);
3504 close($fh) or croak "close: $!";
3505 $c
3506}
3507sub _rev_map_get {
3509 my ($fh, $rev) = @_;
3510 binmode $fh or croak "binmode: $!";
3512 my $size = (stat($fh))[7];
3513 ($size % 24) == 0 or croak "inconsistent size: $size";
3514 if ($size == 0) {
3516 return undef;
3517 }
3518 my ($l, $u) = (0, $size - 24);
3520 my ($r, $c, $buf);
3521 while ($l <= $u) {
3523 my $i = int(($l/24 + $u/24) / 2) * 24;
3524 sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
3525 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
3526 my ($r, $c) = unpack(rev_map_fmt, $buf);
3527 if ($r < $rev) {
3529 $l = $i + 24;
3530 } elsif ($r > $rev) {
3531 $u = $i - 24;
3532 } else { # $r == $rev
3533 return $c eq ('0' x 40) ? undef : $c;
3534 }
3535 }
3536 undef;
3537}
3538# Finds the first svn revision that exists on (if $eq_ok is true) or
3540# before $rev for the current branch. It will not search any lower
3541# than $min_rev. Returns the git commit hash and svn revision number
3542# if found, else (undef, undef).
3543sub find_rev_before {
3544 my ($self, $rev, $eq_ok, $min_rev) = @_;
3545 --$rev unless $eq_ok;
3546 $min_rev ||= 1;
3547 my $max_rev = $self->rev_map_max;
3548 $rev = $max_rev if ($rev > $max_rev);
3549 while ($rev >= $min_rev) {
3550 if (my $c = $self->rev_map_get($rev)) {
3551 return ($rev, $c);
3552 }
3553 --$rev;
3554 }
3555 return (undef, undef);
3556}
3557# Finds the first svn revision that exists on (if $eq_ok is true) or
3559# after $rev for the current branch. It will not search any higher
3560# than $max_rev. Returns the git commit hash and svn revision number
3561# if found, else (undef, undef).
3562sub find_rev_after {
3563 my ($self, $rev, $eq_ok, $max_rev) = @_;
3564 ++$rev unless $eq_ok;
3565 $max_rev ||= $self->rev_map_max;
3566 while ($rev <= $max_rev) {
3567 if (my $c = $self->rev_map_get($rev)) {
3568 return ($rev, $c);
3569 }
3570 ++$rev;
3571 }
3572 return (undef, undef);
3573}
3574sub _new {
3576 my ($class, $repo_id, $ref_id, $path) = @_;
3577 unless (defined $repo_id && length $repo_id) {
3578 $repo_id = $Git::SVN::default_repo_id;
3579 }
3580 unless (defined $ref_id && length $ref_id) {
3581 $_prefix = '' unless defined($_prefix);
3582 $_[2] = $ref_id =
3583 "refs/remotes/$_prefix$Git::SVN::default_ref_id";
3584 }
3585 $_[1] = $repo_id;
3586 my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
3587 # Older repos imported by us used $GIT_DIR/svn/foo instead of
3589 # $GIT_DIR/svn/refs/remotes/foo when tracking refs/remotes/foo
3590 if ($ref_id =~ m{^refs/remotes/(.*)}) {
3591 my $old_dir = "$ENV{GIT_DIR}/svn/$1";
3592 if (-d $old_dir && ! -d $dir) {
3593 $dir = $old_dir;
3594 }
3595 }
3596 $_[3] = $path = '' unless (defined $path);
3598 mkpath([$dir]);
3599 bless {
3600 ref_id => $ref_id, dir => $dir, index => "$dir/index",
3601 path => $path, config => "$ENV{GIT_DIR}/svn/config",
3602 map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
3603}
3604# for read-only access of old .rev_db formats
3606sub unlink_rev_db_symlink {
3607 my ($self) = @_;
3608 my $link = $self->rev_db_path;
3609 $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
3610 if (-l $link) {
3611 unlink $link or croak "unlink: $link failed!";
3612 }
3613}
3614sub rev_db_path {
3616 my ($self, $uuid) = @_;
3617 my $db_path = $self->map_path($uuid);
3618 $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
3619 or croak "map_path: $db_path does not contain '/.rev_map.' !";
3620 $db_path;
3621}
3622# the new replacement for .rev_db
3624sub map_path {
3625 my ($self, $uuid) = @_;
3626 $uuid ||= $self->ra_uuid;
3627 "$self->{map_root}.$uuid";
3628}
3629sub uri_encode {
3631 my ($f) = @_;
3632 $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
3633 $f
3634}
3635sub uri_decode {
3637 my ($f) = @_;
3638 $f =~ s#%([0-9a-fA-F]{2})#chr(hex($1))#eg;
3639 $f
3640}
3641sub remove_username {
3643 $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
3644}
3645package Git::SVN::Prompt;
3647use strict;
3648use warnings;
3649require SVN::Core;
3650use vars qw/$_no_auth_cache $_username/;
3651sub simple {
3653 my ($cred, $realm, $default_username, $may_save, $pool) = @_;
3654 $may_save = undef if $_no_auth_cache;
3655 $default_username = $_username if defined $_username;
3656 if (defined $default_username && length $default_username) {
3657 if (defined $realm && length $realm) {
3658 print STDERR "Authentication realm: $realm\n";
3659 STDERR->flush;
3660 }
3661 $cred->username($default_username);
3662 } else {
3663 username($cred, $realm, $may_save, $pool);
3664 }
3665 $cred->password(_read_password("Password for '" .
3666 $cred->username . "': ", $realm));
3667 $cred->may_save($may_save);
3668 $SVN::_Core::SVN_NO_ERROR;
3669}
3670sub ssl_server_trust {
3672 my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
3673 $may_save = undef if $_no_auth_cache;
3674 print STDERR "Error validating server certificate for '$realm':\n";
3675 {
3676 no warnings 'once';
3677 # All variables SVN::Auth::SSL::* are used only once,
3678 # so we're shutting up Perl warnings about this.
3679 if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
3680 print STDERR " - The certificate is not issued ",
3681 "by a trusted authority. Use the\n",
3682 " fingerprint to validate ",
3683 "the certificate manually!\n";
3684 }
3685 if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
3686 print STDERR " - The certificate hostname ",
3687 "does not match.\n";
3688 }
3689 if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
3690 print STDERR " - The certificate is not yet valid.\n";
3691 }
3692 if ($failures & $SVN::Auth::SSL::EXPIRED) {
3693 print STDERR " - The certificate has expired.\n";
3694 }
3695 if ($failures & $SVN::Auth::SSL::OTHER) {
3696 print STDERR " - The certificate has ",
3697 "an unknown error.\n";
3698 }
3699 } # no warnings 'once'
3700 printf STDERR
3701 "Certificate information:\n".
3702 " - Hostname: %s\n".
3703 " - Valid: from %s until %s\n".
3704 " - Issuer: %s\n".
3705 " - Fingerprint: %s\n",
3706 map $cert_info->$_, qw(hostname valid_from valid_until
3707 issuer_dname fingerprint);
3708 my $choice;
3709prompt:
3710 print STDERR $may_save ?
3711 "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
3712 "(R)eject or accept (t)emporarily? ";
3713 STDERR->flush;
3714 $choice = lc(substr(<STDIN> || 'R', 0, 1));
3715 if ($choice =~ /^t$/i) {
3716 $cred->may_save(undef);
3717 } elsif ($choice =~ /^r$/i) {
3718 return -1;
3719 } elsif ($may_save && $choice =~ /^p$/i) {
3720 $cred->may_save($may_save);
3721 } else {
3722 goto prompt;
3723 }
3724 $cred->accepted_failures($failures);
3725 $SVN::_Core::SVN_NO_ERROR;
3726}
3727sub ssl_client_cert {
3729 my ($cred, $realm, $may_save, $pool) = @_;
3730 $may_save = undef if $_no_auth_cache;
3731 print STDERR "Client certificate filename: ";
3732 STDERR->flush;
3733 chomp(my $filename = <STDIN>);
3734 $cred->cert_file($filename);
3735 $cred->may_save($may_save);
3736 $SVN::_Core::SVN_NO_ERROR;
3737}
3738sub ssl_client_cert_pw {
3740 my ($cred, $realm, $may_save, $pool) = @_;
3741 $may_save = undef if $_no_auth_cache;
3742 $cred->password(_read_password("Password: ", $realm));
3743 $cred->may_save($may_save);
3744 $SVN::_Core::SVN_NO_ERROR;
3745}
3746sub username {
3748 my ($cred, $realm, $may_save, $pool) = @_;
3749 $may_save = undef if $_no_auth_cache;
3750 if (defined $realm && length $realm) {
3751 print STDERR "Authentication realm: $realm\n";
3752 }
3753 my $username;
3754 if (defined $_username) {
3755 $username = $_username;
3756 } else {
3757 print STDERR "Username: ";
3758 STDERR->flush;
3759 chomp($username = <STDIN>);
3760 }
3761 $cred->username($username);
3762 $cred->may_save($may_save);
3763 $SVN::_Core::SVN_NO_ERROR;
3764}
3765sub _read_password {
3767 my ($prompt, $realm) = @_;
3768 print STDERR $prompt;
3769 STDERR->flush;
3770 require Term::ReadKey;
3771 Term::ReadKey::ReadMode('noecho');
3772 my $password = '';
3773 while (defined(my $key = Term::ReadKey::ReadKey(0))) {
3774 last if $key =~ /[\012\015]/; # \n\r
3775 $password .= $key;
3776 }
3777 Term::ReadKey::ReadMode('restore');
3778 print STDERR "\n";
3779 STDERR->flush;
3780 $password;
3781}
3782package SVN::Git::Fetcher;
3784use vars qw/@ISA/;
3785use strict;
3786use warnings;
3787use Carp qw/croak/;
3788use File::Temp qw/tempfile/;
3789use IO::File qw//;
3790use vars qw/$_ignore_regex/;
3791# file baton members: path, mode_a, mode_b, pool, fh, blob, base
3793sub new {
3794 my ($class, $git_svn, $switch_path) = @_;
3795 my $self = SVN::Delta::Editor->new;
3796 bless $self, $class;
3797 if (exists $git_svn->{last_commit}) {
3798 $self->{c} = $git_svn->{last_commit};
3799 $self->{empty_symlinks} =
3800 _mark_empty_symlinks($git_svn, $switch_path);
3801 }
3802 $self->{ignore_regex} = eval { command_oneline('config', '--get',
3803 "svn-remote.$git_svn->{repo_id}.ignore-paths") };
3804 $self->{empty} = {};
3805 $self->{dir_prop} = {};
3806 $self->{file_prop} = {};
3807 $self->{absent_dir} = {};
3808 $self->{absent_file} = {};
3809 $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
3810 $self;
3811}
3812# this uses the Ra object, so it must be called before do_{switch,update},
3814# not inside them (when the Git::SVN::Fetcher object is passed) to
3815# do_{switch,update}
3816sub _mark_empty_symlinks {
3817 my ($git_svn, $switch_path) = @_;
3818 my $bool = Git::config_bool('svn.brokenSymlinkWorkaround');
3819 return {} if (!defined($bool)) || (defined($bool) && ! $bool);
3820 my %ret;
3822 my ($rev, $cmt) = $git_svn->last_rev_commit;
3823 return {} unless ($rev && $cmt);
3824 # allow the warning to be printed for each revision we fetch to
3826 # ensure the user sees it. The user can also disable the workaround
3827 # on the repository even while git svn is running and the next
3828 # revision fetched will skip this expensive function.
3829 my $printed_warning;
3830 chomp(my $empty_blob = `git hash-object -t blob --stdin < /dev/null`);
3831 my ($ls, $ctx) = command_output_pipe(qw/ls-tree -r -z/, $cmt);
3832 local $/ = "\0";
3833 my $pfx = defined($switch_path) ? $switch_path : $git_svn->{path};
3834 $pfx .= '/' if length($pfx);
3835 while (<$ls>) {
3836 chomp;
3837 s/\A100644 blob $empty_blob\t//o or next;
3838 unless ($printed_warning) {
3839 print STDERR "Scanning for empty symlinks, ",
3840 "this may take a while if you have ",
3841 "many empty files\n",
3842 "You may disable this with `",
3843 "git config svn.brokenSymlinkWorkaround ",
3844 "false'.\n",
3845 "This may be done in a different ",
3846 "terminal without restarting ",
3847 "git svn\n";
3848 $printed_warning = 1;
3849 }
3850 my $path = $_;
3851 my (undef, $props) =
3852 $git_svn->ra->get_file($pfx.$path, $rev, undef);
3853 if ($props->{'svn:special'}) {
3854 $ret{$path} = 1;
3855 }
3856 }
3857 command_close_pipe($ls, $ctx);
3858 \%ret;
3859}
3860# returns true if a given path is inside a ".git" directory
3862sub in_dot_git {
3863 $_[0] =~ m{(?:^|/)\.git(?:/|$)};
3864}
3865# return value: 0 -- don't ignore, 1 -- ignore
3867sub is_path_ignored {
3868 my ($self, $path) = @_;
3869 return 1 if in_dot_git($path);
3870 return 1 if defined($self->{ignore_regex}) &&
3871 $path =~ m!$self->{ignore_regex}!;
3872 return 0 unless defined($_ignore_regex);
3873 return 1 if $path =~ m!$_ignore_regex!o;
3874 return 0;
3875}
3876sub set_path_strip {
3878 my ($self, $path) = @_;
3879 $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
3880}
3881sub open_root {
3883 { path => '' };
3884}
3885sub open_directory {
3887 my ($self, $path, $pb, $rev) = @_;
3888 { path => $path };
3889}
3890sub git_path {
3892 my ($self, $path) = @_;
3893 if ($self->{path_strip}) {
3894 $path =~ s!$self->{path_strip}!! or
3895 die "Failed to strip path '$path' ($self->{path_strip})\n";
3896 }
3897 $path;
3898}
3899sub delete_entry {
3901 my ($self, $path, $rev, $pb) = @_;
3902 return undef if $self->is_path_ignored($path);
3903 my $gpath = $self->git_path($path);
3905 return undef if ($gpath eq '');
3906 # remove entire directories.
3908 my ($tree) = (command('ls-tree', '-z', $self->{c}, "./$gpath")
3909 =~ /\A040000 tree ([a-f\d]{40})\t\Q$gpath\E\0/);
3910 if ($tree) {
3911 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3912 -r --name-only -z/,
3913 $tree);
3914 local $/ = "\0";
3915 while (<$ls>) {
3916 chomp;
3917 my $rmpath = "$gpath/$_";
3918 $self->{gii}->remove($rmpath);
3919 print "\tD\t$rmpath\n" unless $::_q;
3920 }
3921 print "\tD\t$gpath/\n" unless $::_q;
3922 command_close_pipe($ls, $ctx);
3923 } else {
3924 $self->{gii}->remove($gpath);
3925 print "\tD\t$gpath\n" unless $::_q;
3926 }
3927 $self->{empty}->{$path} = 0;
3928 undef;
3929}
3930sub open_file {
3932 my ($self, $path, $pb, $rev) = @_;
3933 my ($mode, $blob);
3934 goto out if $self->is_path_ignored($path);
3936 my $gpath = $self->git_path($path);
3938 ($mode, $blob) = (command('ls-tree', '-z', $self->{c}, "./$gpath")
3939 =~ /\A(\d{6}) blob ([a-f\d]{40})\t\Q$gpath\E\0/);
3940 unless (defined $mode && defined $blob) {
3941 die "$path was not found in commit $self->{c} (r$rev)\n";
3942 }
3943 if ($mode eq '100644' && $self->{empty_symlinks}->{$path}) {
3944 $mode = '120000';
3945 }
3946out:
3947 { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3948 pool => SVN::Pool->new, action => 'M' };
3949}
3950sub add_file {
3952 my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3953 my $mode;
3954 if (!$self->is_path_ignored($path)) {
3956 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3957 delete $self->{empty}->{$dir};
3958 $mode = '100644';
3959 }
3960 { path => $path, mode_a => $mode, mode_b => $mode,
3961 pool => SVN::Pool->new, action => 'A' };
3962}
3963sub add_directory {
3965 my ($self, $path, $cp_path, $cp_rev) = @_;
3966 goto out if $self->is_path_ignored($path);
3967 my $gpath = $self->git_path($path);
3968 if ($gpath eq '') {
3969 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3970 -r --name-only -z/,
3971 $self->{c});
3972 local $/ = "\0";
3973 while (<$ls>) {
3974 chomp;
3975 $self->{gii}->remove($_);
3976 print "\tD\t$_\n" unless $::_q;
3977 }
3978 command_close_pipe($ls, $ctx);
3979 $self->{empty}->{$path} = 0;
3980 }
3981 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3982 delete $self->{empty}->{$dir};
3983 $self->{empty}->{$path} = 1;
3984out:
3985 { path => $path };
3986}
3987sub change_dir_prop {
3989 my ($self, $db, $prop, $value) = @_;
3990 return undef if $self->is_path_ignored($db->{path});
3991 $self->{dir_prop}->{$db->{path}} ||= {};
3992 $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3993 undef;
3994}
3995sub absent_directory {
3997 my ($self, $path, $pb) = @_;
3998 return undef if $self->is_path_ignored($path);
3999 $self->{absent_dir}->{$pb->{path}} ||= [];
4000 push @{$self->{absent_dir}->{$pb->{path}}}, $path;
4001 undef;
4002}
4003sub absent_file {
4005 my ($self, $path, $pb) = @_;
4006 return undef if $self->is_path_ignored($path);
4007 $self->{absent_file}->{$pb->{path}} ||= [];
4008 push @{$self->{absent_file}->{$pb->{path}}}, $path;
4009 undef;
4010}
4011sub change_file_prop {
4013 my ($self, $fb, $prop, $value) = @_;
4014 return undef if $self->is_path_ignored($fb->{path});
4015 if ($prop eq 'svn:executable') {
4016 if ($fb->{mode_b} != 120000) {
4017 $fb->{mode_b} = defined $value ? 100755 : 100644;
4018 }
4019 } elsif ($prop eq 'svn:special') {
4020 $fb->{mode_b} = defined $value ? 120000 : 100644;
4021 } else {
4022 $self->{file_prop}->{$fb->{path}} ||= {};
4023 $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
4024 }
4025 undef;
4026}
4027sub apply_textdelta {
4029 my ($self, $fb, $exp) = @_;
4030 return undef if $self->is_path_ignored($fb->{path});
4031 my $fh = $::_repository->temp_acquire('svn_delta');
4032 # $fh gets auto-closed() by SVN::TxDelta::apply(),
4033 # (but $base does not,) so dup() it for reading in close_file
4034 open my $dup, '<&', $fh or croak $!;
4035 my $base = $::_repository->temp_acquire('git_blob');
4036 if ($fb->{blob}) {
4038 my ($base_is_link, $size);
4039 if ($fb->{mode_a} eq '120000' &&
4041 ! $self->{empty_symlinks}->{$fb->{path}}) {
4042 print $base 'link ' or die "print $!\n";
4043 $base_is_link = 1;
4044 }
4045 retry:
4046 $size = $::_repository->cat_blob($fb->{blob}, $base);
4047 die "Failed to read object $fb->{blob}" if ($size < 0);
4048 if (defined $exp) {
4050 seek $base, 0, 0 or croak $!;
4051 my $got = ::md5sum($base);
4052 if ($got ne $exp) {
4053 my $err = "Checksum mismatch: ".
4054 "$fb->{path} $fb->{blob}\n" .
4055 "expected: $exp\n" .
4056 " got: $got\n";
4057 if ($base_is_link) {
4058 warn $err,
4059 "Retrying... (possibly ",
4060 "a bad symlink from SVN)\n";
4061 $::_repository->temp_reset($base);
4062 $base_is_link = 0;
4063 goto retry;
4064 }
4065 die $err;
4066 }
4067 }
4068 }
4069 seek $base, 0, 0 or croak $!;
4070 $fb->{fh} = $fh;
4071 $fb->{base} = $base;
4072 [ SVN::TxDelta::apply($base, $dup, undef, $fb->{path}, $fb->{pool}) ];
4073}
4074sub close_file {
4076 my ($self, $fb, $exp) = @_;
4077 return undef if $self->is_path_ignored($fb->{path});
4078 my $hash;
4080 my $path = $self->git_path($fb->{path});
4081 if (my $fh = $fb->{fh}) {
4082 if (defined $exp) {
4083 seek($fh, 0, 0) or croak $!;
4084 my $got = ::md5sum($fh);
4085 if ($got ne $exp) {
4086 die "Checksum mismatch: $path\n",
4087 "expected: $exp\n got: $got\n";
4088 }
4089 }
4090 if ($fb->{mode_b} == 120000) {
4091 sysseek($fh, 0, 0) or croak $!;
4092 my $rd = sysread($fh, my $buf, 5);
4093 if (!defined $rd) {
4095 croak "sysread: $!\n";
4096 } elsif ($rd == 0) {
4097 warn "$path has mode 120000",
4098 " but it points to nothing\n",
4099 "converting to an empty file with mode",
4100 " 100644\n";
4101 $fb->{mode_b} = '100644';
4102 } elsif ($buf ne 'link ') {
4103 warn "$path has mode 120000",
4104 " but is not a link\n";
4105 } else {
4106 my $tmp_fh = $::_repository->temp_acquire(
4107 'svn_hash');
4108 my $res;
4109 while ($res = sysread($fh, my $str, 1024)) {
4110 my $out = syswrite($tmp_fh, $str, $res);
4111 defined($out) && $out == $res
4112 or croak("write ",
4113 Git::temp_path($tmp_fh),
4114 ": $!\n");
4115 }
4116 defined $res or croak $!;
4117 ($fh, $tmp_fh) = ($tmp_fh, $fh);
4119 Git::temp_release($tmp_fh, 1);
4120 }
4121 }
4122 $hash = $::_repository->hash_and_insert_object(
4124 Git::temp_path($fh));
4125 $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
4126 Git::temp_release($fb->{base}, 1);
4128 Git::temp_release($fh, 1);
4129 } else {
4130 $hash = $fb->{blob} or die "no blob information\n";
4131 }
4132 $fb->{pool}->clear;
4133 $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
4134 print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
4135 undef;
4136}
4137sub abort_edit {
4139 my $self = shift;
4140 $self->{nr} = $self->{gii}->{nr};
4141 delete $self->{gii};
4142 $self->SUPER::abort_edit(@_);
4143}
4144sub close_edit {
4146 my $self = shift;
4147 $self->{git_commit_ok} = 1;
4148 $self->{nr} = $self->{gii}->{nr};
4149 delete $self->{gii};
4150 $self->SUPER::close_edit(@_);
4151}
4152package SVN::Git::Editor;
4154use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
4155use strict;
4156use warnings;
4157use Carp qw/croak/;
4158use IO::File;
4159sub new {
4161 my ($class, $opts) = @_;
4162 foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
4163 die "$_ required!\n" unless (defined $opts->{$_});
4164 }
4165 my $pool = SVN::Pool->new;
4167 my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
4168 my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
4169 $opts->{r}, $mods);
4170 # $opts->{ra} functions should not be used after this:
4172 my @ce = $opts->{ra}->get_commit_editor($opts->{log},
4173 $opts->{editor_cb}, $pool);
4174 my $self = SVN::Delta::Editor->new(@ce, $pool);
4175 bless $self, $class;
4176 foreach (qw/svn_path r tree_a tree_b/) {
4177 $self->{$_} = $opts->{$_};
4178 }
4179 $self->{url} = $opts->{ra}->{url};
4180 $self->{mods} = $mods;
4181 $self->{types} = $types;
4182 $self->{pool} = $pool;
4183 $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
4184 $self->{rm} = { };
4185 $self->{path_prefix} = length $self->{svn_path} ?
4186 "$self->{svn_path}/" : '';
4187 $self->{config} = $opts->{config};
4188 return $self;
4189}
4190sub generate_diff {
4192 my ($tree_a, $tree_b) = @_;
4193 my @diff_tree = qw(diff-tree -z -r);
4194 if ($_cp_similarity) {
4195 push @diff_tree, "-C$_cp_similarity";
4196 } else {
4197 push @diff_tree, '-C';
4198 }
4199 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
4200 push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
4201 push @diff_tree, $tree_a, $tree_b;
4202 my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
4203 local $/ = "\0";
4204 my $state = 'meta';
4205 my @mods;
4206 while (<$diff_fh>) {
4207 chomp $_; # this gets rid of the trailing "\0"
4208 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
4209 ($::sha1)\s($::sha1)\s
4210 ([MTCRAD])\d*$/xo) {
4211 push @mods, { mode_a => $1, mode_b => $2,
4212 sha1_a => $3, sha1_b => $4,
4213 chg => $5 };
4214 if ($5 =~ /^(?:C|R)$/) {
4215 $state = 'file_a';
4216 } else {
4217 $state = 'file_b';
4218 }
4219 } elsif ($state eq 'file_a') {
4220 my $x = $mods[$#mods] or croak "Empty array\n";
4221 if ($x->{chg} !~ /^(?:C|R)$/) {
4222 croak "Error parsing $_, $x->{chg}\n";
4223 }
4224 $x->{file_a} = $_;
4225 $state = 'file_b';
4226 } elsif ($state eq 'file_b') {
4227 my $x = $mods[$#mods] or croak "Empty array\n";
4228 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
4229 croak "Error parsing $_, $x->{chg}\n";
4230 }
4231 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
4232 croak "Error parsing $_, $x->{chg}\n";
4233 }
4234 $x->{file_b} = $_;
4235 $state = 'meta';
4236 } else {
4237 croak "Error parsing $_\n";
4238 }
4239 }
4240 command_close_pipe($diff_fh, $ctx);
4241 \@mods;
4242}
4243sub check_diff_paths {
4245 my ($ra, $pfx, $rev, $mods) = @_;
4246 my %types;
4247 $pfx .= '/' if length $pfx;
4248 sub type_diff_paths {
4250 my ($ra, $types, $path, $rev) = @_;
4251 my @p = split m#/+#, $path;
4252 my $c = shift @p;
4253 unless (defined $types->{$c}) {
4254 $types->{$c} = $ra->check_path($c, $rev);
4255 }
4256 while (@p) {
4257 $c .= '/' . shift @p;
4258 next if defined $types->{$c};
4259 $types->{$c} = $ra->check_path($c, $rev);
4260 }
4261 }
4262 foreach my $m (@$mods) {
4264 foreach my $f (qw/file_a file_b/) {
4265 next unless defined $m->{$f};
4266 my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
4267 if (length $pfx.$dir && ! defined $types{$dir}) {
4268 type_diff_paths($ra, \%types, $pfx.$dir, $rev);
4269 }
4270 }
4271 }
4272 \%types;
4273}
4274sub split_path {
4276 return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
4277}
4278sub repo_path {
4280 my ($self, $path) = @_;
4281 $self->{path_prefix}.(defined $path ? $path : '');
4282}
4283sub url_path {
4285 my ($self, $path) = @_;
4286 if ($self->{url} =~ m#^https?://#) {
4287 $path =~ s!([^~a-zA-Z0-9_./-])!uc sprintf("%%%02x",ord($1))!eg;
4288 }
4289 $self->{url} . '/' . $self->repo_path($path);
4290}
4291sub rmdirs {
4293 my ($self) = @_;
4294 my $rm = $self->{rm};
4295 delete $rm->{''}; # we never delete the url we're tracking
4296 return unless %$rm;
4297 foreach (keys %$rm) {
4299 my @d = split m#/#, $_;
4300 my $c = shift @d;
4301 $rm->{$c} = 1;
4302 while (@d) {
4303 $c .= '/' . shift @d;
4304 $rm->{$c} = 1;
4305 }
4306 }
4307 delete $rm->{$self->{svn_path}};
4308 delete $rm->{''}; # we never delete the url we're tracking
4309 return unless %$rm;
4310 my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
4312 $self->{tree_b});
4313 local $/ = "\0";
4314 while (<$fh>) {
4315 chomp;
4316 my @dn = split m#/#, $_;
4317 while (pop @dn) {
4318 delete $rm->{join '/', @dn};
4319 }
4320 unless (%$rm) {
4321 close $fh;
4322 return;
4323 }
4324 }
4325 command_close_pipe($fh, $ctx);
4326 my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
4328 foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
4329 $self->close_directory($bat->{$d}, $p);
4330 my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
4331 print "\tD+\t$d/\n" unless $::_q;
4332 $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
4333 delete $bat->{$d};
4334 }
4335}
4336sub open_or_add_dir {
4338 my ($self, $full_path, $baton) = @_;
4339 my $t = $self->{types}->{$full_path};
4340 if (!defined $t) {
4341 die "$full_path not known in r$self->{r} or we have a bug!\n";
4342 }
4343 {
4344 no warnings 'once';
4345 # SVN::Node::none and SVN::Node::file are used only once,
4346 # so we're shutting up Perl's warnings about them.
4347 if ($t == $SVN::Node::none) {
4348 return $self->add_directory($full_path, $baton,
4349 undef, -1, $self->{pool});
4350 } elsif ($t == $SVN::Node::dir) {
4351 return $self->open_directory($full_path, $baton,
4352 $self->{r}, $self->{pool});
4353 } # no warnings 'once'
4354 print STDERR "$full_path already exists in repository at ",
4355 "r$self->{r} and it is not a directory (",
4356 ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
4357 } # no warnings 'once'
4358 exit 1;
4359}
4360sub ensure_path {
4362 my ($self, $path) = @_;
4363 my $bat = $self->{bat};
4364 my $repo_path = $self->repo_path($path);
4365 return $bat->{''} unless (length $repo_path);
4366 my @p = split m#/+#, $repo_path;
4367 my $c = shift @p;
4368 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
4369 while (@p) {
4370 my $c0 = $c;
4371 $c .= '/' . shift @p;
4372 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
4373 }
4374 return $bat->{$c};
4375}
4376# Subroutine to convert a globbing pattern to a regular expression.
4378# From perl cookbook.
4379sub glob2pat {
4380 my $globstr = shift;
4381 my %patmap = ('*' => '.*', '?' => '.', '[' => '[', ']' => ']');
4382 $globstr =~ s{(.)} { $patmap{$1} || "\Q$1" }ge;
4383 return '^' . $globstr . '$';
4384}
4385sub check_autoprop {
4387 my ($self, $pattern, $properties, $file, $fbat) = @_;
4388 # Convert the globbing pattern to a regular expression.
4389 my $regex = glob2pat($pattern);
4390 # Check if the pattern matches the file name.
4391 if($file =~ m/($regex)/) {
4392 # Parse the list of properties to set.
4393 my @props = split(/;/, $properties);
4394 foreach my $prop (@props) {
4395 # Parse 'name=value' syntax and set the property.
4396 if ($prop =~ /([^=]+)=(.*)/) {
4397 my ($n,$v) = ($1,$2);
4398 for ($n, $v) {
4399 s/^\s+//; s/\s+$//;
4400 }
4401 $self->change_file_prop($fbat, $n, $v);
4402 }
4403 }
4404 }
4405}
4406sub apply_autoprops {
4408 my ($self, $file, $fbat) = @_;
4409 my $conf_t = ${$self->{config}}{'config'};
4410 no warnings 'once';
4411 # Check [miscellany]/enable-auto-props in svn configuration.
4412 if (SVN::_Core::svn_config_get_bool(
4413 $conf_t,
4414 $SVN::_Core::SVN_CONFIG_SECTION_MISCELLANY,
4415 $SVN::_Core::SVN_CONFIG_OPTION_ENABLE_AUTO_PROPS,
4416 0)) {
4417 # Auto-props are enabled. Enumerate them to look for matches.
4418 my $callback = sub {
4419 $self->check_autoprop($_[0], $_[1], $file, $fbat);
4420 };
4421 SVN::_Core::svn_config_enumerate(
4422 $conf_t,
4423 $SVN::_Core::SVN_CONFIG_SECTION_AUTO_PROPS,
4424 $callback);
4425 }
4426}
4427sub A {
4429 my ($self, $m) = @_;
4430 my ($dir, $file) = split_path($m->{file_b});
4431 my $pbat = $self->ensure_path($dir);
4432 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
4433 undef, -1);
4434 print "\tA\t$m->{file_b}\n" unless $::_q;
4435 $self->apply_autoprops($file, $fbat);
4436 $self->chg_file($fbat, $m);
4437 $self->close_file($fbat,undef,$self->{pool});
4438}
4439sub C {
4441 my ($self, $m) = @_;
4442 my ($dir, $file) = split_path($m->{file_b});
4443 my $pbat = $self->ensure_path($dir);
4444 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
4445 $self->url_path($m->{file_a}), $self->{r});
4446 print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
4447 $self->chg_file($fbat, $m);
4448 $self->close_file($fbat,undef,$self->{pool});
4449}
4450sub delete_entry {
4452 my ($self, $path, $pbat) = @_;
4453 my $rpath = $self->repo_path($path);
4454 my ($dir, $file) = split_path($rpath);
4455 $self->{rm}->{$dir} = 1;
4456 $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
4457}
4458sub R {
4460 my ($self, $m) = @_;
4461 my ($dir, $file) = split_path($m->{file_b});
4462 my $pbat = $self->ensure_path($dir);
4463 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
4464 $self->url_path($m->{file_a}), $self->{r});
4465 print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
4466 $self->apply_autoprops($file, $fbat);
4467 $self->chg_file($fbat, $m);
4468 $self->close_file($fbat,undef,$self->{pool});
4469 ($dir, $file) = split_path($m->{file_a});
4471 $pbat = $self->ensure_path($dir);
4472 $self->delete_entry($m->{file_a}, $pbat);
4473}
4474sub M {
4476 my ($self, $m) = @_;
4477 my ($dir, $file) = split_path($m->{file_b});
4478 my $pbat = $self->ensure_path($dir);
4479 my $fbat = $self->open_file($self->repo_path($m->{file_b}),
4480 $pbat,$self->{r},$self->{pool});
4481 print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
4482 $self->chg_file($fbat, $m);
4483 $self->close_file($fbat,undef,$self->{pool});
4484}
4485sub T { shift->M(@_) }
4487sub change_file_prop {
4489 my ($self, $fbat, $pname, $pval) = @_;
4490 $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
4491}
4492sub _chg_file_get_blob ($$$$) {
4494 my ($self, $fbat, $m, $which) = @_;
4495 my $fh = $::_repository->temp_acquire("git_blob_$which");
4496 if ($m->{"mode_$which"} =~ /^120/) {
4497 print $fh 'link ' or croak $!;
4498 $self->change_file_prop($fbat,'svn:special','*');
4499 } elsif ($m->{mode_a} =~ /^120/ && $m->{"mode_$which"} !~ /^120/) {
4500 $self->change_file_prop($fbat,'svn:special',undef);
4501 }
4502 my $blob = $m->{"sha1_$which"};
4503 return ($fh,) if ($blob =~ /^0{40}$/);
4504 my $size = $::_repository->cat_blob($blob, $fh);
4505 croak "Failed to read object $blob" if ($size < 0);
4506 $fh->flush == 0 or croak $!;
4507 seek $fh, 0, 0 or croak $!;
4508 my $exp = ::md5sum($fh);
4510 seek $fh, 0, 0 or croak $!;
4511 return ($fh, $exp);
4512}
4513sub chg_file {
4515 my ($self, $fbat, $m) = @_;
4516 if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
4517 $self->change_file_prop($fbat,'svn:executable','*');
4518 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
4519 $self->change_file_prop($fbat,'svn:executable',undef);
4520 }
4521 my ($fh_a, $exp_a) = _chg_file_get_blob $self, $fbat, $m, 'a';
4522 my ($fh_b, $exp_b) = _chg_file_get_blob $self, $fbat, $m, 'b';
4523 my $pool = SVN::Pool->new;
4524 my $atd = $self->apply_textdelta($fbat, $exp_a, $pool);
4525 if (-s $fh_a) {
4526 my $txstream = SVN::TxDelta::new ($fh_a, $fh_b, $pool);
4527 my $res = SVN::TxDelta::send_txstream($txstream, @$atd, $pool);
4528 if (defined $res) {
4529 die "Unexpected result from send_txstream: $res\n",
4530 "(SVN::Core::VERSION: $SVN::Core::VERSION)\n";
4531 }
4532 } else {
4533 my $got = SVN::TxDelta::send_stream($fh_b, @$atd, $pool);
4534 die "Checksum mismatch\nexpected: $exp_b\ngot: $got\n"
4535 if ($got ne $exp_b);
4536 }
4537 Git::temp_release($fh_b, 1);
4538 Git::temp_release($fh_a, 1);
4539 $pool->clear;
4540}
4541sub D {
4543 my ($self, $m) = @_;
4544 my ($dir, $file) = split_path($m->{file_b});
4545 my $pbat = $self->ensure_path($dir);
4546 print "\tD\t$m->{file_b}\n" unless $::_q;
4547 $self->delete_entry($m->{file_b}, $pbat);
4548}
4549sub close_edit {
4551 my ($self) = @_;
4552 my ($p,$bat) = ($self->{pool}, $self->{bat});
4553 foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
4554 next if $_ eq '';
4555 $self->close_directory($bat->{$_}, $p);
4556 }
4557 $self->close_directory($bat->{''}, $p);
4558 $self->SUPER::close_edit($p);
4559 $p->clear;
4560}
4561sub abort_edit {
4563 my ($self) = @_;
4564 $self->SUPER::abort_edit($self->{pool});
4565}
4566sub DESTROY {
4568 my $self = shift;
4569 $self->SUPER::DESTROY(@_);
4570 $self->{pool}->clear;
4571}
4572# this drives the editor
4574sub apply_diff {
4575 my ($self) = @_;
4576 my $mods = $self->{mods};
4577 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
4578 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
4579 my $f = $m->{chg};
4580 if (defined $o{$f}) {
4581 $self->$f($m);
4582 } else {
4583 fatal("Invalid change type: $f");
4584 }
4585 }
4586 $self->rmdirs if $_rmdir;
4587 if (@$mods == 0) {
4588 $self->abort_edit;
4589 } else {
4590 $self->close_edit;
4591 }
4592 return scalar @$mods;
4593}
4594package Git::SVN::Ra;
4596use vars qw/@ISA $config_dir $_log_window_size/;
4597use strict;
4598use warnings;
4599my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
4600BEGIN {
4602 # enforce temporary pool usage for some simple functions
4603 no strict 'refs';
4604 for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root
4605 get_file/) {
4606 my $SUPER = "SUPER::$f";
4607 *$f = sub {
4608 my $self = shift;
4609 my $pool = SVN::Pool->new;
4610 my @ret = $self->$SUPER(@_,$pool);
4611 $pool->clear;
4612 wantarray ? @ret : $ret[0];
4613 };
4614 }
4615}
4616sub _auth_providers () {
4618 [
4619 SVN::Client::get_simple_provider(),
4620 SVN::Client::get_ssl_server_trust_file_provider(),
4621 SVN::Client::get_simple_prompt_provider(
4622 \&Git::SVN::Prompt::simple, 2),
4623 SVN::Client::get_ssl_client_cert_file_provider(),
4624 SVN::Client::get_ssl_client_cert_prompt_provider(
4625 \&Git::SVN::Prompt::ssl_client_cert, 2),
4626 SVN::Client::get_ssl_client_cert_pw_file_provider(),
4627 SVN::Client::get_ssl_client_cert_pw_prompt_provider(
4628 \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
4629 SVN::Client::get_username_provider(),
4630 SVN::Client::get_ssl_server_trust_prompt_provider(
4631 \&Git::SVN::Prompt::ssl_server_trust),
4632 SVN::Client::get_username_prompt_provider(
4633 \&Git::SVN::Prompt::username, 2)
4634 ]
4635}
4636sub escape_uri_only {
4638 my ($uri) = @_;
4639 my @tmp;
4640 foreach (split m{/}, $uri) {
4641 s/([^~\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
4642 push @tmp, $_;
4643 }
4644 join('/', @tmp);
4645}
4646sub escape_url {
4648 my ($url) = @_;
4649 if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
4650 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
4651 $url = "$scheme://$domain$uri";
4652 }
4653 $url;
4654}
4655sub new {
4657 my ($class, $url) = @_;
4658 $url =~ s!/+$!!;
4659 return $RA if ($RA && $RA->{url} eq $url);
4660 SVN::_Core::svn_config_ensure($config_dir, undef);
4662 my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
4663 my $config = SVN::Core::config_get_config($config_dir);
4664 $RA = undef;
4665 my $dont_store_passwords = 1;
4666 my $conf_t = ${$config}{'config'};
4667 {
4668 no warnings 'once';
4669 # The usage of $SVN::_Core::SVN_CONFIG_* variables
4670 # produces warnings that variables are used only once.
4671 # I had not found the better way to shut them up, so
4672 # the warnings of type 'once' are disabled in this block.
4673 if (SVN::_Core::svn_config_get_bool($conf_t,
4674 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
4675 $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
4676 1) == 0) {
4677 SVN::_Core::svn_auth_set_parameter($baton,
4678 $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
4679 bless (\$dont_store_passwords, "_p_void"));
4680 }
4681 if (SVN::_Core::svn_config_get_bool($conf_t,
4682 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
4683 $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
4684 1) == 0) {
4685 $Git::SVN::Prompt::_no_auth_cache = 1;
4686 }
4687 } # no warnings 'once'
4688 my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
4689 config => $config,
4690 pool => SVN::Pool->new,
4691 auth_provider_callbacks => $callbacks);
4692 $self->{url} = $url;
4693 $self->{svn_path} = $url;
4694 $self->{repos_root} = $self->get_repos_root;
4695 $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
4696 $self->{cache} = { check_path => { r => 0, data => {} },
4697 get_dir => { r => 0, data => {} } };
4698 $RA = bless $self, $class;
4699}
4700sub check_path {
4702 my ($self, $path, $r) = @_;
4703 my $cache = $self->{cache}->{check_path};
4704 if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
4705 return $cache->{data}->{$path};
4706 }
4707 my $pool = SVN::Pool->new;
4708 my $t = $self->SUPER::check_path($path, $r, $pool);
4709 $pool->clear;
4710 if ($r != $cache->{r}) {
4711 %{$cache->{data}} = ();
4712 $cache->{r} = $r;
4713 }
4714 $cache->{data}->{$path} = $t;
4715}
4716sub get_dir {
4718 my ($self, $dir, $r) = @_;
4719 my $cache = $self->{cache}->{get_dir};
4720 if ($r == $cache->{r}) {
4721 if (my $x = $cache->{data}->{$dir}) {
4722 return wantarray ? @$x : $x->[0];
4723 }
4724 }
4725 my $pool = SVN::Pool->new;
4726 my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
4727 my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
4728 $pool->clear;
4729 if ($r != $cache->{r}) {
4730 %{$cache->{data}} = ();
4731 $cache->{r} = $r;
4732 }
4733 $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
4734 wantarray ? (\%dirents, $r, $props) : \%dirents;
4735}
4736sub DESTROY {
4738 # do not call the real DESTROY since we store ourselves in $RA
4739}
4740# get_log(paths, start, end, limit,
4742# discover_changed_paths, strict_node_history, receiver)
4743sub get_log {
4744 my ($self, @args) = @_;
4745 my $pool = SVN::Pool->new;
4746 # svn_log_changed_path_t objects passed to get_log are likely to be
4748 # overwritten even if only the refs are copied to an external variable,
4749 # so we should dup the structures in their entirety. Using an
4750 # externally passed pool (instead of our temporary and quickly cleared
4751 # pool in Git::SVN::Ra) does not help matters at all...
4752 my $receiver = pop @args;
4753 my $prefix = "/".$self->{svn_path};
4754 $prefix =~ s#/+($)##;
4755 my $prefix_regex = qr#^\Q$prefix\E#;
4756 push(@args, sub {
4757 my ($paths) = $_[0];
4758 return &$receiver(@_) unless $paths;
4759 $_[0] = ();
4760 foreach my $p (keys %$paths) {
4761 my $i = $paths->{$p};
4762 # Make path relative to our url, not repos_root
4763 $p =~ s/$prefix_regex//;
4764 my %s = map { $_ => $i->$_; }
4765 qw/copyfrom_path copyfrom_rev action/;
4766 if ($s{'copyfrom_path'}) {
4767 $s{'copyfrom_path'} =~ s/$prefix_regex//;
4768 }
4769 $_[0]{$p} = \%s;
4770 }
4771 &$receiver(@_);
4772 });
4773 # the limit parameter was not supported in SVN 1.1.x, so we
4776 # drop it. Therefore, the receiver callback passed to it
4777 # is made aware of this limitation by being wrapped if
4778 # the limit passed to is being wrapped.
4779 if ($SVN::Core::VERSION le '1.2.0') {
4780 my $limit = splice(@args, 3, 1);
4781 if ($limit > 0) {
4782 my $receiver = pop @args;
4783 push(@args, sub { &$receiver(@_) if (--$limit >= 0) });
4784 }
4785 }
4786 my $ret = $self->SUPER::get_log(@args, $pool);
4787 $pool->clear;
4788 $ret;
4789}
4790sub trees_match {
4792 my ($self, $url1, $rev1, $url2, $rev2) = @_;
4793 my $ctx = SVN::Client->new(auth => _auth_providers);
4794 my $out = IO::File->new_tmpfile;
4795 # older SVN (1.1.x) doesn't take $pool as the last parameter for
4797 # $ctx->diff(), so we'll create a default one
4798 my $pool = SVN::Pool->new_default_sub;
4799 $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
4801 $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
4802 $out->flush;
4803 my $ret = (($out->stat)[7] == 0);
4804 close $out or croak $!;
4805 $ret;
4807}
4808sub get_commit_editor {
4810 my ($self, $log, $cb, $pool) = @_;
4811 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
4812 $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
4813}
4814sub gs_do_update {
4816 my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
4817 my $new = ($rev_a == $rev_b);
4818 my $path = $gs->{path};
4819 if ($new && -e $gs->{index}) {
4821 unlink $gs->{index} or die
4822 "Couldn't unlink index: $gs->{index}: $!\n";
4823 }
4824 my $pool = SVN::Pool->new;
4825 $editor->set_path_strip($path);
4826 my (@pc) = split m#/#, $path;
4827 my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
4828 1, $editor, $pool);
4829 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
4830 # Since we can't rely on svn_ra_reparent being available, we'll
4832 # just have to do some magic with set_path to make it so
4833 # we only want a partial path.
4834 my $sp = '';
4835 my $final = join('/', @pc);
4836 while (@pc) {
4837 $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
4838 $sp .= '/' if length $sp;
4839 $sp .= shift @pc;
4840 }
4841 die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
4842 $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
4844 $reporter->finish_report($pool);
4846 $pool->clear;
4847 $editor->{git_commit_ok};
4848}
4849# this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
4851# svn_ra_reparent didn't work before 1.4)
4852sub gs_do_switch {
4853 my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
4854 my $path = $gs->{path};
4855 my $pool = SVN::Pool->new;
4856 my $full_url = $self->{url};
4858 my $old_url = $full_url;
4859 $full_url .= '/' . $path if length $path;
4860 my ($ra, $reparented);
4861 if ($old_url =~ m#^svn(\+ssh)?://# ||
4863 ($full_url =~ m#^https?://# &&
4864 escape_url($full_url) ne $full_url)) {
4865 $_[0] = undef;
4866 $self = undef;
4867 $RA = undef;
4868 $ra = Git::SVN::Ra->new($full_url);
4869 $ra_invalid = 1;
4870 } elsif ($old_url ne $full_url) {
4871 SVN::_Ra::svn_ra_reparent($self->{session}, $full_url, $pool);
4872 $self->{url} = $full_url;
4873 $reparented = 1;
4874 }
4875 $ra ||= $self;
4877 $url_b = escape_url($url_b);
4878 my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
4879 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
4880 $reporter->set_path('', $rev_a, 0, @lock, $pool);
4881 $reporter->finish_report($pool);
4882 if ($reparented) {
4884 SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
4885 $self->{url} = $old_url;
4886 }
4887 $pool->clear;
4889 $editor->{git_commit_ok};
4890}
4891sub longest_common_path {
4893 my ($gsv, $globs) = @_;
4894 my %common;
4895 my $common_max = scalar @$gsv;
4896 foreach my $gs (@$gsv) {
4898 my @tmp = split m#/#, $gs->{path};
4899 my $p = '';
4900 foreach (@tmp) {
4901 $p .= length($p) ? "/$_" : $_;
4902 $common{$p} ||= 0;
4903 $common{$p}++;
4904 }
4905 }
4906 $globs ||= [];
4907 $common_max += scalar @$globs;
4908 foreach my $glob (@$globs) {
4909 my @tmp = split m#/#, $glob->{path}->{left};
4910 my $p = '';
4911 foreach (@tmp) {
4912 $p .= length($p) ? "/$_" : $_;
4913 $common{$p} ||= 0;
4914 $common{$p}++;
4915 }
4916 }
4917 my $longest_path = '';
4919 foreach (sort {length $b <=> length $a} keys %common) {
4920 if ($common{$_} == $common_max) {
4921 $longest_path = $_;
4922 last;
4923 }
4924 }
4925 $longest_path;
4926}
4927sub gs_fetch_loop_common {
4929 my ($self, $base, $head, $gsv, $globs) = @_;
4930 return if ($base > $head);
4931 my $inc = $_log_window_size;
4932 my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
4933 my $longest_path = longest_common_path($gsv, $globs);
4934 my $ra_url = $self->{url};
4935 my $find_trailing_edge;
4936 while (1) {
4937 my %revs;
4938 my $err;
4939 my $err_handler = $SVN::Error::handler;
4940 $SVN::Error::handler = sub {
4941 ($err) = @_;
4942 skip_unknown_revs($err);
4943 };
4944 sub _cb {
4945 my ($paths, $r, $author, $date, $log) = @_;
4946 [ $paths,
4947 { author => $author, date => $date, log => $log } ];
4948 }
4949 $self->get_log([$longest_path], $min, $max, 0, 1, 1,
4950 sub { $revs{$_[1]} = _cb(@_) });
4951 if ($err) {
4952 print "Checked through r$max\r";
4953 } else {
4954 $find_trailing_edge = 1;
4955 }
4956 if ($err and $find_trailing_edge) {
4957 print STDERR "Path '$longest_path' ",
4958 "was probably deleted:\n",
4959 $err->expanded_message,
4960 "\nWill attempt to follow ",
4961 "revisions r$min .. r$max ",
4962 "committed before the deletion\n";
4963 my $hi = $max;
4964 while (--$hi >= $min) {
4965 my $ok;
4966 $self->get_log([$longest_path], $min, $hi,
4967 0, 1, 1, sub {
4968 $ok = $_[1];
4969 $revs{$_[1]} = _cb(@_) });
4970 if ($ok) {
4971 print STDERR "r$min .. r$ok OK\n";
4972 last;
4973 }
4974 }
4975 $find_trailing_edge = 0;
4976 }
4977 $SVN::Error::handler = $err_handler;
4978 my %exists = map { $_->{path} => $_ } @$gsv;
4980 foreach my $r (sort {$a <=> $b} keys %revs) {
4981 my ($paths, $logged) = @{$revs{$r}};
4982 foreach my $gs ($self->match_globs(\%exists, $paths,
4984 $globs, $r)) {
4985 if ($gs->rev_map_max >= $r) {
4986 next;
4987 }
4988 next unless $gs->match_paths($paths, $r);
4989 $gs->{logged_rev_props} = $logged;
4990 if (my $last_commit = $gs->last_commit) {
4991 $gs->assert_index_clean($last_commit);
4992 }
4993 my $log_entry = $gs->do_fetch($paths, $r);
4994 if ($log_entry) {
4995 $gs->do_git_commit($log_entry);
4996 }
4997 $INDEX_FILES{$gs->{index}} = 1;
4998 }
4999 foreach my $g (@$globs) {
5000 my $k = "svn-remote.$g->{remote}." .
5001 "$g->{t}-maxRev";
5002 Git::SVN::tmp_config($k, $r);
5003 }
5004 if ($ra_invalid) {
5005 $_[0] = undef;
5006 $self = undef;
5007 $RA = undef;
5008 $self = Git::SVN::Ra->new($ra_url);
5009 $ra_invalid = undef;
5010 }
5011 }
5012 # pre-fill the .rev_db since it'll eventually get filled in
5013 # with '0' x40 if something new gets committed
5014 foreach my $gs (@$gsv) {
5015 next if $gs->rev_map_max >= $max;
5016 next if defined $gs->rev_map_get($max);
5017 $gs->rev_map_set($max, 0 x40);
5018 }
5019 foreach my $g (@$globs) {
5020 my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
5021 Git::SVN::tmp_config($k, $max);
5022 }
5023 last if $max >= $head;
5024 $min = $max + 1;
5025 $max += $inc;
5026 $max = $head if ($max > $head);
5027 }
5028 Git::SVN::gc();
5029}
5030sub get_dir_globbed {
5032 my ($self, $left, $depth, $r) = @_;
5033 my @x = eval { $self->get_dir($left, $r) };
5035 return unless scalar @x == 3;
5036 my $dirents = $x[0];
5037 my @finalents;
5038 foreach my $de (keys %$dirents) {
5039 next if $dirents->{$de}->{kind} != $SVN::Node::dir;
5040 if ($depth > 1) {
5041 my @args = ("$left/$de", $depth - 1, $r);
5042 foreach my $dir ($self->get_dir_globbed(@args)) {
5043 push @finalents, "$de/$dir";
5044 }
5045 } else {
5046 push @finalents, $de;
5047 }
5048 }
5049 @finalents;
5050}
5051sub match_globs {
5053 my ($self, $exists, $paths, $globs, $r) = @_;
5054 sub get_dir_check {
5056 my ($self, $exists, $g, $r) = @_;
5057 my @dirs = $self->get_dir_globbed($g->{path}->{left},
5059 $g->{path}->{depth},
5060 $r);
5061 foreach my $de (@dirs) {
5063 my $p = $g->{path}->full_path($de);
5064 next if $exists->{$p};
5065 next if (length $g->{path}->{right} &&
5066 ($self->check_path($p, $r) !=
5067 $SVN::Node::dir));
5068 $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
5069 $g->{ref}->full_path($de), 1);
5070 }
5071 }
5072 foreach my $g (@$globs) {
5073 if (my $path = $paths->{"/$g->{path}->{left}"}) {
5074 if ($path->{action} =~ /^[AR]$/) {
5075 get_dir_check($self, $exists, $g, $r);
5076 }
5077 }
5078 foreach (keys %$paths) {
5079 if (/$g->{path}->{left_regex}/ &&
5080 !/$g->{path}->{regex}/) {
5081 next if $paths->{$_}->{action} !~ /^[AR]$/;
5082 get_dir_check($self, $exists, $g, $r);
5083 }
5084 next unless /$g->{path}->{regex}/;
5085 my $p = $1;
5086 my $pathname = $g->{path}->full_path($p);
5087 next if $exists->{$pathname};
5088 next if ($self->check_path($pathname, $r) !=
5089 $SVN::Node::dir);
5090 $exists->{$pathname} = Git::SVN->init(
5091 $self->{url}, $pathname, undef,
5092 $g->{ref}->full_path($p), 1);
5093 }
5094 my $c = '';
5095 foreach (split m#/#, $g->{path}->{left}) {
5096 $c .= "/$_";
5097 next unless ($paths->{$c} &&
5098 ($paths->{$c}->{action} =~ /^[AR]$/));
5099 get_dir_check($self, $exists, $g, $r);
5100 }
5101 }
5102 values %$exists;
5103}
5104sub minimize_url {
5106 my ($self) = @_;
5107 return $self->{url} if ($self->{url} eq $self->{repos_root});
5108 my $url = $self->{repos_root};
5109 my @components = split(m!/!, $self->{svn_path});
5110 my $c = '';
5111 do {
5112 $url .= "/$c" if length $c;
5113 eval {
5114 my $ra = (ref $self)->new($url);
5115 my $latest = $ra->get_latest_revnum;
5116 $ra->get_log("", $latest, 0, 1, 0, 1, sub {});
5117 };
5118 } while ($@ && ($c = shift @components));
5119 $url;
5120}
5121sub can_do_switch {
5123 my $self = shift;
5124 unless (defined $can_do_switch) {
5125 my $pool = SVN::Pool->new;
5126 my $rep = eval {
5127 $self->do_switch(1, '', 0, $self->{url},
5128 SVN::Delta::Editor->new, $pool);
5129 };
5130 if ($@) {
5131 $can_do_switch = 0;
5132 } else {
5133 $rep->abort_report($pool);
5134 $can_do_switch = 1;
5135 }
5136 $pool->clear;
5137 }
5138 $can_do_switch;
5139}
5140sub skip_unknown_revs {
5142 my ($err) = @_;
5143 my $errno = $err->apr_err();
5144 # Maybe the branch we're tracking didn't
5145 # exist when the repo started, so it's
5146 # not an error if it doesn't, just continue
5147 #
5148 # Wonderfully consistent library, eh?
5149 # 160013 - svn:// and file://
5150 # 175002 - http(s)://
5151 # 175007 - http(s):// (this repo required authorization, too...)
5152 # More codes may be discovered later...
5153 if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
5154 my $err_key = $err->expanded_message;
5155 # revision numbers change every time, filter them out
5156 $err_key =~ s/\d+/\0/g;
5157 $err_key = "$errno\0$err_key";
5158 unless ($ignored_err{$err_key}) {
5159 warn "W: Ignoring error from SVN, path probably ",
5160 "does not exist: ($errno): ",
5161 $err->expanded_message,"\n";
5162 warn "W: Do not be alarmed at the above message ",
5163 "git-svn is just searching aggressively for ",
5164 "old history.\n",
5165 "This may take a while on large repositories\n";
5166 $ignored_err{$err_key} = 1;
5167 }
5168 return;
5169 }
5170 die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
5171}
5172package Git::SVN::Log;
5174use strict;
5175use warnings;
5176use POSIX qw/strftime/;
5177use Time::Local;
5178use constant commit_log_separator => ('-' x 72) . "\n";
5179use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
5180 %rusers $show_commit $incremental/;
5181my $l_fmt;
5182sub cmt_showable {
5184 my ($c) = @_;
5185 return 1 if defined $c->{r};
5186 # big commit message got truncated by the 16k pretty buffer in rev-list
5188 if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
5189 $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
5190 @{$c->{l}} = ();
5191 my @log = command(qw/cat-file commit/, $c->{c});
5192 # shift off the headers
5194 shift @log while ($log[0] ne '');
5195 shift @log;
5196 # TODO: make $c->{l} not have a trailing newline in the future
5198 @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
5199 (undef, $c->{r}, undef) = ::extract_metadata(
5201 (grep(/^git-svn-id: /, @log))[-1]);
5202 }
5203 return defined $c->{r};
5204}
5205sub log_use_color {
5207 return $color || Git->repository->get_colorbool('color.diff');
5208}
5209sub git_svn_log_cmd {
5211 my ($r_min, $r_max, @args) = @_;
5212 my $head = 'HEAD';
5213 my (@files, @log_opts);
5214 foreach my $x (@args) {
5215 if ($x eq '--' || @files) {
5216 push @files, $x;
5217 } else {
5218 if (::verify_ref("$x^0")) {
5219 $head = $x;
5220 } else {
5221 push @log_opts, $x;
5222 }
5223 }
5224 }
5225 my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
5227 $gs ||= Git::SVN->_new;
5228 my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
5229 $gs->refname);
5230 push @cmd, '-r' unless $non_recursive;
5231 push @cmd, qw/--raw --name-status/ if $verbose;
5232 push @cmd, '--color' if log_use_color();
5233 push @cmd, @log_opts;
5234 if (defined $r_max && $r_max == $r_min) {
5235 push @cmd, '--max-count=1';
5236 if (my $c = $gs->rev_map_get($r_max)) {
5237 push @cmd, $c;
5238 }
5239 } elsif (defined $r_max) {
5240 if ($r_max < $r_min) {
5241 ($r_min, $r_max) = ($r_max, $r_min);
5242 }
5243 my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
5244 my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
5245 # If there are no commits in the range, both $c_max and $c_min
5246 # will be undefined. If there is at least 1 commit in the
5247 # range, both will be defined.
5248 return () if !defined $c_min || !defined $c_max;
5249 if ($c_min eq $c_max) {
5250 push @cmd, '--max-count=1', $c_min;
5251 } else {
5252 push @cmd, '--boundary', "$c_min..$c_max";
5253 }
5254 }
5255 return (@cmd, @files);
5256}
5257# adapted from pager.c
5259sub config_pager {
5260 chomp(my $pager = command_oneline(qw(var GIT_PAGER)));
5261 if ($pager eq 'cat') {
5262 $pager = undef;
5263 }
5264 $ENV{GIT_PAGER_IN_USE} = defined($pager);
5265}
5266sub run_pager {
5268 return unless -t *STDOUT && defined $pager;
5269 pipe my ($rfd, $wfd) or return;
5270 defined(my $pid = fork) or ::fatal "Can't fork: $!";
5271 if (!$pid) {
5272 open STDOUT, '>&', $wfd or
5273 ::fatal "Can't redirect to stdout: $!";
5274 return;
5275 }
5276 open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
5277 $ENV{LESS} ||= 'FRSX';
5278 exec $pager or ::fatal "Can't run pager: $! ($pager)";
5279}
5280sub format_svn_date {
5282 # some systmes don't handle or mishandle %z, so be creative.
5283 my $t = shift || time;
5284 my $gm = timelocal(gmtime($t));
5285 my $sign = qw( + + - )[ $t <=> $gm ];
5286 my $gmoff = sprintf("%s%02d%02d", $sign, (gmtime(abs($t - $gm)))[2,1]);
5287 return strftime("%Y-%m-%d %H:%M:%S $gmoff (%a, %d %b %Y)", localtime($t));
5288}
5289sub parse_git_date {
5291 my ($t, $tz) = @_;
5292 # Date::Parse isn't in the standard Perl distro :(
5293 if ($tz =~ s/^\+//) {
5294 $t += tz_to_s_offset($tz);
5295 } elsif ($tz =~ s/^\-//) {
5296 $t -= tz_to_s_offset($tz);
5297 }
5298 return $t;
5299}
5300sub set_local_timezone {
5302 if (defined $TZ) {
5303 $ENV{TZ} = $TZ;
5304 } else {
5305 delete $ENV{TZ};
5306 }
5307}
5308sub tz_to_s_offset {
5310 my ($tz) = @_;
5311 $tz =~ s/(\d\d)$//;
5312 return ($1 * 60) + ($tz * 3600);
5313}
5314sub get_author_info {
5316 my ($dest, $author, $t, $tz) = @_;
5317 $author =~ s/(?:^\s*|\s*$)//g;
5318 $dest->{a_raw} = $author;
5319 my $au;
5320 if ($::_authors) {
5321 $au = $rusers{$author} || undef;
5322 }
5323 if (!$au) {
5324 ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
5325 }
5326 $dest->{t} = $t;
5327 $dest->{tz} = $tz;
5328 $dest->{a} = $au;
5329 $dest->{t_utc} = parse_git_date($t, $tz);
5330}
5331sub process_commit {
5333 my ($c, $r_min, $r_max, $defer) = @_;
5334 if (defined $r_min && defined $r_max) {
5335 if ($r_min == $c->{r} && $r_min == $r_max) {
5336 show_commit($c);
5337 return 0;
5338 }
5339 return 1 if $r_min == $r_max;
5340 if ($r_min < $r_max) {
5341 # we need to reverse the print order
5342 return 0 if (defined $limit && --$limit < 0);
5343 push @$defer, $c;
5344 return 1;
5345 }
5346 if ($r_min != $r_max) {
5347 return 1 if ($r_min < $c->{r});
5348 return 1 if ($r_max > $c->{r});
5349 }
5350 }
5351 return 0 if (defined $limit && --$limit < 0);
5352 show_commit($c);
5353 return 1;
5354}
5355sub show_commit {
5357 my $c = shift;
5358 if ($oneline) {
5359 my $x = "\n";
5360 if (my $l = $c->{l}) {
5361 while ($l->[0] =~ /^\s*$/) { shift @$l }
5362 $x = $l->[0];
5363 }
5364 $l_fmt ||= 'A' . length($c->{r});
5365 print 'r',pack($l_fmt, $c->{r}),' | ';
5366 print "$c->{c} | " if $show_commit;
5367 print $x;
5368 } else {
5369 show_commit_normal($c);
5370 }
5371}
5372sub show_commit_changed_paths {
5374 my ($c) = @_;
5375 return unless $c->{changed};
5376 print "Changed paths:\n", @{$c->{changed}};
5377}
5378sub show_commit_normal {
5380 my ($c) = @_;
5381 print commit_log_separator, "r$c->{r} | ";
5382 print "$c->{c} | " if $show_commit;
5383 print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
5384 my $nr_line = 0;
5385 if (my $l = $c->{l}) {
5387 while ($l->[$#$l] eq "\n" && $#$l > 0
5388 && $l->[($#$l - 1)] eq "\n") {
5389 pop @$l;
5390 }
5391 $nr_line = scalar @$l;
5392 if (!$nr_line) {
5393 print "1 line\n\n\n";
5394 } else {
5395 if ($nr_line == 1) {
5396 $nr_line = '1 line';
5397 } else {
5398 $nr_line .= ' lines';
5399 }
5400 print $nr_line, "\n";
5401 show_commit_changed_paths($c);
5402 print "\n";
5403 print $_ foreach @$l;
5404 }
5405 } else {
5406 print "1 line\n";
5407 show_commit_changed_paths($c);
5408 print "\n";
5409 }
5411 foreach my $x (qw/raw stat diff/) {
5412 if ($c->{$x}) {
5413 print "\n";
5414 print $_ foreach @{$c->{$x}}
5415 }
5416 }
5417}
5418sub cmd_show_log {
5420 my (@args) = @_;
5421 my ($r_min, $r_max);
5422 my $r_last = -1; # prevent dupes
5423 set_local_timezone();
5424 if (defined $::_revision) {
5425 if ($::_revision =~ /^(\d+):(\d+)$/) {
5426 ($r_min, $r_max) = ($1, $2);
5427 } elsif ($::_revision =~ /^\d+$/) {
5428 $r_min = $r_max = $::_revision;
5429 } else {
5430 ::fatal "-r$::_revision is not supported, use ",
5431 "standard 'git log' arguments instead";
5432 }
5433 }
5434 config_pager();
5436 @args = git_svn_log_cmd($r_min, $r_max, @args);
5437 if (!@args) {
5438 print commit_log_separator unless $incremental || $oneline;
5439 return;
5440 }
5441 my $log = command_output_pipe(@args);
5442 run_pager();
5443 my (@k, $c, $d, $stat);
5444 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
5445 while (<$log>) {
5446 if (/^${esc_color}commit -?($::sha1_short)/o) {
5447 my $cmt = $1;
5448 if ($c && cmt_showable($c) && $c->{r} != $r_last) {
5449 $r_last = $c->{r};
5450 process_commit($c, $r_min, $r_max, \@k) or
5451 goto out;
5452 }
5453 $d = undef;
5454 $c = { c => $cmt };
5455 } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
5456 get_author_info($c, $1, $2, $3);
5457 } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
5458 # ignore
5459 } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
5460 push @{$c->{raw}}, $_;
5461 } elsif (/^${esc_color}[ACRMDT]\t/) {
5462 # we could add $SVN->{svn_path} here, but that requires
5463 # remote access at the moment (repo_path_split)...
5464 s#^(${esc_color})([ACRMDT])\t#$1 $2 #o;
5465 push @{$c->{changed}}, $_;
5466 } elsif (/^${esc_color}diff /o) {
5467 $d = 1;
5468 push @{$c->{diff}}, $_;
5469 } elsif ($d) {
5470 push @{$c->{diff}}, $_;
5471 } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
5472 $esc_color*[\+\-]*$esc_color$/x) {
5473 $stat = 1;
5474 push @{$c->{stat}}, $_;
5475 } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
5476 push @{$c->{stat}}, $_;
5477 $stat = undef;
5478 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
5479 ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
5480 } elsif (s/^${esc_color} //o) {
5481 push @{$c->{l}}, $_;
5482 }
5483 }
5484 if ($c && defined $c->{r} && $c->{r} != $r_last) {
5485 $r_last = $c->{r};
5486 process_commit($c, $r_min, $r_max, \@k);
5487 }
5488 if (@k) {
5489 ($r_min, $r_max) = ($r_max, $r_min);
5490 process_commit($_, $r_min, $r_max) foreach reverse @k;
5491 }
5492out:
5493 close $log;
5494 print commit_log_separator unless $incremental || $oneline;
5495}
5496sub cmd_blame {
5498 my $path = pop;
5499 config_pager();
5501 run_pager();
5502 my ($fh, $ctx, $rev);
5504 if ($_git_format) {
5506 ($fh, $ctx) = command_output_pipe('blame', @_, $path);
5507 while (my $line = <$fh>) {
5508 if ($line =~ /^\^?([[:xdigit:]]+)\s/) {
5509 # Uncommitted edits show up as a rev ID of
5510 # all zeros, which we can't look up with
5511 # cmt_metadata
5512 if ($1 !~ /^0+$/) {
5513 (undef, $rev, undef) =
5514 ::cmt_metadata($1);
5515 $rev = '0' if (!$rev);
5516 } else {
5517 $rev = '0';
5518 }
5519 $rev = sprintf('%-10s', $rev);
5520 $line =~ s/^\^?[[:xdigit:]]+(\s)/$rev$1/;
5521 }
5522 print $line;
5523 }
5524 } else {
5525 ($fh, $ctx) = command_output_pipe('blame', '-p', @_, 'HEAD',
5526 '--', $path);
5527 my ($sha1);
5528 my %authors;
5529 my @buffer;
5530 my %dsha; #distinct sha keys
5531 while (my $line = <$fh>) {
5533 push @buffer, $line;
5534 if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
5535 $dsha{$1} = 1;
5536 }
5537 }
5538 my $s2r = ::cmt_sha2rev_batch([keys %dsha]);
5540 foreach my $line (@buffer) {
5542 if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
5543 $rev = $s2r->{$1};
5544 $rev = '0' if (!$rev)
5545 }
5546 elsif ($line =~ /^author (.*)/) {
5547 $authors{$rev} = $1;
5548 $authors{$rev} =~ s/\s/_/g;
5549 }
5550 elsif ($line =~ /^\t(.*)$/) {
5551 printf("%6s %10s %s\n", $rev, $authors{$rev}, $1);
5552 }
5553 }
5554 }
5555 command_close_pipe($fh, $ctx);
5556}
5557package Git::SVN::Migration;
5559# these version numbers do NOT correspond to actual version numbers
5560# of git nor git-svn. They are just relative.
5561#
5562# v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
5563#
5564# v1 layout: .git/$id/info/url, refs/remotes/$id
5565#
5566# v2 layout: .git/svn/$id/info/url, refs/remotes/$id
5567#
5568# v3 layout: .git/svn/$id, refs/remotes/$id
5569# - info/url may remain for backwards compatibility
5570# - this is what we migrate up to this layout automatically,
5571# - this will be used by git svn init on single branches
5572# v3.1 layout (auto migrated):
5573# - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
5574# for backwards compatibility
5575#
5576# v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
5577# - this is only created for newly multi-init-ed
5578# repositories. Similar in spirit to the
5579# --use-separate-remotes option in git-clone (now default)
5580# - we do not automatically migrate to this (following
5581# the example set by core git)
5582#
5583# v5 layout: .rev_db.$UUID => .rev_map.$UUID
5584# - newer, more-efficient format that uses 24-bytes per record
5585# with no filler space.
5586# - use xxd -c24 < .rev_map.$UUID to view and debug
5587# - This is a one-way migration, repositories updated to the
5588# new format will not be able to use old git-svn without
5589# rebuilding the .rev_db. Rebuilding the rev_db is not
5590# possible if noMetadata or useSvmProps are set; but should
5591# be no problem for users that use the (sensible) defaults.
5592use strict;
5593use warnings;
5594use Carp qw/croak/;
5595use File::Path qw/mkpath/;
5596use File::Basename qw/dirname basename/;
5597use vars qw/$_minimize/;
5598sub migrate_from_v0 {
5600 my $git_dir = $ENV{GIT_DIR};
5601 return undef unless -d $git_dir;
5602 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
5603 my $migrated = 0;
5604 while (<$fh>) {
5605 chomp;
5606 my ($id, $orig_ref) = ($_, $_);
5607 next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
5608 next unless -f "$git_dir/$id/info/url";
5609 my $new_ref = "refs/remotes/$id";
5610 if (::verify_ref("$new_ref^0")) {
5611 print STDERR "W: $orig_ref is probably an old ",
5612 "branch used by an ancient version of ",
5613 "git-svn.\n",
5614 "However, $new_ref also exists.\n",
5615 "We will not be able ",
5616 "to use this branch until this ",
5617 "ambiguity is resolved.\n";
5618 next;
5619 }
5620 print STDERR "Migrating from v0 layout...\n" if !$migrated;
5621 print STDERR "Renaming ref: $orig_ref => $new_ref\n";
5622 command_noisy('update-ref', $new_ref, $orig_ref);
5623 command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
5624 $migrated++;
5625 }
5626 command_close_pipe($fh, $ctx);
5627 print STDERR "Done migrating from v0 layout...\n" if $migrated;
5628 $migrated;
5629}
5630sub migrate_from_v1 {
5632 my $git_dir = $ENV{GIT_DIR};
5633 my $migrated = 0;
5634 return $migrated unless -d $git_dir;
5635 my $svn_dir = "$git_dir/svn";
5636 # just in case somebody used 'svn' as their $id at some point...
5638 return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
5639 print STDERR "Migrating from a git-svn v1 layout...\n";
5641 mkpath([$svn_dir]);
5642 print STDERR "Data from a previous version of git-svn exists, but\n\t",
5643 "$svn_dir\n\t(required for this version ",
5644 "($::VERSION) of git-svn) does not exist.\n";
5645 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
5646 while (<$fh>) {
5647 my $x = $_;
5648 next unless $x =~ s#^refs/remotes/##;
5649 chomp $x;
5650 next unless -f "$git_dir/$x/info/url";
5651 my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
5652 next unless $u;
5653 my $dn = dirname("$git_dir/svn/$x");
5654 mkpath([$dn]) unless -d $dn;
5655 if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
5656 mkpath(["$git_dir/svn/svn"]);
5657 print STDERR " - $git_dir/$x/info => ",
5658 "$git_dir/svn/$x/info\n";
5659 rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
5660 croak "$!: $x";
5661 # don't worry too much about these, they probably
5662 # don't exist with repos this old (save for index,
5663 # and we can easily regenerate that)
5664 foreach my $f (qw/unhandled.log index .rev_db/) {
5665 rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
5666 }
5667 } else {
5668 print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
5669 rename "$git_dir/$x", "$git_dir/svn/$x" or
5670 croak "$!: $x";
5671 }
5672 $migrated++;
5673 }
5674 command_close_pipe($fh, $ctx);
5675 print STDERR "Done migrating from a git-svn v1 layout\n";
5676 $migrated;
5677}
5678sub read_old_urls {
5680 my ($l_map, $pfx, $path) = @_;
5681 my @dir;
5682 foreach (<$path/*>) {
5683 if (-r "$_/info/url") {
5684 $pfx .= '/' if $pfx && $pfx !~ m!/$!;
5685 my $ref_id = $pfx . basename $_;
5686 my $url = ::file_to_s("$_/info/url");
5687 $l_map->{$ref_id} = $url;
5688 } elsif (-d $_) {
5689 push @dir, $_;
5690 }
5691 }
5692 foreach (@dir) {
5693 my $x = $_;
5694 $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
5695 read_old_urls($l_map, $x, $_);
5696 }
5697}
5698sub migrate_from_v2 {
5700 my @cfg = command(qw/config -l/);
5701 return if grep /^svn-remote\..+\.url=/, @cfg;
5702 my %l_map;
5703 read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
5704 my $migrated = 0;
5705 foreach my $ref_id (sort keys %l_map) {
5707 eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
5708 if ($@) {
5709 Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
5710 }
5711 $migrated++;
5712 }
5713 $migrated;
5714}
5715sub minimize_connections {
5717 my $r = Git::SVN::read_all_remotes();
5718 my $new_urls = {};
5719 my $root_repos = {};
5720 foreach my $repo_id (keys %$r) {
5721 my $url = $r->{$repo_id}->{url} or next;
5722 my $fetch = $r->{$repo_id}->{fetch} or next;
5723 my $ra = Git::SVN::Ra->new($url);
5724 # skip existing cases where we already connect to the root
5726 if (($ra->{url} eq $ra->{repos_root}) ||
5727 ($ra->{repos_root} eq $repo_id)) {
5728 $root_repos->{$ra->{url}} = $repo_id;
5729 next;
5730 }
5731 my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
5733 my $root_path = $ra->{url};
5734 $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
5735 foreach my $path (keys %$fetch) {
5736 my $ref_id = $fetch->{$path};
5737 my $gs = Git::SVN->new($ref_id, $repo_id, $path);
5738 # make sure we can read when connecting to
5740 # a higher level of a repository
5741 my ($last_rev, undef) = $gs->last_rev_commit;
5742 if (!defined $last_rev) {
5743 $last_rev = eval {
5744 $root_ra->get_latest_revnum;
5745 };
5746 next if $@;
5747 }
5748 my $new = $root_path;
5749 $new .= length $path ? "/$path" : '';
5750 eval {
5751 $root_ra->get_log([$new], $last_rev, $last_rev,
5752 0, 0, 1, sub { });
5753 };
5754 next if $@;
5755 $new_urls->{$ra->{repos_root}}->{$new} =
5756 { ref_id => $ref_id,
5757 old_repo_id => $repo_id,
5758 old_path => $path };
5759 }
5760 }
5761 my @emptied;
5763 foreach my $url (keys %$new_urls) {
5764 # see if we can re-use an existing [svn-remote "repo_id"]
5765 # instead of creating a(n ugly) new section:
5766 my $repo_id = $root_repos->{$url} || $url;
5767 my $fetch = $new_urls->{$url};
5769 foreach my $path (keys %$fetch) {
5770 my $x = $fetch->{$path};
5771 Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
5772 my $pfx = "svn-remote.$x->{old_repo_id}";
5773 my $old_fetch = quotemeta("$x->{old_path}:".
5775 "$x->{ref_id}");
5776 command_noisy(qw/config --unset/,
5777 "$pfx.fetch", '^'. $old_fetch . '$');
5778 delete $r->{$x->{old_repo_id}}->
5779 {fetch}->{$x->{old_path}};
5780 if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
5781 command_noisy(qw/config --unset/,
5782 "$pfx.url");
5783 push @emptied, $x->{old_repo_id}
5784 }
5785 }
5786 }
5787 if (@emptied) {
5788 my $file = $ENV{GIT_CONFIG} || "$ENV{GIT_DIR}/config";
5789 print STDERR <<EOF;
5790The following [svn-remote] sections in your config file ($file) are empty
5791and can be safely removed:
5792EOF
5793 print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
5794 }
5795}
5796sub migration_check {
5798 migrate_from_v0();
5799 migrate_from_v1();
5800 migrate_from_v2();
5801 minimize_connections() if $_minimize;
5802}
5803package Git::IndexInfo;
5805use strict;
5806use warnings;
5807use Git qw/command_input_pipe command_close_pipe/;
5808sub new {
5810 my ($class) = @_;
5811 my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
5812 bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
5813}
5814sub remove {
5816 my ($self, $path) = @_;
5817 if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
5818 return ++$self->{nr};
5819 }
5820 undef;
5821}
5822sub update {
5824 my ($self, $mode, $hash, $path) = @_;
5825 if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
5826 return ++$self->{nr};
5827 }
5828 undef;
5829}
5830sub DESTROY {
5832 my ($self) = @_;
5833 command_close_pipe($self->{gui}, $self->{ctx});
5834}
5835package Git::SVN::GlobSpec;
5837use strict;
5838use warnings;
5839sub new {
5841 my ($class, $glob) = @_;
5842 my $re = $glob;
5843 $re =~ s!/+$!!g; # no need for trailing slashes
5844 $re =~ m!^([^*]*)(\*(?:/\*)*)(.*)$!;
5845 my $temp = $re;
5846 my ($left, $right) = ($1, $3);
5847 $re = $2;
5848 my $depth = $re =~ tr/*/*/;
5849 if ($depth != $temp =~ tr/*/*/) {
5850 die "Only one set of wildcard directories " .
5851 "(e.g. '*' or '*/*/*') is supported: '$glob'\n";
5852 }
5853 if ($depth == 0) {
5854 die "One '*' is needed for glob: '$glob'\n";
5855 }
5856 $re =~ s!\*!\[^/\]*!g;
5857 $re = quotemeta($left) . "($re)" . quotemeta($right);
5858 if (length $left && !($left =~ s!/+$!!g)) {
5859 die "Missing trailing '/' on left side of: '$glob' ($left)\n";
5860 }
5861 if (length $right && !($right =~ s!^/+!!g)) {
5862 die "Missing leading '/' on right side of: '$glob' ($right)\n";
5863 }
5864 my $left_re = qr/^\/\Q$left\E(\/|$)/;
5865 bless { left => $left, right => $right, left_regex => $left_re,
5866 regex => qr/$re/, glob => $glob, depth => $depth }, $class;
5867}
5868sub full_path {
5870 my ($self, $path) = @_;
5871 return (length $self->{left} ? "$self->{left}/" : '') .
5872 $path . (length $self->{right} ? "/$self->{right}" : '');
5873}
5874__END__
5876Data structures:
5878$remotes = { # returned by read_all_remotes()
5881 'svn' => {
5882 # svn-remote.svn.url=https://svn.musicpd.org
5883 url => 'https://svn.musicpd.org',
5884 # svn-remote.svn.fetch=mpd/trunk:trunk
5885 fetch => {
5886 'mpd/trunk' => 'trunk',
5887 },
5888 # svn-remote.svn.tags=mpd/tags/*:tags/*
5889 tags => {
5890 path => {
5891 left => 'mpd/tags',
5892 right => '',
5893 regex => qr!mpd/tags/([^/]+)$!,
5894 glob => 'tags/*',
5895 },
5896 ref => {
5897 left => 'tags',
5898 right => '',
5899 regex => qr!tags/([^/]+)$!,
5900 glob => 'tags/*',
5901 },
5902 }
5903 }
5904};
5905$log_entry hashref as returned by libsvn_log_entry()
5907{
5908 log => 'whitespace-formatted log entry
5909', # trailing newline is preserved
5910 revision => '8', # integer
5911 date => '2004-02-24T17:01:44.108345Z', # commit date
5912 author => 'committer name'
5913};
5914# this is generated by generate_diff();
5917@mods = array of diff-index line hashes, each element represents one line
5918 of diff-index output
5919diff-index line ($m hash)
5921{
5922 mode_a => first column of diff-index output, no leading ':',
5923 mode_b => second column of diff-index output,
5924 sha1_b => sha1sum of the final blob,
5925 chg => change type [MCRADT],
5926 file_a => original file name of a file (iff chg is 'C' or 'R')
5927 file_b => new/current file name of a file (any chg)
5928}
5929;
5930# retval of read_url_paths{,_all}();
5932$l_map = {
5933 # repository root url
5934 'https://svn.musicpd.org' => {
5935 # repository path # GIT_SVN_ID
5936 'mpd/trunk' => 'trunk',
5937 'mpd/tags/0.11.5' => 'tags/0.11.5',
5938 },
5939}
5940Notes:
5942 I don't trust the each() function on unless I created %hash myself
5943 because the internal iterator may not have started at base.