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 $ENV{SVN_SSH} =~ s/(.*)/"$1"/;
30 }
31 }
32}
33$Git::SVN::Log::TZ = $ENV{TZ};
35$ENV{TZ} = 'UTC';
36$| = 1; # unbuffer STDOUT
37sub fatal (@) { print STDERR "@_\n"; exit 1 }
39sub _req_svn {
40 require SVN::Core; # use()-ing this causes segfaults for me... *shrug*
41 require SVN::Ra;
42 require SVN::Delta;
43 if ($SVN::Core::VERSION lt '1.1.0') {
44 fatal "Need SVN::Core 1.1.0 or better (got $SVN::Core::VERSION)";
45 }
46}
47my $can_compress = eval { require Compress::Zlib; 1};
48push @Git::SVN::Ra::ISA, 'SVN::Ra';
49push @SVN::Git::Editor::ISA, 'SVN::Delta::Editor';
50push @SVN::Git::Fetcher::ISA, 'SVN::Delta::Editor';
51use Carp qw/croak/;
52use Digest::MD5;
53use IO::File qw//;
54use File::Basename qw/dirname basename/;
55use File::Path qw/mkpath/;
56use File::Spec;
57use File::Find;
58use Getopt::Long qw/:config gnu_getopt no_ignore_case auto_abbrev/;
59use IPC::Open3;
60use Git;
61BEGIN {
63 # import functions from Git into our packages, en masse
64 no strict 'refs';
65 foreach (qw/command command_oneline command_noisy command_output_pipe
66 command_input_pipe command_close_pipe
67 command_bidi_pipe command_close_bidi_pipe/) {
68 for my $package ( qw(SVN::Git::Editor SVN::Git::Fetcher
69 Git::SVN::Migration Git::SVN::Log Git::SVN),
70 __PACKAGE__) {
71 *{"${package}::$_"} = \&{"Git::$_"};
72 }
73 }
74}
75my ($SVN);
77$sha1 = qr/[a-f\d]{40}/;
79$sha1_short = qr/[a-f\d]{4,40}/;
80my ($_stdin, $_help, $_edit,
81 $_message, $_file, $_branch_dest,
82 $_template, $_shared,
83 $_version, $_fetch_all, $_no_rebase, $_fetch_parent,
84 $_merge, $_strategy, $_dry_run, $_local,
85 $_prefix, $_no_checkout, $_url, $_verbose,
86 $_git_format, $_commit_url, $_tag);
87$Git::SVN::_follow_parent = 1;
88$_q ||= 0;
89my %remote_opts = ( 'username=s' => \$Git::SVN::Prompt::_username,
90 'config-dir=s' => \$Git::SVN::Ra::config_dir,
91 'no-auth-cache' => \$Git::SVN::Prompt::_no_auth_cache,
92 'ignore-paths=s' => \$SVN::Git::Fetcher::_ignore_regex );
93my %fc_opts = ( 'follow-parent|follow!' => \$Git::SVN::_follow_parent,
94 'authors-file|A=s' => \$_authors,
95 'authors-prog=s' => \$_authors_prog,
96 'repack:i' => \$Git::SVN::_repack,
97 'noMetadata' => \$Git::SVN::_no_metadata,
98 'useSvmProps' => \$Git::SVN::_use_svm_props,
99 'useSvnsyncProps' => \$Git::SVN::_use_svnsync_props,
100 'log-window-size=i' => \$Git::SVN::Ra::_log_window_size,
101 'no-checkout' => \$_no_checkout,
102 'quiet|q+' => \$_q,
103 'repack-flags|repack-args|repack-opts=s' =>
104 \$Git::SVN::_repack_flags,
105 'use-log-author' => \$Git::SVN::_use_log_author,
106 'add-author-from' => \$Git::SVN::_add_author_from,
107 'localtime' => \$Git::SVN::_localtime,
108 %remote_opts );
109my ($_trunk, @_tags, @_branches, $_stdlayout);
111my %icv;
112my %init_opts = ( 'template=s' => \$_template, 'shared:s' => \$_shared,
113 'trunk|T=s' => \$_trunk, 'tags|t=s@' => \@_tags,
114 'branches|b=s@' => \@_branches, 'prefix=s' => \$_prefix,
115 'stdlayout|s' => \$_stdlayout,
116 'minimize-url|m!' => \$Git::SVN::_minimize_url,
117 'no-metadata' => sub { $icv{noMetadata} = 1 },
118 'use-svm-props' => sub { $icv{useSvmProps} = 1 },
119 'use-svnsync-props' => sub { $icv{useSvnsyncProps} = 1 },
120 'rewrite-root=s' => sub { $icv{rewriteRoot} = $_[1] },
121 'rewrite-uuid=s' => sub { $icv{rewriteUUID} = $_[1] },
122 %remote_opts );
123my %cmt_opts = ( 'edit|e' => \$_edit,
124 'rmdir' => \$SVN::Git::Editor::_rmdir,
125 'find-copies-harder' => \$SVN::Git::Editor::_find_copies_harder,
126 'l=i' => \$SVN::Git::Editor::_rename_limit,
127 'copy-similarity|C=i'=> \$SVN::Git::Editor::_cp_similarity
128);
129my %cmd = (
131 fetch => [ \&cmd_fetch, "Download new revisions from SVN",
132 { 'revision|r=s' => \$_revision,
133 'fetch-all|all' => \$_fetch_all,
134 'parent|p' => \$_fetch_parent,
135 %fc_opts } ],
136 clone => [ \&cmd_clone, "Initialize and fetch revisions",
137 { 'revision|r=s' => \$_revision,
138 %fc_opts, %init_opts } ],
139 init => [ \&cmd_init, "Initialize a repo for tracking" .
140 " (requires URL argument)",
141 \%init_opts ],
142 'multi-init' => [ \&cmd_multi_init,
143 "Deprecated alias for ".
144 "'$0 init -T<trunk> -b<branches> -t<tags>'",
145 \%init_opts ],
146 dcommit => [ \&cmd_dcommit,
147 'Commit several diffs to merge with upstream',
148 { 'merge|m|M' => \$_merge,
149 'strategy|s=s' => \$_strategy,
150 'verbose|v' => \$_verbose,
151 'dry-run|n' => \$_dry_run,
152 'fetch-all|all' => \$_fetch_all,
153 'commit-url=s' => \$_commit_url,
154 'revision|r=i' => \$_revision,
155 'no-rebase' => \$_no_rebase,
156 %cmt_opts, %fc_opts } ],
157 branch => [ \&cmd_branch,
158 'Create a branch in the SVN repository',
159 { 'message|m=s' => \$_message,
160 'destination|d=s' => \$_branch_dest,
161 'dry-run|n' => \$_dry_run,
162 'tag|t' => \$_tag,
163 'username=s' => \$Git::SVN::Prompt::_username,
164 'commit-url=s' => \$_commit_url } ],
165 tag => [ sub { $_tag = 1; cmd_branch(@_) },
166 'Create a tag in the SVN repository',
167 { 'message|m=s' => \$_message,
168 'destination|d=s' => \$_branch_dest,
169 'dry-run|n' => \$_dry_run,
170 'username=s' => \$Git::SVN::Prompt::_username,
171 'commit-url=s' => \$_commit_url } ],
172 'set-tree' => [ \&cmd_set_tree,
173 "Set an SVN repository to a git tree-ish",
174 { 'stdin' => \$_stdin, %cmt_opts, %fc_opts, } ],
175 'create-ignore' => [ \&cmd_create_ignore,
176 'Create a .gitignore per svn:ignore',
177 { 'revision|r=i' => \$_revision
178 } ],
179 'mkdirs' => [ \&cmd_mkdirs ,
180 "recreate empty directories after a checkout",
181 { 'revision|r=i' => \$_revision } ],
182 'propget' => [ \&cmd_propget,
183 'Print the value of a property on a file or directory',
184 { 'revision|r=i' => \$_revision } ],
185 'proplist' => [ \&cmd_proplist,
186 'List all properties of a file or directory',
187 { 'revision|r=i' => \$_revision } ],
188 'show-ignore' => [ \&cmd_show_ignore, "Show svn:ignore listings",
189 { 'revision|r=i' => \$_revision
190 } ],
191 'show-externals' => [ \&cmd_show_externals, "Show svn:externals listings",
192 { 'revision|r=i' => \$_revision
193 } ],
194 'multi-fetch' => [ \&cmd_multi_fetch,
195 "Deprecated alias for $0 fetch --all",
196 { 'revision|r=s' => \$_revision, %fc_opts } ],
197 'migrate' => [ sub { },
198 # no-op, we automatically run this anyways,
199 'Migrate configuration/metadata/layout from
200 previous versions of git-svn',
201 { 'minimize' => \$Git::SVN::Migration::_minimize,
202 %remote_opts } ],
203 'log' => [ \&Git::SVN::Log::cmd_show_log, 'Show commit logs',
204 { 'limit=i' => \$Git::SVN::Log::limit,
205 'revision|r=s' => \$_revision,
206 'verbose|v' => \$Git::SVN::Log::verbose,
207 'incremental' => \$Git::SVN::Log::incremental,
208 'oneline' => \$Git::SVN::Log::oneline,
209 'show-commit' => \$Git::SVN::Log::show_commit,
210 'non-recursive' => \$Git::SVN::Log::non_recursive,
211 'authors-file|A=s' => \$_authors,
212 'color' => \$Git::SVN::Log::color,
213 'pager=s' => \$Git::SVN::Log::pager
214 } ],
215 'find-rev' => [ \&cmd_find_rev,
216 "Translate between SVN revision numbers and tree-ish",
217 {} ],
218 'rebase' => [ \&cmd_rebase, "Fetch and rebase your working directory",
219 { 'merge|m|M' => \$_merge,
220 'verbose|v' => \$_verbose,
221 'strategy|s=s' => \$_strategy,
222 'local|l' => \$_local,
223 'fetch-all|all' => \$_fetch_all,
224 'dry-run|n' => \$_dry_run,
225 %fc_opts } ],
226 'commit-diff' => [ \&cmd_commit_diff,
227 'Commit a diff between two trees',
228 { 'message|m=s' => \$_message,
229 'file|F=s' => \$_file,
230 'revision|r=s' => \$_revision,
231 %cmt_opts } ],
232 'info' => [ \&cmd_info,
233 "Show info about the latest SVN revision
234 on the current branch",
235 { 'url' => \$_url, } ],
236 'blame' => [ \&Git::SVN::Log::cmd_blame,
237 "Show what revision and author last modified each line of a file",
238 { 'git-format' => \$_git_format } ],
239 'reset' => [ \&cmd_reset,
240 "Undo fetches back to the specified SVN revision",
241 { 'revision|r=s' => \$_revision,
242 'parent|p' => \$_fetch_parent } ],
243 'gc' => [ \&cmd_gc,
244 "Compress unhandled.log files in .git/svn and remove " .
245 "index files in .git/svn",
246 {} ],
247);
248my $cmd;
250for (my $i = 0; $i < @ARGV; $i++) {
251 if (defined $cmd{$ARGV[$i]}) {
252 $cmd = $ARGV[$i];
253 splice @ARGV, $i, 1;
254 last;
255 } elsif ($ARGV[$i] eq 'help') {
256 $cmd = $ARGV[$i+1];
257 usage(0);
258 }
259};
260# make sure we're always running at the top-level working directory
262unless ($cmd && $cmd =~ /(?:clone|init|multi-init)$/) {
263 unless (-d $ENV{GIT_DIR}) {
264 if ($git_dir_user_set) {
265 die "GIT_DIR=$ENV{GIT_DIR} explicitly set, ",
266 "but it is not a directory\n";
267 }
268 my $git_dir = delete $ENV{GIT_DIR};
269 my $cdup = undef;
270 git_cmd_try {
271 $cdup = command_oneline(qw/rev-parse --show-cdup/);
272 $git_dir = '.' unless ($cdup);
273 chomp $cdup if ($cdup);
274 $cdup = "." unless ($cdup && length $cdup);
275 } "Already at toplevel, but $git_dir not found\n";
276 chdir $cdup or die "Unable to chdir up to '$cdup'\n";
277 unless (-d $git_dir) {
278 die "$git_dir still not found after going to ",
279 "'$cdup'\n";
280 }
281 $ENV{GIT_DIR} = $git_dir;
282 }
283 $_repository = Git->repository(Repository => $ENV{GIT_DIR});
284}
285my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
287read_git_config(\%opts);
289if ($cmd && ($cmd eq 'log' || $cmd eq 'blame')) {
290 Getopt::Long::Configure('pass_through');
291}
292my $rv = GetOptions(%opts, 'help|H|h' => \$_help, 'version|V' => \$_version,
293 'minimize-connections' => \$Git::SVN::Migration::_minimize,
294 'id|i=s' => \$Git::SVN::default_ref_id,
295 'svn-remote|remote|R=s' => sub {
296 $Git::SVN::no_reuse_existing = 1;
297 $Git::SVN::default_repo_id = $_[1] });
298exit 1 if (!$rv && $cmd && $cmd ne 'log');
299usage(0) if $_help;
301version() if $_version;
302usage(1) unless defined $cmd;
303load_authors() if $_authors;
304if (defined $_authors_prog) {
305 $_authors_prog = "'" . File::Spec->rel2abs($_authors_prog) . "'";
306}
307unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
309 Git::SVN::Migration::migration_check();
310}
311Git::SVN::init_vars();
312eval {
313 Git::SVN::verify_remotes_sanity();
314 $cmd{$cmd}->[0]->(@ARGV);
315};
316fatal $@ if $@;
317post_fetch_checkout();
318exit 0;
319####################### primary functions ######################
321sub usage {
322 my $exit = shift || 0;
323 my $fd = $exit ? \*STDERR : \*STDOUT;
324 print $fd <<"";
325git-svn - bidirectional operations between a single Subversion tree and git
326Usage: git svn <command> [options] [arguments]\n
327 print $fd "Available commands:\n" unless $cmd;
329 foreach (sort keys %cmd) {
331 next if $cmd && $cmd ne $_;
332 next if /^multi-/; # don't show deprecated commands
333 print $fd ' ',pack('A17',$_),$cmd{$_}->[1],"\n";
334 foreach (sort keys %{$cmd{$_}->[2]}) {
335 # mixed-case options are for .git/config only
336 next if /[A-Z]/ && /^[a-z]+$/i;
337 # prints out arguments as they should be passed:
338 my $x = s#[:=]s$## ? '<arg>' : s#[:=]i$## ? '<num>' : '';
339 print $fd ' ' x 21, join(', ', map { length $_ > 1 ?
340 "--$_" : "-$_" }
341 split /\|/,$_)," $x\n";
342 }
343 }
344 print $fd <<"";
345\nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
346arbitrary identifier if you're tracking multiple SVN branches/repositories in
347one git repository and want to keep them separate. See git-svn(1) for more
348information.
349 exit $exit;
351}
352sub version {
354 ::_req_svn();
355 print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
356 exit 0;
357}
358sub do_git_init_db {
360 unless (-d $ENV{GIT_DIR}) {
361 my @init_db = ('init');
362 push @init_db, "--template=$_template" if defined $_template;
363 if (defined $_shared) {
364 if ($_shared =~ /[a-z]/) {
365 push @init_db, "--shared=$_shared";
366 } else {
367 push @init_db, "--shared";
368 }
369 }
370 command_noisy(@init_db);
371 $_repository = Git->repository(Repository => ".git");
372 }
373 my $set;
374 my $pfx = "svn-remote.$Git::SVN::default_repo_id";
375 foreach my $i (keys %icv) {
376 die "'$set' and '$i' cannot both be set\n" if $set;
377 next unless defined $icv{$i};
378 command_noisy('config', "$pfx.$i", $icv{$i});
379 $set = $i;
380 }
381 my $ignore_regex = \$SVN::Git::Fetcher::_ignore_regex;
382 command_noisy('config', "$pfx.ignore-paths", $$ignore_regex)
383 if defined $$ignore_regex;
384}
385sub init_subdir {
387 my $repo_path = shift or return;
388 mkpath([$repo_path]) unless -d $repo_path;
389 chdir $repo_path or die "Couldn't chdir to $repo_path: $!\n";
390 $ENV{GIT_DIR} = '.git';
391 $_repository = Git->repository(Repository => $ENV{GIT_DIR});
392}
393sub cmd_clone {
395 my ($url, $path) = @_;
396 if (!defined $path &&
397 (defined $_trunk || @_branches || @_tags ||
398 defined $_stdlayout) &&
399 $url !~ m#^[a-z\+]+://#) {
400 $path = $url;
401 }
402 $path = basename($url) if !defined $path || !length $path;
403 my $authors_absolute = $_authors ? File::Spec->rel2abs($_authors) : "";
404 cmd_init($url, $path);
405 command_oneline('config', 'svn.authorsfile', $authors_absolute)
406 if $_authors;
407 Git::SVN::fetch_all($Git::SVN::default_repo_id);
408}
409sub cmd_init {
411 if (defined $_stdlayout) {
412 $_trunk = 'trunk' if (!defined $_trunk);
413 @_tags = 'tags' if (! @_tags);
414 @_branches = 'branches' if (! @_branches);
415 }
416 if (defined $_trunk || @_branches || @_tags) {
417 return cmd_multi_init(@_);
418 }
419 my $url = shift or die "SVN repository location required ",
420 "as a command-line argument\n";
421 $url = canonicalize_url($url);
422 init_subdir(@_);
423 do_git_init_db();
424 if ($Git::SVN::_minimize_url eq 'unset') {
426 $Git::SVN::_minimize_url = 0;
427 }
428 Git::SVN->init($url);
430}
431sub cmd_fetch {
433 if (grep /^\d+=./, @_) {
434 die "'<rev>=<commit>' fetch arguments are ",
435 "no longer supported.\n";
436 }
437 my ($remote) = @_;
438 if (@_ > 1) {
439 die "Usage: $0 fetch [--all] [--parent] [svn-remote]\n";
440 }
441 $Git::SVN::no_reuse_existing = undef;
442 if ($_fetch_parent) {
443 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
444 unless ($gs) {
445 die "Unable to determine upstream SVN information from ",
446 "working tree history\n";
447 }
448 # just fetch, don't checkout.
449 $_no_checkout = 'true';
450 $_fetch_all ? $gs->fetch_all : $gs->fetch;
451 } elsif ($_fetch_all) {
452 cmd_multi_fetch();
453 } else {
454 $remote ||= $Git::SVN::default_repo_id;
455 Git::SVN::fetch_all($remote, Git::SVN::read_all_remotes());
456 }
457}
458sub cmd_set_tree {
460 my (@commits) = @_;
461 if ($_stdin || !@commits) {
462 print "Reading from stdin...\n";
463 @commits = ();
464 while (<STDIN>) {
465 if (/\b($sha1_short)\b/o) {
466 unshift @commits, $1;
467 }
468 }
469 }
470 my @revs;
471 foreach my $c (@commits) {
472 my @tmp = command('rev-parse',$c);
473 if (scalar @tmp == 1) {
474 push @revs, $tmp[0];
475 } elsif (scalar @tmp > 1) {
476 push @revs, reverse(command('rev-list',@tmp));
477 } else {
478 fatal "Failed to rev-parse $c";
479 }
480 }
481 my $gs = Git::SVN->new;
482 my ($r_last, $cmt_last) = $gs->last_rev_commit;
483 $gs->fetch;
484 if (defined $gs->{last_rev} && $r_last != $gs->{last_rev}) {
485 fatal "There are new revisions that were fetched ",
486 "and need to be merged (or acknowledged) ",
487 "before committing.\nlast rev: $r_last\n",
488 " current: $gs->{last_rev}";
489 }
490 $gs->set_tree($_) foreach @revs;
491 print "Done committing ",scalar @revs," revisions to SVN\n";
492 unlink $gs->{index};
493}
494sub cmd_dcommit {
496 my $head = shift;
497 git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
498 'Cannot dcommit with a dirty index. Commit your changes first, '
499 . "or stash them with `git stash'.\n";
500 $head ||= 'HEAD';
501 my $old_head;
503 if ($head ne 'HEAD') {
504 $old_head = eval {
505 command_oneline([qw/symbolic-ref -q HEAD/])
506 };
507 if ($old_head) {
508 $old_head =~ s{^refs/heads/}{};
509 } else {
510 $old_head = eval { command_oneline(qw/rev-parse HEAD/) };
511 }
512 command(['checkout', $head], STDERR => 0);
513 }
514 my @refs;
516 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD', \@refs);
517 unless ($gs) {
518 die "Unable to determine upstream SVN information from ",
519 "$head history.\nPerhaps the repository is empty.";
520 }
521 if (defined $_commit_url) {
523 $url = $_commit_url;
524 } else {
525 $url = eval { command_oneline('config', '--get',
526 "svn-remote.$gs->{repo_id}.commiturl") };
527 if (!$url) {
528 $url = $gs->full_url
529 }
530 }
531 my $last_rev = $_revision if defined $_revision;
533 if ($url) {
534 print "Committing to $url ...\n";
535 }
536 my ($linear_refs, $parents) = linearize_history($gs, \@refs);
537 if ($_no_rebase && scalar(@$linear_refs) > 1) {
538 warn "Attempting to commit more than one change while ",
539 "--no-rebase is enabled.\n",
540 "If these changes depend on each other, re-running ",
541 "without --no-rebase may be required."
542 }
543 my $expect_url = $url;
544 Git::SVN::remove_username($expect_url);
545 while (1) {
546 my $d = shift @$linear_refs or last;
547 unless (defined $last_rev) {
548 (undef, $last_rev, undef) = cmt_metadata("$d~1");
549 unless (defined $last_rev) {
550 fatal "Unable to extract revision information ",
551 "from commit $d~1";
552 }
553 }
554 if ($_dry_run) {
555 print "diff-tree $d~1 $d\n";
556 } else {
557 my $cmt_rev;
558 my %ed_opts = ( r => $last_rev,
559 log => get_commit_entry($d)->{log},
560 ra => Git::SVN::Ra->new($url),
561 config => SVN::Core::config_get_config(
562 $Git::SVN::Ra::config_dir
563 ),
564 tree_a => "$d~1",
565 tree_b => $d,
566 editor_cb => sub {
567 print "Committed r$_[0]\n";
568 $cmt_rev = $_[0];
569 },
570 svn_path => '');
571 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
572 print "No changes\n$d~1 == $d\n";
573 } elsif ($parents->{$d} && @{$parents->{$d}}) {
574 $gs->{inject_parents_dcommit}->{$cmt_rev} =
575 $parents->{$d};
576 }
577 $_fetch_all ? $gs->fetch_all : $gs->fetch;
578 $last_rev = $cmt_rev;
579 next if $_no_rebase;
580 # we always want to rebase against the current HEAD,
582 # not any head that was passed to us
583 my @diff = command('diff-tree', $d,
584 $gs->refname, '--');
585 my @finish;
586 if (@diff) {
587 @finish = rebase_cmd();
588 print STDERR "W: $d and ", $gs->refname,
589 " differ, using @finish:\n",
590 join("\n", @diff), "\n";
591 } else {
592 print "No changes between current HEAD and ",
593 $gs->refname,
594 "\nResetting to the latest ",
595 $gs->refname, "\n";
596 @finish = qw/reset --mixed/;
597 }
598 command_noisy(@finish, $gs->refname);
599 if (@diff) {
600 @refs = ();
601 my ($url_, $rev_, $uuid_, $gs_) =
602 working_head_info('HEAD', \@refs);
603 my ($linear_refs_, $parents_) =
604 linearize_history($gs_, \@refs);
605 if (scalar(@$linear_refs) !=
606 scalar(@$linear_refs_)) {
607 fatal "# of revisions changed ",
608 "\nbefore:\n",
609 join("\n", @$linear_refs),
610 "\n\nafter:\n",
611 join("\n", @$linear_refs_), "\n",
612 'If you are attempting to commit ',
613 "merges, try running:\n\t",
614 'git rebase --interactive',
615 '--preserve-merges ',
616 $gs->refname,
617 "\nBefore dcommitting";
618 }
619 if ($url_ ne $expect_url) {
620 if ($url_ eq $gs->metadata_url) {
621 print
622 "Accepting rewritten URL:",
623 " $url_\n";
624 } else {
625 fatal
626 "URL mismatch after rebase:",
627 " $url_ != $expect_url";
628 }
629 }
630 if ($uuid_ ne $uuid) {
631 fatal "uuid mismatch after rebase: ",
632 "$uuid_ != $uuid";
633 }
634 # remap parents
635 my (%p, @l, $i);
636 for ($i = 0; $i < scalar @$linear_refs; $i++) {
637 my $new = $linear_refs_->[$i] or next;
638 $p{$new} =
639 $parents->{$linear_refs->[$i]};
640 push @l, $new;
641 }
642 $parents = \%p;
643 $linear_refs = \@l;
644 }
645 }
646 }
647 if ($old_head) {
649 my $new_head = command_oneline(qw/rev-parse HEAD/);
650 my $new_is_symbolic = eval {
651 command_oneline(qw/symbolic-ref -q HEAD/);
652 };
653 if ($new_is_symbolic) {
654 print "dcommitted the branch ", $head, "\n";
655 } else {
656 print "dcommitted on a detached HEAD because you gave ",
657 "a revision argument.\n",
658 "The rewritten commit is: ", $new_head, "\n";
659 }
660 command(['checkout', $old_head], STDERR => 0);
661 }
662 unlink $gs->{index};
664}
665sub cmd_branch {
667 my ($branch_name, $head) = @_;
668 unless (defined $branch_name && length $branch_name) {
670 die(($_tag ? "tag" : "branch") . " name required\n");
671 }
672 $head ||= 'HEAD';
673 my (undef, $rev, undef, $gs) = working_head_info($head);
675 my $src = $gs->full_url;
676 my $remote = Git::SVN::read_all_remotes()->{$gs->{repo_id}};
678 my $allglobs = $remote->{ $_tag ? 'tags' : 'branches' };
679 my $glob;
680 if ($#{$allglobs} == 0) {
681 $glob = $allglobs->[0];
682 } else {
683 unless(defined $_branch_dest) {
684 die "Multiple ",
685 $_tag ? "tag" : "branch",
686 " paths defined for Subversion repository.\n",
687 "You must specify where you want to create the ",
688 $_tag ? "tag" : "branch",
689 " with the --destination argument.\n";
690 }
691 foreach my $g (@{$allglobs}) {
692 # SVN::Git::Editor could probably be moved to Git.pm..
693 my $re = SVN::Git::Editor::glob2pat($g->{path}->{left});
694 if ($_branch_dest =~ /$re/) {
695 $glob = $g;
696 last;
697 }
698 }
699 unless (defined $glob) {
700 my $dest_re = qr/\b\Q$_branch_dest\E\b/;
701 foreach my $g (@{$allglobs}) {
702 $g->{path}->{left} =~ /$dest_re/ or next;
703 if (defined $glob) {
704 die "Ambiguous destination: ",
705 $_branch_dest, "\nmatches both '",
706 $glob->{path}->{left}, "' and '",
707 $g->{path}->{left}, "'\n";
708 }
709 $glob = $g;
710 }
711 unless (defined $glob) {
712 die "Unknown ",
713 $_tag ? "tag" : "branch",
714 " destination $_branch_dest\n";
715 }
716 }
717 }
718 my ($lft, $rgt) = @{ $glob->{path} }{qw/left right/};
719 my $url;
720 if (defined $_commit_url) {
721 $url = $_commit_url;
722 } else {
723 $url = eval { command_oneline('config', '--get',
724 "svn-remote.$gs->{repo_id}.commiturl") };
725 if (!$url) {
726 $url = $remote->{url};
727 }
728 }
729 my $dst = join '/', $url, $lft, $branch_name, ($rgt || ());
730 if ($dst =~ /^https:/ && $src =~ /^http:/) {
732 $src=~s/^http:/https:/;
733 }
734 ::_req_svn();
736 my $ctx = SVN::Client->new(
738 auth => Git::SVN::Ra::_auth_providers(),
739 log_msg => sub {
740 ${ $_[0] } = defined $_message
741 ? $_message
742 : 'Create ' . ($_tag ? 'tag ' : 'branch ' )
743 . $branch_name;
744 },
745 );
746 eval {
748 $ctx->ls($dst, 'HEAD', 0);
749 } and die "branch ${branch_name} already exists\n";
750 print "Copying ${src} at r${rev} to ${dst}...\n";
752 $ctx->copy($src, $rev, $dst)
753 unless $_dry_run;
754 $gs->fetch_all;
756}
757sub cmd_find_rev {
759 my $revision_or_hash = shift or die "SVN or git revision required ",
760 "as a command-line argument\n";
761 my $result;
762 if ($revision_or_hash =~ /^r\d+$/) {
763 my $head = shift;
764 $head ||= 'HEAD';
765 my @refs;
766 my (undef, undef, $uuid, $gs) = working_head_info($head, \@refs);
767 unless ($gs) {
768 die "Unable to determine upstream SVN information from ",
769 "$head history\n";
770 }
771 my $desired_revision = substr($revision_or_hash, 1);
772 $result = $gs->rev_map_get($desired_revision, $uuid);
773 } else {
774 my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
775 $result = $rev;
776 }
777 print "$result\n" if $result;
778}
779sub cmd_rebase {
781 command_noisy(qw/update-index --refresh/);
782 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
783 unless ($gs) {
784 die "Unable to determine upstream SVN information from ",
785 "working tree history\n";
786 }
787 if ($_dry_run) {
788 print "Remote Branch: " . $gs->refname . "\n";
789 print "SVN URL: " . $url . "\n";
790 return;
791 }
792 if (command(qw/diff-index HEAD --/)) {
793 print STDERR "Cannot rebase with uncommited changes:\n";
794 command_noisy('status');
795 exit 1;
796 }
797 unless ($_local) {
798 # rebase will checkout for us, so no need to do it explicitly
799 $_no_checkout = 'true';
800 $_fetch_all ? $gs->fetch_all : $gs->fetch;
801 }
802 command_noisy(rebase_cmd(), $gs->refname);
803 $gs->mkemptydirs;
804}
805sub cmd_show_ignore {
807 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
808 $gs ||= Git::SVN->new;
809 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
810 $gs->prop_walk($gs->{path}, $r, sub {
811 my ($gs, $path, $props) = @_;
812 print STDOUT "\n# $path\n";
813 my $s = $props->{'svn:ignore'} or return;
814 $s =~ s/[\r\n]+/\n/g;
815 $s =~ s/^\n+//;
816 chomp $s;
817 $s =~ s#^#$path#gm;
818 print STDOUT "$s\n";
819 });
820}
821sub cmd_show_externals {
823 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
824 $gs ||= Git::SVN->new;
825 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
826 $gs->prop_walk($gs->{path}, $r, sub {
827 my ($gs, $path, $props) = @_;
828 print STDOUT "\n# $path\n";
829 my $s = $props->{'svn:externals'} or return;
830 $s =~ s/[\r\n]+/\n/g;
831 chomp $s;
832 $s =~ s#^#$path#gm;
833 print STDOUT "$s\n";
834 });
835}
836sub cmd_create_ignore {
838 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
839 $gs ||= Git::SVN->new;
840 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
841 $gs->prop_walk($gs->{path}, $r, sub {
842 my ($gs, $path, $props) = @_;
843 # $path is of the form /path/to/dir/
844 $path = '.' . $path;
845 # SVN can have attributes on empty directories,
846 # which git won't track
847 mkpath([$path]) unless -d $path;
848 my $ignore = $path . '.gitignore';
849 my $s = $props->{'svn:ignore'} or return;
850 open(GITIGNORE, '>', $ignore)
851 or fatal("Failed to open `$ignore' for writing: $!");
852 $s =~ s/[\r\n]+/\n/g;
853 $s =~ s/^\n+//;
854 chomp $s;
855 # Prefix all patterns so that the ignore doesn't apply
856 # to sub-directories.
857 $s =~ s#^#/#gm;
858 print GITIGNORE "$s\n";
859 close(GITIGNORE)
860 or fatal("Failed to close `$ignore': $!");
861 command_noisy('add', '-f', $ignore);
862 });
863}
864sub cmd_mkdirs {
866 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
867 $gs ||= Git::SVN->new;
868 $gs->mkemptydirs($_revision);
869}
870sub canonicalize_path {
872 my ($path) = @_;
873 my $dot_slash_added = 0;
874 if (substr($path, 0, 1) ne "/") {
875 $path = "./" . $path;
876 $dot_slash_added = 1;
877 }
878 # File::Spec->canonpath doesn't collapse x/../y into y (for a
879 # good reason), so let's do this manually.
880 $path =~ s#/+#/#g;
881 $path =~ s#/\.(?:/|$)#/#g;
882 $path =~ s#/[^/]+/\.\.##g;
883 $path =~ s#/$##g;
884 $path =~ s#^\./## if $dot_slash_added;
885 $path =~ s#^/##;
886 $path =~ s#^\.$##;
887 return $path;
888}
889sub canonicalize_url {
891 my ($url) = @_;
892 $url =~ s#^([^:]+://[^/]*/)(.*)$#$1 . canonicalize_path($2)#e;
893 return $url;
894}
895# get_svnprops(PATH)
897# ------------------
898# Helper for cmd_propget and cmd_proplist below.
899sub get_svnprops {
900 my $path = shift;
901 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
902 $gs ||= Git::SVN->new;
903 # prefix THE PATH by the sub-directory from which the user
905 # invoked us.
906 $path = $cmd_dir_prefix . $path;
907 fatal("No such file or directory: $path") unless -e $path;
908 my $is_dir = -d $path ? 1 : 0;
909 $path = $gs->{path} . '/' . $path;
910 # canonicalize the path (otherwise libsvn will abort or fail to
912 # find the file)
913 $path = canonicalize_path($path);
914 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
916 my $props;
917 if ($is_dir) {
918 (undef, undef, $props) = $gs->ra->get_dir($path, $r);
919 }
920 else {
921 (undef, $props) = $gs->ra->get_file($path, $r, undef);
922 }
923 return $props;
924}
925# cmd_propget (PROP, PATH)
927# ------------------------
928# Print the SVN property PROP for PATH.
929sub cmd_propget {
930 my ($prop, $path) = @_;
931 $path = '.' if not defined $path;
932 usage(1) if not defined $prop;
933 my $props = get_svnprops($path);
934 if (not defined $props->{$prop}) {
935 fatal("`$path' does not have a `$prop' SVN property.");
936 }
937 print $props->{$prop} . "\n";
938}
939# cmd_proplist (PATH)
941# -------------------
942# Print the list of SVN properties for PATH.
943sub cmd_proplist {
944 my $path = shift;
945 $path = '.' if not defined $path;
946 my $props = get_svnprops($path);
947 print "Properties on '$path':\n";
948 foreach (sort keys %{$props}) {
949 print " $_\n";
950 }
951}
952sub cmd_multi_init {
954 my $url = shift;
955 unless (defined $_trunk || @_branches || @_tags) {
956 usage(1);
957 }
958 $_prefix = '' unless defined $_prefix;
960 if (defined $url) {
961 $url = canonicalize_url($url);
962 init_subdir(@_);
963 }
964 do_git_init_db();
965 if (defined $_trunk) {
966 my $trunk_ref = 'refs/remotes/' . $_prefix . 'trunk';
967 # try both old-style and new-style lookups:
968 my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
969 unless ($gs_trunk) {
970 my ($trunk_url, $trunk_path) =
971 complete_svn_url($url, $_trunk);
972 $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
973 undef, $trunk_ref);
974 }
975 }
976 return unless @_branches || @_tags;
977 my $ra = $url ? Git::SVN::Ra->new($url) : undef;
978 foreach my $path (@_branches) {
979 complete_url_ls_init($ra, $path, '--branches/-b', $_prefix);
980 }
981 foreach my $path (@_tags) {
982 complete_url_ls_init($ra, $path, '--tags/-t', $_prefix.'tags/');
983 }
984}
985sub cmd_multi_fetch {
987 $Git::SVN::no_reuse_existing = undef;
988 my $remotes = Git::SVN::read_all_remotes();
989 foreach my $repo_id (sort keys %$remotes) {
990 if ($remotes->{$repo_id}->{url}) {
991 Git::SVN::fetch_all($repo_id, $remotes);
992 }
993 }
994}
995# this command is special because it requires no metadata
997sub cmd_commit_diff {
998 my ($ta, $tb, $url) = @_;
999 my $usage = "Usage: $0 commit-diff -r<revision> ".
1000 "<tree-ish> <tree-ish> [<URL>]";
1001 fatal($usage) if (!defined $ta || !defined $tb);
1002 my $svn_path = '';
1003 if (!defined $url) {
1004 my $gs = eval { Git::SVN->new };
1005 if (!$gs) {
1006 fatal("Needed URL or usable git-svn --id in ",
1007 "the command-line\n", $usage);
1008 }
1009 $url = $gs->{url};
1010 $svn_path = $gs->{path};
1011 }
1012 unless (defined $_revision) {
1013 fatal("-r|--revision is a required argument\n", $usage);
1014 }
1015 if (defined $_message && defined $_file) {
1016 fatal("Both --message/-m and --file/-F specified ",
1017 "for the commit message.\n",
1018 "I have no idea what you mean");
1019 }
1020 if (defined $_file) {
1021 $_message = file_to_s($_file);
1022 } else {
1023 $_message ||= get_commit_entry($tb)->{log};
1024 }
1025 my $ra ||= Git::SVN::Ra->new($url);
1026 my $r = $_revision;
1027 if ($r eq 'HEAD') {
1028 $r = $ra->get_latest_revnum;
1029 } elsif ($r !~ /^\d+$/) {
1030 die "revision argument: $r not understood by git-svn\n";
1031 }
1032 my %ed_opts = ( r => $r,
1033 log => $_message,
1034 ra => $ra,
1035 tree_a => $ta,
1036 tree_b => $tb,
1037 editor_cb => sub { print "Committed r$_[0]\n" },
1038 svn_path => $svn_path );
1039 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
1040 print "No changes\n$ta == $tb\n";
1041 }
1042}
1043sub escape_uri_only {
1045 my ($uri) = @_;
1046 my @tmp;
1047 foreach (split m{/}, $uri) {
1048 s/([^~\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
1049 push @tmp, $_;
1050 }
1051 join('/', @tmp);
1052}
1053sub escape_url {
1055 my ($url) = @_;
1056 if ($url =~ m#^([^:]+)://([^/]*)(.*)$#) {
1057 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
1058 $url = "$scheme://$domain$uri";
1059 }
1060 $url;
1061}
1062sub cmd_info {
1064 my $path = canonicalize_path(defined($_[0]) ? $_[0] : ".");
1065 my $fullpath = canonicalize_path($cmd_dir_prefix . $path);
1066 if (exists $_[1]) {
1067 die "Too many arguments specified\n";
1068 }
1069 my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
1071 if (!$file_type && !$diff_status) {
1073 print STDERR "svn: '$path' is not under version control\n";
1074 exit 1;
1075 }
1076 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
1078 unless ($gs) {
1079 die "Unable to determine upstream SVN information from ",
1080 "working tree history\n";
1081 }
1082 # canonicalize_path() will return "" to make libsvn 1.5.x happy,
1084 $path = "." if $path eq "";
1085 my $full_url = $url . ($fullpath eq "" ? "" : "/$fullpath");
1087 if ($_url) {
1089 print escape_url($full_url), "\n";
1090 return;
1091 }
1092 my $result = "Path: $path\n";
1094 $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
1095 $result .= "URL: " . escape_url($full_url) . "\n";
1096 eval {
1098 my $repos_root = $gs->repos_root;
1099 Git::SVN::remove_username($repos_root);
1100 $result .= "Repository Root: " . escape_url($repos_root) . "\n";
1101 };
1102 if ($@) {
1103 $result .= "Repository Root: (offline)\n";
1104 }
1105 ::_req_svn();
1106 $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A" &&
1107 ($SVN::Core::VERSION le '1.5.4' || $file_type ne "dir");
1108 $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
1109 $result .= "Node Kind: " .
1111 ($file_type eq "dir" ? "directory" : "file") . "\n";
1112 my $schedule = $diff_status eq "A"
1114 ? "add"
1115 : ($diff_status eq "D" ? "delete" : "normal");
1116 $result .= "Schedule: $schedule\n";
1117 if ($diff_status eq "A") {
1119 print $result, "\n";
1120 return;
1121 }
1122 my ($lc_author, $lc_rev, $lc_date_utc);
1124 my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $fullpath);
1125 my $log = command_output_pipe(@args);
1126 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
1127 while (<$log>) {
1128 if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
1129 $lc_author = $1;
1130 $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
1131 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
1132 (undef, $lc_rev, undef) = ::extract_metadata($1);
1133 }
1134 }
1135 close $log;
1136 Git::SVN::Log::set_local_timezone();
1138 $result .= "Last Changed Author: $lc_author\n";
1140 $result .= "Last Changed Rev: $lc_rev\n";
1141 $result .= "Last Changed Date: " .
1142 Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
1143 if ($file_type ne "dir") {
1145 my $text_last_updated_date =
1146 ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
1147 $result .=
1148 "Text Last Updated: " .
1149 Git::SVN::Log::format_svn_date($text_last_updated_date) .
1150 "\n";
1151 my $checksum;
1152 if ($diff_status eq "D") {
1153 my ($fh, $ctx) =
1154 command_output_pipe(qw(cat-file blob), "HEAD:$path");
1155 if ($file_type eq "link") {
1156 my $file_name = <$fh>;
1157 $checksum = md5sum("link $file_name");
1158 } else {
1159 $checksum = md5sum($fh);
1160 }
1161 command_close_pipe($fh, $ctx);
1162 } elsif ($file_type eq "link") {
1163 my $file_name =
1164 command(qw(cat-file blob), "HEAD:$path");
1165 $checksum =
1166 md5sum("link " . $file_name);
1167 } else {
1168 open FILE, "<", $path or die $!;
1169 $checksum = md5sum(\*FILE);
1170 close FILE or die $!;
1171 }
1172 $result .= "Checksum: " . $checksum . "\n";
1173 }
1174 print $result, "\n";
1176}
1177sub cmd_reset {
1179 my $target = shift || $_revision or die "SVN revision required\n";
1180 $target = $1 if $target =~ /^r(\d+)$/;
1181 $target =~ /^\d+$/ or die "Numeric SVN revision expected\n";
1182 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
1183 unless ($gs) {
1184 die "Unable to determine upstream SVN information from ".
1185 "history\n";
1186 }
1187 my ($r, $c) = $gs->find_rev_before($target, not $_fetch_parent);
1188 die "Cannot find SVN revision $target\n" unless defined($c);
1189 $gs->rev_map_set($r, $c, 'reset', $uuid);
1190 print "r$r = $c ($gs->{ref_id})\n";
1191}
1192sub cmd_gc {
1194 if (!$can_compress) {
1195 warn "Compress::Zlib could not be found; unhandled.log " .
1196 "files will not be compressed.\n";
1197 }
1198 find({ wanted => \&gc_directory, no_chdir => 1}, "$ENV{GIT_DIR}/svn");
1199}
1200########################### utility functions #########################
1202sub rebase_cmd {
1204 my @cmd = qw/rebase/;
1205 push @cmd, '-v' if $_verbose;
1206 push @cmd, qw/--merge/ if $_merge;
1207 push @cmd, "--strategy=$_strategy" if $_strategy;
1208 @cmd;
1209}
1210sub post_fetch_checkout {
1212 return if $_no_checkout;
1213 my $gs = $Git::SVN::_head or return;
1214 return if verify_ref('refs/heads/master^0');
1215 # look for "trunk" ref if it exists
1217 my $remote = Git::SVN::read_all_remotes()->{$gs->{repo_id}};
1218 my $fetch = $remote->{fetch};
1219 if ($fetch) {
1220 foreach my $p (keys %$fetch) {
1221 basename($fetch->{$p}) eq 'trunk' or next;
1222 $gs = Git::SVN->new($fetch->{$p}, $gs->{repo_id}, $p);
1223 last;
1224 }
1225 }
1226 my $valid_head = verify_ref('HEAD^0');
1228 command_noisy(qw(update-ref refs/heads/master), $gs->refname);
1229 return if ($valid_head || !verify_ref('HEAD^0'));
1230 return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
1232 my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
1233 return if -f $index;
1234 return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
1236 return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
1237 command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
1238 print STDERR "Checked out HEAD:\n ",
1239 $gs->full_url, " r", $gs->last_rev, "\n";
1240 $gs->mkemptydirs($gs->last_rev);
1241}
1242sub complete_svn_url {
1244 my ($url, $path) = @_;
1245 $path =~ s#/+$##;
1246 if ($path !~ m#^[a-z\+]+://#) {
1247 if (!defined $url || $url !~ m#^[a-z\+]+://#) {
1248 fatal("E: '$path' is not a complete URL ",
1249 "and a separate URL is not specified");
1250 }
1251 return ($url, $path);
1252 }
1253 return ($path, '');
1254}
1255sub complete_url_ls_init {
1257 my ($ra, $repo_path, $switch, $pfx) = @_;
1258 unless ($repo_path) {
1259 print STDERR "W: $switch not specified\n";
1260 return;
1261 }
1262 $repo_path =~ s#/+$##;
1263 if ($repo_path =~ m#^[a-z\+]+://#) {
1264 $ra = Git::SVN::Ra->new($repo_path);
1265 $repo_path = '';
1266 } else {
1267 $repo_path =~ s#^/+##;
1268 unless ($ra) {
1269 fatal("E: '$repo_path' is not a complete URL ",
1270 "and a separate URL is not specified");
1271 }
1272 }
1273 my $url = $ra->{url};
1274 my $gs = Git::SVN->init($url, undef, undef, undef, 1);
1275 my $k = "svn-remote.$gs->{repo_id}.url";
1276 my $orig_url = eval { command_oneline(qw/config --get/, $k) };
1277 if ($orig_url && ($orig_url ne $gs->{url})) {
1278 die "$k already set: $orig_url\n",
1279 "wanted to set to: $gs->{url}\n";
1280 }
1281 command_oneline('config', $k, $gs->{url}) unless $orig_url;
1282 my $remote_path = "$gs->{path}/$repo_path";
1283 $remote_path =~ s{%([0-9A-F]{2})}{chr hex($1)}ieg;
1284 $remote_path =~ s#/+#/#g;
1285 $remote_path =~ s#^/##g;
1286 $remote_path .= "/*" if $remote_path !~ /\*/;
1287 my ($n) = ($switch =~ /^--(\w+)/);
1288 if (length $pfx && $pfx !~ m#/$#) {
1289 die "--prefix='$pfx' must have a trailing slash '/'\n";
1290 }
1291 command_noisy('config',
1292 '--add',
1293 "svn-remote.$gs->{repo_id}.$n",
1294 "$remote_path:refs/remotes/$pfx*" .
1295 ('/*' x (($remote_path =~ tr/*/*/) - 1)) );
1296}
1297sub verify_ref {
1299 my ($ref) = @_;
1300 eval { command_oneline([ 'rev-parse', '--verify', $ref ],
1301 { STDERR => 0 }); };
1302}
1303sub get_tree_from_treeish {
1305 my ($treeish) = @_;
1306 # $treeish can be a symbolic ref, too:
1307 my $type = command_oneline(qw/cat-file -t/, $treeish);
1308 my $expected;
1309 while ($type eq 'tag') {
1310 ($treeish, $type) = command(qw/cat-file tag/, $treeish);
1311 }
1312 if ($type eq 'commit') {
1313 $expected = (grep /^tree /, command(qw/cat-file commit/,
1314 $treeish))[0];
1315 ($expected) = ($expected =~ /^tree ($sha1)$/o);
1316 die "Unable to get tree from $treeish\n" unless $expected;
1317 } elsif ($type eq 'tree') {
1318 $expected = $treeish;
1319 } else {
1320 die "$treeish is a $type, expected tree, tag or commit\n";
1321 }
1322 return $expected;
1323}
1324sub get_commit_entry {
1326 my ($treeish) = shift;
1327 my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
1328 my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
1329 my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
1330 open my $log_fh, '>', $commit_editmsg or croak $!;
1331 my $type = command_oneline(qw/cat-file -t/, $treeish);
1333 if ($type eq 'commit' || $type eq 'tag') {
1334 my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1335 $type, $treeish);
1336 my $in_msg = 0;
1337 my $author;
1338 my $saw_from = 0;
1339 my $msgbuf = "";
1340 while (<$msg_fh>) {
1341 if (!$in_msg) {
1342 $in_msg = 1 if (/^\s*$/);
1343 $author = $1 if (/^author (.*>)/);
1344 } elsif (/^git-svn-id: /) {
1345 # skip this for now, we regenerate the
1346 # correct one on re-fetch anyways
1347 # TODO: set *:merge properties or like...
1348 } else {
1349 if (/^From:/ || /^Signed-off-by:/) {
1350 $saw_from = 1;
1351 }
1352 $msgbuf .= $_;
1353 }
1354 }
1355 $msgbuf =~ s/\s+$//s;
1356 if ($Git::SVN::_add_author_from && defined($author)
1357 && !$saw_from) {
1358 $msgbuf .= "\n\nFrom: $author";
1359 }
1360 print $log_fh $msgbuf or croak $!;
1361 command_close_pipe($msg_fh, $ctx);
1362 }
1363 close $log_fh or croak $!;
1364 if ($_edit || ($type eq 'tree')) {
1366 chomp(my $editor = command_oneline(qw(var GIT_EDITOR)));
1367 system('sh', '-c', $editor.' "$@"', $editor, $commit_editmsg);
1368 }
1369 rename $commit_editmsg, $commit_msg or croak $!;
1370 {
1371 require Encode;
1372 # SVN requires messages to be UTF-8 when entering the repo
1373 local $/;
1374 open $log_fh, '<', $commit_msg or croak $!;
1375 binmode $log_fh;
1376 chomp($log_entry{log} = <$log_fh>);
1377 my $enc = Git::config('i18n.commitencoding') || 'UTF-8';
1379 my $msg = $log_entry{log};
1380 eval { $msg = Encode::decode($enc, $msg, 1) };
1382 if ($@) {
1383 die "Could not decode as $enc:\n", $msg,
1384 "\nPerhaps you need to set i18n.commitencoding\n";
1385 }
1386 eval { $msg = Encode::encode('UTF-8', $msg, 1) };
1388 die "Could not encode as UTF-8:\n$msg\n" if $@;
1389 $log_entry{log} = $msg;
1391 close $log_fh or croak $!;
1393 }
1394 unlink $commit_msg;
1395 \%log_entry;
1396}
1397sub s_to_file {
1399 my ($str, $file, $mode) = @_;
1400 open my $fd,'>',$file or croak $!;
1401 print $fd $str,"\n" or croak $!;
1402 close $fd or croak $!;
1403 chmod ($mode &~ umask, $file) if (defined $mode);
1404}
1405sub file_to_s {
1407 my $file = shift;
1408 open my $fd,'<',$file or croak "$!: file: $file\n";
1409 local $/;
1410 my $ret = <$fd>;
1411 close $fd or croak $!;
1412 $ret =~ s/\s*$//s;
1413 return $ret;
1414}
1415# '<svn username> = real-name <email address>' mapping based on git-svnimport:
1417sub load_authors {
1418 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1419 my $log = $cmd eq 'log';
1420 while (<$authors>) {
1421 chomp;
1422 next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1423 my ($user, $name, $email) = ($1, $2, $3);
1424 if ($log) {
1425 $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1426 } else {
1427 $users{$user} = [$name, $email];
1428 }
1429 }
1430 close $authors or croak $!;
1431}
1432# convert GetOpt::Long specs for use by git-config
1434sub read_git_config {
1435 my $opts = shift;
1436 my @config_only;
1437 foreach my $o (keys %$opts) {
1438 # if we have mixedCase and a long option-only, then
1439 # it's a config-only variable that we don't need for
1440 # the command-line.
1441 push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1442 my $v = $opts->{$o};
1443 my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1444 $key =~ s/-//g;
1445 my $arg = 'git config';
1446 $arg .= ' --int' if ($o =~ /[:=]i$/);
1447 $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1448 if (ref $v eq 'ARRAY') {
1449 chomp(my @tmp = `$arg --get-all svn.$key`);
1450 @$v = @tmp if @tmp;
1451 } else {
1452 chomp(my $tmp = `$arg --get svn.$key`);
1453 if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1454 $$v = $tmp;
1455 }
1456 }
1457 }
1458 delete @$opts{@config_only} if @config_only;
1459}
1460sub extract_metadata {
1462 my $id = shift or return (undef, undef, undef);
1463 my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1464 \s([a-f\d\-]+)$/ix);
1465 if (!defined $rev || !$uuid || !$url) {
1466 # some of the original repositories I made had
1467 # identifiers like this:
1468 ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/i);
1469 }
1470 return ($url, $rev, $uuid);
1471}
1472sub cmt_metadata {
1474 return extract_metadata((grep(/^git-svn-id: /,
1475 command(qw/cat-file commit/, shift)))[-1]);
1476}
1477sub cmt_sha2rev_batch {
1479 my %s2r;
1480 my ($pid, $in, $out, $ctx) = command_bidi_pipe(qw/cat-file --batch/);
1481 my $list = shift;
1482 foreach my $sha (@{$list}) {
1484 my $first = 1;
1485 my $size = 0;
1486 print $out $sha, "\n";
1487 while (my $line = <$in>) {
1489 if ($first && $line =~ /^[[:xdigit:]]{40}\smissing$/) {
1490 last;
1491 } elsif ($first &&
1492 $line =~ /^[[:xdigit:]]{40}\scommit\s(\d+)$/) {
1493 $first = 0;
1494 $size = $1;
1495 next;
1496 } elsif ($line =~ /^(git-svn-id: )/) {
1497 my (undef, $rev, undef) =
1498 extract_metadata($line);
1499 $s2r{$sha} = $rev;
1500 }
1501 $size -= length($line);
1503 last if ($size == 0);
1504 }
1505 }
1506 command_close_bidi_pipe($pid, $in, $out, $ctx);
1508 return \%s2r;
1510}
1511sub working_head_info {
1513 my ($head, $refs) = @_;
1514 my @args = ('log', '--no-color', '--first-parent', '--pretty=medium');
1515 my ($fh, $ctx) = command_output_pipe(@args, $head);
1516 my $hash;
1517 my %max;
1518 while (<$fh>) {
1519 if ( m{^commit ($::sha1)$} ) {
1520 unshift @$refs, $hash if $hash and $refs;
1521 $hash = $1;
1522 next;
1523 }
1524 next unless s{^\s*(git-svn-id:)}{$1};
1525 my ($url, $rev, $uuid) = extract_metadata($_);
1526 if (defined $url && defined $rev) {
1527 next if $max{$url} and $max{$url} < $rev;
1528 if (my $gs = Git::SVN->find_by_url($url)) {
1529 my $c = $gs->rev_map_get($rev, $uuid);
1530 if ($c && $c eq $hash) {
1531 close $fh; # break the pipe
1532 return ($url, $rev, $uuid, $gs);
1533 } else {
1534 $max{$url} ||= $gs->rev_map_max;
1535 }
1536 }
1537 }
1538 }
1539 command_close_pipe($fh, $ctx);
1540 (undef, undef, undef, undef);
1541}
1542sub read_commit_parents {
1544 my ($parents, $c) = @_;
1545 chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1546 $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1547 @{$parents->{$c}} = split(/ /, $p);
1548}
1549sub linearize_history {
1551 my ($gs, $refs) = @_;
1552 my %parents;
1553 foreach my $c (@$refs) {
1554 read_commit_parents(\%parents, $c);
1555 }
1556 my @linear_refs;
1558 my %skip = ();
1559 my $last_svn_commit = $gs->last_commit;
1560 foreach my $c (reverse @$refs) {
1561 next if $c eq $last_svn_commit;
1562 last if $skip{$c};
1563 unshift @linear_refs, $c;
1565 $skip{$c} = 1;
1566 # we only want the first parent to diff against for linear
1568 # history, we save the rest to inject when we finalize the
1569 # svn commit
1570 my $fp_a = verify_ref("$c~1");
1571 my $fp_b = shift @{$parents{$c}} if $parents{$c};
1572 if (!$fp_a || !$fp_b) {
1573 die "Commit $c\n",
1574 "has no parent commit, and therefore ",
1575 "nothing to diff against.\n",
1576 "You should be working from a repository ",
1577 "originally created by git-svn\n";
1578 }
1579 if ($fp_a ne $fp_b) {
1580 die "$c~1 = $fp_a, however parsing commit $c ",
1581 "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1582 }
1583 foreach my $p (@{$parents{$c}}) {
1585 $skip{$p} = 1;
1586 }
1587 }
1588 (\@linear_refs, \%parents);
1589}
1590sub find_file_type_and_diff_status {
1592 my ($path) = @_;
1593 return ('dir', '') if $path eq '';
1594 my $diff_output =
1596 command_oneline(qw(diff --cached --name-status --), $path) || "";
1597 my $diff_status = (split(' ', $diff_output))[0] || "";
1598 my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1600 return (undef, undef) if !$diff_status && !$ls_tree;
1602 if ($diff_status eq "A") {
1604 return ("link", $diff_status) if -l $path;
1605 return ("dir", $diff_status) if -d $path;
1606 return ("file", $diff_status);
1607 }
1608 my $mode = (split(' ', $ls_tree))[0] || "";
1610 return ("link", $diff_status) if $mode eq "120000";
1612 return ("dir", $diff_status) if $mode eq "040000";
1613 return ("file", $diff_status);
1614}
1615sub md5sum {
1617 my $arg = shift;
1618 my $ref = ref $arg;
1619 my $md5 = Digest::MD5->new();
1620 if ($ref eq 'GLOB' || $ref eq 'IO::File' || $ref eq 'File::Temp') {
1621 $md5->addfile($arg) or croak $!;
1622 } elsif ($ref eq 'SCALAR') {
1623 $md5->add($$arg) or croak $!;
1624 } elsif (!$ref) {
1625 $md5->add($arg) or croak $!;
1626 } else {
1627 ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
1628 }
1629 return $md5->hexdigest();
1630}
1631sub gc_directory {
1633 if ($can_compress && -f $_ && basename($_) eq "unhandled.log") {
1634 my $out_filename = $_ . ".gz";
1635 open my $in_fh, "<", $_ or die "Unable to open $_: $!\n";
1636 binmode $in_fh;
1637 my $gz = Compress::Zlib::gzopen($out_filename, "ab") or
1638 die "Unable to open $out_filename: $!\n";
1639 my $res;
1641 while ($res = sysread($in_fh, my $str, 1024)) {
1642 $gz->gzwrite($str) or
1643 die "Unable to write: ".$gz->gzerror()."!\n";
1644 }
1645 unlink $_ or die "unlink $File::Find::name: $!\n";
1646 } elsif (-f $_ && basename($_) eq "index") {
1647 unlink $_ or die "unlink $_: $!\n";
1648 }
1649}
1650package Git::SVN;
1652use strict;
1653use warnings;
1654use Fcntl qw/:DEFAULT :seek/;
1655use constant rev_map_fmt => 'NH40';
1656use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1657 $_repack $_repack_flags $_use_svm_props $_head
1658 $_use_svnsync_props $no_reuse_existing $_minimize_url
1659 $_use_log_author $_add_author_from $_localtime/;
1660use Carp qw/croak/;
1661use File::Path qw/mkpath/;
1662use File::Copy qw/copy/;
1663use IPC::Open3;
1664use Memoize; # core since 5.8.0, Jul 2002
1665use Memoize::Storable;
1666my ($_gc_nr, $_gc_period);
1668# properties that we do not log:
1670my %SKIP_PROP;
1671BEGIN {
1672 %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1673 svn:special svn:executable
1674 svn:entry:committed-rev
1675 svn:entry:last-author
1676 svn:entry:uuid
1677 svn:entry:committed-date/;
1678 # some options are read globally, but can be overridden locally
1680 # per [svn-remote "..."] section. Command-line options will *NOT*
1681 # override options set in an [svn-remote "..."] section
1682 no strict 'refs';
1683 for my $option (qw/follow_parent no_metadata use_svm_props
1684 use_svnsync_props/) {
1685 my $key = $option;
1686 $key =~ tr/_//d;
1687 my $prop = "-$option";
1688 *$option = sub {
1689 my ($self) = @_;
1690 return $self->{$prop} if exists $self->{$prop};
1691 my $k = "svn-remote.$self->{repo_id}.$key";
1692 eval { command_oneline(qw/config --get/, $k) };
1693 if ($@) {
1694 $self->{$prop} = ${"Git::SVN::_$option"};
1695 } else {
1696 my $v = command_oneline(qw/config --bool/,$k);
1697 $self->{$prop} = $v eq 'false' ? 0 : 1;
1698 }
1699 return $self->{$prop};
1700 }
1701 }
1702}
1703my (%LOCKFILES, %INDEX_FILES);
1706END {
1707 unlink keys %LOCKFILES if %LOCKFILES;
1708 unlink keys %INDEX_FILES if %INDEX_FILES;
1709}
1710sub resolve_local_globs {
1712 my ($url, $fetch, $glob_spec) = @_;
1713 return unless defined $glob_spec;
1714 my $ref = $glob_spec->{ref};
1715 my $path = $glob_spec->{path};
1716 foreach (command(qw#for-each-ref --format=%(refname) refs/#)) {
1717 next unless m#^$ref->{regex}$#;
1718 my $p = $1;
1719 my $pathname = desanitize_refname($path->full_path($p));
1720 my $refname = desanitize_refname($ref->full_path($p));
1721 if (my $existing = $fetch->{$pathname}) {
1722 if ($existing ne $refname) {
1723 die "Refspec conflict:\n",
1724 "existing: $existing\n",
1725 " globbed: $refname\n";
1726 }
1727 my $u = (::cmt_metadata("$refname"))[0];
1728 $u =~ s!^\Q$url\E(/|$)!! or die
1729 "$refname: '$url' not found in '$u'\n";
1730 if ($pathname ne $u) {
1731 warn "W: Refspec glob conflict ",
1732 "(ref: $refname):\n",
1733 "expected path: $pathname\n",
1734 " real path: $u\n",
1735 "Continuing ahead with $u\n";
1736 next;
1737 }
1738 } else {
1739 $fetch->{$pathname} = $refname;
1740 }
1741 }
1742}
1743sub parse_revision_argument {
1745 my ($base, $head) = @_;
1746 if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1747 return ($base, $head);
1748 }
1749 return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1750 return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1751 return ($head, $head) if ($::_revision eq 'HEAD');
1752 return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1753 return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1754 die "revision argument: $::_revision not understood by git-svn\n";
1755}
1756sub fetch_all {
1758 my ($repo_id, $remotes) = @_;
1759 if (ref $repo_id) {
1760 my $gs = $repo_id;
1761 $repo_id = undef;
1762 $repo_id = $gs->{repo_id};
1763 }
1764 $remotes ||= read_all_remotes();
1765 my $remote = $remotes->{$repo_id} or
1766 die "[svn-remote \"$repo_id\"] unknown\n";
1767 my $fetch = $remote->{fetch};
1768 my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1769 my (@gs, @globs);
1770 my $ra = Git::SVN::Ra->new($url);
1771 my $uuid = $ra->get_uuid;
1772 my $head = $ra->get_latest_revnum;
1773 # ignore errors, $head revision may not even exist anymore
1775 eval { $ra->get_log("", $head, 0, 1, 0, 1, sub { $head = $_[1] }) };
1776 warn "W: $@\n" if $@;
1777 my $base = defined $fetch ? $head : 0;
1779 # read the max revs for wildcard expansion (branches/*, tags/*)
1781 foreach my $t (qw/branches tags/) {
1782 defined $remote->{$t} or next;
1783 push @globs, @{$remote->{$t}};
1784 my $max_rev = eval { tmp_config(qw/--int --get/,
1786 "svn-remote.$repo_id.${t}-maxRev") };
1787 if (defined $max_rev && ($max_rev < $base)) {
1788 $base = $max_rev;
1789 } elsif (!defined $max_rev) {
1790 $base = 0;
1791 }
1792 }
1793 if ($fetch) {
1795 foreach my $p (sort keys %$fetch) {
1796 my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1797 my $lr = $gs->rev_map_max;
1798 if (defined $lr) {
1799 $base = $lr if ($lr < $base);
1800 }
1801 push @gs, $gs;
1802 }
1803 }
1804 ($base, $head) = parse_revision_argument($base, $head);
1806 $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1807}
1808sub read_all_remotes {
1810 my $r = {};
1811 my $use_svm_props = eval { command_oneline(qw/config --bool
1812 svn.useSvmProps/) };
1813 $use_svm_props = $use_svm_props eq 'true' if $use_svm_props;
1814 my $svn_refspec = qr{\s*(.*?)\s*:\s*(.+?)\s*};
1815 foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1816 if (m!^(.+)\.fetch=$svn_refspec$!) {
1817 my ($remote, $local_ref, $remote_ref) = ($1, $2, $3);
1818 die("svn-remote.$remote: remote ref '$remote_ref' "
1819 . "must start with 'refs/'\n")
1820 unless $remote_ref =~ m{^refs/};
1821 $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1822 $r->{$remote}->{svm} = {} if $use_svm_props;
1823 } elsif (m!^(.+)\.usesvmprops=\s*(.*)\s*$!) {
1824 $r->{$1}->{svm} = {};
1825 } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1826 $r->{$1}->{url} = $2;
1827 } elsif (m!^(.+)\.(branches|tags)=$svn_refspec$!) {
1828 my ($remote, $t, $local_ref, $remote_ref) =
1829 ($1, $2, $3, $4);
1830 die("svn-remote.$remote: remote ref '$remote_ref' ($t) "
1831 . "must start with 'refs/'\n")
1832 unless $remote_ref =~ m{^refs/};
1833 my $rs = {
1834 t => $t,
1835 remote => $remote,
1836 path => Git::SVN::GlobSpec->new($local_ref, 1),
1837 ref => Git::SVN::GlobSpec->new($remote_ref, 0) };
1838 if (length($rs->{ref}->{right}) != 0) {
1839 die "The '*' glob character must be the last ",
1840 "character of '$remote_ref'\n";
1841 }
1842 push @{ $r->{$remote}->{$t} }, $rs;
1843 }
1844 }
1845 map {
1847 if (defined $r->{$_}->{svm}) {
1848 my $svm;
1849 eval {
1850 my $section = "svn-remote.$_";
1851 $svm = {
1852 source => tmp_config('--get',
1853 "$section.svm-source"),
1854 replace => tmp_config('--get',
1855 "$section.svm-replace"),
1856 }
1857 };
1858 $r->{$_}->{svm} = $svm;
1859 }
1860 } keys %$r;
1861 $r;
1863}
1864sub init_vars {
1866 $_gc_nr = $_gc_period = 1000;
1867 if (defined $_repack || defined $_repack_flags) {
1868 warn "Repack options are obsolete; they have no effect.\n";
1869 }
1870}
1871sub verify_remotes_sanity {
1873 return unless -d $ENV{GIT_DIR};
1874 my %seen;
1875 foreach (command(qw/config -l/)) {
1876 if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1877 if ($seen{$1}) {
1878 die "Remote ref refs/remote/$1 is tracked by",
1879 "\n \"$_\"\nand\n \"$seen{$1}\"\n",
1880 "Please resolve this ambiguity in ",
1881 "your git configuration file before ",
1882 "continuing\n";
1883 }
1884 $seen{$1} = $_;
1885 }
1886 }
1887}
1888sub find_existing_remote {
1890 my ($url, $remotes) = @_;
1891 return undef if $no_reuse_existing;
1892 my $existing;
1893 foreach my $repo_id (keys %$remotes) {
1894 my $u = $remotes->{$repo_id}->{url} or next;
1895 next if $u ne $url;
1896 $existing = $repo_id;
1897 last;
1898 }
1899 $existing;
1900}
1901sub init_remote_config {
1903 my ($self, $url, $no_write) = @_;
1904 $url =~ s!/+$!!; # strip trailing slash
1905 my $r = read_all_remotes();
1906 my $existing = find_existing_remote($url, $r);
1907 if ($existing) {
1908 unless ($no_write) {
1909 print STDERR "Using existing ",
1910 "[svn-remote \"$existing\"]\n";
1911 }
1912 $self->{repo_id} = $existing;
1913 } elsif ($_minimize_url) {
1914 my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1915 $existing = find_existing_remote($min_url, $r);
1916 if ($existing) {
1917 unless ($no_write) {
1918 print STDERR "Using existing ",
1919 "[svn-remote \"$existing\"]\n";
1920 }
1921 $self->{repo_id} = $existing;
1922 }
1923 if ($min_url ne $url) {
1924 unless ($no_write) {
1925 print STDERR "Using higher level of URL: ",
1926 "$url => $min_url\n";
1927 }
1928 my $old_path = $self->{path};
1929 $self->{path} = $url;
1930 $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1931 if (length $old_path) {
1932 $self->{path} .= "/$old_path";
1933 }
1934 $url = $min_url;
1935 }
1936 }
1937 my $orig_url;
1938 if (!$existing) {
1939 # verify that we aren't overwriting anything:
1940 $orig_url = eval {
1941 command_oneline('config', '--get',
1942 "svn-remote.$self->{repo_id}.url")
1943 };
1944 if ($orig_url && ($orig_url ne $url)) {
1945 die "svn-remote.$self->{repo_id}.url already set: ",
1946 "$orig_url\nwanted to set to: $url\n";
1947 }
1948 }
1949 my ($xrepo_id, $xpath) = find_ref($self->refname);
1950 if (!$no_write && defined $xpath) {
1951 die "svn-remote.$xrepo_id.fetch already set to track ",
1952 "$xpath:", $self->refname, "\n";
1953 }
1954 unless ($no_write) {
1955 command_noisy('config',
1956 "svn-remote.$self->{repo_id}.url", $url);
1957 $self->{path} =~ s{^/}{};
1958 $self->{path} =~ s{%([0-9A-F]{2})}{chr hex($1)}ieg;
1959 command_noisy('config', '--add',
1960 "svn-remote.$self->{repo_id}.fetch",
1961 "$self->{path}:".$self->refname);
1962 }
1963 $self->{url} = $url;
1964}
1965sub find_by_url { # repos_root and, path are optional
1967 my ($class, $full_url, $repos_root, $path) = @_;
1968 return undef unless defined $full_url;
1970 remove_username($full_url);
1971 remove_username($repos_root) if defined $repos_root;
1972 my $remotes = read_all_remotes();
1973 if (defined $full_url && defined $repos_root && !defined $path) {
1974 $path = $full_url;
1975 $path =~ s#^\Q$repos_root\E(?:/|$)##;
1976 }
1977 foreach my $repo_id (keys %$remotes) {
1978 my $u = $remotes->{$repo_id}->{url} or next;
1979 remove_username($u);
1980 next if defined $repos_root && $repos_root ne $u;
1981 my $fetch = $remotes->{$repo_id}->{fetch} || {};
1983 foreach my $t (qw/branches tags/) {
1984 foreach my $globspec (@{$remotes->{$repo_id}->{$t}}) {
1985 resolve_local_globs($u, $fetch, $globspec);
1986 }
1987 }
1988 my $p = $path;
1989 my $rwr = rewrite_root({repo_id => $repo_id});
1990 my $svm = $remotes->{$repo_id}->{svm}
1991 if defined $remotes->{$repo_id}->{svm};
1992 unless (defined $p) {
1993 $p = $full_url;
1994 my $z = $u;
1995 my $prefix = '';
1996 if ($rwr) {
1997 $z = $rwr;
1998 remove_username($z);
1999 } elsif (defined $svm) {
2000 $z = $svm->{source};
2001 $prefix = $svm->{replace};
2002 $prefix =~ s#^\Q$u\E(?:/|$)##;
2003 $prefix =~ s#/$##;
2004 }
2005 $p =~ s#^\Q$z\E(?:/|$)#$prefix# or next;
2006 }
2007 foreach my $f (keys %$fetch) {
2008 next if $f ne $p;
2009 return Git::SVN->new($fetch->{$f}, $repo_id, $f);
2010 }
2011 }
2012 undef;
2013}
2014sub init {
2016 my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
2017 my $self = _new($class, $repo_id, $ref_id, $path);
2018 if (defined $url) {
2019 $self->init_remote_config($url, $no_write);
2020 }
2021 $self;
2022}
2023sub find_ref {
2025 my ($ref_id) = @_;
2026 foreach (command(qw/config -l/)) {
2027 next unless m!^svn-remote\.(.+)\.fetch=
2028 \s*(.*?)\s*:\s*(.+?)\s*$!x;
2029 my ($repo_id, $path, $ref) = ($1, $2, $3);
2030 if ($ref eq $ref_id) {
2031 $path = '' if ($path =~ m#^\./?#);
2032 return ($repo_id, $path);
2033 }
2034 }
2035 (undef, undef, undef);
2036}
2037sub new {
2039 my ($class, $ref_id, $repo_id, $path) = @_;
2040 if (defined $ref_id && !defined $repo_id && !defined $path) {
2041 ($repo_id, $path) = find_ref($ref_id);
2042 if (!defined $repo_id) {
2043 die "Could not find a \"svn-remote.*.fetch\" key ",
2044 "in the repository configuration matching: ",
2045 "$ref_id\n";
2046 }
2047 }
2048 my $self = _new($class, $repo_id, $ref_id, $path);
2049 if (!defined $self->{path} || !length $self->{path}) {
2050 my $fetch = command_oneline('config', '--get',
2051 "svn-remote.$repo_id.fetch",
2052 ":$ref_id\$") or
2053 die "Failed to read \"svn-remote.$repo_id.fetch\" ",
2054 "\":$ref_id\$\" in config\n";
2055 ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
2056 }
2057 $self->{url} = command_oneline('config', '--get',
2058 "svn-remote.$repo_id.url") or
2059 die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
2060 $self->rebuild;
2061 $self;
2062}
2063sub refname {
2065 my ($refname) = $_[0]->{ref_id} ;
2066 # It cannot end with a slash /, we'll throw up on this because
2068 # SVN can't have directories with a slash in their name, either:
2069 if ($refname =~ m{/$}) {
2070 die "ref: '$refname' ends with a trailing slash, this is ",
2071 "not permitted by git nor Subversion\n";
2072 }
2073 # It cannot have ASCII control character space, tilde ~, caret ^,
2075 # colon :, question-mark ?, asterisk *, space, or open bracket [
2076 # anywhere.
2077 #
2078 # Additionally, % must be escaped because it is used for escaping
2079 # and we want our escaped refname to be reversible
2080 $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
2081 # no slash-separated component can begin with a dot .
2083 # /.* becomes /%2E*
2084 $refname =~ s{/\.}{/%2E}g;
2085 # It cannot have two consecutive dots .. anywhere
2087 # .. becomes %2E%2E
2088 $refname =~ s{\.\.}{%2E%2E}g;
2089 # trailing dots and .lock are not allowed
2091 # .$ becomes %2E and .lock becomes %2Elock
2092 $refname =~ s{\.(?=$|lock$)}{%2E};
2093 # the sequence @{ is used to access the reflog
2095 # @{ becomes %40{
2096 $refname =~ s{\@\{}{%40\{}g;
2097 return $refname;
2099}
2100sub desanitize_refname {
2102 my ($refname) = @_;
2103 $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
2104 return $refname;
2105}
2106sub svm_uuid {
2108 my ($self) = @_;
2109 return $self->{svm}->{uuid} if $self->svm;
2110 $self->ra;
2111 unless ($self->{svm}) {
2112 die "SVM UUID not cached, and reading remotely failed\n";
2113 }
2114 $self->{svm}->{uuid};
2115}
2116sub svm {
2118 my ($self) = @_;
2119 return $self->{svm} if $self->{svm};
2120 my $svm;
2121 # see if we have it in our config, first:
2122 eval {
2123 my $section = "svn-remote.$self->{repo_id}";
2124 $svm = {
2125 source => tmp_config('--get', "$section.svm-source"),
2126 uuid => tmp_config('--get', "$section.svm-uuid"),
2127 replace => tmp_config('--get', "$section.svm-replace"),
2128 }
2129 };
2130 if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
2131 $self->{svm} = $svm;
2132 }
2133 $self->{svm};
2134}
2135sub _set_svm_vars {
2137 my ($self, $ra) = @_;
2138 return $ra if $self->svm;
2139 my @err = ( "useSvmProps set, but failed to read SVM properties\n",
2141 "(svm:source, svm:uuid) ",
2142 "from the following URLs:\n" );
2143 sub read_svm_props {
2144 my ($self, $ra, $path, $r) = @_;
2145 my $props = ($ra->get_dir($path, $r))[2];
2146 my $src = $props->{'svm:source'};
2147 my $uuid = $props->{'svm:uuid'};
2148 return undef if (!$src || !$uuid);
2149 chomp($src, $uuid);
2151 $uuid =~ m{^[0-9a-f\-]{30,}$}i
2153 or die "doesn't look right - svm:uuid is '$uuid'\n";
2154 # the '!' is used to mark the repos_root!/relative/path
2156 $src =~ s{/?!/?}{/};
2157 $src =~ s{/+$}{}; # no trailing slashes please
2158 # username is of no interest
2159 $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
2160 my $replace = $ra->{url};
2162 $replace .= "/$path" if length $path;
2163 my $section = "svn-remote.$self->{repo_id}";
2165 tmp_config("$section.svm-source", $src);
2166 tmp_config("$section.svm-replace", $replace);
2167 tmp_config("$section.svm-uuid", $uuid);
2168 $self->{svm} = {
2169 source => $src,
2170 uuid => $uuid,
2171 replace => $replace
2172 };
2173 }
2174 my $r = $ra->get_latest_revnum;
2176 my $path = $self->{path};
2177 my %tried;
2178 while (length $path) {
2179 unless ($tried{"$self->{url}/$path"}) {
2180 return $ra if $self->read_svm_props($ra, $path, $r);
2181 $tried{"$self->{url}/$path"} = 1;
2182 }
2183 $path =~ s#/?[^/]+$##;
2184 }
2185 die "Path: '$path' should be ''\n" if $path ne '';
2186 return $ra if $self->read_svm_props($ra, $path, $r);
2187 $tried{"$self->{url}/$path"} = 1;
2188 if ($ra->{repos_root} eq $self->{url}) {
2190 die @err, (map { " $_\n" } keys %tried), "\n";
2191 }
2192 # nope, make sure we're connected to the repository root:
2194 my $ok;
2195 my @tried_b;
2196 $path = $ra->{svn_path};
2197 $ra = Git::SVN::Ra->new($ra->{repos_root});
2198 while (length $path) {
2199 unless ($tried{"$ra->{url}/$path"}) {
2200 $ok = $self->read_svm_props($ra, $path, $r);
2201 last if $ok;
2202 $tried{"$ra->{url}/$path"} = 1;
2203 }
2204 $path =~ s#/?[^/]+$##;
2205 }
2206 die "Path: '$path' should be ''\n" if $path ne '';
2207 $ok ||= $self->read_svm_props($ra, $path, $r);
2208 $tried{"$ra->{url}/$path"} = 1;
2209 if (!$ok) {
2210 die @err, (map { " $_\n" } keys %tried), "\n";
2211 }
2212 Git::SVN::Ra->new($self->{url});
2213}
2214sub svnsync {
2216 my ($self) = @_;
2217 return $self->{svnsync} if $self->{svnsync};
2218 if ($self->no_metadata) {
2220 die "Can't have both 'noMetadata' and ",
2221 "'useSvnsyncProps' options set!\n";
2222 }
2223 if ($self->rewrite_root) {
2224 die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
2225 "options set!\n";
2226 }
2227 if ($self->rewrite_uuid) {
2228 die "Can't have both 'useSvnsyncProps' and 'rewriteUUID' ",
2229 "options set!\n";
2230 }
2231 my $svnsync;
2233 # see if we have it in our config, first:
2234 eval {
2235 my $section = "svn-remote.$self->{repo_id}";
2236 my $url = tmp_config('--get', "$section.svnsync-url");
2238 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
2239 die "doesn't look right - svn:sync-from-url is '$url'\n";
2240 my $uuid = tmp_config('--get', "$section.svnsync-uuid");
2242 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}i) or
2243 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
2244 $svnsync = { url => $url, uuid => $uuid }
2246 };
2247 if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
2248 return $self->{svnsync} = $svnsync;
2249 }
2250 my $err = "useSvnsyncProps set, but failed to read " .
2252 "svnsync property: svn:sync-from-";
2253 my $rp = $self->ra->rev_proplist(0);
2254 my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
2256 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
2257 die "doesn't look right - svn:sync-from-url is '$url'\n";
2258 my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
2260 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}i) or
2261 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
2262 my $section = "svn-remote.$self->{repo_id}";
2264 tmp_config('--add', "$section.svnsync-uuid", $uuid);
2265 tmp_config('--add', "$section.svnsync-url", $url);
2266 return $self->{svnsync} = { url => $url, uuid => $uuid };
2267}
2268# this allows us to memoize our SVN::Ra UUID locally and avoid a
2270# remote lookup (useful for 'git svn log').
2271sub ra_uuid {
2272 my ($self) = @_;
2273 unless ($self->{ra_uuid}) {
2274 my $key = "svn-remote.$self->{repo_id}.uuid";
2275 my $uuid = eval { tmp_config('--get', $key) };
2276 if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/i) {
2277 $self->{ra_uuid} = $uuid;
2278 } else {
2279 die "ra_uuid called without URL\n" unless $self->{url};
2280 $self->{ra_uuid} = $self->ra->get_uuid;
2281 tmp_config('--add', $key, $self->{ra_uuid});
2282 }
2283 }
2284 $self->{ra_uuid};
2285}
2286sub _set_repos_root {
2288 my ($self, $repos_root) = @_;
2289 my $k = "svn-remote.$self->{repo_id}.reposRoot";
2290 $repos_root ||= $self->ra->{repos_root};
2291 tmp_config($k, $repos_root);
2292 $repos_root;
2293}
2294sub repos_root {
2296 my ($self) = @_;
2297 my $k = "svn-remote.$self->{repo_id}.reposRoot";
2298 eval { tmp_config('--get', $k) } || $self->_set_repos_root;
2299}
2300sub ra {
2302 my ($self) = shift;
2303 my $ra = Git::SVN::Ra->new($self->{url});
2304 $self->_set_repos_root($ra->{repos_root});
2305 if ($self->use_svm_props && !$self->{svm}) {
2306 if ($self->no_metadata) {
2307 die "Can't have both 'noMetadata' and ",
2308 "'useSvmProps' options set!\n";
2309 } elsif ($self->use_svnsync_props) {
2310 die "Can't have both 'useSvnsyncProps' and ",
2311 "'useSvmProps' options set!\n";
2312 }
2313 $ra = $self->_set_svm_vars($ra);
2314 $self->{-want_revprops} = 1;
2315 }
2316 $ra;
2317}
2318# prop_walk(PATH, REV, SUB)
2320# -------------------------
2321# Recursively traverse PATH at revision REV and invoke SUB for each
2322# directory that contains a SVN property. SUB will be invoked as
2323# follows: &SUB(gs, path, props); where `gs' is this instance of
2324# Git::SVN, `path' the path to the directory where the properties
2325# `props' were found. The `path' will be relative to point of checkout,
2326# that is, if url://repo/trunk is the current Git branch, and that
2327# directory contains a sub-directory `d', SUB will be invoked with `/d/'
2328# as `path' (note the trailing `/').
2329sub prop_walk {
2330 my ($self, $path, $rev, $sub) = @_;
2331 $path =~ s#^/##;
2333 my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
2334 $path =~ s#^/*#/#g;
2335 my $p = $path;
2336 # Strip the irrelevant part of the path.
2337 $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
2338 # Ensure the path is terminated by a `/'.
2339 $p =~ s#/*$#/#;
2340 # The properties contain all the internal SVN stuff nobody
2342 # (usually) cares about.
2343 my $interesting_props = 0;
2344 foreach (keys %{$props}) {
2345 # If it doesn't start with `svn:', it must be a
2346 # user-defined property.
2347 ++$interesting_props and next if $_ !~ /^svn:/;
2348 # FIXME: Fragile, if SVN adds new public properties,
2349 # this needs to be updated.
2350 ++$interesting_props if /^svn:(?:ignore|keywords|executable
2351 |eol-style|mime-type
2352 |externals|needs-lock)$/x;
2353 }
2354 &$sub($self, $p, $props) if $interesting_props;
2355 foreach (sort keys %$dirent) {
2357 next if $dirent->{$_}->{kind} != $SVN::Node::dir;
2358 $self->prop_walk($self->{path} . $p . $_, $rev, $sub);
2359 }
2360}
2361sub last_rev { ($_[0]->last_rev_commit)[0] }
2363sub last_commit { ($_[0]->last_rev_commit)[1] }
2364# returns the newest SVN revision number and newest commit SHA1
2366sub last_rev_commit {
2367 my ($self) = @_;
2368 if (defined $self->{last_rev} && defined $self->{last_commit}) {
2369 return ($self->{last_rev}, $self->{last_commit});
2370 }
2371 my $c = ::verify_ref($self->refname.'^0');
2372 if ($c && !$self->use_svm_props && !$self->no_metadata) {
2373 my $rev = (::cmt_metadata($c))[1];
2374 if (defined $rev) {
2375 ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
2376 return ($rev, $c);
2377 }
2378 }
2379 my $map_path = $self->map_path;
2380 unless (-e $map_path) {
2381 ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
2382 return (undef, undef);
2383 }
2384 my ($rev, $commit) = $self->rev_map_max(1);
2385 ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
2386 return ($rev, $commit);
2387}
2388sub get_fetch_range {
2390 my ($self, $min, $max) = @_;
2391 $max ||= $self->ra->get_latest_revnum;
2392 $min ||= $self->rev_map_max;
2393 (++$min, $max);
2394}
2395sub tmp_config {
2397 my (@args) = @_;
2398 my $old_def_config = "$ENV{GIT_DIR}/svn/config";
2399 my $config = "$ENV{GIT_DIR}/svn/.metadata";
2400 if (! -f $config && -f $old_def_config) {
2401 rename $old_def_config, $config or
2402 die "Failed rename $old_def_config => $config: $!\n";
2403 }
2404 my $old_config = $ENV{GIT_CONFIG};
2405 $ENV{GIT_CONFIG} = $config;
2406 $@ = undef;
2407 my @ret = eval {
2408 unless (-f $config) {
2409 mkfile($config);
2410 open my $fh, '>', $config or
2411 die "Can't open $config: $!\n";
2412 print $fh "; This file is used internally by ",
2413 "git-svn\n" or die
2414 "Couldn't write to $config: $!\n";
2415 print $fh "; You should not have to edit it\n" or
2416 die "Couldn't write to $config: $!\n";
2417 close $fh or die "Couldn't close $config: $!\n";
2418 }
2419 command('config', @args);
2420 };
2421 my $err = $@;
2422 if (defined $old_config) {
2423 $ENV{GIT_CONFIG} = $old_config;
2424 } else {
2425 delete $ENV{GIT_CONFIG};
2426 }
2427 die $err if $err;
2428 wantarray ? @ret : $ret[0];
2429}
2430sub tmp_index_do {
2432 my ($self, $sub) = @_;
2433 my $old_index = $ENV{GIT_INDEX_FILE};
2434 $ENV{GIT_INDEX_FILE} = $self->{index};
2435 $@ = undef;
2436 my @ret = eval {
2437 my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
2438 mkpath([$dir]) unless -d $dir;
2439 &$sub;
2440 };
2441 my $err = $@;
2442 if (defined $old_index) {
2443 $ENV{GIT_INDEX_FILE} = $old_index;
2444 } else {
2445 delete $ENV{GIT_INDEX_FILE};
2446 }
2447 die $err if $err;
2448 wantarray ? @ret : $ret[0];
2449}
2450sub assert_index_clean {
2452 my ($self, $treeish) = @_;
2453 $self->tmp_index_do(sub {
2455 command_noisy('read-tree', $treeish) unless -e $self->{index};
2456 my $x = command_oneline('write-tree');
2457 my ($y) = (command(qw/cat-file commit/, $treeish) =~
2458 /^tree ($::sha1)/mo);
2459 return if $y eq $x;
2460 warn "Index mismatch: $y != $x\nrereading $treeish\n";
2462 unlink $self->{index} or die "unlink $self->{index}: $!\n";
2463 command_noisy('read-tree', $treeish);
2464 $x = command_oneline('write-tree');
2465 if ($y ne $x) {
2466 ::fatal "trees ($treeish) $y != $x\n",
2467 "Something is seriously wrong...";
2468 }
2469 });
2470}
2471sub get_commit_parents {
2473 my ($self, $log_entry) = @_;
2474 my (%seen, @ret, @tmp);
2475 # legacy support for 'set-tree'; this is only used by set_tree_cb:
2476 if (my $ip = $self->{inject_parents}) {
2477 if (my $commit = delete $ip->{$log_entry->{revision}}) {
2478 push @tmp, $commit;
2479 }
2480 }
2481 if (my $cur = ::verify_ref($self->refname.'^0')) {
2482 push @tmp, $cur;
2483 }
2484 if (my $ipd = $self->{inject_parents_dcommit}) {
2485 if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2486 push @tmp, @$commit;
2487 }
2488 }
2489 push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2490 while (my $p = shift @tmp) {
2491 next if $seen{$p};
2492 $seen{$p} = 1;
2493 push @ret, $p;
2494 }
2495 @ret;
2496}
2497sub rewrite_root {
2499 my ($self) = @_;
2500 return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2501 my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2502 my $rwr = eval { command_oneline(qw/config --get/, $k) };
2503 if ($rwr) {
2504 $rwr =~ s#/+$##;
2505 if ($rwr !~ m#^[a-z\+]+://#) {
2506 die "$rwr is not a valid URL (key: $k)\n";
2507 }
2508 }
2509 $self->{-rewrite_root} = $rwr;
2510}
2511sub rewrite_uuid {
2513 my ($self) = @_;
2514 return $self->{-rewrite_uuid} if exists $self->{-rewrite_uuid};
2515 my $k = "svn-remote.$self->{repo_id}.rewriteUUID";
2516 my $rwid = eval { command_oneline(qw/config --get/, $k) };
2517 if ($rwid) {
2518 $rwid =~ s#/+$##;
2519 if ($rwid !~ m#^[a-f0-9]{8}-(?:[a-f0-9]{4}-){3}[a-f0-9]{12}$#) {
2520 die "$rwid is not a valid UUID (key: $k)\n";
2521 }
2522 }
2523 $self->{-rewrite_uuid} = $rwid;
2524}
2525sub metadata_url {
2527 my ($self) = @_;
2528 ($self->rewrite_root || $self->{url}) .
2529 (length $self->{path} ? '/' . $self->{path} : '');
2530}
2531sub full_url {
2533 my ($self) = @_;
2534 $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2535}
2536sub set_commit_header_env {
2539 my ($log_entry) = @_;
2540 my %env;
2541 foreach my $ned (qw/NAME EMAIL DATE/) {
2542 foreach my $ac (qw/AUTHOR COMMITTER/) {
2543 $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2544 }
2545 }
2546 $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2548 $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2549 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2550 $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2552 ? $log_entry->{commit_name}
2553 : $log_entry->{name};
2554 $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2555 ? $log_entry->{commit_email}
2556 : $log_entry->{email};
2557 \%env;
2558}
2559sub restore_commit_header_env {
2561 my ($env) = @_;
2562 foreach my $ned (qw/NAME EMAIL DATE/) {
2563 foreach my $ac (qw/AUTHOR COMMITTER/) {
2564 my $k = "GIT_${ac}_${ned}";
2565 if (defined $env->{$k}) {
2566 $ENV{$k} = $env->{$k};
2567 } else {
2568 delete $ENV{$k};
2569 }
2570 }
2571 }
2572}
2573sub gc {
2575 command_noisy('gc', '--auto');
2576};
2577sub do_git_commit {
2579 my ($self, $log_entry) = @_;
2580 my $lr = $self->last_rev;
2581 if (defined $lr && $lr >= $log_entry->{revision}) {
2582 die "Last fetched revision of ", $self->refname,
2583 " was r$lr, but we are about to fetch: ",
2584 "r$log_entry->{revision}!\n";
2585 }
2586 if (my $c = $self->rev_map_get($log_entry->{revision})) {
2587 croak "$log_entry->{revision} = $c already exists! ",
2588 "Why are we refetching it?\n";
2589 }
2590 my $old_env = set_commit_header_env($log_entry);
2591 my $tree = $log_entry->{tree};
2592 if (!defined $tree) {
2593 $tree = $self->tmp_index_do(sub {
2594 command_oneline('write-tree') });
2595 }
2596 die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2597 my @exec = ('git', 'commit-tree', $tree);
2599 foreach ($self->get_commit_parents($log_entry)) {
2600 push @exec, '-p', $_;
2601 }
2602 defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2603 or croak $!;
2604 binmode $msg_fh;
2605 # we always get UTF-8 from SVN, but we may want our commits in
2607 # a different encoding.
2608 if (my $enc = Git::config('i18n.commitencoding')) {
2609 require Encode;
2610 Encode::from_to($log_entry->{log}, 'UTF-8', $enc);
2611 }
2612 print $msg_fh $log_entry->{log} or croak $!;
2613 restore_commit_header_env($old_env);
2614 unless ($self->no_metadata) {
2615 print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2616 or croak $!;
2617 }
2618 $msg_fh->flush == 0 or croak $!;
2619 close $msg_fh or croak $!;
2620 chomp(my $commit = do { local $/; <$out_fh> });
2621 close $out_fh or croak $!;
2622 waitpid $pid, 0;
2623 croak $? if $?;
2624 if ($commit !~ /^$::sha1$/o) {
2625 die "Failed to commit, invalid sha1: $commit\n";
2626 }
2627 $self->rev_map_set($log_entry->{revision}, $commit, 1);
2629 $self->{last_rev} = $log_entry->{revision};
2631 $self->{last_commit} = $commit;
2632 print "r$log_entry->{revision}" unless $::_q > 1;
2633 if (defined $log_entry->{svm_revision}) {
2634 print " (\@$log_entry->{svm_revision})" unless $::_q > 1;
2635 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2636 0, $self->svm_uuid);
2637 }
2638 print " = $commit ($self->{ref_id})\n" unless $::_q > 1;
2639 if (--$_gc_nr == 0) {
2640 $_gc_nr = $_gc_period;
2641 gc();
2642 }
2643 return $commit;
2644}
2645sub match_paths {
2647 my ($self, $paths, $r) = @_;
2648 return 1 if $self->{path} eq '';
2649 if (my $path = $paths->{"/$self->{path}"}) {
2650 return ($path->{action} eq 'D') ? 0 : 1;
2651 }
2652 $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
2653 if (grep /$self->{path_regex}/, keys %$paths) {
2654 return 1;
2655 }
2656 my $c = '';
2657 foreach (split m#/#, $self->{path}) {
2658 $c .= "/$_";
2659 next unless ($paths->{$c} &&
2660 ($paths->{$c}->{action} =~ /^[AR]$/));
2661 if ($self->ra->check_path($self->{path}, $r) ==
2662 $SVN::Node::dir) {
2663 return 1;
2664 }
2665 }
2666 return 0;
2667}
2668sub find_parent_branch {
2670 my ($self, $paths, $rev) = @_;
2671 return undef unless $self->follow_parent;
2672 unless (defined $paths) {
2673 my $err_handler = $SVN::Error::handler;
2674 $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2675 $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1,
2676 sub { $paths = $_[0] });
2677 $SVN::Error::handler = $err_handler;
2678 }
2679 return undef unless defined $paths;
2680 # look for a parent from another branch:
2682 my @b_path_components = split m#/#, $self->{path};
2683 my @a_path_components;
2684 my $i;
2685 while (@b_path_components) {
2686 $i = $paths->{'/'.join('/', @b_path_components)};
2687 last if $i && defined $i->{copyfrom_path};
2688 unshift(@a_path_components, pop(@b_path_components));
2689 }
2690 return undef unless defined $i && defined $i->{copyfrom_path};
2691 my $branch_from = $i->{copyfrom_path};
2692 if (@a_path_components) {
2693 print STDERR "branch_from: $branch_from => ";
2694 $branch_from .= '/'.join('/', @a_path_components);
2695 print STDERR $branch_from, "\n";
2696 }
2697 my $r = $i->{copyfrom_rev};
2698 my $repos_root = $self->ra->{repos_root};
2699 my $url = $self->ra->{url};
2700 my $new_url = $url . $branch_from;
2701 print STDERR "Found possible branch point: ",
2702 "$new_url => ", $self->full_url, ", $r\n"
2703 unless $::_q > 1;
2704 $branch_from =~ s#^/##;
2705 my $gs = $self->other_gs($new_url, $url,
2706 $branch_from, $r, $self->{ref_id});
2707 my ($r0, $parent) = $gs->find_rev_before($r, 1);
2708 {
2709 my ($base, $head);
2710 if (!defined $r0 || !defined $parent) {
2711 ($base, $head) = parse_revision_argument(0, $r);
2712 } else {
2713 if ($r0 < $r) {
2714 $gs->ra->get_log([$gs->{path}], $r0 + 1, $r, 1,
2715 0, 1, sub { $base = $_[1] - 1 });
2716 }
2717 }
2718 if (defined $base && $base <= $r) {
2719 $gs->fetch($base, $r);
2720 }
2721 ($r0, $parent) = $gs->find_rev_before($r, 1);
2722 }
2723 if (defined $r0 && defined $parent) {
2724 print STDERR "Found branch parent: ($self->{ref_id}) $parent\n"
2725 unless $::_q > 1;
2726 my $ed;
2727 if ($self->ra->can_do_switch) {
2728 $self->assert_index_clean($parent);
2729 print STDERR "Following parent with do_switch\n"
2730 unless $::_q > 1;
2731 # do_switch works with svn/trunk >= r22312, but that
2732 # is not included with SVN 1.4.3 (the latest version
2733 # at the moment), so we can't rely on it
2734 $self->{last_rev} = $r0;
2735 $self->{last_commit} = $parent;
2736 $ed = SVN::Git::Fetcher->new($self, $gs->{path});
2737 $gs->ra->gs_do_switch($r0, $rev, $gs,
2738 $self->full_url, $ed)
2739 or die "SVN connection failed somewhere...\n";
2740 } elsif ($self->ra->trees_match($new_url, $r0,
2741 $self->full_url, $rev)) {
2742 print STDERR "Trees match:\n",
2743 " $new_url\@$r0\n",
2744 " ${\$self->full_url}\@$rev\n",
2745 "Following parent with no changes\n"
2746 unless $::_q > 1;
2747 $self->tmp_index_do(sub {
2748 command_noisy('read-tree', $parent);
2749 });
2750 $self->{last_commit} = $parent;
2751 } else {
2752 print STDERR "Following parent with do_update\n"
2753 unless $::_q > 1;
2754 $ed = SVN::Git::Fetcher->new($self);
2755 $self->ra->gs_do_update($rev, $rev, $self, $ed)
2756 or die "SVN connection failed somewhere...\n";
2757 }
2758 print STDERR "Successfully followed parent\n" unless $::_q > 1;
2759 return $self->make_log_entry($rev, [$parent], $ed);
2760 }
2761 return undef;
2762}
2763sub do_fetch {
2765 my ($self, $paths, $rev) = @_;
2766 my $ed;
2767 my ($last_rev, @parents);
2768 if (my $lc = $self->last_commit) {
2769 # we can have a branch that was deleted, then re-added
2770 # under the same name but copied from another path, in
2771 # which case we'll have multiple parents (we don't
2772 # want to break the original ref, nor lose copypath info):
2773 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2774 push @{$log_entry->{parents}}, $lc;
2775 return $log_entry;
2776 }
2777 $ed = SVN::Git::Fetcher->new($self);
2778 $last_rev = $self->{last_rev};
2779 $ed->{c} = $lc;
2780 @parents = ($lc);
2781 } else {
2782 $last_rev = $rev;
2783 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2784 return $log_entry;
2785 }
2786 $ed = SVN::Git::Fetcher->new($self);
2787 }
2788 unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2789 die "SVN connection failed somewhere...\n";
2790 }
2791 $self->make_log_entry($rev, \@parents, $ed);
2792}
2793sub mkemptydirs {
2795 my ($self, $r) = @_;
2796 sub scan {
2798 my ($r, $empty_dirs, $line) = @_;
2799 if (defined $r && $line =~ /^r(\d+)$/) {
2800 return 0 if $1 > $r;
2801 } elsif ($line =~ /^ \+empty_dir: (.+)$/) {
2802 $empty_dirs->{$1} = 1;
2803 } elsif ($line =~ /^ \-empty_dir: (.+)$/) {
2804 my @d = grep {m[^\Q$1\E(/|$)]} (keys %$empty_dirs);
2805 delete @$empty_dirs{@d};
2806 }
2807 1; # continue
2808 };
2809 my %empty_dirs = ();
2811 my $gz_file = "$self->{dir}/unhandled.log.gz";
2812 if (-f $gz_file) {
2813 if (!$can_compress) {
2814 warn "Compress::Zlib could not be found; ",
2815 "empty directories in $gz_file will not be read\n";
2816 } else {
2817 my $gz = Compress::Zlib::gzopen($gz_file, "rb") or
2818 die "Unable to open $gz_file: $!\n";
2819 my $line;
2820 while ($gz->gzreadline($line) > 0) {
2821 scan($r, \%empty_dirs, $line) or last;
2822 }
2823 $gz->gzclose;
2824 }
2825 }
2826 if (open my $fh, '<', "$self->{dir}/unhandled.log") {
2828 binmode $fh or croak "binmode: $!";
2829 while (<$fh>) {
2830 scan($r, \%empty_dirs, $_) or last;
2831 }
2832 close $fh;
2833 }
2834 my $strip = qr/\A\Q$self->{path}\E(?:\/|$)/;
2836 foreach my $d (sort keys %empty_dirs) {
2837 $d = uri_decode($d);
2838 $d =~ s/$strip//;
2839 next if -d $d;
2840 if (-e _) {
2841 warn "$d exists but is not a directory\n";
2842 } else {
2843 print "creating empty directory: $d\n";
2844 mkpath([$d]);
2845 }
2846 }
2847}
2848sub get_untracked {
2850 my ($self, $ed) = @_;
2851 my @out;
2852 my $h = $ed->{empty};
2853 foreach (sort keys %$h) {
2854 my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2855 push @out, " $act: " . uri_encode($_);
2856 warn "W: $act: $_\n";
2857 }
2858 foreach my $t (qw/dir_prop file_prop/) {
2859 $h = $ed->{$t} or next;
2860 foreach my $path (sort keys %$h) {
2861 my $ppath = $path eq '' ? '.' : $path;
2862 foreach my $prop (sort keys %{$h->{$path}}) {
2863 next if $SKIP_PROP{$prop};
2864 my $v = $h->{$path}->{$prop};
2865 my $t_ppath_prop = "$t: " .
2866 uri_encode($ppath) . ' ' .
2867 uri_encode($prop);
2868 if (defined $v) {
2869 push @out, " +$t_ppath_prop " .
2870 uri_encode($v);
2871 } else {
2872 push @out, " -$t_ppath_prop";
2873 }
2874 }
2875 }
2876 }
2877 foreach my $t (qw/absent_file absent_directory/) {
2878 $h = $ed->{$t} or next;
2879 foreach my $parent (sort keys %$h) {
2880 foreach my $path (sort @{$h->{$parent}}) {
2881 push @out, " $t: " .
2882 uri_encode("$parent/$path");
2883 warn "W: $t: $parent/$path ",
2884 "Insufficient permissions?\n";
2885 }
2886 }
2887 }
2888 \@out;
2889}
2890# parse_svn_date(DATE)
2892# --------------------
2893# Given a date (in UTC) from Subversion, return a string in the format
2894# "<TZ Offset> <local date/time>" that Git will use.
2895#
2896# By default the parsed date will be in UTC; if $Git::SVN::_localtime
2897# is true we'll convert it to the local timezone instead.
2898sub parse_svn_date {
2899 my $date = shift || return '+0000 1970-01-01 00:00:00';
2900 my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2901 (\d\d)\:(\d\d)\:(\d\d)\.\d*Z$/x) or
2902 croak "Unable to parse date: $date\n";
2903 my $parsed_date; # Set next.
2904 if ($Git::SVN::_localtime) {
2906 # Translate the Subversion datetime to an epoch time.
2907 # Begin by switching ourselves to $date's timezone, UTC.
2908 my $old_env_TZ = $ENV{TZ};
2909 $ENV{TZ} = 'UTC';
2910 my $epoch_in_UTC =
2912 POSIX::strftime('%s', $S, $M, $H, $d, $m - 1, $Y - 1900);
2913 # Determine our local timezone (including DST) at the
2915 # time of $epoch_in_UTC. $Git::SVN::Log::TZ stored the
2916 # value of TZ, if any, at the time we were run.
2917 if (defined $Git::SVN::Log::TZ) {
2918 $ENV{TZ} = $Git::SVN::Log::TZ;
2919 } else {
2920 delete $ENV{TZ};
2921 }
2922 my $our_TZ =
2924 POSIX::strftime('%Z', $S, $M, $H, $d, $m - 1, $Y - 1900);
2925 # This converts $epoch_in_UTC into our local timezone.
2927 my ($sec, $min, $hour, $mday, $mon, $year,
2928 $wday, $yday, $isdst) = localtime($epoch_in_UTC);
2929 $parsed_date = sprintf('%s %04d-%02d-%02d %02d:%02d:%02d',
2931 $our_TZ, $year + 1900, $mon + 1,
2932 $mday, $hour, $min, $sec);
2933 # Reset us to the timezone in effect when we entered
2935 # this routine.
2936 if (defined $old_env_TZ) {
2937 $ENV{TZ} = $old_env_TZ;
2938 } else {
2939 delete $ENV{TZ};
2940 }
2941 } else {
2942 $parsed_date = "+0000 $Y-$m-$d $H:$M:$S";
2943 }
2944 return $parsed_date;
2946}
2947sub other_gs {
2949 my ($self, $new_url, $url,
2950 $branch_from, $r, $old_ref_id) = @_;
2951 my $gs = Git::SVN->find_by_url($new_url, $url, $branch_from);
2952 unless ($gs) {
2953 my $ref_id = $old_ref_id;
2954 $ref_id =~ s/\@\d+$//;
2955 $ref_id .= "\@$r";
2956 # just grow a tail if we're not unique enough :x
2957 $ref_id .= '-' while find_ref($ref_id);
2958 print STDERR "Initializing parent: $ref_id\n" unless $::_q > 1;
2959 my ($u, $p, $repo_id) = ($new_url, '', $ref_id);
2960 if ($u =~ s#^\Q$url\E(/|$)##) {
2961 $p = $u;
2962 $u = $url;
2963 $repo_id = $self->{repo_id};
2964 }
2965 $gs = Git::SVN->init($u, $p, $repo_id, $ref_id, 1);
2966 }
2967 $gs
2968}
2969sub call_authors_prog {
2971 my ($orig_author) = @_;
2972 $orig_author = command_oneline('rev-parse', '--sq-quote', $orig_author);
2973 my $author = `$::_authors_prog $orig_author`;
2974 if ($? != 0) {
2975 die "$::_authors_prog failed with exit code $?\n"
2976 }
2977 if ($author =~ /^\s*(.+?)\s*<(.*)>\s*$/) {
2978 my ($name, $email) = ($1, $2);
2979 $email = undef if length $2 == 0;
2980 return [$name, $email];
2981 } else {
2982 die "Author: $orig_author: $::_authors_prog returned "
2983 . "invalid author format: $author\n";
2984 }
2985}
2986sub check_author {
2988 my ($author) = @_;
2989 if (!defined $author || length $author == 0) {
2990 $author = '(no author)';
2991 }
2992 if (!defined $::users{$author}) {
2993 if (defined $::_authors_prog) {
2994 $::users{$author} = call_authors_prog($author);
2995 } elsif (defined $::_authors) {
2996 die "Author: $author not defined in $::_authors file\n";
2997 }
2998 }
2999 $author;
3000}
3001sub find_extra_svk_parents {
3003 my ($self, $ed, $tickets, $parents) = @_;
3004 # aha! svk:merge property changed...
3005 my @tickets = split "\n", $tickets;
3006 my @known_parents;
3007 for my $ticket ( @tickets ) {
3008 my ($uuid, $path, $rev) = split /:/, $ticket;
3009 if ( $uuid eq $self->ra_uuid ) {
3010 my $url = $self->{url};
3011 my $repos_root = $url;
3012 my $branch_from = $path;
3013 $branch_from =~ s{^/}{};
3014 my $gs = $self->other_gs($repos_root."/".$branch_from,
3015 $url,
3016 $branch_from,
3017 $rev,
3018 $self->{ref_id});
3019 if ( my $commit = $gs->rev_map_get($rev, $uuid) ) {
3020 # wahey! we found it, but it might be
3021 # an old one (!)
3022 push @known_parents, [ $rev, $commit ];
3023 }
3024 }
3025 }
3026 # Ordering matters; highest-numbered commit merge tickets
3027 # first, as they may account for later merge ticket additions
3028 # or changes.
3029 @known_parents = map {$_->[1]} sort {$b->[0] <=> $a->[0]} @known_parents;
3030 for my $parent ( @known_parents ) {
3031 my @cmd = ('rev-list', $parent, map { "^$_" } @$parents );
3032 my ($msg_fh, $ctx) = command_output_pipe(@cmd);
3033 my $new;
3034 while ( <$msg_fh> ) {
3035 $new=1;last;
3036 }
3037 command_close_pipe($msg_fh, $ctx);
3038 if ( $new ) {
3039 print STDERR
3040 "Found merge parent (svk:merge ticket): $parent\n";
3041 push @$parents, $parent;
3042 }
3043 }
3044}
3045sub lookup_svn_merge {
3047 my $uuid = shift;
3048 my $url = shift;
3049 my $merge = shift;
3050 my ($source, $revs) = split ":", $merge;
3052 my $path = $source;
3053 $path =~ s{^/}{};
3054 my $gs = Git::SVN->find_by_url($url.$source, $url, $path);
3055 if ( !$gs ) {
3056 warn "Couldn't find revmap for $url$source\n";
3057 return;
3058 }
3059 my @ranges = split ",", $revs;
3060 my ($tip, $tip_commit);
3061 my @merged_commit_ranges;
3062 # find the tip
3063 for my $range ( @ranges ) {
3064 my ($bottom, $top) = split "-", $range;
3065 $top ||= $bottom;
3066 my $bottom_commit = $gs->find_rev_after( $bottom, 1, $top );
3067 my $top_commit = $gs->find_rev_before( $top, 1, $bottom );
3068 unless ($top_commit and $bottom_commit) {
3070 warn "W:unknown path/rev in svn:mergeinfo "
3071 ."dirprop: $source:$range\n";
3072 next;
3073 }
3074 push @merged_commit_ranges,
3076 "$bottom_commit^..$top_commit";
3077 if ( !defined $tip or $top > $tip ) {
3079 $tip = $top;
3080 $tip_commit = $top_commit;
3081 }
3082 }
3083 return ($tip_commit, @merged_commit_ranges);
3084}
3085sub _rev_list {
3087 my ($msg_fh, $ctx) = command_output_pipe(
3088 "rev-list", @_,
3089 );
3090 my @rv;
3091 while ( <$msg_fh> ) {
3092 chomp;
3093 push @rv, $_;
3094 }
3095 command_close_pipe($msg_fh, $ctx);
3096 @rv;
3097}
3098sub check_cherry_pick {
3100 my $base = shift;
3101 my $tip = shift;
3102 my @ranges = @_;
3103 my %commits = map { $_ => 1 }
3104 _rev_list("--no-merges", $tip, "--not", $base);
3105 for my $range ( @ranges ) {
3106 delete @commits{_rev_list($range)};
3107 }
3108 for my $commit (keys %commits) {
3109 if (has_no_changes($commit)) {
3110 delete $commits{$commit};
3111 }
3112 }
3113 return (keys %commits);
3114}
3115sub has_no_changes {
3117 my $commit = shift;
3118 my @revs = split / /, command_oneline(
3120 qw(rev-list --parents -1 -m), $commit);
3121 # Commits with no parents, e.g. the start of a partial branch,
3123 # have changes by definition.
3124 return 1 if (@revs < 2);
3125 # Commits with multiple parents, e.g a merge, have no changes
3127 # by definition.
3128 return 0 if (@revs > 2);
3129 return (command_oneline("rev-parse", "$commit^{tree}") eq
3131 command_oneline("rev-parse", "$commit~1^{tree}"));
3132}
3133# The GIT_DIR environment variable is not always set until after the command
3135# line arguments are processed, so we can't memoize in a BEGIN block.
3136{
3137 my $memoized = 0;
3138 sub memoize_svn_mergeinfo_functions {
3140 return if $memoized;
3141 $memoized = 1;
3142 my $cache_path = "$ENV{GIT_DIR}/svn/.caches/";
3144 mkpath([$cache_path]) unless -d $cache_path;
3145 tie my %lookup_svn_merge_cache => 'Memoize::Storable',
3147 "$cache_path/lookup_svn_merge.db", 'nstore';
3148 memoize 'lookup_svn_merge',
3149 SCALAR_CACHE => 'FAULT',
3150 LIST_CACHE => ['HASH' => \%lookup_svn_merge_cache],
3151 ;
3152 tie my %check_cherry_pick_cache => 'Memoize::Storable',
3154 "$cache_path/check_cherry_pick.db", 'nstore';
3155 memoize 'check_cherry_pick',
3156 SCALAR_CACHE => 'FAULT',
3157 LIST_CACHE => ['HASH' => \%check_cherry_pick_cache],
3158 ;
3159 tie my %has_no_changes_cache => 'Memoize::Storable',
3161 "$cache_path/has_no_changes.db", 'nstore';
3162 memoize 'has_no_changes',
3163 SCALAR_CACHE => ['HASH' => \%has_no_changes_cache],
3164 LIST_CACHE => 'FAULT',
3165 ;
3166 }
3167}
3168sub parents_exclude {
3170 my $parents = shift;
3171 my @commits = @_;
3172 return unless @commits;
3173 my @excluded;
3175 my $excluded;
3176 do {
3177 my @cmd = ('rev-list', "-1", @commits, "--not", @$parents );
3178 $excluded = command_oneline(@cmd);
3179 if ( $excluded ) {
3180 my @new;
3181 my $found;
3182 for my $commit ( @commits ) {
3183 if ( $commit eq $excluded ) {
3184 push @excluded, $commit;
3185 $found++;
3186 last;
3187 }
3188 else {
3189 push @new, $commit;
3190 }
3191 }
3192 die "saw commit '$excluded' in rev-list output, "
3193 ."but we didn't ask for that commit (wanted: @commits --not @$parents)"
3194 unless $found;
3195 @commits = @new;
3196 }
3197 }
3198 while ($excluded and @commits);
3199 return @excluded;
3201}
3202# note: this function should only be called if the various dirprops
3205# have actually changed
3206sub find_extra_svn_parents {
3207 my ($self, $ed, $mergeinfo, $parents) = @_;
3208 # aha! svk:merge property changed...
3209 memoize_svn_mergeinfo_functions();
3211 # We first search for merged tips which are not in our
3213 # history. Then, we figure out which git revisions are in
3214 # that tip, but not this revision. If all of those revisions
3215 # are now marked as merge, we can add the tip as a parent.
3216 my @merges = split "\n", $mergeinfo;
3217 my @merge_tips;
3218 my $url = $self->{url};
3219 my $uuid = $self->ra_uuid;
3220 my %ranges;
3221 for my $merge ( @merges ) {
3222 my ($tip_commit, @ranges) =
3223 lookup_svn_merge( $uuid, $url, $merge );
3224 unless (!$tip_commit or
3225 grep { $_ eq $tip_commit } @$parents ) {
3226 push @merge_tips, $tip_commit;
3227 $ranges{$tip_commit} = \@ranges;
3228 } else {
3229 push @merge_tips, undef;
3230 }
3231 }
3232 my %excluded = map { $_ => 1 }
3234 parents_exclude($parents, grep { defined } @merge_tips);
3235 # check merge tips for new parents
3237 my @new_parents;
3238 for my $merge_tip ( @merge_tips ) {
3239 my $spec = shift @merges;
3240 next unless $merge_tip and $excluded{$merge_tip};
3241 my $ranges = $ranges{$merge_tip};
3243 # check out 'new' tips
3245 my $merge_base;
3246 eval {
3247 $merge_base = command_oneline(
3248 "merge-base",
3249 @$parents, $merge_tip,
3250 );
3251 };
3252 if ($@) {
3253 die "An error occurred during merge-base"
3254 unless $@->isa("Git::Error::Command");
3255 warn "W: Cannot find common ancestor between ".
3257 "@$parents and $merge_tip. Ignoring merge info.\n";
3258 next;
3259 }
3260 # double check that there are no missing non-merge commits
3262 my (@incomplete) = check_cherry_pick(
3263 $merge_base, $merge_tip,
3264 @$ranges,
3265 );
3266 if ( @incomplete ) {
3268 warn "W:svn cherry-pick ignored ($spec) - missing "
3269 .@incomplete." commit(s) (eg $incomplete[0])\n";
3270 } else {
3271 warn
3272 "Found merge parent (svn:mergeinfo prop): ",
3273 $merge_tip, "\n";
3274 push @new_parents, $merge_tip;
3275 }
3276 }
3277 # cater for merges which merge commits from multiple branches
3279 if ( @new_parents > 1 ) {
3280 for ( my $i = 0; $i <= $#new_parents; $i++ ) {
3281 for ( my $j = 0; $j <= $#new_parents; $j++ ) {
3282 next if $i == $j;
3283 next unless $new_parents[$i];
3284 next unless $new_parents[$j];
3285 my $revs = command_oneline(
3286 "rev-list", "-1",
3287 "$new_parents[$i]..$new_parents[$j]",
3288 );
3289 if ( !$revs ) {
3290 undef($new_parents[$j]);
3291 }
3292 }
3293 }
3294 }
3295 push @$parents, grep { defined } @new_parents;
3296}
3297sub make_log_entry {
3299 my ($self, $rev, $parents, $ed) = @_;
3300 my $untracked = $self->get_untracked($ed);
3301 my @parents = @$parents;
3303 my $ps = $ed->{path_strip} || "";
3304 for my $path ( grep { m/$ps/ } %{$ed->{dir_prop}} ) {
3305 my $props = $ed->{dir_prop}{$path};
3306 if ( $props->{"svk:merge"} ) {
3307 $self->find_extra_svk_parents
3308 ($ed, $props->{"svk:merge"}, \@parents);
3309 }
3310 if ( $props->{"svn:mergeinfo"} ) {
3311 $self->find_extra_svn_parents
3312 ($ed,
3313 $props->{"svn:mergeinfo"},
3314 \@parents);
3315 }
3316 }
3317 open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
3319 print $un "r$rev\n" or croak $!;
3320 print $un $_, "\n" foreach @$untracked;
3321 my %log_entry = ( parents => \@parents, revision => $rev,
3322 log => '');
3323 my $headrev;
3325 my $logged = delete $self->{logged_rev_props};
3326 if (!$logged || $self->{-want_revprops}) {
3327 my $rp = $self->ra->rev_proplist($rev);
3328 foreach (sort keys %$rp) {
3329 my $v = $rp->{$_};
3330 if (/^svn:(author|date|log)$/) {
3331 $log_entry{$1} = $v;
3332 } elsif ($_ eq 'svm:headrev') {
3333 $headrev = $v;
3334 } else {
3335 print $un " rev_prop: ", uri_encode($_), ' ',
3336 uri_encode($v), "\n";
3337 }
3338 }
3339 } else {
3340 map { $log_entry{$_} = $logged->{$_} } keys %$logged;
3341 }
3342 close $un or croak $!;
3343 $log_entry{date} = parse_svn_date($log_entry{date});
3345 $log_entry{log} .= "\n";
3346 my $author = $log_entry{author} = check_author($log_entry{author});
3347 my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
3348 : ($author, undef);
3349 my ($commit_name, $commit_email) = ($name, $email);
3351 if ($_use_log_author) {
3352 my $name_field;
3353 if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
3354 $name_field = $1;
3355 } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
3356 $name_field = $1;
3357 }
3358 if (!defined $name_field) {
3359 if (!defined $email) {
3360 $email = $name;
3361 }
3362 } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
3363 ($name, $email) = ($1, $2);
3364 } elsif ($name_field =~ /(.*)@/) {
3365 ($name, $email) = ($1, $name_field);
3366 } else {
3367 ($name, $email) = ($name_field, $name_field);
3368 }
3369 }
3370 if (defined $headrev && $self->use_svm_props) {
3371 if ($self->rewrite_root) {
3372 die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
3373 "options set!\n";
3374 }
3375 if ($self->rewrite_uuid) {
3376 die "Can't have both 'useSvmProps' and 'rewriteUUID' ",
3377 "options set!\n";
3378 }
3379 my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$}i;
3380 # we don't want "SVM: initializing mirror for junk" ...
3381 return undef if $r == 0;
3382 my $svm = $self->svm;
3383 if ($uuid ne $svm->{uuid}) {
3384 die "UUID mismatch on SVM path:\n",
3385 "expected: $svm->{uuid}\n",
3386 " got: $uuid\n";
3387 }
3388 my $full_url = $self->full_url;
3389 $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
3390 die "Failed to replace '$svm->{replace}' with ",
3391 "'$svm->{source}' in $full_url\n";
3392 # throw away username for storing in records
3393 remove_username($full_url);
3394 $log_entry{metadata} = "$full_url\@$r $uuid";
3395 $log_entry{svm_revision} = $r;
3396 $email ||= "$author\@$uuid";
3397 $commit_email ||= "$author\@$uuid";
3398 } elsif ($self->use_svnsync_props) {
3399 my $full_url = $self->svnsync->{url};
3400 $full_url .= "/$self->{path}" if length $self->{path};
3401 remove_username($full_url);
3402 my $uuid = $self->svnsync->{uuid};
3403 $log_entry{metadata} = "$full_url\@$rev $uuid";
3404 $email ||= "$author\@$uuid";
3405 $commit_email ||= "$author\@$uuid";
3406 } else {
3407 my $url = $self->metadata_url;
3408 remove_username($url);
3409 my $uuid = $self->rewrite_uuid || $self->ra->get_uuid;
3410 $log_entry{metadata} = "$url\@$rev " . $uuid;
3411 $email ||= "$author\@" . $uuid;
3412 $commit_email ||= "$author\@" . $uuid;
3413 }
3414 $log_entry{name} = $name;
3415 $log_entry{email} = $email;
3416 $log_entry{commit_name} = $commit_name;
3417 $log_entry{commit_email} = $commit_email;
3418 \%log_entry;
3419}
3420sub fetch {
3422 my ($self, $min_rev, $max_rev, @parents) = @_;
3423 my ($last_rev, $last_commit) = $self->last_rev_commit;
3424 my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
3425 $self->ra->gs_fetch_loop_common($base, $head, [$self]);
3426}
3427sub set_tree_cb {
3429 my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
3430 $self->{inject_parents} = { $rev => $tree };
3431 $self->fetch(undef, undef);
3432}
3433sub set_tree {
3435 my ($self, $tree) = (shift, shift);
3436 my $log_entry = ::get_commit_entry($tree);
3437 unless ($self->{last_rev}) {
3438 ::fatal("Must have an existing revision to commit");
3439 }
3440 my %ed_opts = ( r => $self->{last_rev},
3441 log => $log_entry->{log},
3442 ra => $self->ra,
3443 tree_a => $self->{last_commit},
3444 tree_b => $tree,
3445 editor_cb => sub {
3446 $self->set_tree_cb($log_entry, $tree, @_) },
3447 svn_path => $self->{path} );
3448 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
3449 print "No changes\nr$self->{last_rev} = $tree\n";
3450 }
3451}
3452sub rebuild_from_rev_db {
3454 my ($self, $path) = @_;
3455 my $r = -1;
3456 open my $fh, '<', $path or croak "open: $!";
3457 binmode $fh or croak "binmode: $!";
3458 while (<$fh>) {
3459 length($_) == 41 or croak "inconsistent size in ($_) != 41";
3460 chomp($_);
3461 ++$r;
3462 next if $_ eq ('0' x 40);
3463 $self->rev_map_set($r, $_);
3464 print "r$r = $_\n";
3465 }
3466 close $fh or croak "close: $!";
3467 unlink $path or croak "unlink: $!";
3468}
3469sub rebuild {
3471 my ($self) = @_;
3472 my $map_path = $self->map_path;
3473 my $partial = (-e $map_path && ! -z $map_path);
3474 return unless ::verify_ref($self->refname.'^0');
3475 if (!$partial && ($self->use_svm_props || $self->no_metadata)) {
3476 my $rev_db = $self->rev_db_path;
3477 $self->rebuild_from_rev_db($rev_db);
3478 if ($self->use_svm_props) {
3479 my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
3480 $self->rebuild_from_rev_db($svm_rev_db);
3481 }
3482 $self->unlink_rev_db_symlink;
3483 return;
3484 }
3485 print "Rebuilding $map_path ...\n" if (!$partial);
3486 my ($base_rev, $head) = ($partial ? $self->rev_map_max_norebuild(1) :
3487 (undef, undef));
3488 my ($log, $ctx) =
3489 command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
3490 ($head ? "$head.." : "") . $self->refname,
3491 '--');
3492 my $metadata_url = $self->metadata_url;
3493 remove_username($metadata_url);
3494 my $svn_uuid = $self->rewrite_uuid || $self->ra_uuid;
3495 my $c;
3496 while (<$log>) {
3497 if ( m{^commit ($::sha1)$} ) {
3498 $c = $1;
3499 next;
3500 }
3501 next unless s{^\s*(git-svn-id:)}{$1};
3502 my ($url, $rev, $uuid) = ::extract_metadata($_);
3503 remove_username($url);
3504 # ignore merges (from set-tree)
3506 next if (!defined $rev || !$uuid);
3507 # if we merged or otherwise started elsewhere, this is
3509 # how we break out of it
3510 if (($uuid ne $svn_uuid) ||
3511 ($metadata_url && $url && ($url ne $metadata_url))) {
3512 next;
3513 }
3514 if ($partial && $head) {
3515 print "Partial-rebuilding $map_path ...\n";
3516 print "Currently at $base_rev = $head\n";
3517 $head = undef;
3518 }
3519 $self->rev_map_set($rev, $c);
3521 print "r$rev = $c\n";
3522 }
3523 command_close_pipe($log, $ctx);
3524 print "Done rebuilding $map_path\n" if (!$partial || !$head);
3525 my $rev_db_path = $self->rev_db_path;
3526 if (-f $self->rev_db_path) {
3527 unlink $self->rev_db_path or croak "unlink: $!";
3528 }
3529 $self->unlink_rev_db_symlink;
3530}
3531# rev_map:
3533# Tie::File seems to be prone to offset errors if revisions get sparse,
3534# it's not that fast, either. Tie::File is also not in Perl 5.6. So
3535# one of my favorite modules is out :< Next up would be one of the DBM
3536# modules, but I'm not sure which is most portable...
3537#
3538# This is the replacement for the rev_db format, which was too big
3539# and inefficient for large repositories with a lot of sparse history
3540# (mainly tags)
3541#
3542# The format is this:
3543# - 24 bytes for every record,
3544# * 4 bytes for the integer representing an SVN revision number
3545# * 20 bytes representing the sha1 of a git commit
3546# - No empty padding records like the old format
3547# (except the last record, which can be overwritten)
3548# - new records are written append-only since SVN revision numbers
3549# increase monotonically
3550# - lookups on SVN revision number are done via a binary search
3551# - Piping the file to xxd -c24 is a good way of dumping it for
3552# viewing or editing (piped back through xxd -r), should the need
3553# ever arise.
3554# - The last record can be padding revision with an all-zero sha1
3555# This is used to optimize fetch performance when using multiple
3556# "fetch" directives in .git/config
3557#
3558# These files are disposable unless noMetadata or useSvmProps is set
3559sub _rev_map_set {
3561 my ($fh, $rev, $commit) = @_;
3562 binmode $fh or croak "binmode: $!";
3564 my $size = (stat($fh))[7];
3565 ($size % 24) == 0 or croak "inconsistent size: $size";
3566 my $wr_offset = 0;
3568 if ($size > 0) {
3569 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
3570 my $read = sysread($fh, my $buf, 24) or croak "read: $!";
3571 $read == 24 or croak "read only $read bytes (!= 24)";
3572 my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
3573 if ($last_commit eq ('0' x40)) {
3574 if ($size >= 48) {
3575 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
3576 $read = sysread($fh, $buf, 24) or
3577 croak "read: $!";
3578 $read == 24 or
3579 croak "read only $read bytes (!= 24)";
3580 ($last_rev, $last_commit) =
3581 unpack(rev_map_fmt, $buf);
3582 if ($last_commit eq ('0' x40)) {
3583 croak "inconsistent .rev_map\n";
3584 }
3585 }
3586 if ($last_rev >= $rev) {
3587 croak "last_rev is higher!: $last_rev >= $rev";
3588 }
3589 $wr_offset = -24;
3590 }
3591 }
3592 sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
3593 syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
3594 croak "write: $!";
3595}
3596sub _rev_map_reset {
3598 my ($fh, $rev, $commit) = @_;
3599 my $c = _rev_map_get($fh, $rev);
3600 $c eq $commit or die "_rev_map_reset(@_) commit $c does not match!\n";
3601 my $offset = sysseek($fh, 0, SEEK_CUR) or croak "seek: $!";
3602 truncate $fh, $offset or croak "truncate: $!";
3603}
3604sub mkfile {
3606 my ($path) = @_;
3607 unless (-e $path) {
3608 my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
3609 mkpath([$dir]) unless -d $dir;
3610 open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
3611 close $fh or die "Couldn't close (create) $path: $!\n";
3612 }
3613}
3614sub rev_map_set {
3616 my ($self, $rev, $commit, $update_ref, $uuid) = @_;
3617 defined $commit or die "missing arg3\n";
3618 length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
3619 my $db = $self->map_path($uuid);
3620 my $db_lock = "$db.lock";
3621 my $sig;
3622 $update_ref ||= 0;
3623 if ($update_ref) {
3624 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
3625 $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
3626 }
3627 mkfile($db);
3628 $LOCKFILES{$db_lock} = 1;
3630 my $sync;
3631 # both of these options make our .rev_db file very, very important
3632 # and we can't afford to lose it because rebuild() won't work
3633 if ($self->use_svm_props || $self->no_metadata) {
3634 $sync = 1;
3635 copy($db, $db_lock) or die "rev_map_set(@_): ",
3636 "Failed to copy: ",
3637 "$db => $db_lock ($!)\n";
3638 } else {
3639 rename $db, $db_lock or die "rev_map_set(@_): ",
3640 "Failed to rename: ",
3641 "$db => $db_lock ($!)\n";
3642 }
3643 sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
3645 or croak "Couldn't open $db_lock: $!\n";
3646 $update_ref eq 'reset' ? _rev_map_reset($fh, $rev, $commit) :
3647 _rev_map_set($fh, $rev, $commit);
3648 if ($sync) {
3649 $fh->flush or die "Couldn't flush $db_lock: $!\n";
3650 $fh->sync or die "Couldn't sync $db_lock: $!\n";
3651 }
3652 close $fh or croak $!;
3653 if ($update_ref) {
3654 $_head = $self;
3655 my $note = "";
3656 $note = " ($update_ref)" if ($update_ref !~ /^\d*$/);
3657 command_noisy('update-ref', '-m', "r$rev$note",
3658 $self->refname, $commit);
3659 }
3660 rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
3661 "$db_lock => $db ($!)\n";
3662 delete $LOCKFILES{$db_lock};
3663 if ($update_ref) {
3664 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
3665 $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
3666 kill $sig, $$ if defined $sig;
3667 }
3668}
3669# If want_commit, this will return an array of (rev, commit) where
3671# commit _must_ be a valid commit in the archive.
3672# Otherwise, it'll return the max revision (whether or not the
3673# commit is valid or just a 0x40 placeholder).
3674sub rev_map_max {
3675 my ($self, $want_commit) = @_;
3676 $self->rebuild;
3677 my ($r, $c) = $self->rev_map_max_norebuild($want_commit);
3678 $want_commit ? ($r, $c) : $r;
3679}
3680sub rev_map_max_norebuild {
3682 my ($self, $want_commit) = @_;
3683 my $map_path = $self->map_path;
3684 stat $map_path or return $want_commit ? (0, undef) : 0;
3685 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
3686 binmode $fh or croak "binmode: $!";
3687 my $size = (stat($fh))[7];
3688 ($size % 24) == 0 or croak "inconsistent size: $size";
3689 if ($size == 0) {
3691 close $fh or croak "close: $!";
3692 return $want_commit ? (0, undef) : 0;
3693 }
3694 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
3696 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
3697 my ($r, $c) = unpack(rev_map_fmt, $buf);
3698 if ($want_commit && $c eq ('0' x40)) {
3699 if ($size < 48) {
3700 return $want_commit ? (0, undef) : 0;
3701 }
3702 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
3703 sysread($fh, $buf, 24) == 24 or croak "read: $!";
3704 ($r, $c) = unpack(rev_map_fmt, $buf);
3705 if ($c eq ('0'x40)) {
3706 croak "Penultimate record is all-zeroes in $map_path";
3707 }
3708 }
3709 close $fh or croak "close: $!";
3710 $want_commit ? ($r, $c) : $r;
3711}
3712sub rev_map_get {
3714 my ($self, $rev, $uuid) = @_;
3715 my $map_path = $self->map_path($uuid);
3716 return undef unless -e $map_path;
3717 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
3719 my $c = _rev_map_get($fh, $rev);
3720 close($fh) or croak "close: $!";
3721 $c
3722}
3723sub _rev_map_get {
3725 my ($fh, $rev) = @_;
3726 binmode $fh or croak "binmode: $!";
3728 my $size = (stat($fh))[7];
3729 ($size % 24) == 0 or croak "inconsistent size: $size";
3730 if ($size == 0) {
3732 return undef;
3733 }
3734 my ($l, $u) = (0, $size - 24);
3736 my ($r, $c, $buf);
3737 while ($l <= $u) {
3739 my $i = int(($l/24 + $u/24) / 2) * 24;
3740 sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
3741 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
3742 my ($r, $c) = unpack(rev_map_fmt, $buf);
3743 if ($r < $rev) {
3745 $l = $i + 24;
3746 } elsif ($r > $rev) {
3747 $u = $i - 24;
3748 } else { # $r == $rev
3749 return $c eq ('0' x 40) ? undef : $c;
3750 }
3751 }
3752 undef;
3753}
3754# Finds the first svn revision that exists on (if $eq_ok is true) or
3756# before $rev for the current branch. It will not search any lower
3757# than $min_rev. Returns the git commit hash and svn revision number
3758# if found, else (undef, undef).
3759sub find_rev_before {
3760 my ($self, $rev, $eq_ok, $min_rev) = @_;
3761 --$rev unless $eq_ok;
3762 $min_rev ||= 1;
3763 my $max_rev = $self->rev_map_max;
3764 $rev = $max_rev if ($rev > $max_rev);
3765 while ($rev >= $min_rev) {
3766 if (my $c = $self->rev_map_get($rev)) {
3767 return ($rev, $c);
3768 }
3769 --$rev;
3770 }
3771 return (undef, undef);
3772}
3773# Finds the first svn revision that exists on (if $eq_ok is true) or
3775# after $rev for the current branch. It will not search any higher
3776# than $max_rev. Returns the git commit hash and svn revision number
3777# if found, else (undef, undef).
3778sub find_rev_after {
3779 my ($self, $rev, $eq_ok, $max_rev) = @_;
3780 ++$rev unless $eq_ok;
3781 $max_rev ||= $self->rev_map_max;
3782 while ($rev <= $max_rev) {
3783 if (my $c = $self->rev_map_get($rev)) {
3784 return ($rev, $c);
3785 }
3786 ++$rev;
3787 }
3788 return (undef, undef);
3789}
3790sub _new {
3792 my ($class, $repo_id, $ref_id, $path) = @_;
3793 unless (defined $repo_id && length $repo_id) {
3794 $repo_id = $Git::SVN::default_repo_id;
3795 }
3796 unless (defined $ref_id && length $ref_id) {
3797 $_prefix = '' unless defined($_prefix);
3798 $_[2] = $ref_id =
3799 "refs/remotes/$_prefix$Git::SVN::default_ref_id";
3800 }
3801 $_[1] = $repo_id;
3802 my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
3803 # Older repos imported by us used $GIT_DIR/svn/foo instead of
3805 # $GIT_DIR/svn/refs/remotes/foo when tracking refs/remotes/foo
3806 if ($ref_id =~ m{^refs/remotes/(.*)}) {
3807 my $old_dir = "$ENV{GIT_DIR}/svn/$1";
3808 if (-d $old_dir && ! -d $dir) {
3809 $dir = $old_dir;
3810 }
3811 }
3812 $_[3] = $path = '' unless (defined $path);
3814 mkpath([$dir]);
3815 bless {
3816 ref_id => $ref_id, dir => $dir, index => "$dir/index",
3817 path => $path, config => "$ENV{GIT_DIR}/svn/config",
3818 map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
3819}
3820# for read-only access of old .rev_db formats
3822sub unlink_rev_db_symlink {
3823 my ($self) = @_;
3824 my $link = $self->rev_db_path;
3825 $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
3826 if (-l $link) {
3827 unlink $link or croak "unlink: $link failed!";
3828 }
3829}
3830sub rev_db_path {
3832 my ($self, $uuid) = @_;
3833 my $db_path = $self->map_path($uuid);
3834 $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
3835 or croak "map_path: $db_path does not contain '/.rev_map.' !";
3836 $db_path;
3837}
3838# the new replacement for .rev_db
3840sub map_path {
3841 my ($self, $uuid) = @_;
3842 $uuid ||= $self->ra_uuid;
3843 "$self->{map_root}.$uuid";
3844}
3845sub uri_encode {
3847 my ($f) = @_;
3848 $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
3849 $f
3850}
3851sub uri_decode {
3853 my ($f) = @_;
3854 $f =~ s#%([0-9a-fA-F]{2})#chr(hex($1))#eg;
3855 $f
3856}
3857sub remove_username {
3859 $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
3860}
3861package Git::SVN::Prompt;
3863use strict;
3864use warnings;
3865require SVN::Core;
3866use vars qw/$_no_auth_cache $_username/;
3867sub simple {
3869 my ($cred, $realm, $default_username, $may_save, $pool) = @_;
3870 $may_save = undef if $_no_auth_cache;
3871 $default_username = $_username if defined $_username;
3872 if (defined $default_username && length $default_username) {
3873 if (defined $realm && length $realm) {
3874 print STDERR "Authentication realm: $realm\n";
3875 STDERR->flush;
3876 }
3877 $cred->username($default_username);
3878 } else {
3879 username($cred, $realm, $may_save, $pool);
3880 }
3881 $cred->password(_read_password("Password for '" .
3882 $cred->username . "': ", $realm));
3883 $cred->may_save($may_save);
3884 $SVN::_Core::SVN_NO_ERROR;
3885}
3886sub ssl_server_trust {
3888 my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
3889 $may_save = undef if $_no_auth_cache;
3890 print STDERR "Error validating server certificate for '$realm':\n";
3891 {
3892 no warnings 'once';
3893 # All variables SVN::Auth::SSL::* are used only once,
3894 # so we're shutting up Perl warnings about this.
3895 if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
3896 print STDERR " - The certificate is not issued ",
3897 "by a trusted authority. Use the\n",
3898 " fingerprint to validate ",
3899 "the certificate manually!\n";
3900 }
3901 if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
3902 print STDERR " - The certificate hostname ",
3903 "does not match.\n";
3904 }
3905 if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
3906 print STDERR " - The certificate is not yet valid.\n";
3907 }
3908 if ($failures & $SVN::Auth::SSL::EXPIRED) {
3909 print STDERR " - The certificate has expired.\n";
3910 }
3911 if ($failures & $SVN::Auth::SSL::OTHER) {
3912 print STDERR " - The certificate has ",
3913 "an unknown error.\n";
3914 }
3915 } # no warnings 'once'
3916 printf STDERR
3917 "Certificate information:\n".
3918 " - Hostname: %s\n".
3919 " - Valid: from %s until %s\n".
3920 " - Issuer: %s\n".
3921 " - Fingerprint: %s\n",
3922 map $cert_info->$_, qw(hostname valid_from valid_until
3923 issuer_dname fingerprint);
3924 my $choice;
3925prompt:
3926 print STDERR $may_save ?
3927 "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
3928 "(R)eject or accept (t)emporarily? ";
3929 STDERR->flush;
3930 $choice = lc(substr(<STDIN> || 'R', 0, 1));
3931 if ($choice =~ /^t$/i) {
3932 $cred->may_save(undef);
3933 } elsif ($choice =~ /^r$/i) {
3934 return -1;
3935 } elsif ($may_save && $choice =~ /^p$/i) {
3936 $cred->may_save($may_save);
3937 } else {
3938 goto prompt;
3939 }
3940 $cred->accepted_failures($failures);
3941 $SVN::_Core::SVN_NO_ERROR;
3942}
3943sub ssl_client_cert {
3945 my ($cred, $realm, $may_save, $pool) = @_;
3946 $may_save = undef if $_no_auth_cache;
3947 print STDERR "Client certificate filename: ";
3948 STDERR->flush;
3949 chomp(my $filename = <STDIN>);
3950 $cred->cert_file($filename);
3951 $cred->may_save($may_save);
3952 $SVN::_Core::SVN_NO_ERROR;
3953}
3954sub ssl_client_cert_pw {
3956 my ($cred, $realm, $may_save, $pool) = @_;
3957 $may_save = undef if $_no_auth_cache;
3958 $cred->password(_read_password("Password: ", $realm));
3959 $cred->may_save($may_save);
3960 $SVN::_Core::SVN_NO_ERROR;
3961}
3962sub username {
3964 my ($cred, $realm, $may_save, $pool) = @_;
3965 $may_save = undef if $_no_auth_cache;
3966 if (defined $realm && length $realm) {
3967 print STDERR "Authentication realm: $realm\n";
3968 }
3969 my $username;
3970 if (defined $_username) {
3971 $username = $_username;
3972 } else {
3973 print STDERR "Username: ";
3974 STDERR->flush;
3975 chomp($username = <STDIN>);
3976 }
3977 $cred->username($username);
3978 $cred->may_save($may_save);
3979 $SVN::_Core::SVN_NO_ERROR;
3980}
3981sub _read_password {
3983 my ($prompt, $realm) = @_;
3984 my $password = '';
3985 if (exists $ENV{GIT_ASKPASS}) {
3986 open(PH, "-|", $ENV{GIT_ASKPASS}, $prompt);
3987 $password = <PH>;
3988 $password =~ s/[\012\015]//; # \n\r
3989 close(PH);
3990 } else {
3991 print STDERR $prompt;
3992 STDERR->flush;
3993 require Term::ReadKey;
3994 Term::ReadKey::ReadMode('noecho');
3995 while (defined(my $key = Term::ReadKey::ReadKey(0))) {
3996 last if $key =~ /[\012\015]/; # \n\r
3997 $password .= $key;
3998 }
3999 Term::ReadKey::ReadMode('restore');
4000 print STDERR "\n";
4001 STDERR->flush;
4002 }
4003 $password;
4004}
4005package SVN::Git::Fetcher;
4007use vars qw/@ISA/;
4008use strict;
4009use warnings;
4010use Carp qw/croak/;
4011use IO::File qw//;
4012use vars qw/$_ignore_regex/;
4013# file baton members: path, mode_a, mode_b, pool, fh, blob, base
4015sub new {
4016 my ($class, $git_svn, $switch_path) = @_;
4017 my $self = SVN::Delta::Editor->new;
4018 bless $self, $class;
4019 if (exists $git_svn->{last_commit}) {
4020 $self->{c} = $git_svn->{last_commit};
4021 $self->{empty_symlinks} =
4022 _mark_empty_symlinks($git_svn, $switch_path);
4023 }
4024 $self->{ignore_regex} = eval { command_oneline('config', '--get',
4025 "svn-remote.$git_svn->{repo_id}.ignore-paths") };
4026 $self->{empty} = {};
4027 $self->{dir_prop} = {};
4028 $self->{file_prop} = {};
4029 $self->{absent_dir} = {};
4030 $self->{absent_file} = {};
4031 $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
4032 $self;
4033}
4034# this uses the Ra object, so it must be called before do_{switch,update},
4036# not inside them (when the Git::SVN::Fetcher object is passed) to
4037# do_{switch,update}
4038sub _mark_empty_symlinks {
4039 my ($git_svn, $switch_path) = @_;
4040 my $bool = Git::config_bool('svn.brokenSymlinkWorkaround');
4041 return {} if (!defined($bool)) || (defined($bool) && ! $bool);
4042 my %ret;
4044 my ($rev, $cmt) = $git_svn->last_rev_commit;
4045 return {} unless ($rev && $cmt);
4046 # allow the warning to be printed for each revision we fetch to
4048 # ensure the user sees it. The user can also disable the workaround
4049 # on the repository even while git svn is running and the next
4050 # revision fetched will skip this expensive function.
4051 my $printed_warning;
4052 chomp(my $empty_blob = `git hash-object -t blob --stdin < /dev/null`);
4053 my ($ls, $ctx) = command_output_pipe(qw/ls-tree -r -z/, $cmt);
4054 local $/ = "\0";
4055 my $pfx = defined($switch_path) ? $switch_path : $git_svn->{path};
4056 $pfx .= '/' if length($pfx);
4057 while (<$ls>) {
4058 chomp;
4059 s/\A100644 blob $empty_blob\t//o or next;
4060 unless ($printed_warning) {
4061 print STDERR "Scanning for empty symlinks, ",
4062 "this may take a while if you have ",
4063 "many empty files\n",
4064 "You may disable this with `",
4065 "git config svn.brokenSymlinkWorkaround ",
4066 "false'.\n",
4067 "This may be done in a different ",
4068 "terminal without restarting ",
4069 "git svn\n";
4070 $printed_warning = 1;
4071 }
4072 my $path = $_;
4073 my (undef, $props) =
4074 $git_svn->ra->get_file($pfx.$path, $rev, undef);
4075 if ($props->{'svn:special'}) {
4076 $ret{$path} = 1;
4077 }
4078 }
4079 command_close_pipe($ls, $ctx);
4080 \%ret;
4081}
4082# returns true if a given path is inside a ".git" directory
4084sub in_dot_git {
4085 $_[0] =~ m{(?:^|/)\.git(?:/|$)};
4086}
4087# return value: 0 -- don't ignore, 1 -- ignore
4089sub is_path_ignored {
4090 my ($self, $path) = @_;
4091 return 1 if in_dot_git($path);
4092 return 1 if defined($self->{ignore_regex}) &&
4093 $path =~ m!$self->{ignore_regex}!;
4094 return 0 unless defined($_ignore_regex);
4095 return 1 if $path =~ m!$_ignore_regex!o;
4096 return 0;
4097}
4098sub set_path_strip {
4100 my ($self, $path) = @_;
4101 $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
4102}
4103sub open_root {
4105 { path => '' };
4106}
4107sub open_directory {
4109 my ($self, $path, $pb, $rev) = @_;
4110 { path => $path };
4111}
4112sub git_path {
4114 my ($self, $path) = @_;
4115 if ($self->{path_strip}) {
4116 $path =~ s!$self->{path_strip}!! or
4117 die "Failed to strip path '$path' ($self->{path_strip})\n";
4118 }
4119 $path;
4120}
4121sub delete_entry {
4123 my ($self, $path, $rev, $pb) = @_;
4124 return undef if $self->is_path_ignored($path);
4125 my $gpath = $self->git_path($path);
4127 return undef if ($gpath eq '');
4128 # remove entire directories.
4130 my ($tree) = (command('ls-tree', '-z', $self->{c}, "./$gpath")
4131 =~ /\A040000 tree ([a-f\d]{40})\t\Q$gpath\E\0/);
4132 if ($tree) {
4133 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
4134 -r --name-only -z/,
4135 $tree);
4136 local $/ = "\0";
4137 while (<$ls>) {
4138 chomp;
4139 my $rmpath = "$gpath/$_";
4140 $self->{gii}->remove($rmpath);
4141 print "\tD\t$rmpath\n" unless $::_q;
4142 }
4143 print "\tD\t$gpath/\n" unless $::_q;
4144 command_close_pipe($ls, $ctx);
4145 } else {
4146 $self->{gii}->remove($gpath);
4147 print "\tD\t$gpath\n" unless $::_q;
4148 }
4149 $self->{empty}->{$path} = 0;
4150 undef;
4151}
4152sub open_file {
4154 my ($self, $path, $pb, $rev) = @_;
4155 my ($mode, $blob);
4156 goto out if $self->is_path_ignored($path);
4158 my $gpath = $self->git_path($path);
4160 ($mode, $blob) = (command('ls-tree', '-z', $self->{c}, "./$gpath")
4161 =~ /\A(\d{6}) blob ([a-f\d]{40})\t\Q$gpath\E\0/);
4162 unless (defined $mode && defined $blob) {
4163 die "$path was not found in commit $self->{c} (r$rev)\n";
4164 }
4165 if ($mode eq '100644' && $self->{empty_symlinks}->{$path}) {
4166 $mode = '120000';
4167 }
4168out:
4169 { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
4170 pool => SVN::Pool->new, action => 'M' };
4171}
4172sub add_file {
4174 my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
4175 my $mode;
4176 if (!$self->is_path_ignored($path)) {
4178 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
4179 delete $self->{empty}->{$dir};
4180 $mode = '100644';
4181 }
4182 { path => $path, mode_a => $mode, mode_b => $mode,
4183 pool => SVN::Pool->new, action => 'A' };
4184}
4185sub add_directory {
4187 my ($self, $path, $cp_path, $cp_rev) = @_;
4188 goto out if $self->is_path_ignored($path);
4189 my $gpath = $self->git_path($path);
4190 if ($gpath eq '') {
4191 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
4192 -r --name-only -z/,
4193 $self->{c});
4194 local $/ = "\0";
4195 while (<$ls>) {
4196 chomp;
4197 $self->{gii}->remove($_);
4198 print "\tD\t$_\n" unless $::_q;
4199 }
4200 command_close_pipe($ls, $ctx);
4201 $self->{empty}->{$path} = 0;
4202 }
4203 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
4204 delete $self->{empty}->{$dir};
4205 $self->{empty}->{$path} = 1;
4206out:
4207 { path => $path };
4208}
4209sub change_dir_prop {
4211 my ($self, $db, $prop, $value) = @_;
4212 return undef if $self->is_path_ignored($db->{path});
4213 $self->{dir_prop}->{$db->{path}} ||= {};
4214 $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
4215 undef;
4216}
4217sub absent_directory {
4219 my ($self, $path, $pb) = @_;
4220 return undef if $self->is_path_ignored($path);
4221 $self->{absent_dir}->{$pb->{path}} ||= [];
4222 push @{$self->{absent_dir}->{$pb->{path}}}, $path;
4223 undef;
4224}
4225sub absent_file {
4227 my ($self, $path, $pb) = @_;
4228 return undef if $self->is_path_ignored($path);
4229 $self->{absent_file}->{$pb->{path}} ||= [];
4230 push @{$self->{absent_file}->{$pb->{path}}}, $path;
4231 undef;
4232}
4233sub change_file_prop {
4235 my ($self, $fb, $prop, $value) = @_;
4236 return undef if $self->is_path_ignored($fb->{path});
4237 if ($prop eq 'svn:executable') {
4238 if ($fb->{mode_b} != 120000) {
4239 $fb->{mode_b} = defined $value ? 100755 : 100644;
4240 }
4241 } elsif ($prop eq 'svn:special') {
4242 $fb->{mode_b} = defined $value ? 120000 : 100644;
4243 } else {
4244 $self->{file_prop}->{$fb->{path}} ||= {};
4245 $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
4246 }
4247 undef;
4248}
4249sub apply_textdelta {
4251 my ($self, $fb, $exp) = @_;
4252 return undef if $self->is_path_ignored($fb->{path});
4253 my $fh = $::_repository->temp_acquire('svn_delta');
4254 # $fh gets auto-closed() by SVN::TxDelta::apply(),
4255 # (but $base does not,) so dup() it for reading in close_file
4256 open my $dup, '<&', $fh or croak $!;
4257 my $base = $::_repository->temp_acquire('git_blob');
4258 if ($fb->{blob}) {
4260 my ($base_is_link, $size);
4261 if ($fb->{mode_a} eq '120000' &&
4263 ! $self->{empty_symlinks}->{$fb->{path}}) {
4264 print $base 'link ' or die "print $!\n";
4265 $base_is_link = 1;
4266 }
4267 retry:
4268 $size = $::_repository->cat_blob($fb->{blob}, $base);
4269 die "Failed to read object $fb->{blob}" if ($size < 0);
4270 if (defined $exp) {
4272 seek $base, 0, 0 or croak $!;
4273 my $got = ::md5sum($base);
4274 if ($got ne $exp) {
4275 my $err = "Checksum mismatch: ".
4276 "$fb->{path} $fb->{blob}\n" .
4277 "expected: $exp\n" .
4278 " got: $got\n";
4279 if ($base_is_link) {
4280 warn $err,
4281 "Retrying... (possibly ",
4282 "a bad symlink from SVN)\n";
4283 $::_repository->temp_reset($base);
4284 $base_is_link = 0;
4285 goto retry;
4286 }
4287 die $err;
4288 }
4289 }
4290 }
4291 seek $base, 0, 0 or croak $!;
4292 $fb->{fh} = $fh;
4293 $fb->{base} = $base;
4294 [ SVN::TxDelta::apply($base, $dup, undef, $fb->{path}, $fb->{pool}) ];
4295}
4296sub close_file {
4298 my ($self, $fb, $exp) = @_;
4299 return undef if $self->is_path_ignored($fb->{path});
4300 my $hash;
4302 my $path = $self->git_path($fb->{path});
4303 if (my $fh = $fb->{fh}) {
4304 if (defined $exp) {
4305 seek($fh, 0, 0) or croak $!;
4306 my $got = ::md5sum($fh);
4307 if ($got ne $exp) {
4308 die "Checksum mismatch: $path\n",
4309 "expected: $exp\n got: $got\n";
4310 }
4311 }
4312 if ($fb->{mode_b} == 120000) {
4313 sysseek($fh, 0, 0) or croak $!;
4314 my $rd = sysread($fh, my $buf, 5);
4315 if (!defined $rd) {
4317 croak "sysread: $!\n";
4318 } elsif ($rd == 0) {
4319 warn "$path has mode 120000",
4320 " but it points to nothing\n",
4321 "converting to an empty file with mode",
4322 " 100644\n";
4323 $fb->{mode_b} = '100644';
4324 } elsif ($buf ne 'link ') {
4325 warn "$path has mode 120000",
4326 " but is not a link\n";
4327 } else {
4328 my $tmp_fh = $::_repository->temp_acquire(
4329 'svn_hash');
4330 my $res;
4331 while ($res = sysread($fh, my $str, 1024)) {
4332 my $out = syswrite($tmp_fh, $str, $res);
4333 defined($out) && $out == $res
4334 or croak("write ",
4335 Git::temp_path($tmp_fh),
4336 ": $!\n");
4337 }
4338 defined $res or croak $!;
4339 ($fh, $tmp_fh) = ($tmp_fh, $fh);
4341 Git::temp_release($tmp_fh, 1);
4342 }
4343 }
4344 $hash = $::_repository->hash_and_insert_object(
4346 Git::temp_path($fh));
4347 $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
4348 Git::temp_release($fb->{base}, 1);
4350 Git::temp_release($fh, 1);
4351 } else {
4352 $hash = $fb->{blob} or die "no blob information\n";
4353 }
4354 $fb->{pool}->clear;
4355 $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
4356 print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
4357 undef;
4358}
4359sub abort_edit {
4361 my $self = shift;
4362 $self->{nr} = $self->{gii}->{nr};
4363 delete $self->{gii};
4364 $self->SUPER::abort_edit(@_);
4365}
4366sub close_edit {
4368 my $self = shift;
4369 $self->{git_commit_ok} = 1;
4370 $self->{nr} = $self->{gii}->{nr};
4371 delete $self->{gii};
4372 $self->SUPER::close_edit(@_);
4373}
4374package SVN::Git::Editor;
4376use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
4377use strict;
4378use warnings;
4379use Carp qw/croak/;
4380use IO::File;
4381sub new {
4383 my ($class, $opts) = @_;
4384 foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
4385 die "$_ required!\n" unless (defined $opts->{$_});
4386 }
4387 my $pool = SVN::Pool->new;
4389 my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
4390 my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
4391 $opts->{r}, $mods);
4392 # $opts->{ra} functions should not be used after this:
4394 my @ce = $opts->{ra}->get_commit_editor($opts->{log},
4395 $opts->{editor_cb}, $pool);
4396 my $self = SVN::Delta::Editor->new(@ce, $pool);
4397 bless $self, $class;
4398 foreach (qw/svn_path r tree_a tree_b/) {
4399 $self->{$_} = $opts->{$_};
4400 }
4401 $self->{url} = $opts->{ra}->{url};
4402 $self->{mods} = $mods;
4403 $self->{types} = $types;
4404 $self->{pool} = $pool;
4405 $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
4406 $self->{rm} = { };
4407 $self->{path_prefix} = length $self->{svn_path} ?
4408 "$self->{svn_path}/" : '';
4409 $self->{config} = $opts->{config};
4410 return $self;
4411}
4412sub generate_diff {
4414 my ($tree_a, $tree_b) = @_;
4415 my @diff_tree = qw(diff-tree -z -r);
4416 if ($_cp_similarity) {
4417 push @diff_tree, "-C$_cp_similarity";
4418 } else {
4419 push @diff_tree, '-C';
4420 }
4421 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
4422 push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
4423 push @diff_tree, $tree_a, $tree_b;
4424 my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
4425 local $/ = "\0";
4426 my $state = 'meta';
4427 my @mods;
4428 while (<$diff_fh>) {
4429 chomp $_; # this gets rid of the trailing "\0"
4430 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
4431 ($::sha1)\s($::sha1)\s
4432 ([MTCRAD])\d*$/xo) {
4433 push @mods, { mode_a => $1, mode_b => $2,
4434 sha1_a => $3, sha1_b => $4,
4435 chg => $5 };
4436 if ($5 =~ /^(?:C|R)$/) {
4437 $state = 'file_a';
4438 } else {
4439 $state = 'file_b';
4440 }
4441 } elsif ($state eq 'file_a') {
4442 my $x = $mods[$#mods] or croak "Empty array\n";
4443 if ($x->{chg} !~ /^(?:C|R)$/) {
4444 croak "Error parsing $_, $x->{chg}\n";
4445 }
4446 $x->{file_a} = $_;
4447 $state = 'file_b';
4448 } elsif ($state eq 'file_b') {
4449 my $x = $mods[$#mods] or croak "Empty array\n";
4450 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
4451 croak "Error parsing $_, $x->{chg}\n";
4452 }
4453 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
4454 croak "Error parsing $_, $x->{chg}\n";
4455 }
4456 $x->{file_b} = $_;
4457 $state = 'meta';
4458 } else {
4459 croak "Error parsing $_\n";
4460 }
4461 }
4462 command_close_pipe($diff_fh, $ctx);
4463 \@mods;
4464}
4465sub check_diff_paths {
4467 my ($ra, $pfx, $rev, $mods) = @_;
4468 my %types;
4469 $pfx .= '/' if length $pfx;
4470 sub type_diff_paths {
4472 my ($ra, $types, $path, $rev) = @_;
4473 my @p = split m#/+#, $path;
4474 my $c = shift @p;
4475 unless (defined $types->{$c}) {
4476 $types->{$c} = $ra->check_path($c, $rev);
4477 }
4478 while (@p) {
4479 $c .= '/' . shift @p;
4480 next if defined $types->{$c};
4481 $types->{$c} = $ra->check_path($c, $rev);
4482 }
4483 }
4484 foreach my $m (@$mods) {
4486 foreach my $f (qw/file_a file_b/) {
4487 next unless defined $m->{$f};
4488 my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
4489 if (length $pfx.$dir && ! defined $types{$dir}) {
4490 type_diff_paths($ra, \%types, $pfx.$dir, $rev);
4491 }
4492 }
4493 }
4494 \%types;
4495}
4496sub split_path {
4498 return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
4499}
4500sub repo_path {
4502 my ($self, $path) = @_;
4503 $self->{path_prefix}.(defined $path ? $path : '');
4504}
4505sub url_path {
4507 my ($self, $path) = @_;
4508 if ($self->{url} =~ m#^https?://#) {
4509 $path =~ s!([^~a-zA-Z0-9_./-])!uc sprintf("%%%02x",ord($1))!eg;
4510 }
4511 $self->{url} . '/' . $self->repo_path($path);
4512}
4513sub rmdirs {
4515 my ($self) = @_;
4516 my $rm = $self->{rm};
4517 delete $rm->{''}; # we never delete the url we're tracking
4518 return unless %$rm;
4519 foreach (keys %$rm) {
4521 my @d = split m#/#, $_;
4522 my $c = shift @d;
4523 $rm->{$c} = 1;
4524 while (@d) {
4525 $c .= '/' . shift @d;
4526 $rm->{$c} = 1;
4527 }
4528 }
4529 delete $rm->{$self->{svn_path}};
4530 delete $rm->{''}; # we never delete the url we're tracking
4531 return unless %$rm;
4532 my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
4534 $self->{tree_b});
4535 local $/ = "\0";
4536 while (<$fh>) {
4537 chomp;
4538 my @dn = split m#/#, $_;
4539 while (pop @dn) {
4540 delete $rm->{join '/', @dn};
4541 }
4542 unless (%$rm) {
4543 close $fh;
4544 return;
4545 }
4546 }
4547 command_close_pipe($fh, $ctx);
4548 my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
4550 foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
4551 $self->close_directory($bat->{$d}, $p);
4552 my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
4553 print "\tD+\t$d/\n" unless $::_q;
4554 $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
4555 delete $bat->{$d};
4556 }
4557}
4558sub open_or_add_dir {
4560 my ($self, $full_path, $baton) = @_;
4561 my $t = $self->{types}->{$full_path};
4562 if (!defined $t) {
4563 die "$full_path not known in r$self->{r} or we have a bug!\n";
4564 }
4565 {
4566 no warnings 'once';
4567 # SVN::Node::none and SVN::Node::file are used only once,
4568 # so we're shutting up Perl's warnings about them.
4569 if ($t == $SVN::Node::none) {
4570 return $self->add_directory($full_path, $baton,
4571 undef, -1, $self->{pool});
4572 } elsif ($t == $SVN::Node::dir) {
4573 return $self->open_directory($full_path, $baton,
4574 $self->{r}, $self->{pool});
4575 } # no warnings 'once'
4576 print STDERR "$full_path already exists in repository at ",
4577 "r$self->{r} and it is not a directory (",
4578 ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
4579 } # no warnings 'once'
4580 exit 1;
4581}
4582sub ensure_path {
4584 my ($self, $path) = @_;
4585 my $bat = $self->{bat};
4586 my $repo_path = $self->repo_path($path);
4587 return $bat->{''} unless (length $repo_path);
4588 my @p = split m#/+#, $repo_path;
4589 my $c = shift @p;
4590 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
4591 while (@p) {
4592 my $c0 = $c;
4593 $c .= '/' . shift @p;
4594 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
4595 }
4596 return $bat->{$c};
4597}
4598# Subroutine to convert a globbing pattern to a regular expression.
4600# From perl cookbook.
4601sub glob2pat {
4602 my $globstr = shift;
4603 my %patmap = ('*' => '.*', '?' => '.', '[' => '[', ']' => ']');
4604 $globstr =~ s{(.)} { $patmap{$1} || "\Q$1" }ge;
4605 return '^' . $globstr . '$';
4606}
4607sub check_autoprop {
4609 my ($self, $pattern, $properties, $file, $fbat) = @_;
4610 # Convert the globbing pattern to a regular expression.
4611 my $regex = glob2pat($pattern);
4612 # Check if the pattern matches the file name.
4613 if($file =~ m/($regex)/) {
4614 # Parse the list of properties to set.
4615 my @props = split(/;/, $properties);
4616 foreach my $prop (@props) {
4617 # Parse 'name=value' syntax and set the property.
4618 if ($prop =~ /([^=]+)=(.*)/) {
4619 my ($n,$v) = ($1,$2);
4620 for ($n, $v) {
4621 s/^\s+//; s/\s+$//;
4622 }
4623 $self->change_file_prop($fbat, $n, $v);
4624 }
4625 }
4626 }
4627}
4628sub apply_autoprops {
4630 my ($self, $file, $fbat) = @_;
4631 my $conf_t = ${$self->{config}}{'config'};
4632 no warnings 'once';
4633 # Check [miscellany]/enable-auto-props in svn configuration.
4634 if (SVN::_Core::svn_config_get_bool(
4635 $conf_t,
4636 $SVN::_Core::SVN_CONFIG_SECTION_MISCELLANY,
4637 $SVN::_Core::SVN_CONFIG_OPTION_ENABLE_AUTO_PROPS,
4638 0)) {
4639 # Auto-props are enabled. Enumerate them to look for matches.
4640 my $callback = sub {
4641 $self->check_autoprop($_[0], $_[1], $file, $fbat);
4642 };
4643 SVN::_Core::svn_config_enumerate(
4644 $conf_t,
4645 $SVN::_Core::SVN_CONFIG_SECTION_AUTO_PROPS,
4646 $callback);
4647 }
4648}
4649sub A {
4651 my ($self, $m) = @_;
4652 my ($dir, $file) = split_path($m->{file_b});
4653 my $pbat = $self->ensure_path($dir);
4654 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
4655 undef, -1);
4656 print "\tA\t$m->{file_b}\n" unless $::_q;
4657 $self->apply_autoprops($file, $fbat);
4658 $self->chg_file($fbat, $m);
4659 $self->close_file($fbat,undef,$self->{pool});
4660}
4661sub C {
4663 my ($self, $m) = @_;
4664 my ($dir, $file) = split_path($m->{file_b});
4665 my $pbat = $self->ensure_path($dir);
4666 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
4667 $self->url_path($m->{file_a}), $self->{r});
4668 print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
4669 $self->chg_file($fbat, $m);
4670 $self->close_file($fbat,undef,$self->{pool});
4671}
4672sub delete_entry {
4674 my ($self, $path, $pbat) = @_;
4675 my $rpath = $self->repo_path($path);
4676 my ($dir, $file) = split_path($rpath);
4677 $self->{rm}->{$dir} = 1;
4678 $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
4679}
4680sub R {
4682 my ($self, $m) = @_;
4683 my ($dir, $file) = split_path($m->{file_b});
4684 my $pbat = $self->ensure_path($dir);
4685 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
4686 $self->url_path($m->{file_a}), $self->{r});
4687 print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
4688 $self->apply_autoprops($file, $fbat);
4689 $self->chg_file($fbat, $m);
4690 $self->close_file($fbat,undef,$self->{pool});
4691 ($dir, $file) = split_path($m->{file_a});
4693 $pbat = $self->ensure_path($dir);
4694 $self->delete_entry($m->{file_a}, $pbat);
4695}
4696sub M {
4698 my ($self, $m) = @_;
4699 my ($dir, $file) = split_path($m->{file_b});
4700 my $pbat = $self->ensure_path($dir);
4701 my $fbat = $self->open_file($self->repo_path($m->{file_b}),
4702 $pbat,$self->{r},$self->{pool});
4703 print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
4704 $self->chg_file($fbat, $m);
4705 $self->close_file($fbat,undef,$self->{pool});
4706}
4707sub T { shift->M(@_) }
4709sub change_file_prop {
4711 my ($self, $fbat, $pname, $pval) = @_;
4712 $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
4713}
4714sub _chg_file_get_blob ($$$$) {
4716 my ($self, $fbat, $m, $which) = @_;
4717 my $fh = $::_repository->temp_acquire("git_blob_$which");
4718 if ($m->{"mode_$which"} =~ /^120/) {
4719 print $fh 'link ' or croak $!;
4720 $self->change_file_prop($fbat,'svn:special','*');
4721 } elsif ($m->{mode_a} =~ /^120/ && $m->{"mode_$which"} !~ /^120/) {
4722 $self->change_file_prop($fbat,'svn:special',undef);
4723 }
4724 my $blob = $m->{"sha1_$which"};
4725 return ($fh,) if ($blob =~ /^0{40}$/);
4726 my $size = $::_repository->cat_blob($blob, $fh);
4727 croak "Failed to read object $blob" if ($size < 0);
4728 $fh->flush == 0 or croak $!;
4729 seek $fh, 0, 0 or croak $!;
4730 my $exp = ::md5sum($fh);
4732 seek $fh, 0, 0 or croak $!;
4733 return ($fh, $exp);
4734}
4735sub chg_file {
4737 my ($self, $fbat, $m) = @_;
4738 if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
4739 $self->change_file_prop($fbat,'svn:executable','*');
4740 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
4741 $self->change_file_prop($fbat,'svn:executable',undef);
4742 }
4743 my ($fh_a, $exp_a) = _chg_file_get_blob $self, $fbat, $m, 'a';
4744 my ($fh_b, $exp_b) = _chg_file_get_blob $self, $fbat, $m, 'b';
4745 my $pool = SVN::Pool->new;
4746 my $atd = $self->apply_textdelta($fbat, $exp_a, $pool);
4747 if (-s $fh_a) {
4748 my $txstream = SVN::TxDelta::new ($fh_a, $fh_b, $pool);
4749 my $res = SVN::TxDelta::send_txstream($txstream, @$atd, $pool);
4750 if (defined $res) {
4751 die "Unexpected result from send_txstream: $res\n",
4752 "(SVN::Core::VERSION: $SVN::Core::VERSION)\n";
4753 }
4754 } else {
4755 my $got = SVN::TxDelta::send_stream($fh_b, @$atd, $pool);
4756 die "Checksum mismatch\nexpected: $exp_b\ngot: $got\n"
4757 if ($got ne $exp_b);
4758 }
4759 Git::temp_release($fh_b, 1);
4760 Git::temp_release($fh_a, 1);
4761 $pool->clear;
4762}
4763sub D {
4765 my ($self, $m) = @_;
4766 my ($dir, $file) = split_path($m->{file_b});
4767 my $pbat = $self->ensure_path($dir);
4768 print "\tD\t$m->{file_b}\n" unless $::_q;
4769 $self->delete_entry($m->{file_b}, $pbat);
4770}
4771sub close_edit {
4773 my ($self) = @_;
4774 my ($p,$bat) = ($self->{pool}, $self->{bat});
4775 foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
4776 next if $_ eq '';
4777 $self->close_directory($bat->{$_}, $p);
4778 }
4779 $self->close_directory($bat->{''}, $p);
4780 $self->SUPER::close_edit($p);
4781 $p->clear;
4782}
4783sub abort_edit {
4785 my ($self) = @_;
4786 $self->SUPER::abort_edit($self->{pool});
4787}
4788sub DESTROY {
4790 my $self = shift;
4791 $self->SUPER::DESTROY(@_);
4792 $self->{pool}->clear;
4793}
4794# this drives the editor
4796sub apply_diff {
4797 my ($self) = @_;
4798 my $mods = $self->{mods};
4799 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
4800 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
4801 my $f = $m->{chg};
4802 if (defined $o{$f}) {
4803 $self->$f($m);
4804 } else {
4805 fatal("Invalid change type: $f");
4806 }
4807 }
4808 $self->rmdirs if $_rmdir;
4809 if (@$mods == 0) {
4810 $self->abort_edit;
4811 } else {
4812 $self->close_edit;
4813 }
4814 return scalar @$mods;
4815}
4816package Git::SVN::Ra;
4818use vars qw/@ISA $config_dir $_log_window_size/;
4819use strict;
4820use warnings;
4821my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
4822BEGIN {
4824 # enforce temporary pool usage for some simple functions
4825 no strict 'refs';
4826 for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root
4827 get_file/) {
4828 my $SUPER = "SUPER::$f";
4829 *$f = sub {
4830 my $self = shift;
4831 my $pool = SVN::Pool->new;
4832 my @ret = $self->$SUPER(@_,$pool);
4833 $pool->clear;
4834 wantarray ? @ret : $ret[0];
4835 };
4836 }
4837}
4838sub _auth_providers () {
4840 [
4841 SVN::Client::get_simple_provider(),
4842 SVN::Client::get_ssl_server_trust_file_provider(),
4843 SVN::Client::get_simple_prompt_provider(
4844 \&Git::SVN::Prompt::simple, 2),
4845 SVN::Client::get_ssl_client_cert_file_provider(),
4846 SVN::Client::get_ssl_client_cert_prompt_provider(
4847 \&Git::SVN::Prompt::ssl_client_cert, 2),
4848 SVN::Client::get_ssl_client_cert_pw_file_provider(),
4849 SVN::Client::get_ssl_client_cert_pw_prompt_provider(
4850 \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
4851 SVN::Client::get_username_provider(),
4852 SVN::Client::get_ssl_server_trust_prompt_provider(
4853 \&Git::SVN::Prompt::ssl_server_trust),
4854 SVN::Client::get_username_prompt_provider(
4855 \&Git::SVN::Prompt::username, 2)
4856 ]
4857}
4858sub escape_uri_only {
4860 my ($uri) = @_;
4861 my @tmp;
4862 foreach (split m{/}, $uri) {
4863 s/([^~\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
4864 push @tmp, $_;
4865 }
4866 join('/', @tmp);
4867}
4868sub escape_url {
4870 my ($url) = @_;
4871 if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
4872 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
4873 $url = "$scheme://$domain$uri";
4874 }
4875 $url;
4876}
4877sub new {
4879 my ($class, $url) = @_;
4880 $url =~ s!/+$!!;
4881 return $RA if ($RA && $RA->{url} eq $url);
4882 ::_req_svn();
4884 SVN::_Core::svn_config_ensure($config_dir, undef);
4886 my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
4887 my $config = SVN::Core::config_get_config($config_dir);
4888 $RA = undef;
4889 my $dont_store_passwords = 1;
4890 my $conf_t = ${$config}{'config'};
4891 {
4892 no warnings 'once';
4893 # The usage of $SVN::_Core::SVN_CONFIG_* variables
4894 # produces warnings that variables are used only once.
4895 # I had not found the better way to shut them up, so
4896 # the warnings of type 'once' are disabled in this block.
4897 if (SVN::_Core::svn_config_get_bool($conf_t,
4898 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
4899 $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
4900 1) == 0) {
4901 SVN::_Core::svn_auth_set_parameter($baton,
4902 $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
4903 bless (\$dont_store_passwords, "_p_void"));
4904 }
4905 if (SVN::_Core::svn_config_get_bool($conf_t,
4906 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
4907 $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
4908 1) == 0) {
4909 $Git::SVN::Prompt::_no_auth_cache = 1;
4910 }
4911 } # no warnings 'once'
4912 my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
4913 config => $config,
4914 pool => SVN::Pool->new,
4915 auth_provider_callbacks => $callbacks);
4916 $self->{url} = $url;
4917 $self->{svn_path} = $url;
4918 $self->{repos_root} = $self->get_repos_root;
4919 $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
4920 $self->{cache} = { check_path => { r => 0, data => {} },
4921 get_dir => { r => 0, data => {} } };
4922 $RA = bless $self, $class;
4923}
4924sub check_path {
4926 my ($self, $path, $r) = @_;
4927 my $cache = $self->{cache}->{check_path};
4928 if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
4929 return $cache->{data}->{$path};
4930 }
4931 my $pool = SVN::Pool->new;
4932 my $t = $self->SUPER::check_path($path, $r, $pool);
4933 $pool->clear;
4934 if ($r != $cache->{r}) {
4935 %{$cache->{data}} = ();
4936 $cache->{r} = $r;
4937 }
4938 $cache->{data}->{$path} = $t;
4939}
4940sub get_dir {
4942 my ($self, $dir, $r) = @_;
4943 my $cache = $self->{cache}->{get_dir};
4944 if ($r == $cache->{r}) {
4945 if (my $x = $cache->{data}->{$dir}) {
4946 return wantarray ? @$x : $x->[0];
4947 }
4948 }
4949 my $pool = SVN::Pool->new;
4950 my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
4951 my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
4952 $pool->clear;
4953 if ($r != $cache->{r}) {
4954 %{$cache->{data}} = ();
4955 $cache->{r} = $r;
4956 }
4957 $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
4958 wantarray ? (\%dirents, $r, $props) : \%dirents;
4959}
4960sub DESTROY {
4962 # do not call the real DESTROY since we store ourselves in $RA
4963}
4964# get_log(paths, start, end, limit,
4966# discover_changed_paths, strict_node_history, receiver)
4967sub get_log {
4968 my ($self, @args) = @_;
4969 my $pool = SVN::Pool->new;
4970 # svn_log_changed_path_t objects passed to get_log are likely to be
4972 # overwritten even if only the refs are copied to an external variable,
4973 # so we should dup the structures in their entirety. Using an
4974 # externally passed pool (instead of our temporary and quickly cleared
4975 # pool in Git::SVN::Ra) does not help matters at all...
4976 my $receiver = pop @args;
4977 my $prefix = "/".$self->{svn_path};
4978 $prefix =~ s#/+($)##;
4979 my $prefix_regex = qr#^\Q$prefix\E#;
4980 push(@args, sub {
4981 my ($paths) = $_[0];
4982 return &$receiver(@_) unless $paths;
4983 $_[0] = ();
4984 foreach my $p (keys %$paths) {
4985 my $i = $paths->{$p};
4986 # Make path relative to our url, not repos_root
4987 $p =~ s/$prefix_regex//;
4988 my %s = map { $_ => $i->$_; }
4989 qw/copyfrom_path copyfrom_rev action/;
4990 if ($s{'copyfrom_path'}) {
4991 $s{'copyfrom_path'} =~ s/$prefix_regex//;
4992 }
4993 $_[0]{$p} = \%s;
4994 }
4995 &$receiver(@_);
4996 });
4997 # the limit parameter was not supported in SVN 1.1.x, so we
5000 # drop it. Therefore, the receiver callback passed to it
5001 # is made aware of this limitation by being wrapped if
5002 # the limit passed to is being wrapped.
5003 if ($SVN::Core::VERSION le '1.2.0') {
5004 my $limit = splice(@args, 3, 1);
5005 if ($limit > 0) {
5006 my $receiver = pop @args;
5007 push(@args, sub { &$receiver(@_) if (--$limit >= 0) });
5008 }
5009 }
5010 my $ret = $self->SUPER::get_log(@args, $pool);
5011 $pool->clear;
5012 $ret;
5013}
5014sub trees_match {
5016 my ($self, $url1, $rev1, $url2, $rev2) = @_;
5017 my $ctx = SVN::Client->new(auth => _auth_providers);
5018 my $out = IO::File->new_tmpfile;
5019 # older SVN (1.1.x) doesn't take $pool as the last parameter for
5021 # $ctx->diff(), so we'll create a default one
5022 my $pool = SVN::Pool->new_default_sub;
5023 $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
5025 $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
5026 $out->flush;
5027 my $ret = (($out->stat)[7] == 0);
5028 close $out or croak $!;
5029 $ret;
5031}
5032sub get_commit_editor {
5034 my ($self, $log, $cb, $pool) = @_;
5035 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
5036 $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
5037}
5038sub gs_do_update {
5040 my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
5041 my $new = ($rev_a == $rev_b);
5042 my $path = $gs->{path};
5043 if ($new && -e $gs->{index}) {
5045 unlink $gs->{index} or die
5046 "Couldn't unlink index: $gs->{index}: $!\n";
5047 }
5048 my $pool = SVN::Pool->new;
5049 $editor->set_path_strip($path);
5050 my (@pc) = split m#/#, $path;
5051 my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
5052 1, $editor, $pool);
5053 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
5054 # Since we can't rely on svn_ra_reparent being available, we'll
5056 # just have to do some magic with set_path to make it so
5057 # we only want a partial path.
5058 my $sp = '';
5059 my $final = join('/', @pc);
5060 while (@pc) {
5061 $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
5062 $sp .= '/' if length $sp;
5063 $sp .= shift @pc;
5064 }
5065 die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
5066 $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
5068 $reporter->finish_report($pool);
5070 $pool->clear;
5071 $editor->{git_commit_ok};
5072}
5073# this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
5075# svn_ra_reparent didn't work before 1.4)
5076sub gs_do_switch {
5077 my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
5078 my $path = $gs->{path};
5079 my $pool = SVN::Pool->new;
5080 my $full_url = $self->{url};
5082 my $old_url = $full_url;
5083 $full_url .= '/' . $path if length $path;
5084 my ($ra, $reparented);
5085 if ($old_url =~ m#^svn(\+ssh)?://# ||
5087 ($full_url =~ m#^https?://# &&
5088 escape_url($full_url) ne $full_url)) {
5089 $_[0] = undef;
5090 $self = undef;
5091 $RA = undef;
5092 $ra = Git::SVN::Ra->new($full_url);
5093 $ra_invalid = 1;
5094 } elsif ($old_url ne $full_url) {
5095 SVN::_Ra::svn_ra_reparent($self->{session}, $full_url, $pool);
5096 $self->{url} = $full_url;
5097 $reparented = 1;
5098 }
5099 $ra ||= $self;
5101 $url_b = escape_url($url_b);
5102 my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
5103 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
5104 $reporter->set_path('', $rev_a, 0, @lock, $pool);
5105 $reporter->finish_report($pool);
5106 if ($reparented) {
5108 SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
5109 $self->{url} = $old_url;
5110 }
5111 $pool->clear;
5113 $editor->{git_commit_ok};
5114}
5115sub longest_common_path {
5117 my ($gsv, $globs) = @_;
5118 my %common;
5119 my $common_max = scalar @$gsv;
5120 foreach my $gs (@$gsv) {
5122 my @tmp = split m#/#, $gs->{path};
5123 my $p = '';
5124 foreach (@tmp) {
5125 $p .= length($p) ? "/$_" : $_;
5126 $common{$p} ||= 0;
5127 $common{$p}++;
5128 }
5129 }
5130 $globs ||= [];
5131 $common_max += scalar @$globs;
5132 foreach my $glob (@$globs) {
5133 my @tmp = split m#/#, $glob->{path}->{left};
5134 my $p = '';
5135 foreach (@tmp) {
5136 $p .= length($p) ? "/$_" : $_;
5137 $common{$p} ||= 0;
5138 $common{$p}++;
5139 }
5140 }
5141 my $longest_path = '';
5143 foreach (sort {length $b <=> length $a} keys %common) {
5144 if ($common{$_} == $common_max) {
5145 $longest_path = $_;
5146 last;
5147 }
5148 }
5149 $longest_path;
5150}
5151sub gs_fetch_loop_common {
5153 my ($self, $base, $head, $gsv, $globs) = @_;
5154 return if ($base > $head);
5155 my $inc = $_log_window_size;
5156 my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
5157 my $longest_path = longest_common_path($gsv, $globs);
5158 my $ra_url = $self->{url};
5159 my $find_trailing_edge;
5160 while (1) {
5161 my %revs;
5162 my $err;
5163 my $err_handler = $SVN::Error::handler;
5164 $SVN::Error::handler = sub {
5165 ($err) = @_;
5166 skip_unknown_revs($err);
5167 };
5168 sub _cb {
5169 my ($paths, $r, $author, $date, $log) = @_;
5170 [ $paths,
5171 { author => $author, date => $date, log => $log } ];
5172 }
5173 $self->get_log([$longest_path], $min, $max, 0, 1, 1,
5174 sub { $revs{$_[1]} = _cb(@_) });
5175 if ($err) {
5176 print "Checked through r$max\r";
5177 } else {
5178 $find_trailing_edge = 1;
5179 }
5180 if ($err and $find_trailing_edge) {
5181 print STDERR "Path '$longest_path' ",
5182 "was probably deleted:\n",
5183 $err->expanded_message,
5184 "\nWill attempt to follow ",
5185 "revisions r$min .. r$max ",
5186 "committed before the deletion\n";
5187 my $hi = $max;
5188 while (--$hi >= $min) {
5189 my $ok;
5190 $self->get_log([$longest_path], $min, $hi,
5191 0, 1, 1, sub {
5192 $ok = $_[1];
5193 $revs{$_[1]} = _cb(@_) });
5194 if ($ok) {
5195 print STDERR "r$min .. r$ok OK\n";
5196 last;
5197 }
5198 }
5199 $find_trailing_edge = 0;
5200 }
5201 $SVN::Error::handler = $err_handler;
5202 my %exists = map { $_->{path} => $_ } @$gsv;
5204 foreach my $r (sort {$a <=> $b} keys %revs) {
5205 my ($paths, $logged) = @{$revs{$r}};
5206 foreach my $gs ($self->match_globs(\%exists, $paths,
5208 $globs, $r)) {
5209 if ($gs->rev_map_max >= $r) {
5210 next;
5211 }
5212 next unless $gs->match_paths($paths, $r);
5213 $gs->{logged_rev_props} = $logged;
5214 if (my $last_commit = $gs->last_commit) {
5215 $gs->assert_index_clean($last_commit);
5216 }
5217 my $log_entry = $gs->do_fetch($paths, $r);
5218 if ($log_entry) {
5219 $gs->do_git_commit($log_entry);
5220 }
5221 $INDEX_FILES{$gs->{index}} = 1;
5222 }
5223 foreach my $g (@$globs) {
5224 my $k = "svn-remote.$g->{remote}." .
5225 "$g->{t}-maxRev";
5226 Git::SVN::tmp_config($k, $r);
5227 }
5228 if ($ra_invalid) {
5229 $_[0] = undef;
5230 $self = undef;
5231 $RA = undef;
5232 $self = Git::SVN::Ra->new($ra_url);
5233 $ra_invalid = undef;
5234 }
5235 }
5236 # pre-fill the .rev_db since it'll eventually get filled in
5237 # with '0' x40 if something new gets committed
5238 foreach my $gs (@$gsv) {
5239 next if $gs->rev_map_max >= $max;
5240 next if defined $gs->rev_map_get($max);
5241 $gs->rev_map_set($max, 0 x40);
5242 }
5243 foreach my $g (@$globs) {
5244 my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
5245 Git::SVN::tmp_config($k, $max);
5246 }
5247 last if $max >= $head;
5248 $min = $max + 1;
5249 $max += $inc;
5250 $max = $head if ($max > $head);
5251 }
5252 Git::SVN::gc();
5253}
5254sub get_dir_globbed {
5256 my ($self, $left, $depth, $r) = @_;
5257 my @x = eval { $self->get_dir($left, $r) };
5259 return unless scalar @x == 3;
5260 my $dirents = $x[0];
5261 my @finalents;
5262 foreach my $de (keys %$dirents) {
5263 next if $dirents->{$de}->{kind} != $SVN::Node::dir;
5264 if ($depth > 1) {
5265 my @args = ("$left/$de", $depth - 1, $r);
5266 foreach my $dir ($self->get_dir_globbed(@args)) {
5267 push @finalents, "$de/$dir";
5268 }
5269 } else {
5270 push @finalents, $de;
5271 }
5272 }
5273 @finalents;
5274}
5275sub match_globs {
5277 my ($self, $exists, $paths, $globs, $r) = @_;
5278 sub get_dir_check {
5280 my ($self, $exists, $g, $r) = @_;
5281 my @dirs = $self->get_dir_globbed($g->{path}->{left},
5283 $g->{path}->{depth},
5284 $r);
5285 foreach my $de (@dirs) {
5287 my $p = $g->{path}->full_path($de);
5288 next if $exists->{$p};
5289 next if (length $g->{path}->{right} &&
5290 ($self->check_path($p, $r) !=
5291 $SVN::Node::dir));
5292 next unless $p =~ /$g->{path}->{regex}/;
5293 $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
5294 $g->{ref}->full_path($de), 1);
5295 }
5296 }
5297 foreach my $g (@$globs) {
5298 if (my $path = $paths->{"/$g->{path}->{left}"}) {
5299 if ($path->{action} =~ /^[AR]$/) {
5300 get_dir_check($self, $exists, $g, $r);
5301 }
5302 }
5303 foreach (keys %$paths) {
5304 if (/$g->{path}->{left_regex}/ &&
5305 !/$g->{path}->{regex}/) {
5306 next if $paths->{$_}->{action} !~ /^[AR]$/;
5307 get_dir_check($self, $exists, $g, $r);
5308 }
5309 next unless /$g->{path}->{regex}/;
5310 my $p = $1;
5311 my $pathname = $g->{path}->full_path($p);
5312 next if $exists->{$pathname};
5313 next if ($self->check_path($pathname, $r) !=
5314 $SVN::Node::dir);
5315 $exists->{$pathname} = Git::SVN->init(
5316 $self->{url}, $pathname, undef,
5317 $g->{ref}->full_path($p), 1);
5318 }
5319 my $c = '';
5320 foreach (split m#/#, $g->{path}->{left}) {
5321 $c .= "/$_";
5322 next unless ($paths->{$c} &&
5323 ($paths->{$c}->{action} =~ /^[AR]$/));
5324 get_dir_check($self, $exists, $g, $r);
5325 }
5326 }
5327 values %$exists;
5328}
5329sub minimize_url {
5331 my ($self) = @_;
5332 return $self->{url} if ($self->{url} eq $self->{repos_root});
5333 my $url = $self->{repos_root};
5334 my @components = split(m!/!, $self->{svn_path});
5335 my $c = '';
5336 do {
5337 $url .= "/$c" if length $c;
5338 eval {
5339 my $ra = (ref $self)->new($url);
5340 my $latest = $ra->get_latest_revnum;
5341 $ra->get_log("", $latest, 0, 1, 0, 1, sub {});
5342 };
5343 } while ($@ && ($c = shift @components));
5344 $url;
5345}
5346sub can_do_switch {
5348 my $self = shift;
5349 unless (defined $can_do_switch) {
5350 my $pool = SVN::Pool->new;
5351 my $rep = eval {
5352 $self->do_switch(1, '', 0, $self->{url},
5353 SVN::Delta::Editor->new, $pool);
5354 };
5355 if ($@) {
5356 $can_do_switch = 0;
5357 } else {
5358 $rep->abort_report($pool);
5359 $can_do_switch = 1;
5360 }
5361 $pool->clear;
5362 }
5363 $can_do_switch;
5364}
5365sub skip_unknown_revs {
5367 my ($err) = @_;
5368 my $errno = $err->apr_err();
5369 # Maybe the branch we're tracking didn't
5370 # exist when the repo started, so it's
5371 # not an error if it doesn't, just continue
5372 #
5373 # Wonderfully consistent library, eh?
5374 # 160013 - svn:// and file://
5375 # 175002 - http(s)://
5376 # 175007 - http(s):// (this repo required authorization, too...)
5377 # More codes may be discovered later...
5378 if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
5379 my $err_key = $err->expanded_message;
5380 # revision numbers change every time, filter them out
5381 $err_key =~ s/\d+/\0/g;
5382 $err_key = "$errno\0$err_key";
5383 unless ($ignored_err{$err_key}) {
5384 warn "W: Ignoring error from SVN, path probably ",
5385 "does not exist: ($errno): ",
5386 $err->expanded_message,"\n";
5387 warn "W: Do not be alarmed at the above message ",
5388 "git-svn is just searching aggressively for ",
5389 "old history.\n",
5390 "This may take a while on large repositories\n";
5391 $ignored_err{$err_key} = 1;
5392 }
5393 return;
5394 }
5395 die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
5396}
5397package Git::SVN::Log;
5399use strict;
5400use warnings;
5401use POSIX qw/strftime/;
5402use Time::Local;
5403use constant commit_log_separator => ('-' x 72) . "\n";
5404use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
5405 %rusers $show_commit $incremental/;
5406my $l_fmt;
5407sub cmt_showable {
5409 my ($c) = @_;
5410 return 1 if defined $c->{r};
5411 # big commit message got truncated by the 16k pretty buffer in rev-list
5413 if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
5414 $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
5415 @{$c->{l}} = ();
5416 my @log = command(qw/cat-file commit/, $c->{c});
5417 # shift off the headers
5419 shift @log while ($log[0] ne '');
5420 shift @log;
5421 # TODO: make $c->{l} not have a trailing newline in the future
5423 @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
5424 (undef, $c->{r}, undef) = ::extract_metadata(
5426 (grep(/^git-svn-id: /, @log))[-1]);
5427 }
5428 return defined $c->{r};
5429}
5430sub log_use_color {
5432 return $color || Git->repository->get_colorbool('color.diff');
5433}
5434sub git_svn_log_cmd {
5436 my ($r_min, $r_max, @args) = @_;
5437 my $head = 'HEAD';
5438 my (@files, @log_opts);
5439 foreach my $x (@args) {
5440 if ($x eq '--' || @files) {
5441 push @files, $x;
5442 } else {
5443 if (::verify_ref("$x^0")) {
5444 $head = $x;
5445 } else {
5446 push @log_opts, $x;
5447 }
5448 }
5449 }
5450 my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
5452 $gs ||= Git::SVN->_new;
5453 my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
5454 $gs->refname);
5455 push @cmd, '-r' unless $non_recursive;
5456 push @cmd, qw/--raw --name-status/ if $verbose;
5457 push @cmd, '--color' if log_use_color();
5458 push @cmd, @log_opts;
5459 if (defined $r_max && $r_max == $r_min) {
5460 push @cmd, '--max-count=1';
5461 if (my $c = $gs->rev_map_get($r_max)) {
5462 push @cmd, $c;
5463 }
5464 } elsif (defined $r_max) {
5465 if ($r_max < $r_min) {
5466 ($r_min, $r_max) = ($r_max, $r_min);
5467 }
5468 my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
5469 my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
5470 # If there are no commits in the range, both $c_max and $c_min
5471 # will be undefined. If there is at least 1 commit in the
5472 # range, both will be defined.
5473 return () if !defined $c_min || !defined $c_max;
5474 if ($c_min eq $c_max) {
5475 push @cmd, '--max-count=1', $c_min;
5476 } else {
5477 push @cmd, '--boundary', "$c_min..$c_max";
5478 }
5479 }
5480 return (@cmd, @files);
5481}
5482# adapted from pager.c
5484sub config_pager {
5485 if (! -t *STDOUT) {
5486 $ENV{GIT_PAGER_IN_USE} = 'false';
5487 $pager = undef;
5488 return;
5489 }
5490 chomp($pager = command_oneline(qw(var GIT_PAGER)));
5491 if ($pager eq 'cat') {
5492 $pager = undef;
5493 }
5494 $ENV{GIT_PAGER_IN_USE} = defined($pager);
5495}
5496sub run_pager {
5498 return unless defined $pager;
5499 pipe my ($rfd, $wfd) or return;
5500 defined(my $pid = fork) or ::fatal "Can't fork: $!";
5501 if (!$pid) {
5502 open STDOUT, '>&', $wfd or
5503 ::fatal "Can't redirect to stdout: $!";
5504 return;
5505 }
5506 open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
5507 $ENV{LESS} ||= 'FRSX';
5508 exec $pager or ::fatal "Can't run pager: $! ($pager)";
5509}
5510sub format_svn_date {
5512 # some systmes don't handle or mishandle %z, so be creative.
5513 my $t = shift || time;
5514 my $gm = timelocal(gmtime($t));
5515 my $sign = qw( + + - )[ $t <=> $gm ];
5516 my $gmoff = sprintf("%s%02d%02d", $sign, (gmtime(abs($t - $gm)))[2,1]);
5517 return strftime("%Y-%m-%d %H:%M:%S $gmoff (%a, %d %b %Y)", localtime($t));
5518}
5519sub parse_git_date {
5521 my ($t, $tz) = @_;
5522 # Date::Parse isn't in the standard Perl distro :(
5523 if ($tz =~ s/^\+//) {
5524 $t += tz_to_s_offset($tz);
5525 } elsif ($tz =~ s/^\-//) {
5526 $t -= tz_to_s_offset($tz);
5527 }
5528 return $t;
5529}
5530sub set_local_timezone {
5532 if (defined $TZ) {
5533 $ENV{TZ} = $TZ;
5534 } else {
5535 delete $ENV{TZ};
5536 }
5537}
5538sub tz_to_s_offset {
5540 my ($tz) = @_;
5541 $tz =~ s/(\d\d)$//;
5542 return ($1 * 60) + ($tz * 3600);
5543}
5544sub get_author_info {
5546 my ($dest, $author, $t, $tz) = @_;
5547 $author =~ s/(?:^\s*|\s*$)//g;
5548 $dest->{a_raw} = $author;
5549 my $au;
5550 if ($::_authors) {
5551 $au = $rusers{$author} || undef;
5552 }
5553 if (!$au) {
5554 ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
5555 }
5556 $dest->{t} = $t;
5557 $dest->{tz} = $tz;
5558 $dest->{a} = $au;
5559 $dest->{t_utc} = parse_git_date($t, $tz);
5560}
5561sub process_commit {
5563 my ($c, $r_min, $r_max, $defer) = @_;
5564 if (defined $r_min && defined $r_max) {
5565 if ($r_min == $c->{r} && $r_min == $r_max) {
5566 show_commit($c);
5567 return 0;
5568 }
5569 return 1 if $r_min == $r_max;
5570 if ($r_min < $r_max) {
5571 # we need to reverse the print order
5572 return 0 if (defined $limit && --$limit < 0);
5573 push @$defer, $c;
5574 return 1;
5575 }
5576 if ($r_min != $r_max) {
5577 return 1 if ($r_min < $c->{r});
5578 return 1 if ($r_max > $c->{r});
5579 }
5580 }
5581 return 0 if (defined $limit && --$limit < 0);
5582 show_commit($c);
5583 return 1;
5584}
5585sub show_commit {
5587 my $c = shift;
5588 if ($oneline) {
5589 my $x = "\n";
5590 if (my $l = $c->{l}) {
5591 while ($l->[0] =~ /^\s*$/) { shift @$l }
5592 $x = $l->[0];
5593 }
5594 $l_fmt ||= 'A' . length($c->{r});
5595 print 'r',pack($l_fmt, $c->{r}),' | ';
5596 print "$c->{c} | " if $show_commit;
5597 print $x;
5598 } else {
5599 show_commit_normal($c);
5600 }
5601}
5602sub show_commit_changed_paths {
5604 my ($c) = @_;
5605 return unless $c->{changed};
5606 print "Changed paths:\n", @{$c->{changed}};
5607}
5608sub show_commit_normal {
5610 my ($c) = @_;
5611 print commit_log_separator, "r$c->{r} | ";
5612 print "$c->{c} | " if $show_commit;
5613 print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
5614 my $nr_line = 0;
5615 if (my $l = $c->{l}) {
5617 while ($l->[$#$l] eq "\n" && $#$l > 0
5618 && $l->[($#$l - 1)] eq "\n") {
5619 pop @$l;
5620 }
5621 $nr_line = scalar @$l;
5622 if (!$nr_line) {
5623 print "1 line\n\n\n";
5624 } else {
5625 if ($nr_line == 1) {
5626 $nr_line = '1 line';
5627 } else {
5628 $nr_line .= ' lines';
5629 }
5630 print $nr_line, "\n";
5631 show_commit_changed_paths($c);
5632 print "\n";
5633 print $_ foreach @$l;
5634 }
5635 } else {
5636 print "1 line\n";
5637 show_commit_changed_paths($c);
5638 print "\n";
5639 }
5641 foreach my $x (qw/raw stat diff/) {
5642 if ($c->{$x}) {
5643 print "\n";
5644 print $_ foreach @{$c->{$x}}
5645 }
5646 }
5647}
5648sub cmd_show_log {
5650 my (@args) = @_;
5651 my ($r_min, $r_max);
5652 my $r_last = -1; # prevent dupes
5653 set_local_timezone();
5654 if (defined $::_revision) {
5655 if ($::_revision =~ /^(\d+):(\d+)$/) {
5656 ($r_min, $r_max) = ($1, $2);
5657 } elsif ($::_revision =~ /^\d+$/) {
5658 $r_min = $r_max = $::_revision;
5659 } else {
5660 ::fatal "-r$::_revision is not supported, use ",
5661 "standard 'git log' arguments instead";
5662 }
5663 }
5664 config_pager();
5666 @args = git_svn_log_cmd($r_min, $r_max, @args);
5667 if (!@args) {
5668 print commit_log_separator unless $incremental || $oneline;
5669 return;
5670 }
5671 my $log = command_output_pipe(@args);
5672 run_pager();
5673 my (@k, $c, $d, $stat);
5674 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
5675 while (<$log>) {
5676 if (/^${esc_color}commit -?($::sha1_short)/o) {
5677 my $cmt = $1;
5678 if ($c && cmt_showable($c) && $c->{r} != $r_last) {
5679 $r_last = $c->{r};
5680 process_commit($c, $r_min, $r_max, \@k) or
5681 goto out;
5682 }
5683 $d = undef;
5684 $c = { c => $cmt };
5685 } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
5686 get_author_info($c, $1, $2, $3);
5687 } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
5688 # ignore
5689 } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
5690 push @{$c->{raw}}, $_;
5691 } elsif (/^${esc_color}[ACRMDT]\t/) {
5692 # we could add $SVN->{svn_path} here, but that requires
5693 # remote access at the moment (repo_path_split)...
5694 s#^(${esc_color})([ACRMDT])\t#$1 $2 #o;
5695 push @{$c->{changed}}, $_;
5696 } elsif (/^${esc_color}diff /o) {
5697 $d = 1;
5698 push @{$c->{diff}}, $_;
5699 } elsif ($d) {
5700 push @{$c->{diff}}, $_;
5701 } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
5702 $esc_color*[\+\-]*$esc_color$/x) {
5703 $stat = 1;
5704 push @{$c->{stat}}, $_;
5705 } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
5706 push @{$c->{stat}}, $_;
5707 $stat = undef;
5708 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
5709 ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
5710 } elsif (s/^${esc_color} //o) {
5711 push @{$c->{l}}, $_;
5712 }
5713 }
5714 if ($c && defined $c->{r} && $c->{r} != $r_last) {
5715 $r_last = $c->{r};
5716 process_commit($c, $r_min, $r_max, \@k);
5717 }
5718 if (@k) {
5719 ($r_min, $r_max) = ($r_max, $r_min);
5720 process_commit($_, $r_min, $r_max) foreach reverse @k;
5721 }
5722out:
5723 close $log;
5724 print commit_log_separator unless $incremental || $oneline;
5725}
5726sub cmd_blame {
5728 my $path = pop;
5729 config_pager();
5731 run_pager();
5732 my ($fh, $ctx, $rev);
5734 if ($_git_format) {
5736 ($fh, $ctx) = command_output_pipe('blame', @_, $path);
5737 while (my $line = <$fh>) {
5738 if ($line =~ /^\^?([[:xdigit:]]+)\s/) {
5739 # Uncommitted edits show up as a rev ID of
5740 # all zeros, which we can't look up with
5741 # cmt_metadata
5742 if ($1 !~ /^0+$/) {
5743 (undef, $rev, undef) =
5744 ::cmt_metadata($1);
5745 $rev = '0' if (!$rev);
5746 } else {
5747 $rev = '0';
5748 }
5749 $rev = sprintf('%-10s', $rev);
5750 $line =~ s/^\^?[[:xdigit:]]+(\s)/$rev$1/;
5751 }
5752 print $line;
5753 }
5754 } else {
5755 ($fh, $ctx) = command_output_pipe('blame', '-p', @_, 'HEAD',
5756 '--', $path);
5757 my ($sha1);
5758 my %authors;
5759 my @buffer;
5760 my %dsha; #distinct sha keys
5761 while (my $line = <$fh>) {
5763 push @buffer, $line;
5764 if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
5765 $dsha{$1} = 1;
5766 }
5767 }
5768 my $s2r = ::cmt_sha2rev_batch([keys %dsha]);
5770 foreach my $line (@buffer) {
5772 if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
5773 $rev = $s2r->{$1};
5774 $rev = '0' if (!$rev)
5775 }
5776 elsif ($line =~ /^author (.*)/) {
5777 $authors{$rev} = $1;
5778 $authors{$rev} =~ s/\s/_/g;
5779 }
5780 elsif ($line =~ /^\t(.*)$/) {
5781 printf("%6s %10s %s\n", $rev, $authors{$rev}, $1);
5782 }
5783 }
5784 }
5785 command_close_pipe($fh, $ctx);
5786}
5787package Git::SVN::Migration;
5789# these version numbers do NOT correspond to actual version numbers
5790# of git nor git-svn. They are just relative.
5791#
5792# v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
5793#
5794# v1 layout: .git/$id/info/url, refs/remotes/$id
5795#
5796# v2 layout: .git/svn/$id/info/url, refs/remotes/$id
5797#
5798# v3 layout: .git/svn/$id, refs/remotes/$id
5799# - info/url may remain for backwards compatibility
5800# - this is what we migrate up to this layout automatically,
5801# - this will be used by git svn init on single branches
5802# v3.1 layout (auto migrated):
5803# - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
5804# for backwards compatibility
5805#
5806# v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
5807# - this is only created for newly multi-init-ed
5808# repositories. Similar in spirit to the
5809# --use-separate-remotes option in git-clone (now default)
5810# - we do not automatically migrate to this (following
5811# the example set by core git)
5812#
5813# v5 layout: .rev_db.$UUID => .rev_map.$UUID
5814# - newer, more-efficient format that uses 24-bytes per record
5815# with no filler space.
5816# - use xxd -c24 < .rev_map.$UUID to view and debug
5817# - This is a one-way migration, repositories updated to the
5818# new format will not be able to use old git-svn without
5819# rebuilding the .rev_db. Rebuilding the rev_db is not
5820# possible if noMetadata or useSvmProps are set; but should
5821# be no problem for users that use the (sensible) defaults.
5822use strict;
5823use warnings;
5824use Carp qw/croak/;
5825use File::Path qw/mkpath/;
5826use File::Basename qw/dirname basename/;
5827use vars qw/$_minimize/;
5828sub migrate_from_v0 {
5830 my $git_dir = $ENV{GIT_DIR};
5831 return undef unless -d $git_dir;
5832 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
5833 my $migrated = 0;
5834 while (<$fh>) {
5835 chomp;
5836 my ($id, $orig_ref) = ($_, $_);
5837 next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
5838 next unless -f "$git_dir/$id/info/url";
5839 my $new_ref = "refs/remotes/$id";
5840 if (::verify_ref("$new_ref^0")) {
5841 print STDERR "W: $orig_ref is probably an old ",
5842 "branch used by an ancient version of ",
5843 "git-svn.\n",
5844 "However, $new_ref also exists.\n",
5845 "We will not be able ",
5846 "to use this branch until this ",
5847 "ambiguity is resolved.\n";
5848 next;
5849 }
5850 print STDERR "Migrating from v0 layout...\n" if !$migrated;
5851 print STDERR "Renaming ref: $orig_ref => $new_ref\n";
5852 command_noisy('update-ref', $new_ref, $orig_ref);
5853 command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
5854 $migrated++;
5855 }
5856 command_close_pipe($fh, $ctx);
5857 print STDERR "Done migrating from v0 layout...\n" if $migrated;
5858 $migrated;
5859}
5860sub migrate_from_v1 {
5862 my $git_dir = $ENV{GIT_DIR};
5863 my $migrated = 0;
5864 return $migrated unless -d $git_dir;
5865 my $svn_dir = "$git_dir/svn";
5866 # just in case somebody used 'svn' as their $id at some point...
5868 return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
5869 print STDERR "Migrating from a git-svn v1 layout...\n";
5871 mkpath([$svn_dir]);
5872 print STDERR "Data from a previous version of git-svn exists, but\n\t",
5873 "$svn_dir\n\t(required for this version ",
5874 "($::VERSION) of git-svn) does not exist.\n";
5875 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
5876 while (<$fh>) {
5877 my $x = $_;
5878 next unless $x =~ s#^refs/remotes/##;
5879 chomp $x;
5880 next unless -f "$git_dir/$x/info/url";
5881 my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
5882 next unless $u;
5883 my $dn = dirname("$git_dir/svn/$x");
5884 mkpath([$dn]) unless -d $dn;
5885 if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
5886 mkpath(["$git_dir/svn/svn"]);
5887 print STDERR " - $git_dir/$x/info => ",
5888 "$git_dir/svn/$x/info\n";
5889 rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
5890 croak "$!: $x";
5891 # don't worry too much about these, they probably
5892 # don't exist with repos this old (save for index,
5893 # and we can easily regenerate that)
5894 foreach my $f (qw/unhandled.log index .rev_db/) {
5895 rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
5896 }
5897 } else {
5898 print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
5899 rename "$git_dir/$x", "$git_dir/svn/$x" or
5900 croak "$!: $x";
5901 }
5902 $migrated++;
5903 }
5904 command_close_pipe($fh, $ctx);
5905 print STDERR "Done migrating from a git-svn v1 layout\n";
5906 $migrated;
5907}
5908sub read_old_urls {
5910 my ($l_map, $pfx, $path) = @_;
5911 my @dir;
5912 foreach (<$path/*>) {
5913 if (-r "$_/info/url") {
5914 $pfx .= '/' if $pfx && $pfx !~ m!/$!;
5915 my $ref_id = $pfx . basename $_;
5916 my $url = ::file_to_s("$_/info/url");
5917 $l_map->{$ref_id} = $url;
5918 } elsif (-d $_) {
5919 push @dir, $_;
5920 }
5921 }
5922 foreach (@dir) {
5923 my $x = $_;
5924 $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
5925 read_old_urls($l_map, $x, $_);
5926 }
5927}
5928sub migrate_from_v2 {
5930 my @cfg = command(qw/config -l/);
5931 return if grep /^svn-remote\..+\.url=/, @cfg;
5932 my %l_map;
5933 read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
5934 my $migrated = 0;
5935 foreach my $ref_id (sort keys %l_map) {
5937 eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
5938 if ($@) {
5939 Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
5940 }
5941 $migrated++;
5942 }
5943 $migrated;
5944}
5945sub minimize_connections {
5947 my $r = Git::SVN::read_all_remotes();
5948 my $new_urls = {};
5949 my $root_repos = {};
5950 foreach my $repo_id (keys %$r) {
5951 my $url = $r->{$repo_id}->{url} or next;
5952 my $fetch = $r->{$repo_id}->{fetch} or next;
5953 my $ra = Git::SVN::Ra->new($url);
5954 # skip existing cases where we already connect to the root
5956 if (($ra->{url} eq $ra->{repos_root}) ||
5957 ($ra->{repos_root} eq $repo_id)) {
5958 $root_repos->{$ra->{url}} = $repo_id;
5959 next;
5960 }
5961 my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
5963 my $root_path = $ra->{url};
5964 $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
5965 foreach my $path (keys %$fetch) {
5966 my $ref_id = $fetch->{$path};
5967 my $gs = Git::SVN->new($ref_id, $repo_id, $path);
5968 # make sure we can read when connecting to
5970 # a higher level of a repository
5971 my ($last_rev, undef) = $gs->last_rev_commit;
5972 if (!defined $last_rev) {
5973 $last_rev = eval {
5974 $root_ra->get_latest_revnum;
5975 };
5976 next if $@;
5977 }
5978 my $new = $root_path;
5979 $new .= length $path ? "/$path" : '';
5980 eval {
5981 $root_ra->get_log([$new], $last_rev, $last_rev,
5982 0, 0, 1, sub { });
5983 };
5984 next if $@;
5985 $new_urls->{$ra->{repos_root}}->{$new} =
5986 { ref_id => $ref_id,
5987 old_repo_id => $repo_id,
5988 old_path => $path };
5989 }
5990 }
5991 my @emptied;
5993 foreach my $url (keys %$new_urls) {
5994 # see if we can re-use an existing [svn-remote "repo_id"]
5995 # instead of creating a(n ugly) new section:
5996 my $repo_id = $root_repos->{$url} || $url;
5997 my $fetch = $new_urls->{$url};
5999 foreach my $path (keys %$fetch) {
6000 my $x = $fetch->{$path};
6001 Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
6002 my $pfx = "svn-remote.$x->{old_repo_id}";
6003 my $old_fetch = quotemeta("$x->{old_path}:".
6005 "$x->{ref_id}");
6006 command_noisy(qw/config --unset/,
6007 "$pfx.fetch", '^'. $old_fetch . '$');
6008 delete $r->{$x->{old_repo_id}}->
6009 {fetch}->{$x->{old_path}};
6010 if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
6011 command_noisy(qw/config --unset/,
6012 "$pfx.url");
6013 push @emptied, $x->{old_repo_id}
6014 }
6015 }
6016 }
6017 if (@emptied) {
6018 my $file = $ENV{GIT_CONFIG} || "$ENV{GIT_DIR}/config";
6019 print STDERR <<EOF;
6020The following [svn-remote] sections in your config file ($file) are empty
6021and can be safely removed:
6022EOF
6023 print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
6024 }
6025}
6026sub migration_check {
6028 migrate_from_v0();
6029 migrate_from_v1();
6030 migrate_from_v2();
6031 minimize_connections() if $_minimize;
6032}
6033package Git::IndexInfo;
6035use strict;
6036use warnings;
6037use Git qw/command_input_pipe command_close_pipe/;
6038sub new {
6040 my ($class) = @_;
6041 my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
6042 bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
6043}
6044sub remove {
6046 my ($self, $path) = @_;
6047 if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
6048 return ++$self->{nr};
6049 }
6050 undef;
6051}
6052sub update {
6054 my ($self, $mode, $hash, $path) = @_;
6055 if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
6056 return ++$self->{nr};
6057 }
6058 undef;
6059}
6060sub DESTROY {
6062 my ($self) = @_;
6063 command_close_pipe($self->{gui}, $self->{ctx});
6064}
6065package Git::SVN::GlobSpec;
6067use strict;
6068use warnings;
6069sub new {
6071 my ($class, $glob, $pattern_ok) = @_;
6072 my $re = $glob;
6073 $re =~ s!/+$!!g; # no need for trailing slashes
6074 my (@left, @right, @patterns);
6075 my $state = "left";
6076 my $die_msg = "Only one set of wildcard directories " .
6077 "(e.g. '*' or '*/*/*') is supported: '$glob'\n";
6078 for my $part (split(m|/|, $glob)) {
6079 if ($part =~ /\*/ && $part ne "*") {
6080 die "Invalid pattern in '$glob': $part\n";
6081 } elsif ($pattern_ok && $part =~ /[{}]/ &&
6082 $part !~ /^\{[^{}]+\}/) {
6083 die "Invalid pattern in '$glob': $part\n";
6084 }
6085 if ($part eq "*") {
6086 die $die_msg if $state eq "right";
6087 $state = "pattern";
6088 push(@patterns, "[^/]*");
6089 } elsif ($pattern_ok && $part =~ /^\{(.*)\}$/) {
6090 die $die_msg if $state eq "right";
6091 $state = "pattern";
6092 my $p = quotemeta($1);
6093 $p =~ s/\\,/|/g;
6094 push(@patterns, "(?:$p)");
6095 } else {
6096 if ($state eq "left") {
6097 push(@left, $part);
6098 } else {
6099 push(@right, $part);
6100 $state = "right";
6101 }
6102 }
6103 }
6104 my $depth = @patterns;
6105 if ($depth == 0) {
6106 die "One '*' is needed in glob: '$glob'\n";
6107 }
6108 my $left = join('/', @left);
6109 my $right = join('/', @right);
6110 $re = join('/', @patterns);
6111 $re = join('\/',
6112 grep(length, quotemeta($left), "($re)", quotemeta($right)));
6113 my $left_re = qr/^\/\Q$left\E(\/|$)/;
6114 bless { left => $left, right => $right, left_regex => $left_re,
6115 regex => qr/$re/, glob => $glob, depth => $depth }, $class;
6116}
6117sub full_path {
6119 my ($self, $path) = @_;
6120 return (length $self->{left} ? "$self->{left}/" : '') .
6121 $path . (length $self->{right} ? "/$self->{right}" : '');
6122}
6123__END__
6125Data structures:
6127$remotes = { # returned by read_all_remotes()
6130 'svn' => {
6131 # svn-remote.svn.url=https://svn.musicpd.org
6132 url => 'https://svn.musicpd.org',
6133 # svn-remote.svn.fetch=mpd/trunk:trunk
6134 fetch => {
6135 'mpd/trunk' => 'trunk',
6136 },
6137 # svn-remote.svn.tags=mpd/tags/*:tags/*
6138 tags => {
6139 path => {
6140 left => 'mpd/tags',
6141 right => '',
6142 regex => qr!mpd/tags/([^/]+)$!,
6143 glob => 'tags/*',
6144 },
6145 ref => {
6146 left => 'tags',
6147 right => '',
6148 regex => qr!tags/([^/]+)$!,
6149 glob => 'tags/*',
6150 },
6151 }
6152 }
6153};
6154$log_entry hashref as returned by libsvn_log_entry()
6156{
6157 log => 'whitespace-formatted log entry
6158', # trailing newline is preserved
6159 revision => '8', # integer
6160 date => '2004-02-24T17:01:44.108345Z', # commit date
6161 author => 'committer name'
6162};
6163# this is generated by generate_diff();
6166@mods = array of diff-index line hashes, each element represents one line
6167 of diff-index output
6168diff-index line ($m hash)
6170{
6171 mode_a => first column of diff-index output, no leading ':',
6172 mode_b => second column of diff-index output,
6173 sha1_b => sha1sum of the final blob,
6174 chg => change type [MCRADT],
6175 file_a => original file name of a file (iff chg is 'C' or 'R')
6176 file_b => new/current file name of a file (any chg)
6177}
6178;
6179# retval of read_url_paths{,_all}();
6181$l_map = {
6182 # repository root url
6183 'https://svn.musicpd.org' => {
6184 # repository path # GIT_SVN_ID
6185 'mpd/trunk' => 'trunk',
6186 'mpd/tags/0.11.5' => 'tags/0.11.5',
6187 },
6188}
6189Notes:
6191 I don't trust the each() function on unless I created %hash myself
6192 because the internal iterator may not have started at base.