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::Log::TZ = $ENV{TZ};
24$ENV{TZ} = 'UTC';
25$| = 1; # unbuffer STDOUT
26sub fatal (@) { print STDERR "@_\n"; exit 1 }
28require SVN::Core; # use()-ing this causes segfaults for me... *shrug*
29require SVN::Ra;
30require SVN::Delta;
31if ($SVN::Core::VERSION lt '1.1.0') {
32 fatal "Need SVN::Core 1.1.0 or better (got $SVN::Core::VERSION)";
33}
34push @Git::SVN::Ra::ISA, 'SVN::Ra';
35push @SVN::Git::Editor::ISA, 'SVN::Delta::Editor';
36push @SVN::Git::Fetcher::ISA, 'SVN::Delta::Editor';
37use Carp qw/croak/;
38use Digest::MD5;
39use IO::File qw//;
40use File::Basename qw/dirname basename/;
41use File::Path qw/mkpath/;
42use File::Spec;
43use Getopt::Long qw/:config gnu_getopt no_ignore_case auto_abbrev/;
44use IPC::Open3;
45use Git;
46BEGIN {
48 # import functions from Git into our packages, en masse
49 no strict 'refs';
50 foreach (qw/command command_oneline command_noisy command_output_pipe
51 command_input_pipe command_close_pipe
52 command_bidi_pipe command_close_bidi_pipe/) {
53 for my $package ( qw(SVN::Git::Editor SVN::Git::Fetcher
54 Git::SVN::Migration Git::SVN::Log Git::SVN),
55 __PACKAGE__) {
56 *{"${package}::$_"} = \&{"Git::$_"};
57 }
58 }
59}
60my ($SVN);
62$sha1 = qr/[a-f\d]{40}/;
64$sha1_short = qr/[a-f\d]{4,40}/;
65my ($_stdin, $_help, $_edit,
66 $_message, $_file,
67 $_template, $_shared,
68 $_version, $_fetch_all, $_no_rebase, $_fetch_parent,
69 $_merge, $_strategy, $_dry_run, $_local,
70 $_prefix, $_no_checkout, $_url, $_verbose,
71 $_git_format, $_commit_url, $_tag);
72$Git::SVN::_follow_parent = 1;
73$_q ||= 0;
74my %remote_opts = ( 'username=s' => \$Git::SVN::Prompt::_username,
75 'config-dir=s' => \$Git::SVN::Ra::config_dir,
76 'no-auth-cache' => \$Git::SVN::Prompt::_no_auth_cache,
77 'ignore-paths=s' => \$SVN::Git::Fetcher::_ignore_regex );
78my %fc_opts = ( 'follow-parent|follow!' => \$Git::SVN::_follow_parent,
79 'authors-file|A=s' => \$_authors,
80 'authors-prog=s' => \$_authors_prog,
81 'repack:i' => \$Git::SVN::_repack,
82 'noMetadata' => \$Git::SVN::_no_metadata,
83 'useSvmProps' => \$Git::SVN::_use_svm_props,
84 'useSvnsyncProps' => \$Git::SVN::_use_svnsync_props,
85 'log-window-size=i' => \$Git::SVN::Ra::_log_window_size,
86 'no-checkout' => \$_no_checkout,
87 'quiet|q+' => \$_q,
88 'repack-flags|repack-args|repack-opts=s' =>
89 \$Git::SVN::_repack_flags,
90 'use-log-author' => \$Git::SVN::_use_log_author,
91 'add-author-from' => \$Git::SVN::_add_author_from,
92 'localtime' => \$Git::SVN::_localtime,
93 %remote_opts );
94my ($_trunk, $_tags, $_branches, $_stdlayout);
96my %icv;
97my %init_opts = ( 'template=s' => \$_template, 'shared:s' => \$_shared,
98 'trunk|T=s' => \$_trunk, 'tags|t=s' => \$_tags,
99 'branches|b=s' => \$_branches, 'prefix=s' => \$_prefix,
100 'stdlayout|s' => \$_stdlayout,
101 'minimize-url|m' => \$Git::SVN::_minimize_url,
102 'no-metadata' => sub { $icv{noMetadata} = 1 },
103 'use-svm-props' => sub { $icv{useSvmProps} = 1 },
104 'use-svnsync-props' => sub { $icv{useSvnsyncProps} = 1 },
105 'rewrite-root=s' => sub { $icv{rewriteRoot} = $_[1] },
106 %remote_opts );
107my %cmt_opts = ( 'edit|e' => \$_edit,
108 'rmdir' => \$SVN::Git::Editor::_rmdir,
109 'find-copies-harder' => \$SVN::Git::Editor::_find_copies_harder,
110 'l=i' => \$SVN::Git::Editor::_rename_limit,
111 'copy-similarity|C=i'=> \$SVN::Git::Editor::_cp_similarity
112);
113my %cmd = (
115 fetch => [ \&cmd_fetch, "Download new revisions from SVN",
116 { 'revision|r=s' => \$_revision,
117 'fetch-all|all' => \$_fetch_all,
118 'parent|p' => \$_fetch_parent,
119 %fc_opts } ],
120 clone => [ \&cmd_clone, "Initialize and fetch revisions",
121 { 'revision|r=s' => \$_revision,
122 %fc_opts, %init_opts } ],
123 init => [ \&cmd_init, "Initialize a repo for tracking" .
124 " (requires URL argument)",
125 \%init_opts ],
126 'multi-init' => [ \&cmd_multi_init,
127 "Deprecated alias for ".
128 "'$0 init -T<trunk> -b<branches> -t<tags>'",
129 \%init_opts ],
130 dcommit => [ \&cmd_dcommit,
131 'Commit several diffs to merge with upstream',
132 { 'merge|m|M' => \$_merge,
133 'strategy|s=s' => \$_strategy,
134 'verbose|v' => \$_verbose,
135 'dry-run|n' => \$_dry_run,
136 'fetch-all|all' => \$_fetch_all,
137 'commit-url=s' => \$_commit_url,
138 'revision|r=i' => \$_revision,
139 'no-rebase' => \$_no_rebase,
140 %cmt_opts, %fc_opts } ],
141 branch => [ \&cmd_branch,
142 'Create a branch in the SVN repository',
143 { 'message|m=s' => \$_message,
144 'dry-run|n' => \$_dry_run,
145 'tag|t' => \$_tag } ],
146 tag => [ sub { $_tag = 1; cmd_branch(@_) },
147 'Create a tag in the SVN repository',
148 { 'message|m=s' => \$_message,
149 'dry-run|n' => \$_dry_run } ],
150 'set-tree' => [ \&cmd_set_tree,
151 "Set an SVN repository to a git tree-ish",
152 { 'stdin' => \$_stdin, %cmt_opts, %fc_opts, } ],
153 'create-ignore' => [ \&cmd_create_ignore,
154 'Create a .gitignore per svn:ignore',
155 { 'revision|r=i' => \$_revision
156 } ],
157 'propget' => [ \&cmd_propget,
158 'Print the value of a property on a file or directory',
159 { 'revision|r=i' => \$_revision } ],
160 'proplist' => [ \&cmd_proplist,
161 'List all properties of a file or directory',
162 { 'revision|r=i' => \$_revision } ],
163 'show-ignore' => [ \&cmd_show_ignore, "Show svn:ignore listings",
164 { 'revision|r=i' => \$_revision
165 } ],
166 'show-externals' => [ \&cmd_show_externals, "Show svn:externals listings",
167 { 'revision|r=i' => \$_revision
168 } ],
169 'multi-fetch' => [ \&cmd_multi_fetch,
170 "Deprecated alias for $0 fetch --all",
171 { 'revision|r=s' => \$_revision, %fc_opts } ],
172 'migrate' => [ sub { },
173 # no-op, we automatically run this anyways,
174 'Migrate configuration/metadata/layout from
175 previous versions of git-svn',
176 { 'minimize' => \$Git::SVN::Migration::_minimize,
177 %remote_opts } ],
178 'log' => [ \&Git::SVN::Log::cmd_show_log, 'Show commit logs',
179 { 'limit=i' => \$Git::SVN::Log::limit,
180 'revision|r=s' => \$_revision,
181 'verbose|v' => \$Git::SVN::Log::verbose,
182 'incremental' => \$Git::SVN::Log::incremental,
183 'oneline' => \$Git::SVN::Log::oneline,
184 'show-commit' => \$Git::SVN::Log::show_commit,
185 'non-recursive' => \$Git::SVN::Log::non_recursive,
186 'authors-file|A=s' => \$_authors,
187 'color' => \$Git::SVN::Log::color,
188 'pager=s' => \$Git::SVN::Log::pager
189 } ],
190 'find-rev' => [ \&cmd_find_rev,
191 "Translate between SVN revision numbers and tree-ish",
192 {} ],
193 'rebase' => [ \&cmd_rebase, "Fetch and rebase your working directory",
194 { 'merge|m|M' => \$_merge,
195 'verbose|v' => \$_verbose,
196 'strategy|s=s' => \$_strategy,
197 'local|l' => \$_local,
198 'fetch-all|all' => \$_fetch_all,
199 'dry-run|n' => \$_dry_run,
200 %fc_opts } ],
201 'commit-diff' => [ \&cmd_commit_diff,
202 'Commit a diff between two trees',
203 { 'message|m=s' => \$_message,
204 'file|F=s' => \$_file,
205 'revision|r=s' => \$_revision,
206 %cmt_opts } ],
207 'info' => [ \&cmd_info,
208 "Show info about the latest SVN revision
209 on the current branch",
210 { 'url' => \$_url, } ],
211 'blame' => [ \&Git::SVN::Log::cmd_blame,
212 "Show what revision and author last modified each line of a file",
213 { 'git-format' => \$_git_format } ],
214);
215my $cmd;
217for (my $i = 0; $i < @ARGV; $i++) {
218 if (defined $cmd{$ARGV[$i]}) {
219 $cmd = $ARGV[$i];
220 splice @ARGV, $i, 1;
221 last;
222 } elsif ($ARGV[$i] eq 'help') {
223 $cmd = $ARGV[$i+1];
224 usage(0);
225 }
226};
227# make sure we're always running at the top-level working directory
229unless ($cmd && $cmd =~ /(?:clone|init|multi-init)$/) {
230 unless (-d $ENV{GIT_DIR}) {
231 if ($git_dir_user_set) {
232 die "GIT_DIR=$ENV{GIT_DIR} explicitly set, ",
233 "but it is not a directory\n";
234 }
235 my $git_dir = delete $ENV{GIT_DIR};
236 my $cdup = undef;
237 git_cmd_try {
238 $cdup = command_oneline(qw/rev-parse --show-cdup/);
239 $git_dir = '.' unless ($cdup);
240 chomp $cdup if ($cdup);
241 $cdup = "." unless ($cdup && length $cdup);
242 } "Already at toplevel, but $git_dir not found\n";
243 chdir $cdup or die "Unable to chdir up to '$cdup'\n";
244 unless (-d $git_dir) {
245 die "$git_dir still not found after going to ",
246 "'$cdup'\n";
247 }
248 $ENV{GIT_DIR} = $git_dir;
249 }
250 $_repository = Git->repository(Repository => $ENV{GIT_DIR});
251}
252my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
254read_repo_config(\%opts);
256if ($cmd && ($cmd eq 'log' || $cmd eq 'blame')) {
257 Getopt::Long::Configure('pass_through');
258}
259my $rv = GetOptions(%opts, 'help|H|h' => \$_help, 'version|V' => \$_version,
260 'minimize-connections' => \$Git::SVN::Migration::_minimize,
261 'id|i=s' => \$Git::SVN::default_ref_id,
262 'svn-remote|remote|R=s' => sub {
263 $Git::SVN::no_reuse_existing = 1;
264 $Git::SVN::default_repo_id = $_[1] });
265exit 1 if (!$rv && $cmd && $cmd ne 'log');
266usage(0) if $_help;
268version() if $_version;
269usage(1) unless defined $cmd;
270load_authors() if $_authors;
271if (defined $_authors_prog) {
272 $_authors_prog = "'" . File::Spec->rel2abs($_authors_prog) . "'";
273}
274unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
276 Git::SVN::Migration::migration_check();
277}
278Git::SVN::init_vars();
279eval {
280 Git::SVN::verify_remotes_sanity();
281 $cmd{$cmd}->[0]->(@ARGV);
282};
283fatal $@ if $@;
284post_fetch_checkout();
285exit 0;
286####################### primary functions ######################
288sub usage {
289 my $exit = shift || 0;
290 my $fd = $exit ? \*STDERR : \*STDOUT;
291 print $fd <<"";
292git-svn - bidirectional operations between a single Subversion tree and git
293Usage: git svn <command> [options] [arguments]\n
294 print $fd "Available commands:\n" unless $cmd;
296 foreach (sort keys %cmd) {
298 next if $cmd && $cmd ne $_;
299 next if /^multi-/; # don't show deprecated commands
300 print $fd ' ',pack('A17',$_),$cmd{$_}->[1],"\n";
301 foreach (sort keys %{$cmd{$_}->[2]}) {
302 # mixed-case options are for .git/config only
303 next if /[A-Z]/ && /^[a-z]+$/i;
304 # prints out arguments as they should be passed:
305 my $x = s#[:=]s$## ? '<arg>' : s#[:=]i$## ? '<num>' : '';
306 print $fd ' ' x 21, join(', ', map { length $_ > 1 ?
307 "--$_" : "-$_" }
308 split /\|/,$_)," $x\n";
309 }
310 }
311 print $fd <<"";
312\nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
313arbitrary identifier if you're tracking multiple SVN branches/repositories in
314one git repository and want to keep them separate. See git-svn(1) for more
315information.
316 exit $exit;
318}
319sub version {
321 print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
322 exit 0;
323}
324sub do_git_init_db {
326 unless (-d $ENV{GIT_DIR}) {
327 my @init_db = ('init');
328 push @init_db, "--template=$_template" if defined $_template;
329 if (defined $_shared) {
330 if ($_shared =~ /[a-z]/) {
331 push @init_db, "--shared=$_shared";
332 } else {
333 push @init_db, "--shared";
334 }
335 }
336 command_noisy(@init_db);
337 $_repository = Git->repository(Repository => ".git");
338 }
339 command_noisy('config', 'core.autocrlf', 'false');
340 my $set;
341 my $pfx = "svn-remote.$Git::SVN::default_repo_id";
342 foreach my $i (keys %icv) {
343 die "'$set' and '$i' cannot both be set\n" if $set;
344 next unless defined $icv{$i};
345 command_noisy('config', "$pfx.$i", $icv{$i});
346 $set = $i;
347 }
348 my $ignore_regex = \$SVN::Git::Fetcher::_ignore_regex;
349 command_noisy('config', "$pfx.ignore-paths", $$ignore_regex)
350 if defined $$ignore_regex;
351}
352sub init_subdir {
354 my $repo_path = shift or return;
355 mkpath([$repo_path]) unless -d $repo_path;
356 chdir $repo_path or die "Couldn't chdir to $repo_path: $!\n";
357 $ENV{GIT_DIR} = '.git';
358 $_repository = Git->repository(Repository => $ENV{GIT_DIR});
359}
360sub cmd_clone {
362 my ($url, $path) = @_;
363 if (!defined $path &&
364 (defined $_trunk || defined $_branches || defined $_tags ||
365 defined $_stdlayout) &&
366 $url !~ m#^[a-z\+]+://#) {
367 $path = $url;
368 }
369 $path = basename($url) if !defined $path || !length $path;
370 cmd_init($url, $path);
371 Git::SVN::fetch_all($Git::SVN::default_repo_id);
372 command_oneline('config', 'svn.authorsfile', $_authors) if $_authors;
373}
374sub cmd_init {
376 if (defined $_stdlayout) {
377 $_trunk = 'trunk' if (!defined $_trunk);
378 $_tags = 'tags' if (!defined $_tags);
379 $_branches = 'branches' if (!defined $_branches);
380 }
381 if (defined $_trunk || defined $_branches || defined $_tags) {
382 return cmd_multi_init(@_);
383 }
384 my $url = shift or die "SVN repository location required ",
385 "as a command-line argument\n";
386 init_subdir(@_);
387 do_git_init_db();
388 Git::SVN->init($url);
390}
391sub cmd_fetch {
393 if (grep /^\d+=./, @_) {
394 die "'<rev>=<commit>' fetch arguments are ",
395 "no longer supported.\n";
396 }
397 my ($remote) = @_;
398 if (@_ > 1) {
399 die "Usage: $0 fetch [--all] [--parent] [svn-remote]\n";
400 }
401 if ($_fetch_parent) {
402 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
403 unless ($gs) {
404 die "Unable to determine upstream SVN information from ",
405 "working tree history\n";
406 }
407 # just fetch, don't checkout.
408 $_no_checkout = 'true';
409 $_fetch_all ? $gs->fetch_all : $gs->fetch;
410 } elsif ($_fetch_all) {
411 cmd_multi_fetch();
412 } else {
413 $remote ||= $Git::SVN::default_repo_id;
414 Git::SVN::fetch_all($remote, Git::SVN::read_all_remotes());
415 }
416}
417sub cmd_set_tree {
419 my (@commits) = @_;
420 if ($_stdin || !@commits) {
421 print "Reading from stdin...\n";
422 @commits = ();
423 while (<STDIN>) {
424 if (/\b($sha1_short)\b/o) {
425 unshift @commits, $1;
426 }
427 }
428 }
429 my @revs;
430 foreach my $c (@commits) {
431 my @tmp = command('rev-parse',$c);
432 if (scalar @tmp == 1) {
433 push @revs, $tmp[0];
434 } elsif (scalar @tmp > 1) {
435 push @revs, reverse(command('rev-list',@tmp));
436 } else {
437 fatal "Failed to rev-parse $c";
438 }
439 }
440 my $gs = Git::SVN->new;
441 my ($r_last, $cmt_last) = $gs->last_rev_commit;
442 $gs->fetch;
443 if (defined $gs->{last_rev} && $r_last != $gs->{last_rev}) {
444 fatal "There are new revisions that were fetched ",
445 "and need to be merged (or acknowledged) ",
446 "before committing.\nlast rev: $r_last\n",
447 " current: $gs->{last_rev}";
448 }
449 $gs->set_tree($_) foreach @revs;
450 print "Done committing ",scalar @revs," revisions to SVN\n";
451 unlink $gs->{index};
452}
453sub cmd_dcommit {
455 my $head = shift;
456 git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
457 'Cannot dcommit with a dirty index. Commit your changes first, '
458 . "or stash them with `git stash'.\n";
459 $head ||= 'HEAD';
460 my $old_head;
462 if ($head ne 'HEAD') {
463 $old_head = eval {
464 command_oneline([qw/symbolic-ref -q HEAD/])
465 };
466 if ($old_head) {
467 $old_head =~ s{^refs/heads/}{};
468 } else {
469 $old_head = eval { command_oneline(qw/rev-parse HEAD/) };
470 }
471 command(['checkout', $head], STDERR => 0);
472 }
473 my @refs;
475 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD', \@refs);
476 unless ($gs) {
477 die "Unable to determine upstream SVN information from ",
478 "$head history.\nPerhaps the repository is empty.";
479 }
480 if (defined $_commit_url) {
482 $url = $_commit_url;
483 } else {
484 $url = eval { command_oneline('config', '--get',
485 "svn-remote.$gs->{repo_id}.commiturl") };
486 if (!$url) {
487 $url = $gs->full_url
488 }
489 }
490 my $last_rev = $_revision if defined $_revision;
492 if ($url) {
493 print "Committing to $url ...\n";
494 }
495 my ($linear_refs, $parents) = linearize_history($gs, \@refs);
496 if ($_no_rebase && scalar(@$linear_refs) > 1) {
497 warn "Attempting to commit more than one change while ",
498 "--no-rebase is enabled.\n",
499 "If these changes depend on each other, re-running ",
500 "without --no-rebase may be required."
501 }
502 my $expect_url = $url;
503 Git::SVN::remove_username($expect_url);
504 while (1) {
505 my $d = shift @$linear_refs or last;
506 unless (defined $last_rev) {
507 (undef, $last_rev, undef) = cmt_metadata("$d~1");
508 unless (defined $last_rev) {
509 fatal "Unable to extract revision information ",
510 "from commit $d~1";
511 }
512 }
513 if ($_dry_run) {
514 print "diff-tree $d~1 $d\n";
515 } else {
516 my $cmt_rev;
517 my %ed_opts = ( r => $last_rev,
518 log => get_commit_entry($d)->{log},
519 ra => Git::SVN::Ra->new($url),
520 config => SVN::Core::config_get_config(
521 $Git::SVN::Ra::config_dir
522 ),
523 tree_a => "$d~1",
524 tree_b => $d,
525 editor_cb => sub {
526 print "Committed r$_[0]\n";
527 $cmt_rev = $_[0];
528 },
529 svn_path => '');
530 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
531 print "No changes\n$d~1 == $d\n";
532 } elsif ($parents->{$d} && @{$parents->{$d}}) {
533 $gs->{inject_parents_dcommit}->{$cmt_rev} =
534 $parents->{$d};
535 }
536 $_fetch_all ? $gs->fetch_all : $gs->fetch;
537 $last_rev = $cmt_rev;
538 next if $_no_rebase;
539 # we always want to rebase against the current HEAD,
541 # not any head that was passed to us
542 my @diff = command('diff-tree', $d,
543 $gs->refname, '--');
544 my @finish;
545 if (@diff) {
546 @finish = rebase_cmd();
547 print STDERR "W: $d and ", $gs->refname,
548 " differ, using @finish:\n",
549 join("\n", @diff), "\n";
550 } else {
551 print "No changes between current HEAD and ",
552 $gs->refname,
553 "\nResetting to the latest ",
554 $gs->refname, "\n";
555 @finish = qw/reset --mixed/;
556 }
557 command_noisy(@finish, $gs->refname);
558 if (@diff) {
559 @refs = ();
560 my ($url_, $rev_, $uuid_, $gs_) =
561 working_head_info('HEAD', \@refs);
562 my ($linear_refs_, $parents_) =
563 linearize_history($gs_, \@refs);
564 if (scalar(@$linear_refs) !=
565 scalar(@$linear_refs_)) {
566 fatal "# of revisions changed ",
567 "\nbefore:\n",
568 join("\n", @$linear_refs),
569 "\n\nafter:\n",
570 join("\n", @$linear_refs_), "\n",
571 'If you are attempting to commit ',
572 "merges, try running:\n\t",
573 'git rebase --interactive',
574 '--preserve-merges ',
575 $gs->refname,
576 "\nBefore dcommitting";
577 }
578 if ($url_ ne $expect_url) {
579 fatal "URL mismatch after rebase: ",
580 "$url_ != $expect_url";
581 }
582 if ($uuid_ ne $uuid) {
583 fatal "uuid mismatch after rebase: ",
584 "$uuid_ != $uuid";
585 }
586 # remap parents
587 my (%p, @l, $i);
588 for ($i = 0; $i < scalar @$linear_refs; $i++) {
589 my $new = $linear_refs_->[$i] or next;
590 $p{$new} =
591 $parents->{$linear_refs->[$i]};
592 push @l, $new;
593 }
594 $parents = \%p;
595 $linear_refs = \@l;
596 }
597 }
598 }
599 if ($old_head) {
601 my $new_head = command_oneline(qw/rev-parse HEAD/);
602 my $new_is_symbolic = eval {
603 command_oneline(qw/symbolic-ref -q HEAD/);
604 };
605 if ($new_is_symbolic) {
606 print "dcommitted the branch ", $head, "\n";
607 } else {
608 print "dcommitted on a detached HEAD because you gave ",
609 "a revision argument.\n",
610 "The rewritten commit is: ", $new_head, "\n";
611 }
612 command(['checkout', $old_head], STDERR => 0);
613 }
614 unlink $gs->{index};
616}
617sub cmd_branch {
619 my ($branch_name, $head) = @_;
620 unless (defined $branch_name && length $branch_name) {
622 die(($_tag ? "tag" : "branch") . " name required\n");
623 }
624 $head ||= 'HEAD';
625 my ($src, $rev, undef, $gs) = working_head_info($head);
627 my $remote = Git::SVN::read_all_remotes()->{$gs->{repo_id}};
629 my $glob = $remote->{ $_tag ? 'tags' : 'branches' };
630 my ($lft, $rgt) = @{ $glob->{path} }{qw/left right/};
631 my $dst = join '/', $remote->{url}, $lft, $branch_name, ($rgt || ());
632 my $ctx = SVN::Client->new(
634 auth => Git::SVN::Ra::_auth_providers(),
635 log_msg => sub {
636 ${ $_[0] } = defined $_message
637 ? $_message
638 : 'Create ' . ($_tag ? 'tag ' : 'branch ' )
639 . $branch_name;
640 },
641 );
642 eval {
644 $ctx->ls($dst, 'HEAD', 0);
645 } and die "branch ${branch_name} already exists\n";
646 print "Copying ${src} at r${rev} to ${dst}...\n";
648 $ctx->copy($src, $rev, $dst)
649 unless $_dry_run;
650 $gs->fetch_all;
652}
653sub cmd_find_rev {
655 my $revision_or_hash = shift or die "SVN or git revision required ",
656 "as a command-line argument\n";
657 my $result;
658 if ($revision_or_hash =~ /^r\d+$/) {
659 my $head = shift;
660 $head ||= 'HEAD';
661 my @refs;
662 my (undef, undef, $uuid, $gs) = working_head_info($head, \@refs);
663 unless ($gs) {
664 die "Unable to determine upstream SVN information from ",
665 "$head history\n";
666 }
667 my $desired_revision = substr($revision_or_hash, 1);
668 $result = $gs->rev_map_get($desired_revision, $uuid);
669 } else {
670 my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
671 $result = $rev;
672 }
673 print "$result\n" if $result;
674}
675sub cmd_rebase {
677 command_noisy(qw/update-index --refresh/);
678 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
679 unless ($gs) {
680 die "Unable to determine upstream SVN information from ",
681 "working tree history\n";
682 }
683 if ($_dry_run) {
684 print "Remote Branch: " . $gs->refname . "\n";
685 print "SVN URL: " . $url . "\n";
686 return;
687 }
688 if (command(qw/diff-index HEAD --/)) {
689 print STDERR "Cannot rebase with uncommited changes:\n";
690 command_noisy('status');
691 exit 1;
692 }
693 unless ($_local) {
694 # rebase will checkout for us, so no need to do it explicitly
695 $_no_checkout = 'true';
696 $_fetch_all ? $gs->fetch_all : $gs->fetch;
697 }
698 command_noisy(rebase_cmd(), $gs->refname);
699}
700sub cmd_show_ignore {
702 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
703 $gs ||= Git::SVN->new;
704 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
705 $gs->prop_walk($gs->{path}, $r, sub {
706 my ($gs, $path, $props) = @_;
707 print STDOUT "\n# $path\n";
708 my $s = $props->{'svn:ignore'} or return;
709 $s =~ s/[\r\n]+/\n/g;
710 chomp $s;
711 $s =~ s#^#$path#gm;
712 print STDOUT "$s\n";
713 });
714}
715sub cmd_show_externals {
717 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
718 $gs ||= Git::SVN->new;
719 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
720 $gs->prop_walk($gs->{path}, $r, sub {
721 my ($gs, $path, $props) = @_;
722 print STDOUT "\n# $path\n";
723 my $s = $props->{'svn:externals'} or return;
724 $s =~ s/[\r\n]+/\n/g;
725 chomp $s;
726 $s =~ s#^#$path#gm;
727 print STDOUT "$s\n";
728 });
729}
730sub cmd_create_ignore {
732 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
733 $gs ||= Git::SVN->new;
734 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
735 $gs->prop_walk($gs->{path}, $r, sub {
736 my ($gs, $path, $props) = @_;
737 # $path is of the form /path/to/dir/
738 $path = '.' . $path;
739 # SVN can have attributes on empty directories,
740 # which git won't track
741 mkpath([$path]) unless -d $path;
742 my $ignore = $path . '.gitignore';
743 my $s = $props->{'svn:ignore'} or return;
744 open(GITIGNORE, '>', $ignore)
745 or fatal("Failed to open `$ignore' for writing: $!");
746 $s =~ s/[\r\n]+/\n/g;
747 chomp $s;
748 # Prefix all patterns so that the ignore doesn't apply
749 # to sub-directories.
750 $s =~ s#^#/#gm;
751 print GITIGNORE "$s\n";
752 close(GITIGNORE)
753 or fatal("Failed to close `$ignore': $!");
754 command_noisy('add', '-f', $ignore);
755 });
756}
757sub canonicalize_path {
759 my ($path) = @_;
760 my $dot_slash_added = 0;
761 if (substr($path, 0, 1) ne "/") {
762 $path = "./" . $path;
763 $dot_slash_added = 1;
764 }
765 # File::Spec->canonpath doesn't collapse x/../y into y (for a
766 # good reason), so let's do this manually.
767 $path =~ s#/+#/#g;
768 $path =~ s#/\.(?:/|$)#/#g;
769 $path =~ s#/[^/]+/\.\.##g;
770 $path =~ s#/$##g;
771 $path =~ s#^\./## if $dot_slash_added;
772 $path =~ s#^/##;
773 $path =~ s#^\.$##;
774 return $path;
775}
776# get_svnprops(PATH)
778# ------------------
779# Helper for cmd_propget and cmd_proplist below.
780sub get_svnprops {
781 my $path = shift;
782 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
783 $gs ||= Git::SVN->new;
784 # prefix THE PATH by the sub-directory from which the user
786 # invoked us.
787 $path = $cmd_dir_prefix . $path;
788 fatal("No such file or directory: $path") unless -e $path;
789 my $is_dir = -d $path ? 1 : 0;
790 $path = $gs->{path} . '/' . $path;
791 # canonicalize the path (otherwise libsvn will abort or fail to
793 # find the file)
794 $path = canonicalize_path($path);
795 my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
797 my $props;
798 if ($is_dir) {
799 (undef, undef, $props) = $gs->ra->get_dir($path, $r);
800 }
801 else {
802 (undef, $props) = $gs->ra->get_file($path, $r, undef);
803 }
804 return $props;
805}
806# cmd_propget (PROP, PATH)
808# ------------------------
809# Print the SVN property PROP for PATH.
810sub cmd_propget {
811 my ($prop, $path) = @_;
812 $path = '.' if not defined $path;
813 usage(1) if not defined $prop;
814 my $props = get_svnprops($path);
815 if (not defined $props->{$prop}) {
816 fatal("`$path' does not have a `$prop' SVN property.");
817 }
818 print $props->{$prop} . "\n";
819}
820# cmd_proplist (PATH)
822# -------------------
823# Print the list of SVN properties for PATH.
824sub cmd_proplist {
825 my $path = shift;
826 $path = '.' if not defined $path;
827 my $props = get_svnprops($path);
828 print "Properties on '$path':\n";
829 foreach (sort keys %{$props}) {
830 print " $_\n";
831 }
832}
833sub cmd_multi_init {
835 my $url = shift;
836 unless (defined $_trunk || defined $_branches || defined $_tags) {
837 usage(1);
838 }
839 # there are currently some bugs that prevent multi-init/multi-fetch
841 # setups from working well without this.
842 $Git::SVN::_minimize_url = 1;
843 $_prefix = '' unless defined $_prefix;
845 if (defined $url) {
846 $url =~ s#/+$##;
847 init_subdir(@_);
848 }
849 do_git_init_db();
850 if (defined $_trunk) {
851 my $trunk_ref = $_prefix . 'trunk';
852 # try both old-style and new-style lookups:
853 my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
854 unless ($gs_trunk) {
855 my ($trunk_url, $trunk_path) =
856 complete_svn_url($url, $_trunk);
857 $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
858 undef, $trunk_ref);
859 }
860 }
861 return unless defined $_branches || defined $_tags;
862 my $ra = $url ? Git::SVN::Ra->new($url) : undef;
863 complete_url_ls_init($ra, $_branches, '--branches/-b', $_prefix);
864 complete_url_ls_init($ra, $_tags, '--tags/-t', $_prefix . 'tags/');
865}
866sub cmd_multi_fetch {
868 my $remotes = Git::SVN::read_all_remotes();
869 foreach my $repo_id (sort keys %$remotes) {
870 if ($remotes->{$repo_id}->{url}) {
871 Git::SVN::fetch_all($repo_id, $remotes);
872 }
873 }
874}
875# this command is special because it requires no metadata
877sub cmd_commit_diff {
878 my ($ta, $tb, $url) = @_;
879 my $usage = "Usage: $0 commit-diff -r<revision> ".
880 "<tree-ish> <tree-ish> [<URL>]";
881 fatal($usage) if (!defined $ta || !defined $tb);
882 my $svn_path = '';
883 if (!defined $url) {
884 my $gs = eval { Git::SVN->new };
885 if (!$gs) {
886 fatal("Needed URL or usable git-svn --id in ",
887 "the command-line\n", $usage);
888 }
889 $url = $gs->{url};
890 $svn_path = $gs->{path};
891 }
892 unless (defined $_revision) {
893 fatal("-r|--revision is a required argument\n", $usage);
894 }
895 if (defined $_message && defined $_file) {
896 fatal("Both --message/-m and --file/-F specified ",
897 "for the commit message.\n",
898 "I have no idea what you mean");
899 }
900 if (defined $_file) {
901 $_message = file_to_s($_file);
902 } else {
903 $_message ||= get_commit_entry($tb)->{log};
904 }
905 my $ra ||= Git::SVN::Ra->new($url);
906 my $r = $_revision;
907 if ($r eq 'HEAD') {
908 $r = $ra->get_latest_revnum;
909 } elsif ($r !~ /^\d+$/) {
910 die "revision argument: $r not understood by git-svn\n";
911 }
912 my %ed_opts = ( r => $r,
913 log => $_message,
914 ra => $ra,
915 tree_a => $ta,
916 tree_b => $tb,
917 editor_cb => sub { print "Committed r$_[0]\n" },
918 svn_path => $svn_path );
919 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
920 print "No changes\n$ta == $tb\n";
921 }
922}
923sub escape_uri_only {
925 my ($uri) = @_;
926 my @tmp;
927 foreach (split m{/}, $uri) {
928 s/([^~\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
929 push @tmp, $_;
930 }
931 join('/', @tmp);
932}
933sub escape_url {
935 my ($url) = @_;
936 if ($url =~ m#^([^:]+)://([^/]*)(.*)$#) {
937 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
938 $url = "$scheme://$domain$uri";
939 }
940 $url;
941}
942sub cmd_info {
944 my $path = canonicalize_path(defined($_[0]) ? $_[0] : ".");
945 my $fullpath = canonicalize_path($cmd_dir_prefix . $path);
946 if (exists $_[1]) {
947 die "Too many arguments specified\n";
948 }
949 my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
951 if (!$file_type && !$diff_status) {
953 print STDERR "svn: '$path' is not under version control\n";
954 exit 1;
955 }
956 my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
958 unless ($gs) {
959 die "Unable to determine upstream SVN information from ",
960 "working tree history\n";
961 }
962 # canonicalize_path() will return "" to make libsvn 1.5.x happy,
964 $path = "." if $path eq "";
965 my $full_url = $url . ($fullpath eq "" ? "" : "/$fullpath");
967 if ($_url) {
969 print escape_url($full_url), "\n";
970 return;
971 }
972 my $result = "Path: $path\n";
974 $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
975 $result .= "URL: " . escape_url($full_url) . "\n";
976 eval {
978 my $repos_root = $gs->repos_root;
979 Git::SVN::remove_username($repos_root);
980 $result .= "Repository Root: " . escape_url($repos_root) . "\n";
981 };
982 if ($@) {
983 $result .= "Repository Root: (offline)\n";
984 }
985 $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A" &&
986 ($SVN::Core::VERSION le '1.5.4' || $file_type ne "dir");
987 $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
988 $result .= "Node Kind: " .
990 ($file_type eq "dir" ? "directory" : "file") . "\n";
991 my $schedule = $diff_status eq "A"
993 ? "add"
994 : ($diff_status eq "D" ? "delete" : "normal");
995 $result .= "Schedule: $schedule\n";
996 if ($diff_status eq "A") {
998 print $result, "\n";
999 return;
1000 }
1001 my ($lc_author, $lc_rev, $lc_date_utc);
1003 my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $fullpath);
1004 my $log = command_output_pipe(@args);
1005 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
1006 while (<$log>) {
1007 if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
1008 $lc_author = $1;
1009 $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
1010 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
1011 (undef, $lc_rev, undef) = ::extract_metadata($1);
1012 }
1013 }
1014 close $log;
1015 Git::SVN::Log::set_local_timezone();
1017 $result .= "Last Changed Author: $lc_author\n";
1019 $result .= "Last Changed Rev: $lc_rev\n";
1020 $result .= "Last Changed Date: " .
1021 Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
1022 if ($file_type ne "dir") {
1024 my $text_last_updated_date =
1025 ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
1026 $result .=
1027 "Text Last Updated: " .
1028 Git::SVN::Log::format_svn_date($text_last_updated_date) .
1029 "\n";
1030 my $checksum;
1031 if ($diff_status eq "D") {
1032 my ($fh, $ctx) =
1033 command_output_pipe(qw(cat-file blob), "HEAD:$path");
1034 if ($file_type eq "link") {
1035 my $file_name = <$fh>;
1036 $checksum = md5sum("link $file_name");
1037 } else {
1038 $checksum = md5sum($fh);
1039 }
1040 command_close_pipe($fh, $ctx);
1041 } elsif ($file_type eq "link") {
1042 my $file_name =
1043 command(qw(cat-file blob), "HEAD:$path");
1044 $checksum =
1045 md5sum("link " . $file_name);
1046 } else {
1047 open FILE, "<", $path or die $!;
1048 $checksum = md5sum(\*FILE);
1049 close FILE or die $!;
1050 }
1051 $result .= "Checksum: " . $checksum . "\n";
1052 }
1053 print $result, "\n";
1055}
1056########################### utility functions #########################
1058sub rebase_cmd {
1060 my @cmd = qw/rebase/;
1061 push @cmd, '-v' if $_verbose;
1062 push @cmd, qw/--merge/ if $_merge;
1063 push @cmd, "--strategy=$_strategy" if $_strategy;
1064 @cmd;
1065}
1066sub post_fetch_checkout {
1068 return if $_no_checkout;
1069 my $gs = $Git::SVN::_head or return;
1070 return if verify_ref('refs/heads/master^0');
1071 my $valid_head = verify_ref('HEAD^0');
1073 command_noisy(qw(update-ref refs/heads/master), $gs->refname);
1074 return if ($valid_head || !verify_ref('HEAD^0'));
1075 return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
1077 my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
1078 return if -f $index;
1079 return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
1081 return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
1082 command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
1083 print STDERR "Checked out HEAD:\n ",
1084 $gs->full_url, " r", $gs->last_rev, "\n";
1085}
1086sub complete_svn_url {
1088 my ($url, $path) = @_;
1089 $path =~ s#/+$##;
1090 if ($path !~ m#^[a-z\+]+://#) {
1091 if (!defined $url || $url !~ m#^[a-z\+]+://#) {
1092 fatal("E: '$path' is not a complete URL ",
1093 "and a separate URL is not specified");
1094 }
1095 return ($url, $path);
1096 }
1097 return ($path, '');
1098}
1099sub complete_url_ls_init {
1101 my ($ra, $repo_path, $switch, $pfx) = @_;
1102 unless ($repo_path) {
1103 print STDERR "W: $switch not specified\n";
1104 return;
1105 }
1106 $repo_path =~ s#/+$##;
1107 if ($repo_path =~ m#^[a-z\+]+://#) {
1108 $ra = Git::SVN::Ra->new($repo_path);
1109 $repo_path = '';
1110 } else {
1111 $repo_path =~ s#^/+##;
1112 unless ($ra) {
1113 fatal("E: '$repo_path' is not a complete URL ",
1114 "and a separate URL is not specified");
1115 }
1116 }
1117 my $url = $ra->{url};
1118 my $gs = Git::SVN->init($url, undef, undef, undef, 1);
1119 my $k = "svn-remote.$gs->{repo_id}.url";
1120 my $orig_url = eval { command_oneline(qw/config --get/, $k) };
1121 if ($orig_url && ($orig_url ne $gs->{url})) {
1122 die "$k already set: $orig_url\n",
1123 "wanted to set to: $gs->{url}\n";
1124 }
1125 command_oneline('config', $k, $gs->{url}) unless $orig_url;
1126 my $remote_path = "$ra->{svn_path}/$repo_path";
1127 $remote_path =~ s#/+#/#g;
1128 $remote_path =~ s#^/##g;
1129 $remote_path .= "/*" if $remote_path !~ /\*/;
1130 my ($n) = ($switch =~ /^--(\w+)/);
1131 if (length $pfx && $pfx !~ m#/$#) {
1132 die "--prefix='$pfx' must have a trailing slash '/'\n";
1133 }
1134 command_noisy('config',
1135 "svn-remote.$gs->{repo_id}.$n",
1136 "$remote_path:refs/remotes/$pfx*" .
1137 ('/*' x (($remote_path =~ tr/*/*/) - 1)) );
1138}
1139sub verify_ref {
1141 my ($ref) = @_;
1142 eval { command_oneline([ 'rev-parse', '--verify', $ref ],
1143 { STDERR => 0 }); };
1144}
1145sub get_tree_from_treeish {
1147 my ($treeish) = @_;
1148 # $treeish can be a symbolic ref, too:
1149 my $type = command_oneline(qw/cat-file -t/, $treeish);
1150 my $expected;
1151 while ($type eq 'tag') {
1152 ($treeish, $type) = command(qw/cat-file tag/, $treeish);
1153 }
1154 if ($type eq 'commit') {
1155 $expected = (grep /^tree /, command(qw/cat-file commit/,
1156 $treeish))[0];
1157 ($expected) = ($expected =~ /^tree ($sha1)$/o);
1158 die "Unable to get tree from $treeish\n" unless $expected;
1159 } elsif ($type eq 'tree') {
1160 $expected = $treeish;
1161 } else {
1162 die "$treeish is a $type, expected tree, tag or commit\n";
1163 }
1164 return $expected;
1165}
1166sub get_commit_entry {
1168 my ($treeish) = shift;
1169 my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
1170 my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
1171 my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
1172 open my $log_fh, '>', $commit_editmsg or croak $!;
1173 my $type = command_oneline(qw/cat-file -t/, $treeish);
1175 if ($type eq 'commit' || $type eq 'tag') {
1176 my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1177 $type, $treeish);
1178 my $in_msg = 0;
1179 my $author;
1180 my $saw_from = 0;
1181 my $msgbuf = "";
1182 while (<$msg_fh>) {
1183 if (!$in_msg) {
1184 $in_msg = 1 if (/^\s*$/);
1185 $author = $1 if (/^author (.*>)/);
1186 } elsif (/^git-svn-id: /) {
1187 # skip this for now, we regenerate the
1188 # correct one on re-fetch anyways
1189 # TODO: set *:merge properties or like...
1190 } else {
1191 if (/^From:/ || /^Signed-off-by:/) {
1192 $saw_from = 1;
1193 }
1194 $msgbuf .= $_;
1195 }
1196 }
1197 $msgbuf =~ s/\s+$//s;
1198 if ($Git::SVN::_add_author_from && defined($author)
1199 && !$saw_from) {
1200 $msgbuf .= "\n\nFrom: $author";
1201 }
1202 print $log_fh $msgbuf or croak $!;
1203 command_close_pipe($msg_fh, $ctx);
1204 }
1205 close $log_fh or croak $!;
1206 if ($_edit || ($type eq 'tree')) {
1208 my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
1209 # TODO: strip out spaces, comments, like git-commit.sh
1210 system($editor, $commit_editmsg);
1211 }
1212 rename $commit_editmsg, $commit_msg or croak $!;
1213 {
1214 require Encode;
1215 # SVN requires messages to be UTF-8 when entering the repo
1216 local $/;
1217 open $log_fh, '<', $commit_msg or croak $!;
1218 binmode $log_fh;
1219 chomp($log_entry{log} = <$log_fh>);
1220 my $enc = Git::config('i18n.commitencoding') || 'UTF-8';
1222 my $msg = $log_entry{log};
1223 eval { $msg = Encode::decode($enc, $msg, 1) };
1225 if ($@) {
1226 die "Could not decode as $enc:\n", $msg,
1227 "\nPerhaps you need to set i18n.commitencoding\n";
1228 }
1229 eval { $msg = Encode::encode('UTF-8', $msg, 1) };
1231 die "Could not encode as UTF-8:\n$msg\n" if $@;
1232 $log_entry{log} = $msg;
1234 close $log_fh or croak $!;
1236 }
1237 unlink $commit_msg;
1238 \%log_entry;
1239}
1240sub s_to_file {
1242 my ($str, $file, $mode) = @_;
1243 open my $fd,'>',$file or croak $!;
1244 print $fd $str,"\n" or croak $!;
1245 close $fd or croak $!;
1246 chmod ($mode &~ umask, $file) if (defined $mode);
1247}
1248sub file_to_s {
1250 my $file = shift;
1251 open my $fd,'<',$file or croak "$!: file: $file\n";
1252 local $/;
1253 my $ret = <$fd>;
1254 close $fd or croak $!;
1255 $ret =~ s/\s*$//s;
1256 return $ret;
1257}
1258# '<svn username> = real-name <email address>' mapping based on git-svnimport:
1260sub load_authors {
1261 open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1262 my $log = $cmd eq 'log';
1263 while (<$authors>) {
1264 chomp;
1265 next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1266 my ($user, $name, $email) = ($1, $2, $3);
1267 if ($log) {
1268 $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1269 } else {
1270 $users{$user} = [$name, $email];
1271 }
1272 }
1273 close $authors or croak $!;
1274}
1275# convert GetOpt::Long specs for use by git-config
1277sub read_repo_config {
1278 return unless -d $ENV{GIT_DIR};
1279 my $opts = shift;
1280 my @config_only;
1281 foreach my $o (keys %$opts) {
1282 # if we have mixedCase and a long option-only, then
1283 # it's a config-only variable that we don't need for
1284 # the command-line.
1285 push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1286 my $v = $opts->{$o};
1287 my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1288 $key =~ s/-//g;
1289 my $arg = 'git config';
1290 $arg .= ' --int' if ($o =~ /[:=]i$/);
1291 $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1292 if (ref $v eq 'ARRAY') {
1293 chomp(my @tmp = `$arg --get-all svn.$key`);
1294 @$v = @tmp if @tmp;
1295 } else {
1296 chomp(my $tmp = `$arg --get svn.$key`);
1297 if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1298 $$v = $tmp;
1299 }
1300 }
1301 }
1302 delete @$opts{@config_only} if @config_only;
1303}
1304sub extract_metadata {
1306 my $id = shift or return (undef, undef, undef);
1307 my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1308 \s([a-f\d\-]+)$/x);
1309 if (!defined $rev || !$uuid || !$url) {
1310 # some of the original repositories I made had
1311 # identifiers like this:
1312 ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/);
1313 }
1314 return ($url, $rev, $uuid);
1315}
1316sub cmt_metadata {
1318 return extract_metadata((grep(/^git-svn-id: /,
1319 command(qw/cat-file commit/, shift)))[-1]);
1320}
1321sub cmt_sha2rev_batch {
1323 my %s2r;
1324 my ($pid, $in, $out, $ctx) = command_bidi_pipe(qw/cat-file --batch/);
1325 my $list = shift;
1326 foreach my $sha (@{$list}) {
1328 my $first = 1;
1329 my $size = 0;
1330 print $out $sha, "\n";
1331 while (my $line = <$in>) {
1333 if ($first && $line =~ /^[[:xdigit:]]{40}\smissing$/) {
1334 last;
1335 } elsif ($first &&
1336 $line =~ /^[[:xdigit:]]{40}\scommit\s(\d+)$/) {
1337 $first = 0;
1338 $size = $1;
1339 next;
1340 } elsif ($line =~ /^(git-svn-id: )/) {
1341 my (undef, $rev, undef) =
1342 extract_metadata($line);
1343 $s2r{$sha} = $rev;
1344 }
1345 $size -= length($line);
1347 last if ($size == 0);
1348 }
1349 }
1350 command_close_bidi_pipe($pid, $in, $out, $ctx);
1352 return \%s2r;
1354}
1355sub working_head_info {
1357 my ($head, $refs) = @_;
1358 my @args = ('log', '--no-color', '--first-parent', '--pretty=medium');
1359 my ($fh, $ctx) = command_output_pipe(@args, $head);
1360 my $hash;
1361 my %max;
1362 while (<$fh>) {
1363 if ( m{^commit ($::sha1)$} ) {
1364 unshift @$refs, $hash if $hash and $refs;
1365 $hash = $1;
1366 next;
1367 }
1368 next unless s{^\s*(git-svn-id:)}{$1};
1369 my ($url, $rev, $uuid) = extract_metadata($_);
1370 if (defined $url && defined $rev) {
1371 next if $max{$url} and $max{$url} < $rev;
1372 if (my $gs = Git::SVN->find_by_url($url)) {
1373 my $c = $gs->rev_map_get($rev, $uuid);
1374 if ($c && $c eq $hash) {
1375 close $fh; # break the pipe
1376 return ($url, $rev, $uuid, $gs);
1377 } else {
1378 $max{$url} ||= $gs->rev_map_max;
1379 }
1380 }
1381 }
1382 }
1383 command_close_pipe($fh, $ctx);
1384 (undef, undef, undef, undef);
1385}
1386sub read_commit_parents {
1388 my ($parents, $c) = @_;
1389 chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1390 $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1391 @{$parents->{$c}} = split(/ /, $p);
1392}
1393sub linearize_history {
1395 my ($gs, $refs) = @_;
1396 my %parents;
1397 foreach my $c (@$refs) {
1398 read_commit_parents(\%parents, $c);
1399 }
1400 my @linear_refs;
1402 my %skip = ();
1403 my $last_svn_commit = $gs->last_commit;
1404 foreach my $c (reverse @$refs) {
1405 next if $c eq $last_svn_commit;
1406 last if $skip{$c};
1407 unshift @linear_refs, $c;
1409 $skip{$c} = 1;
1410 # we only want the first parent to diff against for linear
1412 # history, we save the rest to inject when we finalize the
1413 # svn commit
1414 my $fp_a = verify_ref("$c~1");
1415 my $fp_b = shift @{$parents{$c}} if $parents{$c};
1416 if (!$fp_a || !$fp_b) {
1417 die "Commit $c\n",
1418 "has no parent commit, and therefore ",
1419 "nothing to diff against.\n",
1420 "You should be working from a repository ",
1421 "originally created by git-svn\n";
1422 }
1423 if ($fp_a ne $fp_b) {
1424 die "$c~1 = $fp_a, however parsing commit $c ",
1425 "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1426 }
1427 foreach my $p (@{$parents{$c}}) {
1429 $skip{$p} = 1;
1430 }
1431 }
1432 (\@linear_refs, \%parents);
1433}
1434sub find_file_type_and_diff_status {
1436 my ($path) = @_;
1437 return ('dir', '') if $path eq '';
1438 my $diff_output =
1440 command_oneline(qw(diff --cached --name-status --), $path) || "";
1441 my $diff_status = (split(' ', $diff_output))[0] || "";
1442 my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1444 return (undef, undef) if !$diff_status && !$ls_tree;
1446 if ($diff_status eq "A") {
1448 return ("link", $diff_status) if -l $path;
1449 return ("dir", $diff_status) if -d $path;
1450 return ("file", $diff_status);
1451 }
1452 my $mode = (split(' ', $ls_tree))[0] || "";
1454 return ("link", $diff_status) if $mode eq "120000";
1456 return ("dir", $diff_status) if $mode eq "040000";
1457 return ("file", $diff_status);
1458}
1459sub md5sum {
1461 my $arg = shift;
1462 my $ref = ref $arg;
1463 my $md5 = Digest::MD5->new();
1464 if ($ref eq 'GLOB' || $ref eq 'IO::File' || $ref eq 'File::Temp') {
1465 $md5->addfile($arg) or croak $!;
1466 } elsif ($ref eq 'SCALAR') {
1467 $md5->add($$arg) or croak $!;
1468 } elsif (!$ref) {
1469 $md5->add($arg) or croak $!;
1470 } else {
1471 ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
1472 }
1473 return $md5->hexdigest();
1474}
1475package Git::SVN;
1477use strict;
1478use warnings;
1479use Fcntl qw/:DEFAULT :seek/;
1480use constant rev_map_fmt => 'NH40';
1481use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1482 $_repack $_repack_flags $_use_svm_props $_head
1483 $_use_svnsync_props $no_reuse_existing $_minimize_url
1484 $_use_log_author $_add_author_from $_localtime/;
1485use Carp qw/croak/;
1486use File::Path qw/mkpath/;
1487use File::Copy qw/copy/;
1488use IPC::Open3;
1489my ($_gc_nr, $_gc_period);
1491# properties that we do not log:
1493my %SKIP_PROP;
1494BEGIN {
1495 %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1496 svn:special svn:executable
1497 svn:entry:committed-rev
1498 svn:entry:last-author
1499 svn:entry:uuid
1500 svn:entry:committed-date/;
1501 # some options are read globally, but can be overridden locally
1503 # per [svn-remote "..."] section. Command-line options will *NOT*
1504 # override options set in an [svn-remote "..."] section
1505 no strict 'refs';
1506 for my $option (qw/follow_parent no_metadata use_svm_props
1507 use_svnsync_props/) {
1508 my $key = $option;
1509 $key =~ tr/_//d;
1510 my $prop = "-$option";
1511 *$option = sub {
1512 my ($self) = @_;
1513 return $self->{$prop} if exists $self->{$prop};
1514 my $k = "svn-remote.$self->{repo_id}.$key";
1515 eval { command_oneline(qw/config --get/, $k) };
1516 if ($@) {
1517 $self->{$prop} = ${"Git::SVN::_$option"};
1518 } else {
1519 my $v = command_oneline(qw/config --bool/,$k);
1520 $self->{$prop} = $v eq 'false' ? 0 : 1;
1521 }
1522 return $self->{$prop};
1523 }
1524 }
1525}
1526my (%LOCKFILES, %INDEX_FILES);
1529END {
1530 unlink keys %LOCKFILES if %LOCKFILES;
1531 unlink keys %INDEX_FILES if %INDEX_FILES;
1532}
1533sub resolve_local_globs {
1535 my ($url, $fetch, $glob_spec) = @_;
1536 return unless defined $glob_spec;
1537 my $ref = $glob_spec->{ref};
1538 my $path = $glob_spec->{path};
1539 foreach (command(qw#for-each-ref --format=%(refname) refs/remotes#)) {
1540 next unless m#^refs/remotes/$ref->{regex}$#;
1541 my $p = $1;
1542 my $pathname = desanitize_refname($path->full_path($p));
1543 my $refname = desanitize_refname($ref->full_path($p));
1544 if (my $existing = $fetch->{$pathname}) {
1545 if ($existing ne $refname) {
1546 die "Refspec conflict:\n",
1547 "existing: refs/remotes/$existing\n",
1548 " globbed: refs/remotes/$refname\n";
1549 }
1550 my $u = (::cmt_metadata("refs/remotes/$refname"))[0];
1551 $u =~ s!^\Q$url\E(/|$)!! or die
1552 "refs/remotes/$refname: '$url' not found in '$u'\n";
1553 if ($pathname ne $u) {
1554 warn "W: Refspec glob conflict ",
1555 "(ref: refs/remotes/$refname):\n",
1556 "expected path: $pathname\n",
1557 " real path: $u\n",
1558 "Continuing ahead with $u\n";
1559 next;
1560 }
1561 } else {
1562 $fetch->{$pathname} = $refname;
1563 }
1564 }
1565}
1566sub parse_revision_argument {
1568 my ($base, $head) = @_;
1569 if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1570 return ($base, $head);
1571 }
1572 return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1573 return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1574 return ($head, $head) if ($::_revision eq 'HEAD');
1575 return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1576 return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1577 die "revision argument: $::_revision not understood by git-svn\n";
1578}
1579sub fetch_all {
1581 my ($repo_id, $remotes) = @_;
1582 if (ref $repo_id) {
1583 my $gs = $repo_id;
1584 $repo_id = undef;
1585 $repo_id = $gs->{repo_id};
1586 }
1587 $remotes ||= read_all_remotes();
1588 my $remote = $remotes->{$repo_id} or
1589 die "[svn-remote \"$repo_id\"] unknown\n";
1590 my $fetch = $remote->{fetch};
1591 my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1592 my (@gs, @globs);
1593 my $ra = Git::SVN::Ra->new($url);
1594 my $uuid = $ra->get_uuid;
1595 my $head = $ra->get_latest_revnum;
1596 my $base = defined $fetch ? $head : 0;
1597 # read the max revs for wildcard expansion (branches/*, tags/*)
1599 foreach my $t (qw/branches tags/) {
1600 defined $remote->{$t} or next;
1601 push @globs, $remote->{$t};
1602 my $max_rev = eval { tmp_config(qw/--int --get/,
1603 "svn-remote.$repo_id.${t}-maxRev") };
1604 if (defined $max_rev && ($max_rev < $base)) {
1605 $base = $max_rev;
1606 } elsif (!defined $max_rev) {
1607 $base = 0;
1608 }
1609 }
1610 if ($fetch) {
1612 foreach my $p (sort keys %$fetch) {
1613 my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1614 my $lr = $gs->rev_map_max;
1615 if (defined $lr) {
1616 $base = $lr if ($lr < $base);
1617 }
1618 push @gs, $gs;
1619 }
1620 }
1621 ($base, $head) = parse_revision_argument($base, $head);
1623 $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1624}
1625sub read_all_remotes {
1627 my $r = {};
1628 my $use_svm_props = eval { command_oneline(qw/config --bool
1629 svn.useSvmProps/) };
1630 $use_svm_props = $use_svm_props eq 'true' if $use_svm_props;
1631 foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1632 if (m!^(.+)\.fetch=\s*(.*)\s*:\s*(.+)\s*$!) {
1633 my ($remote, $local_ref, $_remote_ref) = ($1, $2, $3);
1634 die("svn-remote.$remote: remote ref '$_remote_ref' "
1635 . "must start with 'refs/remotes/'\n")
1636 unless $_remote_ref =~ m{^refs/remotes/(.+)};
1637 my $remote_ref = $1;
1638 $local_ref =~ s{^/}{};
1639 $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1640 $r->{$remote}->{svm} = {} if $use_svm_props;
1641 } elsif (m!^(.+)\.usesvmprops=\s*(.*)\s*$!) {
1642 $r->{$1}->{svm} = {};
1643 } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1644 $r->{$1}->{url} = $2;
1645 } elsif (m!^(.+)\.(branches|tags)=
1646 (.*):refs/remotes/(.+)\s*$/!x) {
1647 my ($p, $g) = ($3, $4);
1648 my $rs = $r->{$1}->{$2} = {
1649 t => $2,
1650 remote => $1,
1651 path => Git::SVN::GlobSpec->new($p),
1652 ref => Git::SVN::GlobSpec->new($g) };
1653 if (length($rs->{ref}->{right}) != 0) {
1654 die "The '*' glob character must be the last ",
1655 "character of '$g'\n";
1656 }
1657 }
1658 }
1659 map {
1661 if (defined $r->{$_}->{svm}) {
1662 my $svm;
1663 eval {
1664 my $section = "svn-remote.$_";
1665 $svm = {
1666 source => tmp_config('--get',
1667 "$section.svm-source"),
1668 replace => tmp_config('--get',
1669 "$section.svm-replace"),
1670 }
1671 };
1672 $r->{$_}->{svm} = $svm;
1673 }
1674 } keys %$r;
1675 $r;
1677}
1678sub init_vars {
1680 $_gc_nr = $_gc_period = 1000;
1681 if (defined $_repack || defined $_repack_flags) {
1682 warn "Repack options are obsolete; they have no effect.\n";
1683 }
1684}
1685sub verify_remotes_sanity {
1687 return unless -d $ENV{GIT_DIR};
1688 my %seen;
1689 foreach (command(qw/config -l/)) {
1690 if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1691 if ($seen{$1}) {
1692 die "Remote ref refs/remote/$1 is tracked by",
1693 "\n \"$_\"\nand\n \"$seen{$1}\"\n",
1694 "Please resolve this ambiguity in ",
1695 "your git configuration file before ",
1696 "continuing\n";
1697 }
1698 $seen{$1} = $_;
1699 }
1700 }
1701}
1702sub find_existing_remote {
1704 my ($url, $remotes) = @_;
1705 return undef if $no_reuse_existing;
1706 my $existing;
1707 foreach my $repo_id (keys %$remotes) {
1708 my $u = $remotes->{$repo_id}->{url} or next;
1709 next if $u ne $url;
1710 $existing = $repo_id;
1711 last;
1712 }
1713 $existing;
1714}
1715sub init_remote_config {
1717 my ($self, $url, $no_write) = @_;
1718 $url =~ s!/+$!!; # strip trailing slash
1719 my $r = read_all_remotes();
1720 my $existing = find_existing_remote($url, $r);
1721 if ($existing) {
1722 unless ($no_write) {
1723 print STDERR "Using existing ",
1724 "[svn-remote \"$existing\"]\n";
1725 }
1726 $self->{repo_id} = $existing;
1727 } elsif ($_minimize_url) {
1728 my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1729 $existing = find_existing_remote($min_url, $r);
1730 if ($existing) {
1731 unless ($no_write) {
1732 print STDERR "Using existing ",
1733 "[svn-remote \"$existing\"]\n";
1734 }
1735 $self->{repo_id} = $existing;
1736 }
1737 if ($min_url ne $url) {
1738 unless ($no_write) {
1739 print STDERR "Using higher level of URL: ",
1740 "$url => $min_url\n";
1741 }
1742 my $old_path = $self->{path};
1743 $self->{path} = $url;
1744 $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1745 if (length $old_path) {
1746 $self->{path} .= "/$old_path";
1747 }
1748 $url = $min_url;
1749 }
1750 }
1751 my $orig_url;
1752 if (!$existing) {
1753 # verify that we aren't overwriting anything:
1754 $orig_url = eval {
1755 command_oneline('config', '--get',
1756 "svn-remote.$self->{repo_id}.url")
1757 };
1758 if ($orig_url && ($orig_url ne $url)) {
1759 die "svn-remote.$self->{repo_id}.url already set: ",
1760 "$orig_url\nwanted to set to: $url\n";
1761 }
1762 }
1763 my ($xrepo_id, $xpath) = find_ref($self->refname);
1764 if (defined $xpath) {
1765 die "svn-remote.$xrepo_id.fetch already set to track ",
1766 "$xpath:refs/remotes/", $self->refname, "\n";
1767 }
1768 unless ($no_write) {
1769 command_noisy('config',
1770 "svn-remote.$self->{repo_id}.url", $url);
1771 $self->{path} =~ s{^/}{};
1772 command_noisy('config', '--add',
1773 "svn-remote.$self->{repo_id}.fetch",
1774 "$self->{path}:".$self->refname);
1775 }
1776 $self->{url} = $url;
1777}
1778sub find_by_url { # repos_root and, path are optional
1780 my ($class, $full_url, $repos_root, $path) = @_;
1781 return undef unless defined $full_url;
1783 remove_username($full_url);
1784 remove_username($repos_root) if defined $repos_root;
1785 my $remotes = read_all_remotes();
1786 if (defined $full_url && defined $repos_root && !defined $path) {
1787 $path = $full_url;
1788 $path =~ s#^\Q$repos_root\E(?:/|$)##;
1789 }
1790 foreach my $repo_id (keys %$remotes) {
1791 my $u = $remotes->{$repo_id}->{url} or next;
1792 remove_username($u);
1793 next if defined $repos_root && $repos_root ne $u;
1794 my $fetch = $remotes->{$repo_id}->{fetch} || {};
1796 foreach (qw/branches tags/) {
1797 resolve_local_globs($u, $fetch,
1798 $remotes->{$repo_id}->{$_});
1799 }
1800 my $p = $path;
1801 my $rwr = rewrite_root({repo_id => $repo_id});
1802 my $svm = $remotes->{$repo_id}->{svm}
1803 if defined $remotes->{$repo_id}->{svm};
1804 unless (defined $p) {
1805 $p = $full_url;
1806 my $z = $u;
1807 my $prefix = '';
1808 if ($rwr) {
1809 $z = $rwr;
1810 remove_username($z);
1811 } elsif (defined $svm) {
1812 $z = $svm->{source};
1813 $prefix = $svm->{replace};
1814 $prefix =~ s#^\Q$u\E(?:/|$)##;
1815 $prefix =~ s#/$##;
1816 }
1817 $p =~ s#^\Q$z\E(?:/|$)#$prefix# or next;
1818 }
1819 foreach my $f (keys %$fetch) {
1820 next if $f ne $p;
1821 return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1822 }
1823 }
1824 undef;
1825}
1826sub init {
1828 my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1829 my $self = _new($class, $repo_id, $ref_id, $path);
1830 if (defined $url) {
1831 $self->init_remote_config($url, $no_write);
1832 }
1833 $self;
1834}
1835sub find_ref {
1837 my ($ref_id) = @_;
1838 foreach (command(qw/config -l/)) {
1839 next unless m!^svn-remote\.(.+)\.fetch=
1840 \s*(.*)\s*:\s*refs/remotes/(.+)\s*$!x;
1841 my ($repo_id, $path, $ref) = ($1, $2, $3);
1842 if ($ref eq $ref_id) {
1843 $path = '' if ($path =~ m#^\./?#);
1844 return ($repo_id, $path);
1845 }
1846 }
1847 (undef, undef, undef);
1848}
1849sub new {
1851 my ($class, $ref_id, $repo_id, $path) = @_;
1852 if (defined $ref_id && !defined $repo_id && !defined $path) {
1853 ($repo_id, $path) = find_ref($ref_id);
1854 if (!defined $repo_id) {
1855 die "Could not find a \"svn-remote.*.fetch\" key ",
1856 "in the repository configuration matching: ",
1857 "refs/remotes/$ref_id\n";
1858 }
1859 }
1860 my $self = _new($class, $repo_id, $ref_id, $path);
1861 if (!defined $self->{path} || !length $self->{path}) {
1862 my $fetch = command_oneline('config', '--get',
1863 "svn-remote.$repo_id.fetch",
1864 ":refs/remotes/$ref_id\$") or
1865 die "Failed to read \"svn-remote.$repo_id.fetch\" ",
1866 "\":refs/remotes/$ref_id\$\" in config\n";
1867 ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
1868 }
1869 $self->{url} = command_oneline('config', '--get',
1870 "svn-remote.$repo_id.url") or
1871 die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
1872 $self->rebuild;
1873 $self;
1874}
1875sub refname {
1877 my ($refname) = "refs/remotes/$_[0]->{ref_id}" ;
1878 # It cannot end with a slash /, we'll throw up on this because
1880 # SVN can't have directories with a slash in their name, either:
1881 if ($refname =~ m{/$}) {
1882 die "ref: '$refname' ends with a trailing slash, this is ",
1883 "not permitted by git nor Subversion\n";
1884 }
1885 # It cannot have ASCII control character space, tilde ~, caret ^,
1887 # colon :, question-mark ?, asterisk *, space, or open bracket [
1888 # anywhere.
1889 #
1890 # Additionally, % must be escaped because it is used for escaping
1891 # and we want our escaped refname to be reversible
1892 $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
1893 # no slash-separated component can begin with a dot .
1895 # /.* becomes /%2E*
1896 $refname =~ s{/\.}{/%2E}g;
1897 # It cannot have two consecutive dots .. anywhere
1899 # .. becomes %2E%2E
1900 $refname =~ s{\.\.}{%2E%2E}g;
1901 return $refname;
1903}
1904sub desanitize_refname {
1906 my ($refname) = @_;
1907 $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
1908 return $refname;
1909}
1910sub svm_uuid {
1912 my ($self) = @_;
1913 return $self->{svm}->{uuid} if $self->svm;
1914 $self->ra;
1915 unless ($self->{svm}) {
1916 die "SVM UUID not cached, and reading remotely failed\n";
1917 }
1918 $self->{svm}->{uuid};
1919}
1920sub svm {
1922 my ($self) = @_;
1923 return $self->{svm} if $self->{svm};
1924 my $svm;
1925 # see if we have it in our config, first:
1926 eval {
1927 my $section = "svn-remote.$self->{repo_id}";
1928 $svm = {
1929 source => tmp_config('--get', "$section.svm-source"),
1930 uuid => tmp_config('--get', "$section.svm-uuid"),
1931 replace => tmp_config('--get', "$section.svm-replace"),
1932 }
1933 };
1934 if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
1935 $self->{svm} = $svm;
1936 }
1937 $self->{svm};
1938}
1939sub _set_svm_vars {
1941 my ($self, $ra) = @_;
1942 return $ra if $self->svm;
1943 my @err = ( "useSvmProps set, but failed to read SVM properties\n",
1945 "(svm:source, svm:uuid) ",
1946 "from the following URLs:\n" );
1947 sub read_svm_props {
1948 my ($self, $ra, $path, $r) = @_;
1949 my $props = ($ra->get_dir($path, $r))[2];
1950 my $src = $props->{'svm:source'};
1951 my $uuid = $props->{'svm:uuid'};
1952 return undef if (!$src || !$uuid);
1953 chomp($src, $uuid);
1955 $uuid =~ m{^[0-9a-f\-]{30,}$}
1957 or die "doesn't look right - svm:uuid is '$uuid'\n";
1958 # the '!' is used to mark the repos_root!/relative/path
1960 $src =~ s{/?!/?}{/};
1961 $src =~ s{/+$}{}; # no trailing slashes please
1962 # username is of no interest
1963 $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
1964 my $replace = $ra->{url};
1966 $replace .= "/$path" if length $path;
1967 my $section = "svn-remote.$self->{repo_id}";
1969 tmp_config("$section.svm-source", $src);
1970 tmp_config("$section.svm-replace", $replace);
1971 tmp_config("$section.svm-uuid", $uuid);
1972 $self->{svm} = {
1973 source => $src,
1974 uuid => $uuid,
1975 replace => $replace
1976 };
1977 }
1978 my $r = $ra->get_latest_revnum;
1980 my $path = $self->{path};
1981 my %tried;
1982 while (length $path) {
1983 unless ($tried{"$self->{url}/$path"}) {
1984 return $ra if $self->read_svm_props($ra, $path, $r);
1985 $tried{"$self->{url}/$path"} = 1;
1986 }
1987 $path =~ s#/?[^/]+$##;
1988 }
1989 die "Path: '$path' should be ''\n" if $path ne '';
1990 return $ra if $self->read_svm_props($ra, $path, $r);
1991 $tried{"$self->{url}/$path"} = 1;
1992 if ($ra->{repos_root} eq $self->{url}) {
1994 die @err, (map { " $_\n" } keys %tried), "\n";
1995 }
1996 # nope, make sure we're connected to the repository root:
1998 my $ok;
1999 my @tried_b;
2000 $path = $ra->{svn_path};
2001 $ra = Git::SVN::Ra->new($ra->{repos_root});
2002 while (length $path) {
2003 unless ($tried{"$ra->{url}/$path"}) {
2004 $ok = $self->read_svm_props($ra, $path, $r);
2005 last if $ok;
2006 $tried{"$ra->{url}/$path"} = 1;
2007 }
2008 $path =~ s#/?[^/]+$##;
2009 }
2010 die "Path: '$path' should be ''\n" if $path ne '';
2011 $ok ||= $self->read_svm_props($ra, $path, $r);
2012 $tried{"$ra->{url}/$path"} = 1;
2013 if (!$ok) {
2014 die @err, (map { " $_\n" } keys %tried), "\n";
2015 }
2016 Git::SVN::Ra->new($self->{url});
2017}
2018sub svnsync {
2020 my ($self) = @_;
2021 return $self->{svnsync} if $self->{svnsync};
2022 if ($self->no_metadata) {
2024 die "Can't have both 'noMetadata' and ",
2025 "'useSvnsyncProps' options set!\n";
2026 }
2027 if ($self->rewrite_root) {
2028 die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
2029 "options set!\n";
2030 }
2031 my $svnsync;
2033 # see if we have it in our config, first:
2034 eval {
2035 my $section = "svn-remote.$self->{repo_id}";
2036 my $url = tmp_config('--get', "$section.svnsync-url");
2038 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
2039 die "doesn't look right - svn:sync-from-url is '$url'\n";
2040 my $uuid = tmp_config('--get', "$section.svnsync-uuid");
2042 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
2043 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
2044 $svnsync = { url => $url, uuid => $uuid }
2046 };
2047 if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
2048 return $self->{svnsync} = $svnsync;
2049 }
2050 my $err = "useSvnsyncProps set, but failed to read " .
2052 "svnsync property: svn:sync-from-";
2053 my $rp = $self->ra->rev_proplist(0);
2054 my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
2056 ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
2057 die "doesn't look right - svn:sync-from-url is '$url'\n";
2058 my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
2060 ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
2061 die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
2062 my $section = "svn-remote.$self->{repo_id}";
2064 tmp_config('--add', "$section.svnsync-uuid", $uuid);
2065 tmp_config('--add', "$section.svnsync-url", $url);
2066 return $self->{svnsync} = { url => $url, uuid => $uuid };
2067}
2068# this allows us to memoize our SVN::Ra UUID locally and avoid a
2070# remote lookup (useful for 'git svn log').
2071sub ra_uuid {
2072 my ($self) = @_;
2073 unless ($self->{ra_uuid}) {
2074 my $key = "svn-remote.$self->{repo_id}.uuid";
2075 my $uuid = eval { tmp_config('--get', $key) };
2076 if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/) {
2077 $self->{ra_uuid} = $uuid;
2078 } else {
2079 die "ra_uuid called without URL\n" unless $self->{url};
2080 $self->{ra_uuid} = $self->ra->get_uuid;
2081 tmp_config('--add', $key, $self->{ra_uuid});
2082 }
2083 }
2084 $self->{ra_uuid};
2085}
2086sub _set_repos_root {
2088 my ($self, $repos_root) = @_;
2089 my $k = "svn-remote.$self->{repo_id}.reposRoot";
2090 $repos_root ||= $self->ra->{repos_root};
2091 tmp_config($k, $repos_root);
2092 $repos_root;
2093}
2094sub repos_root {
2096 my ($self) = @_;
2097 my $k = "svn-remote.$self->{repo_id}.reposRoot";
2098 eval { tmp_config('--get', $k) } || $self->_set_repos_root;
2099}
2100sub ra {
2102 my ($self) = shift;
2103 my $ra = Git::SVN::Ra->new($self->{url});
2104 $self->_set_repos_root($ra->{repos_root});
2105 if ($self->use_svm_props && !$self->{svm}) {
2106 if ($self->no_metadata) {
2107 die "Can't have both 'noMetadata' and ",
2108 "'useSvmProps' options set!\n";
2109 } elsif ($self->use_svnsync_props) {
2110 die "Can't have both 'useSvnsyncProps' and ",
2111 "'useSvmProps' options set!\n";
2112 }
2113 $ra = $self->_set_svm_vars($ra);
2114 $self->{-want_revprops} = 1;
2115 }
2116 $ra;
2117}
2118sub rel_path {
2120 my ($self) = @_;
2121 my $repos_root = $self->ra->{repos_root};
2122 return $self->{path} if ($self->{url} eq $repos_root);
2123 my $url = $self->{url} .
2124 (length $self->{path} ? "/$self->{path}" : $self->{path});
2125 $url =~ s!^\Q$repos_root\E(?:/+|$)!!g;
2126 $url;
2127}
2128# prop_walk(PATH, REV, SUB)
2130# -------------------------
2131# Recursively traverse PATH at revision REV and invoke SUB for each
2132# directory that contains a SVN property. SUB will be invoked as
2133# follows: &SUB(gs, path, props); where `gs' is this instance of
2134# Git::SVN, `path' the path to the directory where the properties
2135# `props' were found. The `path' will be relative to point of checkout,
2136# that is, if url://repo/trunk is the current Git branch, and that
2137# directory contains a sub-directory `d', SUB will be invoked with `/d/'
2138# as `path' (note the trailing `/').
2139sub prop_walk {
2140 my ($self, $path, $rev, $sub) = @_;
2141 $path =~ s#^/##;
2143 my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
2144 $path =~ s#^/*#/#g;
2145 my $p = $path;
2146 # Strip the irrelevant part of the path.
2147 $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
2148 # Ensure the path is terminated by a `/'.
2149 $p =~ s#/*$#/#;
2150 # The properties contain all the internal SVN stuff nobody
2152 # (usually) cares about.
2153 my $interesting_props = 0;
2154 foreach (keys %{$props}) {
2155 # If it doesn't start with `svn:', it must be a
2156 # user-defined property.
2157 ++$interesting_props and next if $_ !~ /^svn:/;
2158 # FIXME: Fragile, if SVN adds new public properties,
2159 # this needs to be updated.
2160 ++$interesting_props if /^svn:(?:ignore|keywords|executable
2161 |eol-style|mime-type
2162 |externals|needs-lock)$/x;
2163 }
2164 &$sub($self, $p, $props) if $interesting_props;
2165 foreach (sort keys %$dirent) {
2167 next if $dirent->{$_}->{kind} != $SVN::Node::dir;
2168 $self->prop_walk($self->{path} . $p . $_, $rev, $sub);
2169 }
2170}
2171sub last_rev { ($_[0]->last_rev_commit)[0] }
2173sub last_commit { ($_[0]->last_rev_commit)[1] }
2174# returns the newest SVN revision number and newest commit SHA1
2176sub last_rev_commit {
2177 my ($self) = @_;
2178 if (defined $self->{last_rev} && defined $self->{last_commit}) {
2179 return ($self->{last_rev}, $self->{last_commit});
2180 }
2181 my $c = ::verify_ref($self->refname.'^0');
2182 if ($c && !$self->use_svm_props && !$self->no_metadata) {
2183 my $rev = (::cmt_metadata($c))[1];
2184 if (defined $rev) {
2185 ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
2186 return ($rev, $c);
2187 }
2188 }
2189 my $map_path = $self->map_path;
2190 unless (-e $map_path) {
2191 ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
2192 return (undef, undef);
2193 }
2194 my ($rev, $commit) = $self->rev_map_max(1);
2195 ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
2196 return ($rev, $commit);
2197}
2198sub get_fetch_range {
2200 my ($self, $min, $max) = @_;
2201 $max ||= $self->ra->get_latest_revnum;
2202 $min ||= $self->rev_map_max;
2203 (++$min, $max);
2204}
2205sub tmp_config {
2207 my (@args) = @_;
2208 my $old_def_config = "$ENV{GIT_DIR}/svn/config";
2209 my $config = "$ENV{GIT_DIR}/svn/.metadata";
2210 if (! -f $config && -f $old_def_config) {
2211 rename $old_def_config, $config or
2212 die "Failed rename $old_def_config => $config: $!\n";
2213 }
2214 my $old_config = $ENV{GIT_CONFIG};
2215 $ENV{GIT_CONFIG} = $config;
2216 $@ = undef;
2217 my @ret = eval {
2218 unless (-f $config) {
2219 mkfile($config);
2220 open my $fh, '>', $config or
2221 die "Can't open $config: $!\n";
2222 print $fh "; This file is used internally by ",
2223 "git-svn\n" or die
2224 "Couldn't write to $config: $!\n";
2225 print $fh "; You should not have to edit it\n" or
2226 die "Couldn't write to $config: $!\n";
2227 close $fh or die "Couldn't close $config: $!\n";
2228 }
2229 command('config', @args);
2230 };
2231 my $err = $@;
2232 if (defined $old_config) {
2233 $ENV{GIT_CONFIG} = $old_config;
2234 } else {
2235 delete $ENV{GIT_CONFIG};
2236 }
2237 die $err if $err;
2238 wantarray ? @ret : $ret[0];
2239}
2240sub tmp_index_do {
2242 my ($self, $sub) = @_;
2243 my $old_index = $ENV{GIT_INDEX_FILE};
2244 $ENV{GIT_INDEX_FILE} = $self->{index};
2245 $@ = undef;
2246 my @ret = eval {
2247 my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
2248 mkpath([$dir]) unless -d $dir;
2249 &$sub;
2250 };
2251 my $err = $@;
2252 if (defined $old_index) {
2253 $ENV{GIT_INDEX_FILE} = $old_index;
2254 } else {
2255 delete $ENV{GIT_INDEX_FILE};
2256 }
2257 die $err if $err;
2258 wantarray ? @ret : $ret[0];
2259}
2260sub assert_index_clean {
2262 my ($self, $treeish) = @_;
2263 $self->tmp_index_do(sub {
2265 command_noisy('read-tree', $treeish) unless -e $self->{index};
2266 my $x = command_oneline('write-tree');
2267 my ($y) = (command(qw/cat-file commit/, $treeish) =~
2268 /^tree ($::sha1)/mo);
2269 return if $y eq $x;
2270 warn "Index mismatch: $y != $x\nrereading $treeish\n";
2272 unlink $self->{index} or die "unlink $self->{index}: $!\n";
2273 command_noisy('read-tree', $treeish);
2274 $x = command_oneline('write-tree');
2275 if ($y ne $x) {
2276 ::fatal "trees ($treeish) $y != $x\n",
2277 "Something is seriously wrong...";
2278 }
2279 });
2280}
2281sub get_commit_parents {
2283 my ($self, $log_entry) = @_;
2284 my (%seen, @ret, @tmp);
2285 # legacy support for 'set-tree'; this is only used by set_tree_cb:
2286 if (my $ip = $self->{inject_parents}) {
2287 if (my $commit = delete $ip->{$log_entry->{revision}}) {
2288 push @tmp, $commit;
2289 }
2290 }
2291 if (my $cur = ::verify_ref($self->refname.'^0')) {
2292 push @tmp, $cur;
2293 }
2294 if (my $ipd = $self->{inject_parents_dcommit}) {
2295 if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2296 push @tmp, @$commit;
2297 }
2298 }
2299 push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2300 while (my $p = shift @tmp) {
2301 next if $seen{$p};
2302 $seen{$p} = 1;
2303 push @ret, $p;
2304 # MAXPARENT is defined to 16 in commit-tree.c:
2305 last if @ret >= 16;
2306 }
2307 if (@tmp) {
2308 die "r$log_entry->{revision}: No room for parents:\n\t",
2309 join("\n\t", @tmp), "\n";
2310 }
2311 @ret;
2312}
2313sub rewrite_root {
2315 my ($self) = @_;
2316 return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2317 my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2318 my $rwr = eval { command_oneline(qw/config --get/, $k) };
2319 if ($rwr) {
2320 $rwr =~ s#/+$##;
2321 if ($rwr !~ m#^[a-z\+]+://#) {
2322 die "$rwr is not a valid URL (key: $k)\n";
2323 }
2324 }
2325 $self->{-rewrite_root} = $rwr;
2326}
2327sub metadata_url {
2329 my ($self) = @_;
2330 ($self->rewrite_root || $self->{url}) .
2331 (length $self->{path} ? '/' . $self->{path} : '');
2332}
2333sub full_url {
2335 my ($self) = @_;
2336 $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2337}
2338sub set_commit_header_env {
2341 my ($log_entry) = @_;
2342 my %env;
2343 foreach my $ned (qw/NAME EMAIL DATE/) {
2344 foreach my $ac (qw/AUTHOR COMMITTER/) {
2345 $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2346 }
2347 }
2348 $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2350 $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2351 $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2352 $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2354 ? $log_entry->{commit_name}
2355 : $log_entry->{name};
2356 $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2357 ? $log_entry->{commit_email}
2358 : $log_entry->{email};
2359 \%env;
2360}
2361sub restore_commit_header_env {
2363 my ($env) = @_;
2364 foreach my $ned (qw/NAME EMAIL DATE/) {
2365 foreach my $ac (qw/AUTHOR COMMITTER/) {
2366 my $k = "GIT_${ac}_${ned}";
2367 if (defined $env->{$k}) {
2368 $ENV{$k} = $env->{$k};
2369 } else {
2370 delete $ENV{$k};
2371 }
2372 }
2373 }
2374}
2375sub gc {
2377 command_noisy('gc', '--auto');
2378};
2379sub do_git_commit {
2381 my ($self, $log_entry) = @_;
2382 my $lr = $self->last_rev;
2383 if (defined $lr && $lr >= $log_entry->{revision}) {
2384 die "Last fetched revision of ", $self->refname,
2385 " was r$lr, but we are about to fetch: ",
2386 "r$log_entry->{revision}!\n";
2387 }
2388 if (my $c = $self->rev_map_get($log_entry->{revision})) {
2389 croak "$log_entry->{revision} = $c already exists! ",
2390 "Why are we refetching it?\n";
2391 }
2392 my $old_env = set_commit_header_env($log_entry);
2393 my $tree = $log_entry->{tree};
2394 if (!defined $tree) {
2395 $tree = $self->tmp_index_do(sub {
2396 command_oneline('write-tree') });
2397 }
2398 die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2399 my @exec = ('git', 'commit-tree', $tree);
2401 foreach ($self->get_commit_parents($log_entry)) {
2402 push @exec, '-p', $_;
2403 }
2404 defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2405 or croak $!;
2406 binmode $msg_fh;
2407 # we always get UTF-8 from SVN, but we may want our commits in
2409 # a different encoding.
2410 if (my $enc = Git::config('i18n.commitencoding')) {
2411 require Encode;
2412 Encode::from_to($log_entry->{log}, 'UTF-8', $enc);
2413 }
2414 print $msg_fh $log_entry->{log} or croak $!;
2415 restore_commit_header_env($old_env);
2416 unless ($self->no_metadata) {
2417 print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2418 or croak $!;
2419 }
2420 $msg_fh->flush == 0 or croak $!;
2421 close $msg_fh or croak $!;
2422 chomp(my $commit = do { local $/; <$out_fh> });
2423 close $out_fh or croak $!;
2424 waitpid $pid, 0;
2425 croak $? if $?;
2426 if ($commit !~ /^$::sha1$/o) {
2427 die "Failed to commit, invalid sha1: $commit\n";
2428 }
2429 $self->rev_map_set($log_entry->{revision}, $commit, 1);
2431 $self->{last_rev} = $log_entry->{revision};
2433 $self->{last_commit} = $commit;
2434 print "r$log_entry->{revision}" unless $::_q > 1;
2435 if (defined $log_entry->{svm_revision}) {
2436 print " (\@$log_entry->{svm_revision})" unless $::_q > 1;
2437 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2438 0, $self->svm_uuid);
2439 }
2440 print " = $commit ($self->{ref_id})\n" unless $::_q > 1;
2441 if (--$_gc_nr == 0) {
2442 $_gc_nr = $_gc_period;
2443 gc();
2444 }
2445 return $commit;
2446}
2447sub match_paths {
2449 my ($self, $paths, $r) = @_;
2450 return 1 if $self->{path} eq '';
2451 if (my $path = $paths->{"/$self->{path}"}) {
2452 return ($path->{action} eq 'D') ? 0 : 1;
2453 }
2454 my $repos_root = $self->ra->{repos_root};
2455 my $extended_path = $self->{url} . '/' . $self->{path};
2456 $extended_path =~ s#^\Q$repos_root\E(/|$)##;
2457 $self->{path_regex} ||= qr/^\/\Q$extended_path\E\//;
2458 if (grep /$self->{path_regex}/, keys %$paths) {
2459 return 1;
2460 }
2461 my $c = '';
2462 foreach (split m#/#, $self->{path}) {
2463 $c .= "/$_";
2464 next unless ($paths->{$c} &&
2465 ($paths->{$c}->{action} =~ /^[AR]$/));
2466 if ($self->ra->check_path($self->{path}, $r) ==
2467 $SVN::Node::dir) {
2468 return 1;
2469 }
2470 }
2471 return 0;
2472}
2473sub find_parent_branch {
2475 my ($self, $paths, $rev) = @_;
2476 return undef unless $self->follow_parent;
2477 unless (defined $paths) {
2478 my $err_handler = $SVN::Error::handler;
2479 $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2480 $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1, sub {
2481 $paths =
2482 Git::SVN::Ra::dup_changed_paths($_[0]) });
2483 $SVN::Error::handler = $err_handler;
2484 }
2485 return undef unless defined $paths;
2486 # look for a parent from another branch:
2488 my @b_path_components = split m#/#, $self->rel_path;
2489 my @a_path_components;
2490 my $i;
2491 while (@b_path_components) {
2492 $i = $paths->{'/'.join('/', @b_path_components)};
2493 last if $i && defined $i->{copyfrom_path};
2494 unshift(@a_path_components, pop(@b_path_components));
2495 }
2496 return undef unless defined $i && defined $i->{copyfrom_path};
2497 my $branch_from = $i->{copyfrom_path};
2498 if (@a_path_components) {
2499 print STDERR "branch_from: $branch_from => ";
2500 $branch_from .= '/'.join('/', @a_path_components);
2501 print STDERR $branch_from, "\n";
2502 }
2503 my $r = $i->{copyfrom_rev};
2504 my $repos_root = $self->ra->{repos_root};
2505 my $url = $self->ra->{url};
2506 my $new_url = $repos_root . $branch_from;
2507 print STDERR "Found possible branch point: ",
2508 "$new_url => ", $self->full_url, ", $r\n";
2509 $branch_from =~ s#^/##;
2510 my $gs = $self->other_gs($new_url, $url, $repos_root,
2511 $branch_from, $r, $self->{ref_id});
2512 my ($r0, $parent) = $gs->find_rev_before($r, 1);
2513 {
2514 my ($base, $head);
2515 if (!defined $r0 || !defined $parent) {
2516 ($base, $head) = parse_revision_argument(0, $r);
2517 } else {
2518 if ($r0 < $r) {
2519 $gs->ra->get_log([$gs->{path}], $r0 + 1, $r, 1,
2520 0, 1, sub { $base = $_[1] - 1 });
2521 }
2522 }
2523 if (defined $base && $base <= $r) {
2524 $gs->fetch($base, $r);
2525 }
2526 ($r0, $parent) = $gs->find_rev_before($r, 1);
2527 }
2528 if (defined $r0 && defined $parent) {
2529 print STDERR "Found branch parent: ($self->{ref_id}) $parent\n";
2530 my $ed;
2531 if ($self->ra->can_do_switch) {
2532 $self->assert_index_clean($parent);
2533 print STDERR "Following parent with do_switch\n";
2534 # do_switch works with svn/trunk >= r22312, but that
2535 # is not included with SVN 1.4.3 (the latest version
2536 # at the moment), so we can't rely on it
2537 $self->{last_rev} = $r0;
2538 $self->{last_commit} = $parent;
2539 $ed = SVN::Git::Fetcher->new($self, $gs->{path});
2540 $gs->ra->gs_do_switch($r0, $rev, $gs,
2541 $self->full_url, $ed)
2542 or die "SVN connection failed somewhere...\n";
2543 } elsif ($self->ra->trees_match($new_url, $r0,
2544 $self->full_url, $rev)) {
2545 print STDERR "Trees match:\n",
2546 " $new_url\@$r0\n",
2547 " ${\$self->full_url}\@$rev\n",
2548 "Following parent with no changes\n";
2549 $self->tmp_index_do(sub {
2550 command_noisy('read-tree', $parent);
2551 });
2552 $self->{last_commit} = $parent;
2553 } else {
2554 print STDERR "Following parent with do_update\n";
2555 $ed = SVN::Git::Fetcher->new($self);
2556 $self->ra->gs_do_update($rev, $rev, $self, $ed)
2557 or die "SVN connection failed somewhere...\n";
2558 }
2559 print STDERR "Successfully followed parent\n";
2560 return $self->make_log_entry($rev, [$parent], $ed);
2561 }
2562 return undef;
2563}
2564sub do_fetch {
2566 my ($self, $paths, $rev) = @_;
2567 my $ed;
2568 my ($last_rev, @parents);
2569 if (my $lc = $self->last_commit) {
2570 # we can have a branch that was deleted, then re-added
2571 # under the same name but copied from another path, in
2572 # which case we'll have multiple parents (we don't
2573 # want to break the original ref, nor lose copypath info):
2574 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2575 push @{$log_entry->{parents}}, $lc;
2576 return $log_entry;
2577 }
2578 $ed = SVN::Git::Fetcher->new($self);
2579 $last_rev = $self->{last_rev};
2580 $ed->{c} = $lc;
2581 @parents = ($lc);
2582 } else {
2583 $last_rev = $rev;
2584 if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2585 return $log_entry;
2586 }
2587 $ed = SVN::Git::Fetcher->new($self);
2588 }
2589 unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2590 die "SVN connection failed somewhere...\n";
2591 }
2592 $self->make_log_entry($rev, \@parents, $ed);
2593}
2594sub get_untracked {
2596 my ($self, $ed) = @_;
2597 my @out;
2598 my $h = $ed->{empty};
2599 foreach (sort keys %$h) {
2600 my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2601 push @out, " $act: " . uri_encode($_);
2602 warn "W: $act: $_\n";
2603 }
2604 foreach my $t (qw/dir_prop file_prop/) {
2605 $h = $ed->{$t} or next;
2606 foreach my $path (sort keys %$h) {
2607 my $ppath = $path eq '' ? '.' : $path;
2608 foreach my $prop (sort keys %{$h->{$path}}) {
2609 next if $SKIP_PROP{$prop};
2610 my $v = $h->{$path}->{$prop};
2611 my $t_ppath_prop = "$t: " .
2612 uri_encode($ppath) . ' ' .
2613 uri_encode($prop);
2614 if (defined $v) {
2615 push @out, " +$t_ppath_prop " .
2616 uri_encode($v);
2617 } else {
2618 push @out, " -$t_ppath_prop";
2619 }
2620 }
2621 }
2622 }
2623 foreach my $t (qw/absent_file absent_directory/) {
2624 $h = $ed->{$t} or next;
2625 foreach my $parent (sort keys %$h) {
2626 foreach my $path (sort @{$h->{$parent}}) {
2627 push @out, " $t: " .
2628 uri_encode("$parent/$path");
2629 warn "W: $t: $parent/$path ",
2630 "Insufficient permissions?\n";
2631 }
2632 }
2633 }
2634 \@out;
2635}
2636# parse_svn_date(DATE)
2638# --------------------
2639# Given a date (in UTC) from Subversion, return a string in the format
2640# "<TZ Offset> <local date/time>" that Git will use.
2641#
2642# By default the parsed date will be in UTC; if $Git::SVN::_localtime
2643# is true we'll convert it to the local timezone instead.
2644sub parse_svn_date {
2645 my $date = shift || return '+0000 1970-01-01 00:00:00';
2646 my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2647 (\d\d)\:(\d\d)\:(\d\d)\.\d*Z$/x) or
2648 croak "Unable to parse date: $date\n";
2649 my $parsed_date; # Set next.
2650 if ($Git::SVN::_localtime) {
2652 # Translate the Subversion datetime to an epoch time.
2653 # Begin by switching ourselves to $date's timezone, UTC.
2654 my $old_env_TZ = $ENV{TZ};
2655 $ENV{TZ} = 'UTC';
2656 my $epoch_in_UTC =
2658 POSIX::strftime('%s', $S, $M, $H, $d, $m - 1, $Y - 1900);
2659 # Determine our local timezone (including DST) at the
2661 # time of $epoch_in_UTC. $Git::SVN::Log::TZ stored the
2662 # value of TZ, if any, at the time we were run.
2663 if (defined $Git::SVN::Log::TZ) {
2664 $ENV{TZ} = $Git::SVN::Log::TZ;
2665 } else {
2666 delete $ENV{TZ};
2667 }
2668 my $our_TZ =
2670 POSIX::strftime('%Z', $S, $M, $H, $d, $m - 1, $Y - 1900);
2671 # This converts $epoch_in_UTC into our local timezone.
2673 my ($sec, $min, $hour, $mday, $mon, $year,
2674 $wday, $yday, $isdst) = localtime($epoch_in_UTC);
2675 $parsed_date = sprintf('%s %04d-%02d-%02d %02d:%02d:%02d',
2677 $our_TZ, $year + 1900, $mon + 1,
2678 $mday, $hour, $min, $sec);
2679 # Reset us to the timezone in effect when we entered
2681 # this routine.
2682 if (defined $old_env_TZ) {
2683 $ENV{TZ} = $old_env_TZ;
2684 } else {
2685 delete $ENV{TZ};
2686 }
2687 } else {
2688 $parsed_date = "+0000 $Y-$m-$d $H:$M:$S";
2689 }
2690 return $parsed_date;
2692}
2693sub other_gs {
2695 my ($self, $new_url, $url, $repos_root,
2696 $branch_from, $r, $old_ref_id) = @_;
2697 my $gs = Git::SVN->find_by_url($new_url, $repos_root, $branch_from);
2698 unless ($gs) {
2699 my $ref_id = $old_ref_id;
2700 $ref_id =~ s/\@\d+$//;
2701 $ref_id .= "\@$r";
2702 # just grow a tail if we're not unique enough :x
2703 $ref_id .= '-' while find_ref($ref_id);
2704 print STDERR "Initializing parent: $ref_id\n";
2705 my ($u, $p, $repo_id) = ($new_url, '', $ref_id);
2706 if ($u =~ s#^\Q$url\E(/|$)##) {
2707 $p = $u;
2708 $u = $url;
2709 $repo_id = $self->{repo_id};
2710 }
2711 $gs = Git::SVN->init($u, $p, $repo_id, $ref_id, 1);
2712 }
2713 $gs
2714}
2715sub call_authors_prog {
2717 my ($orig_author) = @_;
2718 my $author = `$::_authors_prog $orig_author`;
2719 if ($? != 0) {
2720 die "$::_authors_prog failed with exit code $?\n"
2721 }
2722 if ($author =~ /^\s*(.+?)\s*<(.*)>\s*$/) {
2723 my ($name, $email) = ($1, $2);
2724 $email = undef if length $2 == 0;
2725 return [$name, $email];
2726 } else {
2727 die "Author: $orig_author: $::_authors_prog returned "
2728 . "invalid author format: $author\n";
2729 }
2730}
2731sub check_author {
2733 my ($author) = @_;
2734 if (!defined $author || length $author == 0) {
2735 $author = '(no author)';
2736 }
2737 if (!defined $::users{$author}) {
2738 if (defined $::_authors_prog) {
2739 $::users{$author} = call_authors_prog($author);
2740 } elsif (defined $::_authors) {
2741 die "Author: $author not defined in $::_authors file\n";
2742 }
2743 }
2744 $author;
2745}
2746sub make_log_entry {
2748 my ($self, $rev, $parents, $ed) = @_;
2749 my $untracked = $self->get_untracked($ed);
2750 open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
2752 print $un "r$rev\n" or croak $!;
2753 print $un $_, "\n" foreach @$untracked;
2754 my %log_entry = ( parents => $parents || [], revision => $rev,
2755 log => '');
2756 my $headrev;
2758 my $logged = delete $self->{logged_rev_props};
2759 if (!$logged || $self->{-want_revprops}) {
2760 my $rp = $self->ra->rev_proplist($rev);
2761 foreach (sort keys %$rp) {
2762 my $v = $rp->{$_};
2763 if (/^svn:(author|date|log)$/) {
2764 $log_entry{$1} = $v;
2765 } elsif ($_ eq 'svm:headrev') {
2766 $headrev = $v;
2767 } else {
2768 print $un " rev_prop: ", uri_encode($_), ' ',
2769 uri_encode($v), "\n";
2770 }
2771 }
2772 } else {
2773 map { $log_entry{$_} = $logged->{$_} } keys %$logged;
2774 }
2775 close $un or croak $!;
2776 $log_entry{date} = parse_svn_date($log_entry{date});
2778 $log_entry{log} .= "\n";
2779 my $author = $log_entry{author} = check_author($log_entry{author});
2780 my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
2781 : ($author, undef);
2782 my ($commit_name, $commit_email) = ($name, $email);
2784 if ($_use_log_author) {
2785 my $name_field;
2786 if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
2787 $name_field = $1;
2788 } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
2789 $name_field = $1;
2790 }
2791 if (!defined $name_field) {
2792 if (!defined $email) {
2793 $email = $name;
2794 }
2795 } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
2796 ($name, $email) = ($1, $2);
2797 } elsif ($name_field =~ /(.*)@/) {
2798 ($name, $email) = ($1, $name_field);
2799 } else {
2800 ($name, $email) = ($name_field, $name_field);
2801 }
2802 }
2803 if (defined $headrev && $self->use_svm_props) {
2804 if ($self->rewrite_root) {
2805 die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
2806 "options set!\n";
2807 }
2808 my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$};
2809 # we don't want "SVM: initializing mirror for junk" ...
2810 return undef if $r == 0;
2811 my $svm = $self->svm;
2812 if ($uuid ne $svm->{uuid}) {
2813 die "UUID mismatch on SVM path:\n",
2814 "expected: $svm->{uuid}\n",
2815 " got: $uuid\n";
2816 }
2817 my $full_url = $self->full_url;
2818 $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
2819 die "Failed to replace '$svm->{replace}' with ",
2820 "'$svm->{source}' in $full_url\n";
2821 # throw away username for storing in records
2822 remove_username($full_url);
2823 $log_entry{metadata} = "$full_url\@$r $uuid";
2824 $log_entry{svm_revision} = $r;
2825 $email ||= "$author\@$uuid";
2826 $commit_email ||= "$author\@$uuid";
2827 } elsif ($self->use_svnsync_props) {
2828 my $full_url = $self->svnsync->{url};
2829 $full_url .= "/$self->{path}" if length $self->{path};
2830 remove_username($full_url);
2831 my $uuid = $self->svnsync->{uuid};
2832 $log_entry{metadata} = "$full_url\@$rev $uuid";
2833 $email ||= "$author\@$uuid";
2834 $commit_email ||= "$author\@$uuid";
2835 } else {
2836 my $url = $self->metadata_url;
2837 remove_username($url);
2838 $log_entry{metadata} = "$url\@$rev " .
2839 $self->ra->get_uuid;
2840 $email ||= "$author\@" . $self->ra->get_uuid;
2841 $commit_email ||= "$author\@" . $self->ra->get_uuid;
2842 }
2843 $log_entry{name} = $name;
2844 $log_entry{email} = $email;
2845 $log_entry{commit_name} = $commit_name;
2846 $log_entry{commit_email} = $commit_email;
2847 \%log_entry;
2848}
2849sub fetch {
2851 my ($self, $min_rev, $max_rev, @parents) = @_;
2852 my ($last_rev, $last_commit) = $self->last_rev_commit;
2853 my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
2854 $self->ra->gs_fetch_loop_common($base, $head, [$self]);
2855}
2856sub set_tree_cb {
2858 my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
2859 $self->{inject_parents} = { $rev => $tree };
2860 $self->fetch(undef, undef);
2861}
2862sub set_tree {
2864 my ($self, $tree) = (shift, shift);
2865 my $log_entry = ::get_commit_entry($tree);
2866 unless ($self->{last_rev}) {
2867 ::fatal("Must have an existing revision to commit");
2868 }
2869 my %ed_opts = ( r => $self->{last_rev},
2870 log => $log_entry->{log},
2871 ra => $self->ra,
2872 tree_a => $self->{last_commit},
2873 tree_b => $tree,
2874 editor_cb => sub {
2875 $self->set_tree_cb($log_entry, $tree, @_) },
2876 svn_path => $self->{path} );
2877 if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
2878 print "No changes\nr$self->{last_rev} = $tree\n";
2879 }
2880}
2881sub rebuild_from_rev_db {
2883 my ($self, $path) = @_;
2884 my $r = -1;
2885 open my $fh, '<', $path or croak "open: $!";
2886 binmode $fh or croak "binmode: $!";
2887 while (<$fh>) {
2888 length($_) == 41 or croak "inconsistent size in ($_) != 41";
2889 chomp($_);
2890 ++$r;
2891 next if $_ eq ('0' x 40);
2892 $self->rev_map_set($r, $_);
2893 print "r$r = $_\n";
2894 }
2895 close $fh or croak "close: $!";
2896 unlink $path or croak "unlink: $!";
2897}
2898sub rebuild {
2900 my ($self) = @_;
2901 my $map_path = $self->map_path;
2902 my $partial = (-e $map_path && ! -z $map_path);
2903 return unless ::verify_ref($self->refname.'^0');
2904 if (!$partial && ($self->use_svm_props || $self->no_metadata)) {
2905 my $rev_db = $self->rev_db_path;
2906 $self->rebuild_from_rev_db($rev_db);
2907 if ($self->use_svm_props) {
2908 my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
2909 $self->rebuild_from_rev_db($svm_rev_db);
2910 }
2911 $self->unlink_rev_db_symlink;
2912 return;
2913 }
2914 print "Rebuilding $map_path ...\n" if (!$partial);
2915 my ($base_rev, $head) = ($partial ? $self->rev_map_max_norebuild(1) :
2916 (undef, undef));
2917 my ($log, $ctx) =
2918 command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
2919 ($head ? "$head.." : "") . $self->refname,
2920 '--');
2921 my $metadata_url = $self->metadata_url;
2922 remove_username($metadata_url);
2923 my $svn_uuid = $self->ra_uuid;
2924 my $c;
2925 while (<$log>) {
2926 if ( m{^commit ($::sha1)$} ) {
2927 $c = $1;
2928 next;
2929 }
2930 next unless s{^\s*(git-svn-id:)}{$1};
2931 my ($url, $rev, $uuid) = ::extract_metadata($_);
2932 remove_username($url);
2933 # ignore merges (from set-tree)
2935 next if (!defined $rev || !$uuid);
2936 # if we merged or otherwise started elsewhere, this is
2938 # how we break out of it
2939 if (($uuid ne $svn_uuid) ||
2940 ($metadata_url && $url && ($url ne $metadata_url))) {
2941 next;
2942 }
2943 if ($partial && $head) {
2944 print "Partial-rebuilding $map_path ...\n";
2945 print "Currently at $base_rev = $head\n";
2946 $head = undef;
2947 }
2948 $self->rev_map_set($rev, $c);
2950 print "r$rev = $c\n";
2951 }
2952 command_close_pipe($log, $ctx);
2953 print "Done rebuilding $map_path\n" if (!$partial || !$head);
2954 my $rev_db_path = $self->rev_db_path;
2955 if (-f $self->rev_db_path) {
2956 unlink $self->rev_db_path or croak "unlink: $!";
2957 }
2958 $self->unlink_rev_db_symlink;
2959}
2960# rev_map:
2962# Tie::File seems to be prone to offset errors if revisions get sparse,
2963# it's not that fast, either. Tie::File is also not in Perl 5.6. So
2964# one of my favorite modules is out :< Next up would be one of the DBM
2965# modules, but I'm not sure which is most portable...
2966#
2967# This is the replacement for the rev_db format, which was too big
2968# and inefficient for large repositories with a lot of sparse history
2969# (mainly tags)
2970#
2971# The format is this:
2972# - 24 bytes for every record,
2973# * 4 bytes for the integer representing an SVN revision number
2974# * 20 bytes representing the sha1 of a git commit
2975# - No empty padding records like the old format
2976# (except the last record, which can be overwritten)
2977# - new records are written append-only since SVN revision numbers
2978# increase monotonically
2979# - lookups on SVN revision number are done via a binary search
2980# - Piping the file to xxd -c24 is a good way of dumping it for
2981# viewing or editing (piped back through xxd -r), should the need
2982# ever arise.
2983# - The last record can be padding revision with an all-zero sha1
2984# This is used to optimize fetch performance when using multiple
2985# "fetch" directives in .git/config
2986#
2987# These files are disposable unless noMetadata or useSvmProps is set
2988sub _rev_map_set {
2990 my ($fh, $rev, $commit) = @_;
2991 binmode $fh or croak "binmode: $!";
2993 my $size = (stat($fh))[7];
2994 ($size % 24) == 0 or croak "inconsistent size: $size";
2995 my $wr_offset = 0;
2997 if ($size > 0) {
2998 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2999 my $read = sysread($fh, my $buf, 24) or croak "read: $!";
3000 $read == 24 or croak "read only $read bytes (!= 24)";
3001 my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
3002 if ($last_commit eq ('0' x40)) {
3003 if ($size >= 48) {
3004 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
3005 $read = sysread($fh, $buf, 24) or
3006 croak "read: $!";
3007 $read == 24 or
3008 croak "read only $read bytes (!= 24)";
3009 ($last_rev, $last_commit) =
3010 unpack(rev_map_fmt, $buf);
3011 if ($last_commit eq ('0' x40)) {
3012 croak "inconsistent .rev_map\n";
3013 }
3014 }
3015 if ($last_rev >= $rev) {
3016 croak "last_rev is higher!: $last_rev >= $rev";
3017 }
3018 $wr_offset = -24;
3019 }
3020 }
3021 sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
3022 syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
3023 croak "write: $!";
3024}
3025sub mkfile {
3027 my ($path) = @_;
3028 unless (-e $path) {
3029 my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
3030 mkpath([$dir]) unless -d $dir;
3031 open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
3032 close $fh or die "Couldn't close (create) $path: $!\n";
3033 }
3034}
3035sub rev_map_set {
3037 my ($self, $rev, $commit, $update_ref, $uuid) = @_;
3038 length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
3039 my $db = $self->map_path($uuid);
3040 my $db_lock = "$db.lock";
3041 my $sig;
3042 if ($update_ref) {
3043 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
3044 $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
3045 }
3046 mkfile($db);
3047 $LOCKFILES{$db_lock} = 1;
3049 my $sync;
3050 # both of these options make our .rev_db file very, very important
3051 # and we can't afford to lose it because rebuild() won't work
3052 if ($self->use_svm_props || $self->no_metadata) {
3053 $sync = 1;
3054 copy($db, $db_lock) or die "rev_map_set(@_): ",
3055 "Failed to copy: ",
3056 "$db => $db_lock ($!)\n";
3057 } else {
3058 rename $db, $db_lock or die "rev_map_set(@_): ",
3059 "Failed to rename: ",
3060 "$db => $db_lock ($!)\n";
3061 }
3062 sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
3064 or croak "Couldn't open $db_lock: $!\n";
3065 _rev_map_set($fh, $rev, $commit);
3066 if ($sync) {
3067 $fh->flush or die "Couldn't flush $db_lock: $!\n";
3068 $fh->sync or die "Couldn't sync $db_lock: $!\n";
3069 }
3070 close $fh or croak $!;
3071 if ($update_ref) {
3072 $_head = $self;
3073 command_noisy('update-ref', '-m', "r$rev",
3074 $self->refname, $commit);
3075 }
3076 rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
3077 "$db_lock => $db ($!)\n";
3078 delete $LOCKFILES{$db_lock};
3079 if ($update_ref) {
3080 $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
3081 $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
3082 kill $sig, $$ if defined $sig;
3083 }
3084}
3085# If want_commit, this will return an array of (rev, commit) where
3087# commit _must_ be a valid commit in the archive.
3088# Otherwise, it'll return the max revision (whether or not the
3089# commit is valid or just a 0x40 placeholder).
3090sub rev_map_max {
3091 my ($self, $want_commit) = @_;
3092 $self->rebuild;
3093 my ($r, $c) = $self->rev_map_max_norebuild($want_commit);
3094 $want_commit ? ($r, $c) : $r;
3095}
3096sub rev_map_max_norebuild {
3098 my ($self, $want_commit) = @_;
3099 my $map_path = $self->map_path;
3100 stat $map_path or return $want_commit ? (0, undef) : 0;
3101 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
3102 binmode $fh or croak "binmode: $!";
3103 my $size = (stat($fh))[7];
3104 ($size % 24) == 0 or croak "inconsistent size: $size";
3105 if ($size == 0) {
3107 close $fh or croak "close: $!";
3108 return $want_commit ? (0, undef) : 0;
3109 }
3110 sysseek($fh, -24, SEEK_END) or croak "seek: $!";
3112 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
3113 my ($r, $c) = unpack(rev_map_fmt, $buf);
3114 if ($want_commit && $c eq ('0' x40)) {
3115 if ($size < 48) {
3116 return $want_commit ? (0, undef) : 0;
3117 }
3118 sysseek($fh, -48, SEEK_END) or croak "seek: $!";
3119 sysread($fh, $buf, 24) == 24 or croak "read: $!";
3120 ($r, $c) = unpack(rev_map_fmt, $buf);
3121 if ($c eq ('0'x40)) {
3122 croak "Penultimate record is all-zeroes in $map_path";
3123 }
3124 }
3125 close $fh or croak "close: $!";
3126 $want_commit ? ($r, $c) : $r;
3127}
3128sub rev_map_get {
3130 my ($self, $rev, $uuid) = @_;
3131 my $map_path = $self->map_path($uuid);
3132 return undef unless -e $map_path;
3133 sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
3135 binmode $fh or croak "binmode: $!";
3136 my $size = (stat($fh))[7];
3137 ($size % 24) == 0 or croak "inconsistent size: $size";
3138 if ($size == 0) {
3140 close $fh or croak "close: $fh";
3141 return undef;
3142 }
3143 my ($l, $u) = (0, $size - 24);
3145 my ($r, $c, $buf);
3146 while ($l <= $u) {
3148 my $i = int(($l/24 + $u/24) / 2) * 24;
3149 sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
3150 sysread($fh, my $buf, 24) == 24 or croak "read: $!";
3151 my ($r, $c) = unpack('NH40', $buf);
3152 if ($r < $rev) {
3154 $l = $i + 24;
3155 } elsif ($r > $rev) {
3156 $u = $i - 24;
3157 } else { # $r == $rev
3158 close($fh) or croak "close: $!";
3159 return $c eq ('0' x 40) ? undef : $c;
3160 }
3161 }
3162 close($fh) or croak "close: $!";
3163 undef;
3164}
3165# Finds the first svn revision that exists on (if $eq_ok is true) or
3167# before $rev for the current branch. It will not search any lower
3168# than $min_rev. Returns the git commit hash and svn revision number
3169# if found, else (undef, undef).
3170sub find_rev_before {
3171 my ($self, $rev, $eq_ok, $min_rev) = @_;
3172 --$rev unless $eq_ok;
3173 $min_rev ||= 1;
3174 while ($rev >= $min_rev) {
3175 if (my $c = $self->rev_map_get($rev)) {
3176 return ($rev, $c);
3177 }
3178 --$rev;
3179 }
3180 return (undef, undef);
3181}
3182# Finds the first svn revision that exists on (if $eq_ok is true) or
3184# after $rev for the current branch. It will not search any higher
3185# than $max_rev. Returns the git commit hash and svn revision number
3186# if found, else (undef, undef).
3187sub find_rev_after {
3188 my ($self, $rev, $eq_ok, $max_rev) = @_;
3189 ++$rev unless $eq_ok;
3190 $max_rev ||= $self->rev_map_max;
3191 while ($rev <= $max_rev) {
3192 if (my $c = $self->rev_map_get($rev)) {
3193 return ($rev, $c);
3194 }
3195 ++$rev;
3196 }
3197 return (undef, undef);
3198}
3199sub _new {
3201 my ($class, $repo_id, $ref_id, $path) = @_;
3202 unless (defined $repo_id && length $repo_id) {
3203 $repo_id = $Git::SVN::default_repo_id;
3204 }
3205 unless (defined $ref_id && length $ref_id) {
3206 $_[2] = $ref_id = $Git::SVN::default_ref_id;
3207 }
3208 $_[1] = $repo_id;
3209 my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
3210 $_[3] = $path = '' unless (defined $path);
3211 mkpath(["$ENV{GIT_DIR}/svn"]);
3212 bless {
3213 ref_id => $ref_id, dir => $dir, index => "$dir/index",
3214 path => $path, config => "$ENV{GIT_DIR}/svn/config",
3215 map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
3216}
3217# for read-only access of old .rev_db formats
3219sub unlink_rev_db_symlink {
3220 my ($self) = @_;
3221 my $link = $self->rev_db_path;
3222 $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
3223 if (-l $link) {
3224 unlink $link or croak "unlink: $link failed!";
3225 }
3226}
3227sub rev_db_path {
3229 my ($self, $uuid) = @_;
3230 my $db_path = $self->map_path($uuid);
3231 $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
3232 or croak "map_path: $db_path does not contain '/.rev_map.' !";
3233 $db_path;
3234}
3235# the new replacement for .rev_db
3237sub map_path {
3238 my ($self, $uuid) = @_;
3239 $uuid ||= $self->ra_uuid;
3240 "$self->{map_root}.$uuid";
3241}
3242sub uri_encode {
3244 my ($f) = @_;
3245 $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
3246 $f
3247}
3248sub remove_username {
3250 $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
3251}
3252package Git::SVN::Prompt;
3254use strict;
3255use warnings;
3256require SVN::Core;
3257use vars qw/$_no_auth_cache $_username/;
3258sub simple {
3260 my ($cred, $realm, $default_username, $may_save, $pool) = @_;
3261 $may_save = undef if $_no_auth_cache;
3262 $default_username = $_username if defined $_username;
3263 if (defined $default_username && length $default_username) {
3264 if (defined $realm && length $realm) {
3265 print STDERR "Authentication realm: $realm\n";
3266 STDERR->flush;
3267 }
3268 $cred->username($default_username);
3269 } else {
3270 username($cred, $realm, $may_save, $pool);
3271 }
3272 $cred->password(_read_password("Password for '" .
3273 $cred->username . "': ", $realm));
3274 $cred->may_save($may_save);
3275 $SVN::_Core::SVN_NO_ERROR;
3276}
3277sub ssl_server_trust {
3279 my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
3280 $may_save = undef if $_no_auth_cache;
3281 print STDERR "Error validating server certificate for '$realm':\n";
3282 {
3283 no warnings 'once';
3284 # All variables SVN::Auth::SSL::* are used only once,
3285 # so we're shutting up Perl warnings about this.
3286 if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
3287 print STDERR " - The certificate is not issued ",
3288 "by a trusted authority. Use the\n",
3289 " fingerprint to validate ",
3290 "the certificate manually!\n";
3291 }
3292 if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
3293 print STDERR " - The certificate hostname ",
3294 "does not match.\n";
3295 }
3296 if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
3297 print STDERR " - The certificate is not yet valid.\n";
3298 }
3299 if ($failures & $SVN::Auth::SSL::EXPIRED) {
3300 print STDERR " - The certificate has expired.\n";
3301 }
3302 if ($failures & $SVN::Auth::SSL::OTHER) {
3303 print STDERR " - The certificate has ",
3304 "an unknown error.\n";
3305 }
3306 } # no warnings 'once'
3307 printf STDERR
3308 "Certificate information:\n".
3309 " - Hostname: %s\n".
3310 " - Valid: from %s until %s\n".
3311 " - Issuer: %s\n".
3312 " - Fingerprint: %s\n",
3313 map $cert_info->$_, qw(hostname valid_from valid_until
3314 issuer_dname fingerprint);
3315 my $choice;
3316prompt:
3317 print STDERR $may_save ?
3318 "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
3319 "(R)eject or accept (t)emporarily? ";
3320 STDERR->flush;
3321 $choice = lc(substr(<STDIN> || 'R', 0, 1));
3322 if ($choice =~ /^t$/i) {
3323 $cred->may_save(undef);
3324 } elsif ($choice =~ /^r$/i) {
3325 return -1;
3326 } elsif ($may_save && $choice =~ /^p$/i) {
3327 $cred->may_save($may_save);
3328 } else {
3329 goto prompt;
3330 }
3331 $cred->accepted_failures($failures);
3332 $SVN::_Core::SVN_NO_ERROR;
3333}
3334sub ssl_client_cert {
3336 my ($cred, $realm, $may_save, $pool) = @_;
3337 $may_save = undef if $_no_auth_cache;
3338 print STDERR "Client certificate filename: ";
3339 STDERR->flush;
3340 chomp(my $filename = <STDIN>);
3341 $cred->cert_file($filename);
3342 $cred->may_save($may_save);
3343 $SVN::_Core::SVN_NO_ERROR;
3344}
3345sub ssl_client_cert_pw {
3347 my ($cred, $realm, $may_save, $pool) = @_;
3348 $may_save = undef if $_no_auth_cache;
3349 $cred->password(_read_password("Password: ", $realm));
3350 $cred->may_save($may_save);
3351 $SVN::_Core::SVN_NO_ERROR;
3352}
3353sub username {
3355 my ($cred, $realm, $may_save, $pool) = @_;
3356 $may_save = undef if $_no_auth_cache;
3357 if (defined $realm && length $realm) {
3358 print STDERR "Authentication realm: $realm\n";
3359 }
3360 my $username;
3361 if (defined $_username) {
3362 $username = $_username;
3363 } else {
3364 print STDERR "Username: ";
3365 STDERR->flush;
3366 chomp($username = <STDIN>);
3367 }
3368 $cred->username($username);
3369 $cred->may_save($may_save);
3370 $SVN::_Core::SVN_NO_ERROR;
3371}
3372sub _read_password {
3374 my ($prompt, $realm) = @_;
3375 print STDERR $prompt;
3376 STDERR->flush;
3377 require Term::ReadKey;
3378 Term::ReadKey::ReadMode('noecho');
3379 my $password = '';
3380 while (defined(my $key = Term::ReadKey::ReadKey(0))) {
3381 last if $key =~ /[\012\015]/; # \n\r
3382 $password .= $key;
3383 }
3384 Term::ReadKey::ReadMode('restore');
3385 print STDERR "\n";
3386 STDERR->flush;
3387 $password;
3388}
3389package SVN::Git::Fetcher;
3391use vars qw/@ISA/;
3392use strict;
3393use warnings;
3394use Carp qw/croak/;
3395use File::Temp qw/tempfile/;
3396use IO::File qw//;
3397use vars qw/$_ignore_regex/;
3398# file baton members: path, mode_a, mode_b, pool, fh, blob, base
3400sub new {
3401 my ($class, $git_svn, $switch_path) = @_;
3402 my $self = SVN::Delta::Editor->new;
3403 bless $self, $class;
3404 if (exists $git_svn->{last_commit}) {
3405 $self->{c} = $git_svn->{last_commit};
3406 $self->{empty_symlinks} =
3407 _mark_empty_symlinks($git_svn, $switch_path);
3408 }
3409 $self->{ignore_regex} = eval { command_oneline('config', '--get',
3410 "svn-remote.$git_svn->{repo_id}.ignore-paths") };
3411 $self->{empty} = {};
3412 $self->{dir_prop} = {};
3413 $self->{file_prop} = {};
3414 $self->{absent_dir} = {};
3415 $self->{absent_file} = {};
3416 $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
3417 $self;
3418}
3419# this uses the Ra object, so it must be called before do_{switch,update},
3421# not inside them (when the Git::SVN::Fetcher object is passed) to
3422# do_{switch,update}
3423sub _mark_empty_symlinks {
3424 my ($git_svn, $switch_path) = @_;
3425 my $bool = Git::config_bool('svn.brokenSymlinkWorkaround');
3426 return {} if (!defined($bool)) || (defined($bool) && ! $bool);
3427 my %ret;
3429 my ($rev, $cmt) = $git_svn->last_rev_commit;
3430 return {} unless ($rev && $cmt);
3431 # allow the warning to be printed for each revision we fetch to
3433 # ensure the user sees it. The user can also disable the workaround
3434 # on the repository even while git svn is running and the next
3435 # revision fetched will skip this expensive function.
3436 my $printed_warning;
3437 chomp(my $empty_blob = `git hash-object -t blob --stdin < /dev/null`);
3438 my ($ls, $ctx) = command_output_pipe(qw/ls-tree -r -z/, $cmt);
3439 local $/ = "\0";
3440 my $pfx = defined($switch_path) ? $switch_path : $git_svn->{path};
3441 $pfx .= '/' if length($pfx);
3442 while (<$ls>) {
3443 chomp;
3444 s/\A100644 blob $empty_blob\t//o or next;
3445 unless ($printed_warning) {
3446 print STDERR "Scanning for empty symlinks, ",
3447 "this may take a while if you have ",
3448 "many empty files\n",
3449 "You may disable this with `",
3450 "git config svn.brokenSymlinkWorkaround ",
3451 "false'.\n",
3452 "This may be done in a different ",
3453 "terminal without restarting ",
3454 "git svn\n";
3455 $printed_warning = 1;
3456 }
3457 my $path = $_;
3458 my (undef, $props) =
3459 $git_svn->ra->get_file($pfx.$path, $rev, undef);
3460 if ($props->{'svn:special'}) {
3461 $ret{$path} = 1;
3462 }
3463 }
3464 command_close_pipe($ls, $ctx);
3465 \%ret;
3466}
3467# returns true if a given path is inside a ".git" directory
3469sub in_dot_git {
3470 $_[0] =~ m{(?:^|/)\.git(?:/|$)};
3471}
3472# return value: 0 -- don't ignore, 1 -- ignore
3474sub is_path_ignored {
3475 my ($self, $path) = @_;
3476 return 1 if in_dot_git($path);
3477 return 1 if defined($self->{ignore_regex}) &&
3478 $path =~ m!$self->{ignore_regex}!;
3479 return 0 unless defined($_ignore_regex);
3480 return 1 if $path =~ m!$_ignore_regex!o;
3481 return 0;
3482}
3483sub set_path_strip {
3485 my ($self, $path) = @_;
3486 $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
3487}
3488sub open_root {
3490 { path => '' };
3491}
3492sub open_directory {
3494 my ($self, $path, $pb, $rev) = @_;
3495 { path => $path };
3496}
3497sub git_path {
3499 my ($self, $path) = @_;
3500 if ($self->{path_strip}) {
3501 $path =~ s!$self->{path_strip}!! or
3502 die "Failed to strip path '$path' ($self->{path_strip})\n";
3503 }
3504 $path;
3505}
3506sub delete_entry {
3508 my ($self, $path, $rev, $pb) = @_;
3509 return undef if $self->is_path_ignored($path);
3510 my $gpath = $self->git_path($path);
3512 return undef if ($gpath eq '');
3513 # remove entire directories.
3515 my ($tree) = (command('ls-tree', '-z', $self->{c}, "./$gpath")
3516 =~ /\A040000 tree ([a-f\d]{40})\t\Q$gpath\E\0/);
3517 if ($tree) {
3518 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3519 -r --name-only -z/,
3520 $tree);
3521 local $/ = "\0";
3522 while (<$ls>) {
3523 chomp;
3524 my $rmpath = "$gpath/$_";
3525 $self->{gii}->remove($rmpath);
3526 print "\tD\t$rmpath\n" unless $::_q;
3527 }
3528 print "\tD\t$gpath/\n" unless $::_q;
3529 command_close_pipe($ls, $ctx);
3530 $self->{empty}->{$path} = 0
3531 } else {
3532 $self->{gii}->remove($gpath);
3533 print "\tD\t$gpath\n" unless $::_q;
3534 }
3535 undef;
3536}
3537sub open_file {
3539 my ($self, $path, $pb, $rev) = @_;
3540 my ($mode, $blob);
3541 goto out if $self->is_path_ignored($path);
3543 my $gpath = $self->git_path($path);
3545 ($mode, $blob) = (command('ls-tree', '-z', $self->{c}, "./$gpath")
3546 =~ /\A(\d{6}) blob ([a-f\d]{40})\t\Q$gpath\E\0/);
3547 unless (defined $mode && defined $blob) {
3548 die "$path was not found in commit $self->{c} (r$rev)\n";
3549 }
3550 if ($mode eq '100644' && $self->{empty_symlinks}->{$path}) {
3551 $mode = '120000';
3552 }
3553out:
3554 { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3555 pool => SVN::Pool->new, action => 'M' };
3556}
3557sub add_file {
3559 my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3560 my $mode;
3561 if (!$self->is_path_ignored($path)) {
3563 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3564 delete $self->{empty}->{$dir};
3565 $mode = '100644';
3566 }
3567 { path => $path, mode_a => $mode, mode_b => $mode,
3568 pool => SVN::Pool->new, action => 'A' };
3569}
3570sub add_directory {
3572 my ($self, $path, $cp_path, $cp_rev) = @_;
3573 goto out if $self->is_path_ignored($path);
3574 my $gpath = $self->git_path($path);
3575 if ($gpath eq '') {
3576 my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3577 -r --name-only -z/,
3578 $self->{c});
3579 local $/ = "\0";
3580 while (<$ls>) {
3581 chomp;
3582 $self->{gii}->remove($_);
3583 print "\tD\t$_\n" unless $::_q;
3584 }
3585 command_close_pipe($ls, $ctx);
3586 $self->{empty}->{$path} = 0;
3587 }
3588 my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3589 delete $self->{empty}->{$dir};
3590 $self->{empty}->{$path} = 1;
3591out:
3592 { path => $path };
3593}
3594sub change_dir_prop {
3596 my ($self, $db, $prop, $value) = @_;
3597 return undef if $self->is_path_ignored($db->{path});
3598 $self->{dir_prop}->{$db->{path}} ||= {};
3599 $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3600 undef;
3601}
3602sub absent_directory {
3604 my ($self, $path, $pb) = @_;
3605 return undef if $self->is_path_ignored($path);
3606 $self->{absent_dir}->{$pb->{path}} ||= [];
3607 push @{$self->{absent_dir}->{$pb->{path}}}, $path;
3608 undef;
3609}
3610sub absent_file {
3612 my ($self, $path, $pb) = @_;
3613 return undef if $self->is_path_ignored($path);
3614 $self->{absent_file}->{$pb->{path}} ||= [];
3615 push @{$self->{absent_file}->{$pb->{path}}}, $path;
3616 undef;
3617}
3618sub change_file_prop {
3620 my ($self, $fb, $prop, $value) = @_;
3621 return undef if $self->is_path_ignored($fb->{path});
3622 if ($prop eq 'svn:executable') {
3623 if ($fb->{mode_b} != 120000) {
3624 $fb->{mode_b} = defined $value ? 100755 : 100644;
3625 }
3626 } elsif ($prop eq 'svn:special') {
3627 $fb->{mode_b} = defined $value ? 120000 : 100644;
3628 } else {
3629 $self->{file_prop}->{$fb->{path}} ||= {};
3630 $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
3631 }
3632 undef;
3633}
3634sub apply_textdelta {
3636 my ($self, $fb, $exp) = @_;
3637 return undef if $self->is_path_ignored($fb->{path});
3638 my $fh = $::_repository->temp_acquire('svn_delta');
3639 # $fh gets auto-closed() by SVN::TxDelta::apply(),
3640 # (but $base does not,) so dup() it for reading in close_file
3641 open my $dup, '<&', $fh or croak $!;
3642 my $base = $::_repository->temp_acquire('git_blob');
3643 if ($fb->{blob}) {
3645 my ($base_is_link, $size);
3646 if ($fb->{mode_a} eq '120000' &&
3648 ! $self->{empty_symlinks}->{$fb->{path}}) {
3649 print $base 'link ' or die "print $!\n";
3650 $base_is_link = 1;
3651 }
3652 retry:
3653 $size = $::_repository->cat_blob($fb->{blob}, $base);
3654 die "Failed to read object $fb->{blob}" if ($size < 0);
3655 if (defined $exp) {
3657 seek $base, 0, 0 or croak $!;
3658 my $got = ::md5sum($base);
3659 if ($got ne $exp) {
3660 my $err = "Checksum mismatch: ".
3661 "$fb->{path} $fb->{blob}\n" .
3662 "expected: $exp\n" .
3663 " got: $got\n";
3664 if ($base_is_link) {
3665 warn $err,
3666 "Retrying... (possibly ",
3667 "a bad symlink from SVN)\n";
3668 $::_repository->temp_reset($base);
3669 $base_is_link = 0;
3670 goto retry;
3671 }
3672 die $err;
3673 }
3674 }
3675 }
3676 seek $base, 0, 0 or croak $!;
3677 $fb->{fh} = $fh;
3678 $fb->{base} = $base;
3679 [ SVN::TxDelta::apply($base, $dup, undef, $fb->{path}, $fb->{pool}) ];
3680}
3681sub close_file {
3683 my ($self, $fb, $exp) = @_;
3684 return undef if $self->is_path_ignored($fb->{path});
3685 my $hash;
3687 my $path = $self->git_path($fb->{path});
3688 if (my $fh = $fb->{fh}) {
3689 if (defined $exp) {
3690 seek($fh, 0, 0) or croak $!;
3691 my $got = ::md5sum($fh);
3692 if ($got ne $exp) {
3693 die "Checksum mismatch: $path\n",
3694 "expected: $exp\n got: $got\n";
3695 }
3696 }
3697 if ($fb->{mode_b} == 120000) {
3698 sysseek($fh, 0, 0) or croak $!;
3699 my $rd = sysread($fh, my $buf, 5);
3700 if (!defined $rd) {
3702 croak "sysread: $!\n";
3703 } elsif ($rd == 0) {
3704 warn "$path has mode 120000",
3705 " but it points to nothing\n",
3706 "converting to an empty file with mode",
3707 " 100644\n";
3708 $fb->{mode_b} = '100644';
3709 } elsif ($buf ne 'link ') {
3710 warn "$path has mode 120000",
3711 " but is not a link\n";
3712 } else {
3713 my $tmp_fh = $::_repository->temp_acquire(
3714 'svn_hash');
3715 my $res;
3716 while ($res = sysread($fh, my $str, 1024)) {
3717 my $out = syswrite($tmp_fh, $str, $res);
3718 defined($out) && $out == $res
3719 or croak("write ",
3720 Git::temp_path($tmp_fh),
3721 ": $!\n");
3722 }
3723 defined $res or croak $!;
3724 ($fh, $tmp_fh) = ($tmp_fh, $fh);
3726 Git::temp_release($tmp_fh, 1);
3727 }
3728 }
3729 $hash = $::_repository->hash_and_insert_object(
3731 Git::temp_path($fh));
3732 $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
3733 Git::temp_release($fb->{base}, 1);
3735 Git::temp_release($fh, 1);
3736 } else {
3737 $hash = $fb->{blob} or die "no blob information\n";
3738 }
3739 $fb->{pool}->clear;
3740 $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
3741 print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
3742 undef;
3743}
3744sub abort_edit {
3746 my $self = shift;
3747 $self->{nr} = $self->{gii}->{nr};
3748 delete $self->{gii};
3749 $self->SUPER::abort_edit(@_);
3750}
3751sub close_edit {
3753 my $self = shift;
3754 $self->{git_commit_ok} = 1;
3755 $self->{nr} = $self->{gii}->{nr};
3756 delete $self->{gii};
3757 $self->SUPER::close_edit(@_);
3758}
3759package SVN::Git::Editor;
3761use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
3762use strict;
3763use warnings;
3764use Carp qw/croak/;
3765use IO::File;
3766sub new {
3768 my ($class, $opts) = @_;
3769 foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
3770 die "$_ required!\n" unless (defined $opts->{$_});
3771 }
3772 my $pool = SVN::Pool->new;
3774 my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
3775 my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
3776 $opts->{r}, $mods);
3777 # $opts->{ra} functions should not be used after this:
3779 my @ce = $opts->{ra}->get_commit_editor($opts->{log},
3780 $opts->{editor_cb}, $pool);
3781 my $self = SVN::Delta::Editor->new(@ce, $pool);
3782 bless $self, $class;
3783 foreach (qw/svn_path r tree_a tree_b/) {
3784 $self->{$_} = $opts->{$_};
3785 }
3786 $self->{url} = $opts->{ra}->{url};
3787 $self->{mods} = $mods;
3788 $self->{types} = $types;
3789 $self->{pool} = $pool;
3790 $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
3791 $self->{rm} = { };
3792 $self->{path_prefix} = length $self->{svn_path} ?
3793 "$self->{svn_path}/" : '';
3794 $self->{config} = $opts->{config};
3795 return $self;
3796}
3797sub generate_diff {
3799 my ($tree_a, $tree_b) = @_;
3800 my @diff_tree = qw(diff-tree -z -r);
3801 if ($_cp_similarity) {
3802 push @diff_tree, "-C$_cp_similarity";
3803 } else {
3804 push @diff_tree, '-C';
3805 }
3806 push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
3807 push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
3808 push @diff_tree, $tree_a, $tree_b;
3809 my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
3810 local $/ = "\0";
3811 my $state = 'meta';
3812 my @mods;
3813 while (<$diff_fh>) {
3814 chomp $_; # this gets rid of the trailing "\0"
3815 if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
3816 ($::sha1)\s($::sha1)\s
3817 ([MTCRAD])\d*$/xo) {
3818 push @mods, { mode_a => $1, mode_b => $2,
3819 sha1_a => $3, sha1_b => $4,
3820 chg => $5 };
3821 if ($5 =~ /^(?:C|R)$/) {
3822 $state = 'file_a';
3823 } else {
3824 $state = 'file_b';
3825 }
3826 } elsif ($state eq 'file_a') {
3827 my $x = $mods[$#mods] or croak "Empty array\n";
3828 if ($x->{chg} !~ /^(?:C|R)$/) {
3829 croak "Error parsing $_, $x->{chg}\n";
3830 }
3831 $x->{file_a} = $_;
3832 $state = 'file_b';
3833 } elsif ($state eq 'file_b') {
3834 my $x = $mods[$#mods] or croak "Empty array\n";
3835 if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
3836 croak "Error parsing $_, $x->{chg}\n";
3837 }
3838 if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
3839 croak "Error parsing $_, $x->{chg}\n";
3840 }
3841 $x->{file_b} = $_;
3842 $state = 'meta';
3843 } else {
3844 croak "Error parsing $_\n";
3845 }
3846 }
3847 command_close_pipe($diff_fh, $ctx);
3848 \@mods;
3849}
3850sub check_diff_paths {
3852 my ($ra, $pfx, $rev, $mods) = @_;
3853 my %types;
3854 $pfx .= '/' if length $pfx;
3855 sub type_diff_paths {
3857 my ($ra, $types, $path, $rev) = @_;
3858 my @p = split m#/+#, $path;
3859 my $c = shift @p;
3860 unless (defined $types->{$c}) {
3861 $types->{$c} = $ra->check_path($c, $rev);
3862 }
3863 while (@p) {
3864 $c .= '/' . shift @p;
3865 next if defined $types->{$c};
3866 $types->{$c} = $ra->check_path($c, $rev);
3867 }
3868 }
3869 foreach my $m (@$mods) {
3871 foreach my $f (qw/file_a file_b/) {
3872 next unless defined $m->{$f};
3873 my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
3874 if (length $pfx.$dir && ! defined $types{$dir}) {
3875 type_diff_paths($ra, \%types, $pfx.$dir, $rev);
3876 }
3877 }
3878 }
3879 \%types;
3880}
3881sub split_path {
3883 return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
3884}
3885sub repo_path {
3887 my ($self, $path) = @_;
3888 $self->{path_prefix}.(defined $path ? $path : '');
3889}
3890sub url_path {
3892 my ($self, $path) = @_;
3893 if ($self->{url} =~ m#^https?://#) {
3894 $path =~ s/([^~a-zA-Z0-9_.-])/uc sprintf("%%%02x",ord($1))/eg;
3895 }
3896 $self->{url} . '/' . $self->repo_path($path);
3897}
3898sub rmdirs {
3900 my ($self) = @_;
3901 my $rm = $self->{rm};
3902 delete $rm->{''}; # we never delete the url we're tracking
3903 return unless %$rm;
3904 foreach (keys %$rm) {
3906 my @d = split m#/#, $_;
3907 my $c = shift @d;
3908 $rm->{$c} = 1;
3909 while (@d) {
3910 $c .= '/' . shift @d;
3911 $rm->{$c} = 1;
3912 }
3913 }
3914 delete $rm->{$self->{svn_path}};
3915 delete $rm->{''}; # we never delete the url we're tracking
3916 return unless %$rm;
3917 my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
3919 $self->{tree_b});
3920 local $/ = "\0";
3921 while (<$fh>) {
3922 chomp;
3923 my @dn = split m#/#, $_;
3924 while (pop @dn) {
3925 delete $rm->{join '/', @dn};
3926 }
3927 unless (%$rm) {
3928 close $fh;
3929 return;
3930 }
3931 }
3932 command_close_pipe($fh, $ctx);
3933 my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
3935 foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
3936 $self->close_directory($bat->{$d}, $p);
3937 my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
3938 print "\tD+\t$d/\n" unless $::_q;
3939 $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
3940 delete $bat->{$d};
3941 }
3942}
3943sub open_or_add_dir {
3945 my ($self, $full_path, $baton) = @_;
3946 my $t = $self->{types}->{$full_path};
3947 if (!defined $t) {
3948 die "$full_path not known in r$self->{r} or we have a bug!\n";
3949 }
3950 {
3951 no warnings 'once';
3952 # SVN::Node::none and SVN::Node::file are used only once,
3953 # so we're shutting up Perl's warnings about them.
3954 if ($t == $SVN::Node::none) {
3955 return $self->add_directory($full_path, $baton,
3956 undef, -1, $self->{pool});
3957 } elsif ($t == $SVN::Node::dir) {
3958 return $self->open_directory($full_path, $baton,
3959 $self->{r}, $self->{pool});
3960 } # no warnings 'once'
3961 print STDERR "$full_path already exists in repository at ",
3962 "r$self->{r} and it is not a directory (",
3963 ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
3964 } # no warnings 'once'
3965 exit 1;
3966}
3967sub ensure_path {
3969 my ($self, $path) = @_;
3970 my $bat = $self->{bat};
3971 my $repo_path = $self->repo_path($path);
3972 return $bat->{''} unless (length $repo_path);
3973 my @p = split m#/+#, $repo_path;
3974 my $c = shift @p;
3975 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
3976 while (@p) {
3977 my $c0 = $c;
3978 $c .= '/' . shift @p;
3979 $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
3980 }
3981 return $bat->{$c};
3982}
3983# Subroutine to convert a globbing pattern to a regular expression.
3985# From perl cookbook.
3986sub glob2pat {
3987 my $globstr = shift;
3988 my %patmap = ('*' => '.*', '?' => '.', '[' => '[', ']' => ']');
3989 $globstr =~ s{(.)} { $patmap{$1} || "\Q$1" }ge;
3990 return '^' . $globstr . '$';
3991}
3992sub check_autoprop {
3994 my ($self, $pattern, $properties, $file, $fbat) = @_;
3995 # Convert the globbing pattern to a regular expression.
3996 my $regex = glob2pat($pattern);
3997 # Check if the pattern matches the file name.
3998 if($file =~ m/($regex)/) {
3999 # Parse the list of properties to set.
4000 my @props = split(/;/, $properties);
4001 foreach my $prop (@props) {
4002 # Parse 'name=value' syntax and set the property.
4003 if ($prop =~ /([^=]+)=(.*)/) {
4004 my ($n,$v) = ($1,$2);
4005 for ($n, $v) {
4006 s/^\s+//; s/\s+$//;
4007 }
4008 $self->change_file_prop($fbat, $n, $v);
4009 }
4010 }
4011 }
4012}
4013sub apply_autoprops {
4015 my ($self, $file, $fbat) = @_;
4016 my $conf_t = ${$self->{config}}{'config'};
4017 no warnings 'once';
4018 # Check [miscellany]/enable-auto-props in svn configuration.
4019 if (SVN::_Core::svn_config_get_bool(
4020 $conf_t,
4021 $SVN::_Core::SVN_CONFIG_SECTION_MISCELLANY,
4022 $SVN::_Core::SVN_CONFIG_OPTION_ENABLE_AUTO_PROPS,
4023 0)) {
4024 # Auto-props are enabled. Enumerate them to look for matches.
4025 my $callback = sub {
4026 $self->check_autoprop($_[0], $_[1], $file, $fbat);
4027 };
4028 SVN::_Core::svn_config_enumerate(
4029 $conf_t,
4030 $SVN::_Core::SVN_CONFIG_SECTION_AUTO_PROPS,
4031 $callback);
4032 }
4033}
4034sub A {
4036 my ($self, $m) = @_;
4037 my ($dir, $file) = split_path($m->{file_b});
4038 my $pbat = $self->ensure_path($dir);
4039 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
4040 undef, -1);
4041 print "\tA\t$m->{file_b}\n" unless $::_q;
4042 $self->apply_autoprops($file, $fbat);
4043 $self->chg_file($fbat, $m);
4044 $self->close_file($fbat,undef,$self->{pool});
4045}
4046sub C {
4048 my ($self, $m) = @_;
4049 my ($dir, $file) = split_path($m->{file_b});
4050 my $pbat = $self->ensure_path($dir);
4051 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
4052 $self->url_path($m->{file_a}), $self->{r});
4053 print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
4054 $self->chg_file($fbat, $m);
4055 $self->close_file($fbat,undef,$self->{pool});
4056}
4057sub delete_entry {
4059 my ($self, $path, $pbat) = @_;
4060 my $rpath = $self->repo_path($path);
4061 my ($dir, $file) = split_path($rpath);
4062 $self->{rm}->{$dir} = 1;
4063 $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
4064}
4065sub R {
4067 my ($self, $m) = @_;
4068 my ($dir, $file) = split_path($m->{file_b});
4069 my $pbat = $self->ensure_path($dir);
4070 my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
4071 $self->url_path($m->{file_a}), $self->{r});
4072 print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
4073 $self->apply_autoprops($file, $fbat);
4074 $self->chg_file($fbat, $m);
4075 $self->close_file($fbat,undef,$self->{pool});
4076 ($dir, $file) = split_path($m->{file_a});
4078 $pbat = $self->ensure_path($dir);
4079 $self->delete_entry($m->{file_a}, $pbat);
4080}
4081sub M {
4083 my ($self, $m) = @_;
4084 my ($dir, $file) = split_path($m->{file_b});
4085 my $pbat = $self->ensure_path($dir);
4086 my $fbat = $self->open_file($self->repo_path($m->{file_b}),
4087 $pbat,$self->{r},$self->{pool});
4088 print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
4089 $self->chg_file($fbat, $m);
4090 $self->close_file($fbat,undef,$self->{pool});
4091}
4092sub T { shift->M(@_) }
4094sub change_file_prop {
4096 my ($self, $fbat, $pname, $pval) = @_;
4097 $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
4098}
4099sub _chg_file_get_blob ($$$$) {
4101 my ($self, $fbat, $m, $which) = @_;
4102 my $fh = $::_repository->temp_acquire("git_blob_$which");
4103 if ($m->{"mode_$which"} =~ /^120/) {
4104 print $fh 'link ' or croak $!;
4105 $self->change_file_prop($fbat,'svn:special','*');
4106 } elsif ($m->{mode_a} =~ /^120/ && $m->{"mode_$which"} !~ /^120/) {
4107 $self->change_file_prop($fbat,'svn:special',undef);
4108 }
4109 my $blob = $m->{"sha1_$which"};
4110 return ($fh,) if ($blob =~ /^0{40}$/);
4111 my $size = $::_repository->cat_blob($blob, $fh);
4112 croak "Failed to read object $blob" if ($size < 0);
4113 $fh->flush == 0 or croak $!;
4114 seek $fh, 0, 0 or croak $!;
4115 my $exp = ::md5sum($fh);
4117 seek $fh, 0, 0 or croak $!;
4118 return ($fh, $exp);
4119}
4120sub chg_file {
4122 my ($self, $fbat, $m) = @_;
4123 if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
4124 $self->change_file_prop($fbat,'svn:executable','*');
4125 } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
4126 $self->change_file_prop($fbat,'svn:executable',undef);
4127 }
4128 my ($fh_a, $exp_a) = _chg_file_get_blob $self, $fbat, $m, 'a';
4129 my ($fh_b, $exp_b) = _chg_file_get_blob $self, $fbat, $m, 'b';
4130 my $pool = SVN::Pool->new;
4131 my $atd = $self->apply_textdelta($fbat, $exp_a, $pool);
4132 if (-s $fh_a) {
4133 my $txstream = SVN::TxDelta::new ($fh_a, $fh_b, $pool);
4134 my $res = SVN::TxDelta::send_txstream($txstream, @$atd, $pool);
4135 if (defined $res) {
4136 die "Unexpected result from send_txstream: $res\n",
4137 "(SVN::Core::VERSION: $SVN::Core::VERSION)\n";
4138 }
4139 } else {
4140 my $got = SVN::TxDelta::send_stream($fh_b, @$atd, $pool);
4141 die "Checksum mismatch\nexpected: $exp_b\ngot: $got\n"
4142 if ($got ne $exp_b);
4143 }
4144 Git::temp_release($fh_b, 1);
4145 Git::temp_release($fh_a, 1);
4146 $pool->clear;
4147}
4148sub D {
4150 my ($self, $m) = @_;
4151 my ($dir, $file) = split_path($m->{file_b});
4152 my $pbat = $self->ensure_path($dir);
4153 print "\tD\t$m->{file_b}\n" unless $::_q;
4154 $self->delete_entry($m->{file_b}, $pbat);
4155}
4156sub close_edit {
4158 my ($self) = @_;
4159 my ($p,$bat) = ($self->{pool}, $self->{bat});
4160 foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
4161 next if $_ eq '';
4162 $self->close_directory($bat->{$_}, $p);
4163 }
4164 $self->close_directory($bat->{''}, $p);
4165 $self->SUPER::close_edit($p);
4166 $p->clear;
4167}
4168sub abort_edit {
4170 my ($self) = @_;
4171 $self->SUPER::abort_edit($self->{pool});
4172}
4173sub DESTROY {
4175 my $self = shift;
4176 $self->SUPER::DESTROY(@_);
4177 $self->{pool}->clear;
4178}
4179# this drives the editor
4181sub apply_diff {
4182 my ($self) = @_;
4183 my $mods = $self->{mods};
4184 my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
4185 foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
4186 my $f = $m->{chg};
4187 if (defined $o{$f}) {
4188 $self->$f($m);
4189 } else {
4190 fatal("Invalid change type: $f");
4191 }
4192 }
4193 $self->rmdirs if $_rmdir;
4194 if (@$mods == 0) {
4195 $self->abort_edit;
4196 } else {
4197 $self->close_edit;
4198 }
4199 return scalar @$mods;
4200}
4201package Git::SVN::Ra;
4203use vars qw/@ISA $config_dir $_log_window_size/;
4204use strict;
4205use warnings;
4206my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
4207BEGIN {
4209 # enforce temporary pool usage for some simple functions
4210 no strict 'refs';
4211 for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root
4212 get_file/) {
4213 my $SUPER = "SUPER::$f";
4214 *$f = sub {
4215 my $self = shift;
4216 my $pool = SVN::Pool->new;
4217 my @ret = $self->$SUPER(@_,$pool);
4218 $pool->clear;
4219 wantarray ? @ret : $ret[0];
4220 };
4221 }
4222}
4223sub _auth_providers () {
4225 [
4226 SVN::Client::get_simple_provider(),
4227 SVN::Client::get_ssl_server_trust_file_provider(),
4228 SVN::Client::get_simple_prompt_provider(
4229 \&Git::SVN::Prompt::simple, 2),
4230 SVN::Client::get_ssl_client_cert_file_provider(),
4231 SVN::Client::get_ssl_client_cert_prompt_provider(
4232 \&Git::SVN::Prompt::ssl_client_cert, 2),
4233 SVN::Client::get_ssl_client_cert_pw_file_provider(),
4234 SVN::Client::get_ssl_client_cert_pw_prompt_provider(
4235 \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
4236 SVN::Client::get_username_provider(),
4237 SVN::Client::get_ssl_server_trust_prompt_provider(
4238 \&Git::SVN::Prompt::ssl_server_trust),
4239 SVN::Client::get_username_prompt_provider(
4240 \&Git::SVN::Prompt::username, 2)
4241 ]
4242}
4243sub escape_uri_only {
4245 my ($uri) = @_;
4246 my @tmp;
4247 foreach (split m{/}, $uri) {
4248 s/([^~\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
4249 push @tmp, $_;
4250 }
4251 join('/', @tmp);
4252}
4253sub escape_url {
4255 my ($url) = @_;
4256 if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
4257 my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
4258 $url = "$scheme://$domain$uri";
4259 }
4260 $url;
4261}
4262sub new {
4264 my ($class, $url) = @_;
4265 $url =~ s!/+$!!;
4266 return $RA if ($RA && $RA->{url} eq $url);
4267 SVN::_Core::svn_config_ensure($config_dir, undef);
4269 my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
4270 my $config = SVN::Core::config_get_config($config_dir);
4271 $RA = undef;
4272 my $dont_store_passwords = 1;
4273 my $conf_t = ${$config}{'config'};
4274 {
4275 no warnings 'once';
4276 # The usage of $SVN::_Core::SVN_CONFIG_* variables
4277 # produces warnings that variables are used only once.
4278 # I had not found the better way to shut them up, so
4279 # the warnings of type 'once' are disabled in this block.
4280 if (SVN::_Core::svn_config_get_bool($conf_t,
4281 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
4282 $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
4283 1) == 0) {
4284 SVN::_Core::svn_auth_set_parameter($baton,
4285 $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
4286 bless (\$dont_store_passwords, "_p_void"));
4287 }
4288 if (SVN::_Core::svn_config_get_bool($conf_t,
4289 $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
4290 $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
4291 1) == 0) {
4292 $Git::SVN::Prompt::_no_auth_cache = 1;
4293 }
4294 } # no warnings 'once'
4295 my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
4296 config => $config,
4297 pool => SVN::Pool->new,
4298 auth_provider_callbacks => $callbacks);
4299 $self->{url} = $url;
4300 $self->{svn_path} = $url;
4301 $self->{repos_root} = $self->get_repos_root;
4302 $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
4303 $self->{cache} = { check_path => { r => 0, data => {} },
4304 get_dir => { r => 0, data => {} } };
4305 $RA = bless $self, $class;
4306}
4307sub check_path {
4309 my ($self, $path, $r) = @_;
4310 my $cache = $self->{cache}->{check_path};
4311 if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
4312 return $cache->{data}->{$path};
4313 }
4314 my $pool = SVN::Pool->new;
4315 my $t = $self->SUPER::check_path($path, $r, $pool);
4316 $pool->clear;
4317 if ($r != $cache->{r}) {
4318 %{$cache->{data}} = ();
4319 $cache->{r} = $r;
4320 }
4321 $cache->{data}->{$path} = $t;
4322}
4323sub get_dir {
4325 my ($self, $dir, $r) = @_;
4326 my $cache = $self->{cache}->{get_dir};
4327 if ($r == $cache->{r}) {
4328 if (my $x = $cache->{data}->{$dir}) {
4329 return wantarray ? @$x : $x->[0];
4330 }
4331 }
4332 my $pool = SVN::Pool->new;
4333 my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
4334 my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
4335 $pool->clear;
4336 if ($r != $cache->{r}) {
4337 %{$cache->{data}} = ();
4338 $cache->{r} = $r;
4339 }
4340 $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
4341 wantarray ? (\%dirents, $r, $props) : \%dirents;
4342}
4343sub DESTROY {
4345 # do not call the real DESTROY since we store ourselves in $RA
4346}
4347# get_log(paths, start, end, limit,
4349# discover_changed_paths, strict_node_history, receiver)
4350sub get_log {
4351 my ($self, @args) = @_;
4352 my $pool = SVN::Pool->new;
4353 # the limit parameter was not supported in SVN 1.1.x, so we
4355 # drop it. Therefore, the receiver callback passed to it
4356 # is made aware of this limitation by being wrapped if
4357 # the limit passed to is being wrapped.
4358 if ($SVN::Core::VERSION le '1.2.0') {
4359 my $limit = splice(@args, 3, 1);
4360 if ($limit > 0) {
4361 my $receiver = pop @args;
4362 push(@args, sub { &$receiver(@_) if (--$limit >= 0) });
4363 }
4364 }
4365 my $ret = $self->SUPER::get_log(@args, $pool);
4366 $pool->clear;
4367 $ret;
4368}
4369sub trees_match {
4371 my ($self, $url1, $rev1, $url2, $rev2) = @_;
4372 my $ctx = SVN::Client->new(auth => _auth_providers);
4373 my $out = IO::File->new_tmpfile;
4374 # older SVN (1.1.x) doesn't take $pool as the last parameter for
4376 # $ctx->diff(), so we'll create a default one
4377 my $pool = SVN::Pool->new_default_sub;
4378 $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
4380 $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
4381 $out->flush;
4382 my $ret = (($out->stat)[7] == 0);
4383 close $out or croak $!;
4384 $ret;
4386}
4387sub get_commit_editor {
4389 my ($self, $log, $cb, $pool) = @_;
4390 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
4391 $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
4392}
4393sub gs_do_update {
4395 my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
4396 my $new = ($rev_a == $rev_b);
4397 my $path = $gs->{path};
4398 if ($new && -e $gs->{index}) {
4400 unlink $gs->{index} or die
4401 "Couldn't unlink index: $gs->{index}: $!\n";
4402 }
4403 my $pool = SVN::Pool->new;
4404 $editor->set_path_strip($path);
4405 my (@pc) = split m#/#, $path;
4406 my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
4407 1, $editor, $pool);
4408 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
4409 # Since we can't rely on svn_ra_reparent being available, we'll
4411 # just have to do some magic with set_path to make it so
4412 # we only want a partial path.
4413 my $sp = '';
4414 my $final = join('/', @pc);
4415 while (@pc) {
4416 $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
4417 $sp .= '/' if length $sp;
4418 $sp .= shift @pc;
4419 }
4420 die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
4421 $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
4423 $reporter->finish_report($pool);
4425 $pool->clear;
4426 $editor->{git_commit_ok};
4427}
4428# this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
4430# svn_ra_reparent didn't work before 1.4)
4431sub gs_do_switch {
4432 my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
4433 my $path = $gs->{path};
4434 my $pool = SVN::Pool->new;
4435 my $full_url = $self->{url};
4437 my $old_url = $full_url;
4438 $full_url .= '/' . escape_uri_only($path) if length $path;
4439 my ($ra, $reparented);
4440 if ($old_url =~ m#^svn(\+ssh)?://#) {
4442 $_[0] = undef;
4443 $self = undef;
4444 $RA = undef;
4445 $ra = Git::SVN::Ra->new($full_url);
4446 $ra_invalid = 1;
4447 } elsif ($old_url ne $full_url) {
4448 SVN::_Ra::svn_ra_reparent($self->{session}, $full_url, $pool);
4449 $self->{url} = $full_url;
4450 $reparented = 1;
4451 }
4452 $ra ||= $self;
4454 $url_b = escape_url($url_b);
4455 my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
4456 my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
4457 $reporter->set_path('', $rev_a, 0, @lock, $pool);
4458 $reporter->finish_report($pool);
4459 if ($reparented) {
4461 SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
4462 $self->{url} = $old_url;
4463 }
4464 $pool->clear;
4466 $editor->{git_commit_ok};
4467}
4468sub longest_common_path {
4470 my ($gsv, $globs) = @_;
4471 my %common;
4472 my $common_max = scalar @$gsv;
4473 foreach my $gs (@$gsv) {
4475 my @tmp = split m#/#, $gs->{path};
4476 my $p = '';
4477 foreach (@tmp) {
4478 $p .= length($p) ? "/$_" : $_;
4479 $common{$p} ||= 0;
4480 $common{$p}++;
4481 }
4482 }
4483 $globs ||= [];
4484 $common_max += scalar @$globs;
4485 foreach my $glob (@$globs) {
4486 my @tmp = split m#/#, $glob->{path}->{left};
4487 my $p = '';
4488 foreach (@tmp) {
4489 $p .= length($p) ? "/$_" : $_;
4490 $common{$p} ||= 0;
4491 $common{$p}++;
4492 }
4493 }
4494 my $longest_path = '';
4496 foreach (sort {length $b <=> length $a} keys %common) {
4497 if ($common{$_} == $common_max) {
4498 $longest_path = $_;
4499 last;
4500 }
4501 }
4502 $longest_path;
4503}
4504sub gs_fetch_loop_common {
4506 my ($self, $base, $head, $gsv, $globs) = @_;
4507 return if ($base > $head);
4508 my $inc = $_log_window_size;
4509 my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
4510 my $longest_path = longest_common_path($gsv, $globs);
4511 my $ra_url = $self->{url};
4512 my $find_trailing_edge;
4513 while (1) {
4514 my %revs;
4515 my $err;
4516 my $err_handler = $SVN::Error::handler;
4517 $SVN::Error::handler = sub {
4518 ($err) = @_;
4519 skip_unknown_revs($err);
4520 };
4521 sub _cb {
4522 my ($paths, $r, $author, $date, $log) = @_;
4523 [ dup_changed_paths($paths),
4524 { author => $author, date => $date, log => $log } ];
4525 }
4526 $self->get_log([$longest_path], $min, $max, 0, 1, 1,
4527 sub { $revs{$_[1]} = _cb(@_) });
4528 if ($err) {
4529 print "Checked through r$max\r";
4530 } else {
4531 $find_trailing_edge = 1;
4532 }
4533 if ($err and $find_trailing_edge) {
4534 print STDERR "Path '$longest_path' ",
4535 "was probably deleted:\n",
4536 $err->expanded_message,
4537 "\nWill attempt to follow ",
4538 "revisions r$min .. r$max ",
4539 "committed before the deletion\n";
4540 my $hi = $max;
4541 while (--$hi >= $min) {
4542 my $ok;
4543 $self->get_log([$longest_path], $min, $hi,
4544 0, 1, 1, sub {
4545 $ok = $_[1];
4546 $revs{$_[1]} = _cb(@_) });
4547 if ($ok) {
4548 print STDERR "r$min .. r$ok OK\n";
4549 last;
4550 }
4551 }
4552 $find_trailing_edge = 0;
4553 }
4554 $SVN::Error::handler = $err_handler;
4555 my %exists = map { $_->{path} => $_ } @$gsv;
4557 foreach my $r (sort {$a <=> $b} keys %revs) {
4558 my ($paths, $logged) = @{$revs{$r}};
4559 foreach my $gs ($self->match_globs(\%exists, $paths,
4561 $globs, $r)) {
4562 if ($gs->rev_map_max >= $r) {
4563 next;
4564 }
4565 next unless $gs->match_paths($paths, $r);
4566 $gs->{logged_rev_props} = $logged;
4567 if (my $last_commit = $gs->last_commit) {
4568 $gs->assert_index_clean($last_commit);
4569 }
4570 my $log_entry = $gs->do_fetch($paths, $r);
4571 if ($log_entry) {
4572 $gs->do_git_commit($log_entry);
4573 }
4574 $INDEX_FILES{$gs->{index}} = 1;
4575 }
4576 foreach my $g (@$globs) {
4577 my $k = "svn-remote.$g->{remote}." .
4578 "$g->{t}-maxRev";
4579 Git::SVN::tmp_config($k, $r);
4580 }
4581 if ($ra_invalid) {
4582 $_[0] = undef;
4583 $self = undef;
4584 $RA = undef;
4585 $self = Git::SVN::Ra->new($ra_url);
4586 $ra_invalid = undef;
4587 }
4588 }
4589 # pre-fill the .rev_db since it'll eventually get filled in
4590 # with '0' x40 if something new gets committed
4591 foreach my $gs (@$gsv) {
4592 next if $gs->rev_map_max >= $max;
4593 next if defined $gs->rev_map_get($max);
4594 $gs->rev_map_set($max, 0 x40);
4595 }
4596 foreach my $g (@$globs) {
4597 my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
4598 Git::SVN::tmp_config($k, $max);
4599 }
4600 last if $max >= $head;
4601 $min = $max + 1;
4602 $max += $inc;
4603 $max = $head if ($max > $head);
4604 }
4605 Git::SVN::gc();
4606}
4607sub get_dir_globbed {
4609 my ($self, $left, $depth, $r) = @_;
4610 my @x = eval { $self->get_dir($left, $r) };
4612 return unless scalar @x == 3;
4613 my $dirents = $x[0];
4614 my @finalents;
4615 foreach my $de (keys %$dirents) {
4616 next if $dirents->{$de}->{kind} != $SVN::Node::dir;
4617 if ($depth > 1) {
4618 my @args = ("$left/$de", $depth - 1, $r);
4619 foreach my $dir ($self->get_dir_globbed(@args)) {
4620 push @finalents, "$de/$dir";
4621 }
4622 } else {
4623 push @finalents, $de;
4624 }
4625 }
4626 @finalents;
4627}
4628sub match_globs {
4630 my ($self, $exists, $paths, $globs, $r) = @_;
4631 sub get_dir_check {
4633 my ($self, $exists, $g, $r) = @_;
4634 my @dirs = $self->get_dir_globbed($g->{path}->{left},
4636 $g->{path}->{depth},
4637 $r);
4638 foreach my $de (@dirs) {
4640 my $p = $g->{path}->full_path($de);
4641 next if $exists->{$p};
4642 next if (length $g->{path}->{right} &&
4643 ($self->check_path($p, $r) !=
4644 $SVN::Node::dir));
4645 $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
4646 $g->{ref}->full_path($de), 1);
4647 }
4648 }
4649 foreach my $g (@$globs) {
4650 if (my $path = $paths->{"/$g->{path}->{left}"}) {
4651 if ($path->{action} =~ /^[AR]$/) {
4652 get_dir_check($self, $exists, $g, $r);
4653 }
4654 }
4655 foreach (keys %$paths) {
4656 if (/$g->{path}->{left_regex}/ &&
4657 !/$g->{path}->{regex}/) {
4658 next if $paths->{$_}->{action} !~ /^[AR]$/;
4659 get_dir_check($self, $exists, $g, $r);
4660 }
4661 next unless /$g->{path}->{regex}/;
4662 my $p = $1;
4663 my $pathname = $g->{path}->full_path($p);
4664 next if $exists->{$pathname};
4665 next if ($self->check_path($pathname, $r) !=
4666 $SVN::Node::dir);
4667 $exists->{$pathname} = Git::SVN->init(
4668 $self->{url}, $pathname, undef,
4669 $g->{ref}->full_path($p), 1);
4670 }
4671 my $c = '';
4672 foreach (split m#/#, $g->{path}->{left}) {
4673 $c .= "/$_";
4674 next unless ($paths->{$c} &&
4675 ($paths->{$c}->{action} =~ /^[AR]$/));
4676 get_dir_check($self, $exists, $g, $r);
4677 }
4678 }
4679 values %$exists;
4680}
4681sub minimize_url {
4683 my ($self) = @_;
4684 return $self->{url} if ($self->{url} eq $self->{repos_root});
4685 my $url = $self->{repos_root};
4686 my @components = split(m!/!, $self->{svn_path});
4687 my $c = '';
4688 do {
4689 $url .= "/$c" if length $c;
4690 eval { (ref $self)->new($url)->get_latest_revnum };
4691 } while ($@ && ($c = shift @components));
4692 $url;
4693}
4694sub can_do_switch {
4696 my $self = shift;
4697 unless (defined $can_do_switch) {
4698 my $pool = SVN::Pool->new;
4699 my $rep = eval {
4700 $self->do_switch(1, '', 0, $self->{url},
4701 SVN::Delta::Editor->new, $pool);
4702 };
4703 if ($@) {
4704 $can_do_switch = 0;
4705 } else {
4706 $rep->abort_report($pool);
4707 $can_do_switch = 1;
4708 }
4709 $pool->clear;
4710 }
4711 $can_do_switch;
4712}
4713sub skip_unknown_revs {
4715 my ($err) = @_;
4716 my $errno = $err->apr_err();
4717 # Maybe the branch we're tracking didn't
4718 # exist when the repo started, so it's
4719 # not an error if it doesn't, just continue
4720 #
4721 # Wonderfully consistent library, eh?
4722 # 160013 - svn:// and file://
4723 # 175002 - http(s)://
4724 # 175007 - http(s):// (this repo required authorization, too...)
4725 # More codes may be discovered later...
4726 if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
4727 my $err_key = $err->expanded_message;
4728 # revision numbers change every time, filter them out
4729 $err_key =~ s/\d+/\0/g;
4730 $err_key = "$errno\0$err_key";
4731 unless ($ignored_err{$err_key}) {
4732 warn "W: Ignoring error from SVN, path probably ",
4733 "does not exist: ($errno): ",
4734 $err->expanded_message,"\n";
4735 warn "W: Do not be alarmed at the above message ",
4736 "git-svn is just searching aggressively for ",
4737 "old history.\n",
4738 "This may take a while on large repositories\n";
4739 $ignored_err{$err_key} = 1;
4740 }
4741 return;
4742 }
4743 die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
4744}
4745# svn_log_changed_path_t objects passed to get_log are likely to be
4747# overwritten even if only the refs are copied to an external variable,
4748# so we should dup the structures in their entirety. Using an externally
4749# passed pool (instead of our temporary and quickly cleared pool in
4750# Git::SVN::Ra) does not help matters at all...
4751sub dup_changed_paths {
4752 my ($paths) = @_;
4753 return undef unless $paths;
4754 my %ret;
4755 foreach my $p (keys %$paths) {
4756 my $i = $paths->{$p};
4757 my %s = map { $_ => $i->$_ }
4758 qw/copyfrom_path copyfrom_rev action/;
4759 $ret{$p} = \%s;
4760 }
4761 \%ret;
4762}
4763package Git::SVN::Log;
4765use strict;
4766use warnings;
4767use POSIX qw/strftime/;
4768use Time::Local;
4769use constant commit_log_separator => ('-' x 72) . "\n";
4770use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
4771 %rusers $show_commit $incremental/;
4772my $l_fmt;
4773sub cmt_showable {
4775 my ($c) = @_;
4776 return 1 if defined $c->{r};
4777 # big commit message got truncated by the 16k pretty buffer in rev-list
4779 if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
4780 $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
4781 @{$c->{l}} = ();
4782 my @log = command(qw/cat-file commit/, $c->{c});
4783 # shift off the headers
4785 shift @log while ($log[0] ne '');
4786 shift @log;
4787 # TODO: make $c->{l} not have a trailing newline in the future
4789 @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
4790 (undef, $c->{r}, undef) = ::extract_metadata(
4792 (grep(/^git-svn-id: /, @log))[-1]);
4793 }
4794 return defined $c->{r};
4795}
4796sub log_use_color {
4798 return $color || Git->repository->get_colorbool('color.diff');
4799}
4800sub git_svn_log_cmd {
4802 my ($r_min, $r_max, @args) = @_;
4803 my $head = 'HEAD';
4804 my (@files, @log_opts);
4805 foreach my $x (@args) {
4806 if ($x eq '--' || @files) {
4807 push @files, $x;
4808 } else {
4809 if (::verify_ref("$x^0")) {
4810 $head = $x;
4811 } else {
4812 push @log_opts, $x;
4813 }
4814 }
4815 }
4816 my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
4818 $gs ||= Git::SVN->_new;
4819 my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
4820 $gs->refname);
4821 push @cmd, '-r' unless $non_recursive;
4822 push @cmd, qw/--raw --name-status/ if $verbose;
4823 push @cmd, '--color' if log_use_color();
4824 push @cmd, @log_opts;
4825 if (defined $r_max && $r_max == $r_min) {
4826 push @cmd, '--max-count=1';
4827 if (my $c = $gs->rev_map_get($r_max)) {
4828 push @cmd, $c;
4829 }
4830 } elsif (defined $r_max) {
4831 if ($r_max < $r_min) {
4832 ($r_min, $r_max) = ($r_max, $r_min);
4833 }
4834 my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
4835 my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
4836 # If there are no commits in the range, both $c_max and $c_min
4837 # will be undefined. If there is at least 1 commit in the
4838 # range, both will be defined.
4839 return () if !defined $c_min || !defined $c_max;
4840 if ($c_min eq $c_max) {
4841 push @cmd, '--max-count=1', $c_min;
4842 } else {
4843 push @cmd, '--boundary', "$c_min..$c_max";
4844 }
4845 }
4846 return (@cmd, @files);
4847}
4848# adapted from pager.c
4850sub config_pager {
4851 $pager ||= $ENV{GIT_PAGER} || $ENV{PAGER};
4852 if (!defined $pager) {
4853 $pager = 'less';
4854 } elsif (length $pager == 0 || $pager eq 'cat') {
4855 $pager = undef;
4856 }
4857 $ENV{GIT_PAGER_IN_USE} = defined($pager);
4858}
4859sub run_pager {
4861 return unless -t *STDOUT && defined $pager;
4862 pipe my ($rfd, $wfd) or return;
4863 defined(my $pid = fork) or ::fatal "Can't fork: $!";
4864 if (!$pid) {
4865 open STDOUT, '>&', $wfd or
4866 ::fatal "Can't redirect to stdout: $!";
4867 return;
4868 }
4869 open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
4870 $ENV{LESS} ||= 'FRSX';
4871 exec $pager or ::fatal "Can't run pager: $! ($pager)";
4872}
4873sub format_svn_date {
4875 # some systmes don't handle or mishandle %z, so be creative.
4876 my $t = shift || time;
4877 my $gm = timelocal(gmtime($t));
4878 my $sign = qw( + + - )[ $t <=> $gm ];
4879 my $gmoff = sprintf("%s%02d%02d", $sign, (gmtime(abs($t - $gm)))[2,1]);
4880 return strftime("%Y-%m-%d %H:%M:%S $gmoff (%a, %d %b %Y)", localtime($t));
4881}
4882sub parse_git_date {
4884 my ($t, $tz) = @_;
4885 # Date::Parse isn't in the standard Perl distro :(
4886 if ($tz =~ s/^\+//) {
4887 $t += tz_to_s_offset($tz);
4888 } elsif ($tz =~ s/^\-//) {
4889 $t -= tz_to_s_offset($tz);
4890 }
4891 return $t;
4892}
4893sub set_local_timezone {
4895 if (defined $TZ) {
4896 $ENV{TZ} = $TZ;
4897 } else {
4898 delete $ENV{TZ};
4899 }
4900}
4901sub tz_to_s_offset {
4903 my ($tz) = @_;
4904 $tz =~ s/(\d\d)$//;
4905 return ($1 * 60) + ($tz * 3600);
4906}
4907sub get_author_info {
4909 my ($dest, $author, $t, $tz) = @_;
4910 $author =~ s/(?:^\s*|\s*$)//g;
4911 $dest->{a_raw} = $author;
4912 my $au;
4913 if ($::_authors) {
4914 $au = $rusers{$author} || undef;
4915 }
4916 if (!$au) {
4917 ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
4918 }
4919 $dest->{t} = $t;
4920 $dest->{tz} = $tz;
4921 $dest->{a} = $au;
4922 $dest->{t_utc} = parse_git_date($t, $tz);
4923}
4924sub process_commit {
4926 my ($c, $r_min, $r_max, $defer) = @_;
4927 if (defined $r_min && defined $r_max) {
4928 if ($r_min == $c->{r} && $r_min == $r_max) {
4929 show_commit($c);
4930 return 0;
4931 }
4932 return 1 if $r_min == $r_max;
4933 if ($r_min < $r_max) {
4934 # we need to reverse the print order
4935 return 0 if (defined $limit && --$limit < 0);
4936 push @$defer, $c;
4937 return 1;
4938 }
4939 if ($r_min != $r_max) {
4940 return 1 if ($r_min < $c->{r});
4941 return 1 if ($r_max > $c->{r});
4942 }
4943 }
4944 return 0 if (defined $limit && --$limit < 0);
4945 show_commit($c);
4946 return 1;
4947}
4948sub show_commit {
4950 my $c = shift;
4951 if ($oneline) {
4952 my $x = "\n";
4953 if (my $l = $c->{l}) {
4954 while ($l->[0] =~ /^\s*$/) { shift @$l }
4955 $x = $l->[0];
4956 }
4957 $l_fmt ||= 'A' . length($c->{r});
4958 print 'r',pack($l_fmt, $c->{r}),' | ';
4959 print "$c->{c} | " if $show_commit;
4960 print $x;
4961 } else {
4962 show_commit_normal($c);
4963 }
4964}
4965sub show_commit_changed_paths {
4967 my ($c) = @_;
4968 return unless $c->{changed};
4969 print "Changed paths:\n", @{$c->{changed}};
4970}
4971sub show_commit_normal {
4973 my ($c) = @_;
4974 print commit_log_separator, "r$c->{r} | ";
4975 print "$c->{c} | " if $show_commit;
4976 print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
4977 my $nr_line = 0;
4978 if (my $l = $c->{l}) {
4980 while ($l->[$#$l] eq "\n" && $#$l > 0
4981 && $l->[($#$l - 1)] eq "\n") {
4982 pop @$l;
4983 }
4984 $nr_line = scalar @$l;
4985 if (!$nr_line) {
4986 print "1 line\n\n\n";
4987 } else {
4988 if ($nr_line == 1) {
4989 $nr_line = '1 line';
4990 } else {
4991 $nr_line .= ' lines';
4992 }
4993 print $nr_line, "\n";
4994 show_commit_changed_paths($c);
4995 print "\n";
4996 print $_ foreach @$l;
4997 }
4998 } else {
4999 print "1 line\n";
5000 show_commit_changed_paths($c);
5001 print "\n";
5002 }
5004 foreach my $x (qw/raw stat diff/) {
5005 if ($c->{$x}) {
5006 print "\n";
5007 print $_ foreach @{$c->{$x}}
5008 }
5009 }
5010}
5011sub cmd_show_log {
5013 my (@args) = @_;
5014 my ($r_min, $r_max);
5015 my $r_last = -1; # prevent dupes
5016 set_local_timezone();
5017 if (defined $::_revision) {
5018 if ($::_revision =~ /^(\d+):(\d+)$/) {
5019 ($r_min, $r_max) = ($1, $2);
5020 } elsif ($::_revision =~ /^\d+$/) {
5021 $r_min = $r_max = $::_revision;
5022 } else {
5023 ::fatal "-r$::_revision is not supported, use ",
5024 "standard 'git log' arguments instead";
5025 }
5026 }
5027 config_pager();
5029 @args = git_svn_log_cmd($r_min, $r_max, @args);
5030 if (!@args) {
5031 print commit_log_separator unless $incremental || $oneline;
5032 return;
5033 }
5034 my $log = command_output_pipe(@args);
5035 run_pager();
5036 my (@k, $c, $d, $stat);
5037 my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
5038 while (<$log>) {
5039 if (/^${esc_color}commit -?($::sha1_short)/o) {
5040 my $cmt = $1;
5041 if ($c && cmt_showable($c) && $c->{r} != $r_last) {
5042 $r_last = $c->{r};
5043 process_commit($c, $r_min, $r_max, \@k) or
5044 goto out;
5045 }
5046 $d = undef;
5047 $c = { c => $cmt };
5048 } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
5049 get_author_info($c, $1, $2, $3);
5050 } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
5051 # ignore
5052 } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
5053 push @{$c->{raw}}, $_;
5054 } elsif (/^${esc_color}[ACRMDT]\t/) {
5055 # we could add $SVN->{svn_path} here, but that requires
5056 # remote access at the moment (repo_path_split)...
5057 s#^(${esc_color})([ACRMDT])\t#$1 $2 #o;
5058 push @{$c->{changed}}, $_;
5059 } elsif (/^${esc_color}diff /o) {
5060 $d = 1;
5061 push @{$c->{diff}}, $_;
5062 } elsif ($d) {
5063 push @{$c->{diff}}, $_;
5064 } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
5065 $esc_color*[\+\-]*$esc_color$/x) {
5066 $stat = 1;
5067 push @{$c->{stat}}, $_;
5068 } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
5069 push @{$c->{stat}}, $_;
5070 $stat = undef;
5071 } elsif (/^${esc_color} (git-svn-id:.+)$/o) {
5072 ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
5073 } elsif (s/^${esc_color} //o) {
5074 push @{$c->{l}}, $_;
5075 }
5076 }
5077 if ($c && defined $c->{r} && $c->{r} != $r_last) {
5078 $r_last = $c->{r};
5079 process_commit($c, $r_min, $r_max, \@k);
5080 }
5081 if (@k) {
5082 ($r_min, $r_max) = ($r_max, $r_min);
5083 process_commit($_, $r_min, $r_max) foreach reverse @k;
5084 }
5085out:
5086 close $log;
5087 print commit_log_separator unless $incremental || $oneline;
5088}
5089sub cmd_blame {
5091 my $path = pop;
5092 config_pager();
5094 run_pager();
5095 my ($fh, $ctx, $rev);
5097 if ($_git_format) {
5099 ($fh, $ctx) = command_output_pipe('blame', @_, $path);
5100 while (my $line = <$fh>) {
5101 if ($line =~ /^\^?([[:xdigit:]]+)\s/) {
5102 # Uncommitted edits show up as a rev ID of
5103 # all zeros, which we can't look up with
5104 # cmt_metadata
5105 if ($1 !~ /^0+$/) {
5106 (undef, $rev, undef) =
5107 ::cmt_metadata($1);
5108 $rev = '0' if (!$rev);
5109 } else {
5110 $rev = '0';
5111 }
5112 $rev = sprintf('%-10s', $rev);
5113 $line =~ s/^\^?[[:xdigit:]]+(\s)/$rev$1/;
5114 }
5115 print $line;
5116 }
5117 } else {
5118 ($fh, $ctx) = command_output_pipe('blame', '-p', @_, 'HEAD',
5119 '--', $path);
5120 my ($sha1);
5121 my %authors;
5122 my @buffer;
5123 my %dsha; #distinct sha keys
5124 while (my $line = <$fh>) {
5126 push @buffer, $line;
5127 if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
5128 $dsha{$1} = 1;
5129 }
5130 }
5131 my $s2r = ::cmt_sha2rev_batch([keys %dsha]);
5133 foreach my $line (@buffer) {
5135 if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
5136 $rev = $s2r->{$1};
5137 $rev = '0' if (!$rev)
5138 }
5139 elsif ($line =~ /^author (.*)/) {
5140 $authors{$rev} = $1;
5141 $authors{$rev} =~ s/\s/_/g;
5142 }
5143 elsif ($line =~ /^\t(.*)$/) {
5144 printf("%6s %10s %s\n", $rev, $authors{$rev}, $1);
5145 }
5146 }
5147 }
5148 command_close_pipe($fh, $ctx);
5149}
5150package Git::SVN::Migration;
5152# these version numbers do NOT correspond to actual version numbers
5153# of git nor git-svn. They are just relative.
5154#
5155# v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
5156#
5157# v1 layout: .git/$id/info/url, refs/remotes/$id
5158#
5159# v2 layout: .git/svn/$id/info/url, refs/remotes/$id
5160#
5161# v3 layout: .git/svn/$id, refs/remotes/$id
5162# - info/url may remain for backwards compatibility
5163# - this is what we migrate up to this layout automatically,
5164# - this will be used by git svn init on single branches
5165# v3.1 layout (auto migrated):
5166# - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
5167# for backwards compatibility
5168#
5169# v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
5170# - this is only created for newly multi-init-ed
5171# repositories. Similar in spirit to the
5172# --use-separate-remotes option in git-clone (now default)
5173# - we do not automatically migrate to this (following
5174# the example set by core git)
5175#
5176# v5 layout: .rev_db.$UUID => .rev_map.$UUID
5177# - newer, more-efficient format that uses 24-bytes per record
5178# with no filler space.
5179# - use xxd -c24 < .rev_map.$UUID to view and debug
5180# - This is a one-way migration, repositories updated to the
5181# new format will not be able to use old git-svn without
5182# rebuilding the .rev_db. Rebuilding the rev_db is not
5183# possible if noMetadata or useSvmProps are set; but should
5184# be no problem for users that use the (sensible) defaults.
5185use strict;
5186use warnings;
5187use Carp qw/croak/;
5188use File::Path qw/mkpath/;
5189use File::Basename qw/dirname basename/;
5190use vars qw/$_minimize/;
5191sub migrate_from_v0 {
5193 my $git_dir = $ENV{GIT_DIR};
5194 return undef unless -d $git_dir;
5195 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
5196 my $migrated = 0;
5197 while (<$fh>) {
5198 chomp;
5199 my ($id, $orig_ref) = ($_, $_);
5200 next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
5201 next unless -f "$git_dir/$id/info/url";
5202 my $new_ref = "refs/remotes/$id";
5203 if (::verify_ref("$new_ref^0")) {
5204 print STDERR "W: $orig_ref is probably an old ",
5205 "branch used by an ancient version of ",
5206 "git-svn.\n",
5207 "However, $new_ref also exists.\n",
5208 "We will not be able ",
5209 "to use this branch until this ",
5210 "ambiguity is resolved.\n";
5211 next;
5212 }
5213 print STDERR "Migrating from v0 layout...\n" if !$migrated;
5214 print STDERR "Renaming ref: $orig_ref => $new_ref\n";
5215 command_noisy('update-ref', $new_ref, $orig_ref);
5216 command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
5217 $migrated++;
5218 }
5219 command_close_pipe($fh, $ctx);
5220 print STDERR "Done migrating from v0 layout...\n" if $migrated;
5221 $migrated;
5222}
5223sub migrate_from_v1 {
5225 my $git_dir = $ENV{GIT_DIR};
5226 my $migrated = 0;
5227 return $migrated unless -d $git_dir;
5228 my $svn_dir = "$git_dir/svn";
5229 # just in case somebody used 'svn' as their $id at some point...
5231 return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
5232 print STDERR "Migrating from a git-svn v1 layout...\n";
5234 mkpath([$svn_dir]);
5235 print STDERR "Data from a previous version of git-svn exists, but\n\t",
5236 "$svn_dir\n\t(required for this version ",
5237 "($::VERSION) of git-svn) does not exist.\n";
5238 my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
5239 while (<$fh>) {
5240 my $x = $_;
5241 next unless $x =~ s#^refs/remotes/##;
5242 chomp $x;
5243 next unless -f "$git_dir/$x/info/url";
5244 my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
5245 next unless $u;
5246 my $dn = dirname("$git_dir/svn/$x");
5247 mkpath([$dn]) unless -d $dn;
5248 if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
5249 mkpath(["$git_dir/svn/svn"]);
5250 print STDERR " - $git_dir/$x/info => ",
5251 "$git_dir/svn/$x/info\n";
5252 rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
5253 croak "$!: $x";
5254 # don't worry too much about these, they probably
5255 # don't exist with repos this old (save for index,
5256 # and we can easily regenerate that)
5257 foreach my $f (qw/unhandled.log index .rev_db/) {
5258 rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
5259 }
5260 } else {
5261 print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
5262 rename "$git_dir/$x", "$git_dir/svn/$x" or
5263 croak "$!: $x";
5264 }
5265 $migrated++;
5266 }
5267 command_close_pipe($fh, $ctx);
5268 print STDERR "Done migrating from a git-svn v1 layout\n";
5269 $migrated;
5270}
5271sub read_old_urls {
5273 my ($l_map, $pfx, $path) = @_;
5274 my @dir;
5275 foreach (<$path/*>) {
5276 if (-r "$_/info/url") {
5277 $pfx .= '/' if $pfx && $pfx !~ m!/$!;
5278 my $ref_id = $pfx . basename $_;
5279 my $url = ::file_to_s("$_/info/url");
5280 $l_map->{$ref_id} = $url;
5281 } elsif (-d $_) {
5282 push @dir, $_;
5283 }
5284 }
5285 foreach (@dir) {
5286 my $x = $_;
5287 $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
5288 read_old_urls($l_map, $x, $_);
5289 }
5290}
5291sub migrate_from_v2 {
5293 my @cfg = command(qw/config -l/);
5294 return if grep /^svn-remote\..+\.url=/, @cfg;
5295 my %l_map;
5296 read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
5297 my $migrated = 0;
5298 foreach my $ref_id (sort keys %l_map) {
5300 eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
5301 if ($@) {
5302 Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
5303 }
5304 $migrated++;
5305 }
5306 $migrated;
5307}
5308sub minimize_connections {
5310 my $r = Git::SVN::read_all_remotes();
5311 my $new_urls = {};
5312 my $root_repos = {};
5313 foreach my $repo_id (keys %$r) {
5314 my $url = $r->{$repo_id}->{url} or next;
5315 my $fetch = $r->{$repo_id}->{fetch} or next;
5316 my $ra = Git::SVN::Ra->new($url);
5317 # skip existing cases where we already connect to the root
5319 if (($ra->{url} eq $ra->{repos_root}) ||
5320 ($ra->{repos_root} eq $repo_id)) {
5321 $root_repos->{$ra->{url}} = $repo_id;
5322 next;
5323 }
5324 my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
5326 my $root_path = $ra->{url};
5327 $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
5328 foreach my $path (keys %$fetch) {
5329 my $ref_id = $fetch->{$path};
5330 my $gs = Git::SVN->new($ref_id, $repo_id, $path);
5331 # make sure we can read when connecting to
5333 # a higher level of a repository
5334 my ($last_rev, undef) = $gs->last_rev_commit;
5335 if (!defined $last_rev) {
5336 $last_rev = eval {
5337 $root_ra->get_latest_revnum;
5338 };
5339 next if $@;
5340 }
5341 my $new = $root_path;
5342 $new .= length $path ? "/$path" : '';
5343 eval {
5344 $root_ra->get_log([$new], $last_rev, $last_rev,
5345 0, 0, 1, sub { });
5346 };
5347 next if $@;
5348 $new_urls->{$ra->{repos_root}}->{$new} =
5349 { ref_id => $ref_id,
5350 old_repo_id => $repo_id,
5351 old_path => $path };
5352 }
5353 }
5354 my @emptied;
5356 foreach my $url (keys %$new_urls) {
5357 # see if we can re-use an existing [svn-remote "repo_id"]
5358 # instead of creating a(n ugly) new section:
5359 my $repo_id = $root_repos->{$url} || $url;
5360 my $fetch = $new_urls->{$url};
5362 foreach my $path (keys %$fetch) {
5363 my $x = $fetch->{$path};
5364 Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
5365 my $pfx = "svn-remote.$x->{old_repo_id}";
5366 my $old_fetch = quotemeta("$x->{old_path}:".
5368 "refs/remotes/$x->{ref_id}");
5369 command_noisy(qw/config --unset/,
5370 "$pfx.fetch", '^'. $old_fetch . '$');
5371 delete $r->{$x->{old_repo_id}}->
5372 {fetch}->{$x->{old_path}};
5373 if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
5374 command_noisy(qw/config --unset/,
5375 "$pfx.url");
5376 push @emptied, $x->{old_repo_id}
5377 }
5378 }
5379 }
5380 if (@emptied) {
5381 my $file = $ENV{GIT_CONFIG} || "$ENV{GIT_DIR}/config";
5382 print STDERR <<EOF;
5383The following [svn-remote] sections in your config file ($file) are empty
5384and can be safely removed:
5385EOF
5386 print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
5387 }
5388}
5389sub migration_check {
5391 migrate_from_v0();
5392 migrate_from_v1();
5393 migrate_from_v2();
5394 minimize_connections() if $_minimize;
5395}
5396package Git::IndexInfo;
5398use strict;
5399use warnings;
5400use Git qw/command_input_pipe command_close_pipe/;
5401sub new {
5403 my ($class) = @_;
5404 my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
5405 bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
5406}
5407sub remove {
5409 my ($self, $path) = @_;
5410 if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
5411 return ++$self->{nr};
5412 }
5413 undef;
5414}
5415sub update {
5417 my ($self, $mode, $hash, $path) = @_;
5418 if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
5419 return ++$self->{nr};
5420 }
5421 undef;
5422}
5423sub DESTROY {
5425 my ($self) = @_;
5426 command_close_pipe($self->{gui}, $self->{ctx});
5427}
5428package Git::SVN::GlobSpec;
5430use strict;
5431use warnings;
5432sub new {
5434 my ($class, $glob) = @_;
5435 my $re = $glob;
5436 $re =~ s!/+$!!g; # no need for trailing slashes
5437 $re =~ m!^([^*]*)(\*(?:/\*)*)([^*]*)$!;
5438 my $temp = $re;
5439 my ($left, $right) = ($1, $3);
5440 $re = $2;
5441 my $depth = $re =~ tr/*/*/;
5442 if ($depth != $temp =~ tr/*/*/) {
5443 die "Only one set of wildcard directories " .
5444 "(e.g. '*' or '*/*/*') is supported: '$glob'\n";
5445 }
5446 if ($depth == 0) {
5447 die "One '*' is needed for glob: '$glob'\n";
5448 }
5449 $re =~ s!\*!\[^/\]*!g;
5450 $re = quotemeta($left) . "($re)" . quotemeta($right);
5451 if (length $left && !($left =~ s!/+$!!g)) {
5452 die "Missing trailing '/' on left side of: '$glob' ($left)\n";
5453 }
5454 if (length $right && !($right =~ s!^/+!!g)) {
5455 die "Missing leading '/' on right side of: '$glob' ($right)\n";
5456 }
5457 my $left_re = qr/^\/\Q$left\E(\/|$)/;
5458 bless { left => $left, right => $right, left_regex => $left_re,
5459 regex => qr/$re/, glob => $glob, depth => $depth }, $class;
5460}
5461sub full_path {
5463 my ($self, $path) = @_;
5464 return (length $self->{left} ? "$self->{left}/" : '') .
5465 $path . (length $self->{right} ? "/$self->{right}" : '');
5466}
5467__END__
5469Data structures:
5471$remotes = { # returned by read_all_remotes()
5474 'svn' => {
5475 # svn-remote.svn.url=https://svn.musicpd.org
5476 url => 'https://svn.musicpd.org',
5477 # svn-remote.svn.fetch=mpd/trunk:trunk
5478 fetch => {
5479 'mpd/trunk' => 'trunk',
5480 },
5481 # svn-remote.svn.tags=mpd/tags/*:tags/*
5482 tags => {
5483 path => {
5484 left => 'mpd/tags',
5485 right => '',
5486 regex => qr!mpd/tags/([^/]+)$!,
5487 glob => 'tags/*',
5488 },
5489 ref => {
5490 left => 'tags',
5491 right => '',
5492 regex => qr!tags/([^/]+)$!,
5493 glob => 'tags/*',
5494 },
5495 }
5496 }
5497};
5498$log_entry hashref as returned by libsvn_log_entry()
5500{
5501 log => 'whitespace-formatted log entry
5502', # trailing newline is preserved
5503 revision => '8', # integer
5504 date => '2004-02-24T17:01:44.108345Z', # commit date
5505 author => 'committer name'
5506};
5507# this is generated by generate_diff();
5510@mods = array of diff-index line hashes, each element represents one line
5511 of diff-index output
5512diff-index line ($m hash)
5514{
5515 mode_a => first column of diff-index output, no leading ':',
5516 mode_b => second column of diff-index output,
5517 sha1_b => sha1sum of the final blob,
5518 chg => change type [MCRADT],
5519 file_a => original file name of a file (iff chg is 'C' or 'R')
5520 file_b => new/current file name of a file (any chg)
5521}
5522;
5523# retval of read_url_paths{,_all}();
5525$l_map = {
5526 # repository root url
5527 'https://svn.musicpd.org' => {
5528 # repository path # GIT_SVN_ID
5529 'mpd/trunk' => 'trunk',
5530 'mpd/tags/0.11.5' => 'tags/0.11.5',
5531 },
5532}
5533Notes:
5535 I don't trust the each() function on unless I created %hash myself
5536 because the internal iterator may not have started at base.