git-svn.perlon commit Docs: send-email: --chain_reply_to -> --[no-]chain-reply-to (c306e07)
   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 $_authors %users/;
   9$AUTHOR = 'Eric Wong <normalperson@yhbt.net>';
  10$VERSION = '@@GIT_VERSION@@';
  11
  12# From which subdir have we been invoked?
  13my $cmd_dir_prefix = eval {
  14        command_oneline([qw/rev-parse --show-prefix/], STDERR => 0)
  15} || '';
  16
  17my $git_dir_user_set = 1 if defined $ENV{GIT_DIR};
  18$ENV{GIT_DIR} ||= '.git';
  19$Git::SVN::default_repo_id = 'svn';
  20$Git::SVN::default_ref_id = $ENV{GIT_SVN_ID} || 'git-svn';
  21$Git::SVN::Ra::_log_window_size = 100;
  22
  23$Git::SVN::Log::TZ = $ENV{TZ};
  24$ENV{TZ} = 'UTC';
  25$| = 1; # unbuffer STDOUT
  26
  27sub fatal (@) { print STDERR "@_\n"; exit 1 }
  28require SVN::Core; # use()-ing this causes segfaults for me... *shrug*
  29require SVN::Ra;
  30require SVN::Delta;
  31if ($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 Getopt::Long qw/:config gnu_getopt no_ignore_case auto_abbrev/;
  43use IPC::Open3;
  44use Git;
  45
  46BEGIN {
  47        # import functions from Git into our packages, en masse
  48        no strict 'refs';
  49        foreach (qw/command command_oneline command_noisy command_output_pipe
  50                    command_input_pipe command_close_pipe/) {
  51                for my $package ( qw(SVN::Git::Editor SVN::Git::Fetcher
  52                        Git::SVN::Migration Git::SVN::Log Git::SVN),
  53                        __PACKAGE__) {
  54                        *{"${package}::$_"} = \&{"Git::$_"};
  55                }
  56        }
  57}
  58
  59my ($SVN);
  60
  61$sha1 = qr/[a-f\d]{40}/;
  62$sha1_short = qr/[a-f\d]{4,40}/;
  63my ($_stdin, $_help, $_edit,
  64        $_message, $_file,
  65        $_template, $_shared,
  66        $_version, $_fetch_all, $_no_rebase,
  67        $_merge, $_strategy, $_dry_run, $_local,
  68        $_prefix, $_no_checkout, $_url, $_verbose,
  69        $_git_format, $_commit_url);
  70$Git::SVN::_follow_parent = 1;
  71my %remote_opts = ( 'username=s' => \$Git::SVN::Prompt::_username,
  72                    'config-dir=s' => \$Git::SVN::Ra::config_dir,
  73                    'no-auth-cache' => \$Git::SVN::Prompt::_no_auth_cache );
  74my %fc_opts = ( 'follow-parent|follow!' => \$Git::SVN::_follow_parent,
  75                'authors-file|A=s' => \$_authors,
  76                'repack:i' => \$Git::SVN::_repack,
  77                'noMetadata' => \$Git::SVN::_no_metadata,
  78                'useSvmProps' => \$Git::SVN::_use_svm_props,
  79                'useSvnsyncProps' => \$Git::SVN::_use_svnsync_props,
  80                'log-window-size=i' => \$Git::SVN::Ra::_log_window_size,
  81                'no-checkout' => \$_no_checkout,
  82                'quiet|q' => \$_q,
  83                'repack-flags|repack-args|repack-opts=s' =>
  84                   \$Git::SVN::_repack_flags,
  85                'use-log-author' => \$Git::SVN::_use_log_author,
  86                'add-author-from' => \$Git::SVN::_add_author_from,
  87                %remote_opts );
  88
  89my ($_trunk, $_tags, $_branches, $_stdlayout);
  90my %icv;
  91my %init_opts = ( 'template=s' => \$_template, 'shared:s' => \$_shared,
  92                  'trunk|T=s' => \$_trunk, 'tags|t=s' => \$_tags,
  93                  'branches|b=s' => \$_branches, 'prefix=s' => \$_prefix,
  94                  'stdlayout|s' => \$_stdlayout,
  95                  'minimize-url|m' => \$Git::SVN::_minimize_url,
  96                  'no-metadata' => sub { $icv{noMetadata} = 1 },
  97                  'use-svm-props' => sub { $icv{useSvmProps} = 1 },
  98                  'use-svnsync-props' => sub { $icv{useSvnsyncProps} = 1 },
  99                  'rewrite-root=s' => sub { $icv{rewriteRoot} = $_[1] },
 100                  %remote_opts );
 101my %cmt_opts = ( 'edit|e' => \$_edit,
 102                'rmdir' => \$SVN::Git::Editor::_rmdir,
 103                'find-copies-harder' => \$SVN::Git::Editor::_find_copies_harder,
 104                'l=i' => \$SVN::Git::Editor::_rename_limit,
 105                'copy-similarity|C=i'=> \$SVN::Git::Editor::_cp_similarity
 106);
 107
 108my %cmd = (
 109        fetch => [ \&cmd_fetch, "Download new revisions from SVN",
 110                        { 'revision|r=s' => \$_revision,
 111                          'fetch-all|all' => \$_fetch_all,
 112                           %fc_opts } ],
 113        clone => [ \&cmd_clone, "Initialize and fetch revisions",
 114                        { 'revision|r=s' => \$_revision,
 115                           %fc_opts, %init_opts } ],
 116        init => [ \&cmd_init, "Initialize a repo for tracking" .
 117                          " (requires URL argument)",
 118                          \%init_opts ],
 119        'multi-init' => [ \&cmd_multi_init,
 120                          "Deprecated alias for ".
 121                          "'$0 init -T<trunk> -b<branches> -t<tags>'",
 122                          \%init_opts ],
 123        dcommit => [ \&cmd_dcommit,
 124                     'Commit several diffs to merge with upstream',
 125                        { 'merge|m|M' => \$_merge,
 126                          'strategy|s=s' => \$_strategy,
 127                          'verbose|v' => \$_verbose,
 128                          'dry-run|n' => \$_dry_run,
 129                          'fetch-all|all' => \$_fetch_all,
 130                          'commit-url=s' => \$_commit_url,
 131                          'revision|r=i' => \$_revision,
 132                          'no-rebase' => \$_no_rebase,
 133                        %cmt_opts, %fc_opts } ],
 134        'set-tree' => [ \&cmd_set_tree,
 135                        "Set an SVN repository to a git tree-ish",
 136                        { 'stdin|' => \$_stdin, %cmt_opts, %fc_opts, } ],
 137        'create-ignore' => [ \&cmd_create_ignore,
 138                             'Create a .gitignore per svn:ignore',
 139                             { 'revision|r=i' => \$_revision
 140                             } ],
 141        'propget' => [ \&cmd_propget,
 142                       'Print the value of a property on a file or directory',
 143                       { 'revision|r=i' => \$_revision } ],
 144        'proplist' => [ \&cmd_proplist,
 145                       'List all properties of a file or directory',
 146                       { 'revision|r=i' => \$_revision } ],
 147        'show-ignore' => [ \&cmd_show_ignore, "Show svn:ignore listings",
 148                        { 'revision|r=i' => \$_revision
 149                        } ],
 150        'show-externals' => [ \&cmd_show_externals, "Show svn:externals listings",
 151                        { 'revision|r=i' => \$_revision
 152                        } ],
 153        'multi-fetch' => [ \&cmd_multi_fetch,
 154                           "Deprecated alias for $0 fetch --all",
 155                           { 'revision|r=s' => \$_revision, %fc_opts } ],
 156        'migrate' => [ sub { },
 157                       # no-op, we automatically run this anyways,
 158                       'Migrate configuration/metadata/layout from
 159                        previous versions of git-svn',
 160                       { 'minimize' => \$Git::SVN::Migration::_minimize,
 161                         %remote_opts } ],
 162        'log' => [ \&Git::SVN::Log::cmd_show_log, 'Show commit logs',
 163                        { 'limit=i' => \$Git::SVN::Log::limit,
 164                          'revision|r=s' => \$_revision,
 165                          'verbose|v' => \$Git::SVN::Log::verbose,
 166                          'incremental' => \$Git::SVN::Log::incremental,
 167                          'oneline' => \$Git::SVN::Log::oneline,
 168                          'show-commit' => \$Git::SVN::Log::show_commit,
 169                          'non-recursive' => \$Git::SVN::Log::non_recursive,
 170                          'authors-file|A=s' => \$_authors,
 171                          'color' => \$Git::SVN::Log::color,
 172                          'pager=s' => \$Git::SVN::Log::pager
 173                        } ],
 174        'find-rev' => [ \&cmd_find_rev,
 175                        "Translate between SVN revision numbers and tree-ish",
 176                        {} ],
 177        'rebase' => [ \&cmd_rebase, "Fetch and rebase your working directory",
 178                        { 'merge|m|M' => \$_merge,
 179                          'verbose|v' => \$_verbose,
 180                          'strategy|s=s' => \$_strategy,
 181                          'local|l' => \$_local,
 182                          'fetch-all|all' => \$_fetch_all,
 183                          'dry-run|n' => \$_dry_run,
 184                          %fc_opts } ],
 185        'commit-diff' => [ \&cmd_commit_diff,
 186                           'Commit a diff between two trees',
 187                        { 'message|m=s' => \$_message,
 188                          'file|F=s' => \$_file,
 189                          'revision|r=s' => \$_revision,
 190                        %cmt_opts } ],
 191        'info' => [ \&cmd_info,
 192                    "Show info about the latest SVN revision
 193                     on the current branch",
 194                    { 'url' => \$_url, } ],
 195        'blame' => [ \&Git::SVN::Log::cmd_blame,
 196                    "Show what revision and author last modified each line of a file",
 197                    { 'git-format' => \$_git_format } ],
 198);
 199
 200my $cmd;
 201for (my $i = 0; $i < @ARGV; $i++) {
 202        if (defined $cmd{$ARGV[$i]}) {
 203                $cmd = $ARGV[$i];
 204                splice @ARGV, $i, 1;
 205                last;
 206        }
 207};
 208
 209# make sure we're always running at the top-level working directory
 210unless ($cmd && $cmd =~ /(?:clone|init|multi-init)$/) {
 211        unless (-d $ENV{GIT_DIR}) {
 212                if ($git_dir_user_set) {
 213                        die "GIT_DIR=$ENV{GIT_DIR} explicitly set, ",
 214                            "but it is not a directory\n";
 215                }
 216                my $git_dir = delete $ENV{GIT_DIR};
 217                chomp(my $cdup = command_oneline(qw/rev-parse --show-cdup/));
 218                unless (length $cdup) {
 219                        die "Already at toplevel, but $git_dir ",
 220                            "not found '$cdup'\n";
 221                }
 222                chdir $cdup or die "Unable to chdir up to '$cdup'\n";
 223                unless (-d $git_dir) {
 224                        die "$git_dir still not found after going to ",
 225                            "'$cdup'\n";
 226                }
 227                $ENV{GIT_DIR} = $git_dir;
 228        }
 229        $_repository = Git->repository(Repository => $ENV{GIT_DIR});
 230}
 231
 232my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
 233
 234read_repo_config(\%opts);
 235if ($cmd && ($cmd eq 'log' || $cmd eq 'blame')) {
 236        Getopt::Long::Configure('pass_through');
 237}
 238my $rv = GetOptions(%opts, 'help|H|h' => \$_help, 'version|V' => \$_version,
 239                    'minimize-connections' => \$Git::SVN::Migration::_minimize,
 240                    'id|i=s' => \$Git::SVN::default_ref_id,
 241                    'svn-remote|remote|R=s' => sub {
 242                       $Git::SVN::no_reuse_existing = 1;
 243                       $Git::SVN::default_repo_id = $_[1] });
 244exit 1 if (!$rv && $cmd && $cmd ne 'log');
 245
 246usage(0) if $_help;
 247version() if $_version;
 248usage(1) unless defined $cmd;
 249load_authors() if $_authors;
 250
 251unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
 252        Git::SVN::Migration::migration_check();
 253}
 254Git::SVN::init_vars();
 255eval {
 256        Git::SVN::verify_remotes_sanity();
 257        $cmd{$cmd}->[0]->(@ARGV);
 258};
 259fatal $@ if $@;
 260post_fetch_checkout();
 261exit 0;
 262
 263####################### primary functions ######################
 264sub usage {
 265        my $exit = shift || 0;
 266        my $fd = $exit ? \*STDERR : \*STDOUT;
 267        print $fd <<"";
 268git-svn - bidirectional operations between a single Subversion tree and git
 269Usage: git svn <command> [options] [arguments]\n
 270
 271        print $fd "Available commands:\n" unless $cmd;
 272
 273        foreach (sort keys %cmd) {
 274                next if $cmd && $cmd ne $_;
 275                next if /^multi-/; # don't show deprecated commands
 276                print $fd '  ',pack('A17',$_),$cmd{$_}->[1],"\n";
 277                foreach (sort keys %{$cmd{$_}->[2]}) {
 278                        # mixed-case options are for .git/config only
 279                        next if /[A-Z]/ && /^[a-z]+$/i;
 280                        # prints out arguments as they should be passed:
 281                        my $x = s#[:=]s$## ? '<arg>' : s#[:=]i$## ? '<num>' : '';
 282                        print $fd ' ' x 21, join(', ', map { length $_ > 1 ?
 283                                                        "--$_" : "-$_" }
 284                                                split /\|/,$_)," $x\n";
 285                }
 286        }
 287        print $fd <<"";
 288\nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
 289arbitrary identifier if you're tracking multiple SVN branches/repositories in
 290one git repository and want to keep them separate.  See git-svn(1) for more
 291information.
 292
 293        exit $exit;
 294}
 295
 296sub version {
 297        print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
 298        exit 0;
 299}
 300
 301sub do_git_init_db {
 302        unless (-d $ENV{GIT_DIR}) {
 303                my @init_db = ('init');
 304                push @init_db, "--template=$_template" if defined $_template;
 305                if (defined $_shared) {
 306                        if ($_shared =~ /[a-z]/) {
 307                                push @init_db, "--shared=$_shared";
 308                        } else {
 309                                push @init_db, "--shared";
 310                        }
 311                }
 312                command_noisy(@init_db);
 313                $_repository = Git->repository(Repository => ".git");
 314        }
 315        my $set;
 316        my $pfx = "svn-remote.$Git::SVN::default_repo_id";
 317        foreach my $i (keys %icv) {
 318                die "'$set' and '$i' cannot both be set\n" if $set;
 319                next unless defined $icv{$i};
 320                command_noisy('config', "$pfx.$i", $icv{$i});
 321                $set = $i;
 322        }
 323}
 324
 325sub init_subdir {
 326        my $repo_path = shift or return;
 327        mkpath([$repo_path]) unless -d $repo_path;
 328        chdir $repo_path or die "Couldn't chdir to $repo_path: $!\n";
 329        $ENV{GIT_DIR} = '.git';
 330        $_repository = Git->repository(Repository => $ENV{GIT_DIR});
 331}
 332
 333sub cmd_clone {
 334        my ($url, $path) = @_;
 335        if (!defined $path &&
 336            (defined $_trunk || defined $_branches || defined $_tags ||
 337             defined $_stdlayout) &&
 338            $url !~ m#^[a-z\+]+://#) {
 339                $path = $url;
 340        }
 341        $path = basename($url) if !defined $path || !length $path;
 342        cmd_init($url, $path);
 343        Git::SVN::fetch_all($Git::SVN::default_repo_id);
 344}
 345
 346sub cmd_init {
 347        if (defined $_stdlayout) {
 348                $_trunk = 'trunk' if (!defined $_trunk);
 349                $_tags = 'tags' if (!defined $_tags);
 350                $_branches = 'branches' if (!defined $_branches);
 351        }
 352        if (defined $_trunk || defined $_branches || defined $_tags) {
 353                return cmd_multi_init(@_);
 354        }
 355        my $url = shift or die "SVN repository location required ",
 356                               "as a command-line argument\n";
 357        init_subdir(@_);
 358        do_git_init_db();
 359
 360        Git::SVN->init($url);
 361}
 362
 363sub cmd_fetch {
 364        if (grep /^\d+=./, @_) {
 365                die "'<rev>=<commit>' fetch arguments are ",
 366                    "no longer supported.\n";
 367        }
 368        my ($remote) = @_;
 369        if (@_ > 1) {
 370                die "Usage: $0 fetch [--all] [svn-remote]\n";
 371        }
 372        $remote ||= $Git::SVN::default_repo_id;
 373        if ($_fetch_all) {
 374                cmd_multi_fetch();
 375        } else {
 376                Git::SVN::fetch_all($remote, Git::SVN::read_all_remotes());
 377        }
 378}
 379
 380sub cmd_set_tree {
 381        my (@commits) = @_;
 382        if ($_stdin || !@commits) {
 383                print "Reading from stdin...\n";
 384                @commits = ();
 385                while (<STDIN>) {
 386                        if (/\b($sha1_short)\b/o) {
 387                                unshift @commits, $1;
 388                        }
 389                }
 390        }
 391        my @revs;
 392        foreach my $c (@commits) {
 393                my @tmp = command('rev-parse',$c);
 394                if (scalar @tmp == 1) {
 395                        push @revs, $tmp[0];
 396                } elsif (scalar @tmp > 1) {
 397                        push @revs, reverse(command('rev-list',@tmp));
 398                } else {
 399                        fatal "Failed to rev-parse $c";
 400                }
 401        }
 402        my $gs = Git::SVN->new;
 403        my ($r_last, $cmt_last) = $gs->last_rev_commit;
 404        $gs->fetch;
 405        if (defined $gs->{last_rev} && $r_last != $gs->{last_rev}) {
 406                fatal "There are new revisions that were fetched ",
 407                      "and need to be merged (or acknowledged) ",
 408                      "before committing.\nlast rev: $r_last\n",
 409                      " current: $gs->{last_rev}";
 410        }
 411        $gs->set_tree($_) foreach @revs;
 412        print "Done committing ",scalar @revs," revisions to SVN\n";
 413        unlink $gs->{index};
 414}
 415
 416sub cmd_dcommit {
 417        my $head = shift;
 418        git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
 419                'Cannot dcommit with a dirty index.  Commit your changes first, '
 420                . "or stash them with `git stash'.\n";
 421        $head ||= 'HEAD';
 422        my @refs;
 423        my ($url, $rev, $uuid, $gs) = working_head_info($head, \@refs);
 424        unless ($gs) {
 425                die "Unable to determine upstream SVN information from ",
 426                    "$head history.\nPerhaps the repository is empty.";
 427        }
 428        $url = defined $_commit_url ? $_commit_url : $gs->full_url;
 429        my $last_rev = $_revision if defined $_revision;
 430        if ($url) {
 431                print "Committing to $url ...\n";
 432        }
 433        my ($linear_refs, $parents) = linearize_history($gs, \@refs);
 434        if ($_no_rebase && scalar(@$linear_refs) > 1) {
 435                warn "Attempting to commit more than one change while ",
 436                     "--no-rebase is enabled.\n",
 437                     "If these changes depend on each other, re-running ",
 438                     "without --no-rebase may be required."
 439        }
 440        my $expect_url = $url;
 441        Git::SVN::remove_username($expect_url);
 442        while (1) {
 443                my $d = shift @$linear_refs or last;
 444                unless (defined $last_rev) {
 445                        (undef, $last_rev, undef) = cmt_metadata("$d~1");
 446                        unless (defined $last_rev) {
 447                                fatal "Unable to extract revision information ",
 448                                      "from commit $d~1";
 449                        }
 450                }
 451                if ($_dry_run) {
 452                        print "diff-tree $d~1 $d\n";
 453                } else {
 454                        my $cmt_rev;
 455                        my %ed_opts = ( r => $last_rev,
 456                                        log => get_commit_entry($d)->{log},
 457                                        ra => Git::SVN::Ra->new($url),
 458                                        config => SVN::Core::config_get_config(
 459                                                $Git::SVN::Ra::config_dir
 460                                        ),
 461                                        tree_a => "$d~1",
 462                                        tree_b => $d,
 463                                        editor_cb => sub {
 464                                               print "Committed r$_[0]\n";
 465                                               $cmt_rev = $_[0];
 466                                        },
 467                                        svn_path => '');
 468                        if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
 469                                print "No changes\n$d~1 == $d\n";
 470                        } elsif ($parents->{$d} && @{$parents->{$d}}) {
 471                                $gs->{inject_parents_dcommit}->{$cmt_rev} =
 472                                                               $parents->{$d};
 473                        }
 474                        $_fetch_all ? $gs->fetch_all : $gs->fetch;
 475                        $last_rev = $cmt_rev;
 476                        next if $_no_rebase;
 477
 478                        # we always want to rebase against the current HEAD,
 479                        # not any head that was passed to us
 480                        my @diff = command('diff-tree', $d,
 481                                           $gs->refname, '--');
 482                        my @finish;
 483                        if (@diff) {
 484                                @finish = rebase_cmd();
 485                                print STDERR "W: $d and ", $gs->refname,
 486                                             " differ, using @finish:\n",
 487                                             join("\n", @diff), "\n";
 488                        } else {
 489                                print "No changes between current HEAD and ",
 490                                      $gs->refname,
 491                                      "\nResetting to the latest ",
 492                                      $gs->refname, "\n";
 493                                @finish = qw/reset --mixed/;
 494                        }
 495                        command_noisy(@finish, $gs->refname);
 496                        if (@diff) {
 497                                @refs = ();
 498                                my ($url_, $rev_, $uuid_, $gs_) =
 499                                              working_head_info($head, \@refs);
 500                                my ($linear_refs_, $parents_) =
 501                                              linearize_history($gs_, \@refs);
 502                                if (scalar(@$linear_refs) !=
 503                                    scalar(@$linear_refs_)) {
 504                                        fatal "# of revisions changed ",
 505                                          "\nbefore:\n",
 506                                          join("\n", @$linear_refs),
 507                                          "\n\nafter:\n",
 508                                          join("\n", @$linear_refs_), "\n",
 509                                          'If you are attempting to commit ',
 510                                          "merges, try running:\n\t",
 511                                          'git rebase --interactive',
 512                                          '--preserve-merges ',
 513                                          $gs->refname,
 514                                          "\nBefore dcommitting";
 515                                }
 516                                if ($url_ ne $expect_url) {
 517                                        fatal "URL mismatch after rebase: ",
 518                                              "$url_ != $expect_url";
 519                                }
 520                                if ($uuid_ ne $uuid) {
 521                                        fatal "uuid mismatch after rebase: ",
 522                                              "$uuid_ != $uuid";
 523                                }
 524                                # remap parents
 525                                my (%p, @l, $i);
 526                                for ($i = 0; $i < scalar @$linear_refs; $i++) {
 527                                        my $new = $linear_refs_->[$i] or next;
 528                                        $p{$new} =
 529                                                $parents->{$linear_refs->[$i]};
 530                                        push @l, $new;
 531                                }
 532                                $parents = \%p;
 533                                $linear_refs = \@l;
 534                        }
 535                }
 536        }
 537        unlink $gs->{index};
 538}
 539
 540sub cmd_find_rev {
 541        my $revision_or_hash = shift or die "SVN or git revision required ",
 542                                            "as a command-line argument\n";
 543        my $result;
 544        if ($revision_or_hash =~ /^r\d+$/) {
 545                my $head = shift;
 546                $head ||= 'HEAD';
 547                my @refs;
 548                my (undef, undef, $uuid, $gs) = working_head_info($head, \@refs);
 549                unless ($gs) {
 550                        die "Unable to determine upstream SVN information from ",
 551                            "$head history\n";
 552                }
 553                my $desired_revision = substr($revision_or_hash, 1);
 554                $result = $gs->rev_map_get($desired_revision, $uuid);
 555        } else {
 556                my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
 557                $result = $rev;
 558        }
 559        print "$result\n" if $result;
 560}
 561
 562sub cmd_rebase {
 563        command_noisy(qw/update-index --refresh/);
 564        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 565        unless ($gs) {
 566                die "Unable to determine upstream SVN information from ",
 567                    "working tree history\n";
 568        }
 569        if ($_dry_run) {
 570                print "Remote Branch: " . $gs->refname . "\n";
 571                print "SVN URL: " . $url . "\n";
 572                return;
 573        }
 574        if (command(qw/diff-index HEAD --/)) {
 575                print STDERR "Cannot rebase with uncommited changes:\n";
 576                command_noisy('status');
 577                exit 1;
 578        }
 579        unless ($_local) {
 580                # rebase will checkout for us, so no need to do it explicitly
 581                $_no_checkout = 'true';
 582                $_fetch_all ? $gs->fetch_all : $gs->fetch;
 583        }
 584        command_noisy(rebase_cmd(), $gs->refname);
 585}
 586
 587sub cmd_show_ignore {
 588        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 589        $gs ||= Git::SVN->new;
 590        my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
 591        $gs->prop_walk($gs->{path}, $r, sub {
 592                my ($gs, $path, $props) = @_;
 593                print STDOUT "\n# $path\n";
 594                my $s = $props->{'svn:ignore'} or return;
 595                $s =~ s/[\r\n]+/\n/g;
 596                chomp $s;
 597                $s =~ s#^#$path#gm;
 598                print STDOUT "$s\n";
 599        });
 600}
 601
 602sub cmd_show_externals {
 603        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 604        $gs ||= Git::SVN->new;
 605        my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
 606        $gs->prop_walk($gs->{path}, $r, sub {
 607                my ($gs, $path, $props) = @_;
 608                print STDOUT "\n# $path\n";
 609                my $s = $props->{'svn:externals'} or return;
 610                $s =~ s/[\r\n]+/\n/g;
 611                chomp $s;
 612                $s =~ s#^#$path#gm;
 613                print STDOUT "$s\n";
 614        });
 615}
 616
 617sub cmd_create_ignore {
 618        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 619        $gs ||= Git::SVN->new;
 620        my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
 621        $gs->prop_walk($gs->{path}, $r, sub {
 622                my ($gs, $path, $props) = @_;
 623                # $path is of the form /path/to/dir/
 624                my $ignore = '.' . $path . '.gitignore';
 625                my $s = $props->{'svn:ignore'} or return;
 626                open(GITIGNORE, '>', $ignore)
 627                  or fatal("Failed to open `$ignore' for writing: $!");
 628                $s =~ s/[\r\n]+/\n/g;
 629                chomp $s;
 630                # Prefix all patterns so that the ignore doesn't apply
 631                # to sub-directories.
 632                $s =~ s#^#/#gm;
 633                print GITIGNORE "$s\n";
 634                close(GITIGNORE)
 635                  or fatal("Failed to close `$ignore': $!");
 636                command_noisy('add', '-f', $ignore);
 637        });
 638}
 639
 640sub canonicalize_path {
 641        my ($path) = @_;
 642        my $dot_slash_added = 0;
 643        if (substr($path, 0, 1) ne "/") {
 644                $path = "./" . $path;
 645                $dot_slash_added = 1;
 646        }
 647        # File::Spec->canonpath doesn't collapse x/../y into y (for a
 648        # good reason), so let's do this manually.
 649        $path =~ s#/+#/#g;
 650        $path =~ s#/\.(?:/|$)#/#g;
 651        $path =~ s#/[^/]+/\.\.##g;
 652        $path =~ s#/$##g;
 653        $path =~ s#^\./## if $dot_slash_added;
 654        $path =~ s#^/##;
 655        $path =~ s#^\.$##;
 656        return $path;
 657}
 658
 659# get_svnprops(PATH)
 660# ------------------
 661# Helper for cmd_propget and cmd_proplist below.
 662sub get_svnprops {
 663        my $path = shift;
 664        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 665        $gs ||= Git::SVN->new;
 666
 667        # prefix THE PATH by the sub-directory from which the user
 668        # invoked us.
 669        $path = $cmd_dir_prefix . $path;
 670        fatal("No such file or directory: $path") unless -e $path;
 671        my $is_dir = -d $path ? 1 : 0;
 672        $path = $gs->{path} . '/' . $path;
 673
 674        # canonicalize the path (otherwise libsvn will abort or fail to
 675        # find the file)
 676        $path = canonicalize_path($path);
 677
 678        my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
 679        my $props;
 680        if ($is_dir) {
 681                (undef, undef, $props) = $gs->ra->get_dir($path, $r);
 682        }
 683        else {
 684                (undef, $props) = $gs->ra->get_file($path, $r, undef);
 685        }
 686        return $props;
 687}
 688
 689# cmd_propget (PROP, PATH)
 690# ------------------------
 691# Print the SVN property PROP for PATH.
 692sub cmd_propget {
 693        my ($prop, $path) = @_;
 694        $path = '.' if not defined $path;
 695        usage(1) if not defined $prop;
 696        my $props = get_svnprops($path);
 697        if (not defined $props->{$prop}) {
 698                fatal("`$path' does not have a `$prop' SVN property.");
 699        }
 700        print $props->{$prop} . "\n";
 701}
 702
 703# cmd_proplist (PATH)
 704# -------------------
 705# Print the list of SVN properties for PATH.
 706sub cmd_proplist {
 707        my $path = shift;
 708        $path = '.' if not defined $path;
 709        my $props = get_svnprops($path);
 710        print "Properties on '$path':\n";
 711        foreach (sort keys %{$props}) {
 712                print "  $_\n";
 713        }
 714}
 715
 716sub cmd_multi_init {
 717        my $url = shift;
 718        unless (defined $_trunk || defined $_branches || defined $_tags) {
 719                usage(1);
 720        }
 721
 722        # there are currently some bugs that prevent multi-init/multi-fetch
 723        # setups from working well without this.
 724        $Git::SVN::_minimize_url = 1;
 725
 726        $_prefix = '' unless defined $_prefix;
 727        if (defined $url) {
 728                $url =~ s#/+$##;
 729                init_subdir(@_);
 730        }
 731        do_git_init_db();
 732        if (defined $_trunk) {
 733                my $trunk_ref = $_prefix . 'trunk';
 734                # try both old-style and new-style lookups:
 735                my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
 736                unless ($gs_trunk) {
 737                        my ($trunk_url, $trunk_path) =
 738                                              complete_svn_url($url, $_trunk);
 739                        $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
 740                                                   undef, $trunk_ref);
 741                }
 742        }
 743        return unless defined $_branches || defined $_tags;
 744        my $ra = $url ? Git::SVN::Ra->new($url) : undef;
 745        complete_url_ls_init($ra, $_branches, '--branches/-b', $_prefix);
 746        complete_url_ls_init($ra, $_tags, '--tags/-t', $_prefix . 'tags/');
 747}
 748
 749sub cmd_multi_fetch {
 750        my $remotes = Git::SVN::read_all_remotes();
 751        foreach my $repo_id (sort keys %$remotes) {
 752                if ($remotes->{$repo_id}->{url}) {
 753                        Git::SVN::fetch_all($repo_id, $remotes);
 754                }
 755        }
 756}
 757
 758# this command is special because it requires no metadata
 759sub cmd_commit_diff {
 760        my ($ta, $tb, $url) = @_;
 761        my $usage = "Usage: $0 commit-diff -r<revision> ".
 762                    "<tree-ish> <tree-ish> [<URL>]";
 763        fatal($usage) if (!defined $ta || !defined $tb);
 764        my $svn_path = '';
 765        if (!defined $url) {
 766                my $gs = eval { Git::SVN->new };
 767                if (!$gs) {
 768                        fatal("Needed URL or usable git-svn --id in ",
 769                              "the command-line\n", $usage);
 770                }
 771                $url = $gs->{url};
 772                $svn_path = $gs->{path};
 773        }
 774        unless (defined $_revision) {
 775                fatal("-r|--revision is a required argument\n", $usage);
 776        }
 777        if (defined $_message && defined $_file) {
 778                fatal("Both --message/-m and --file/-F specified ",
 779                      "for the commit message.\n",
 780                      "I have no idea what you mean");
 781        }
 782        if (defined $_file) {
 783                $_message = file_to_s($_file);
 784        } else {
 785                $_message ||= get_commit_entry($tb)->{log};
 786        }
 787        my $ra ||= Git::SVN::Ra->new($url);
 788        my $r = $_revision;
 789        if ($r eq 'HEAD') {
 790                $r = $ra->get_latest_revnum;
 791        } elsif ($r !~ /^\d+$/) {
 792                die "revision argument: $r not understood by git-svn\n";
 793        }
 794        my %ed_opts = ( r => $r,
 795                        log => $_message,
 796                        ra => $ra,
 797                        tree_a => $ta,
 798                        tree_b => $tb,
 799                        editor_cb => sub { print "Committed r$_[0]\n" },
 800                        svn_path => $svn_path );
 801        if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
 802                print "No changes\n$ta == $tb\n";
 803        }
 804}
 805
 806sub escape_uri_only {
 807        my ($uri) = @_;
 808        my @tmp;
 809        foreach (split m{/}, $uri) {
 810                s/([^\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
 811                push @tmp, $_;
 812        }
 813        join('/', @tmp);
 814}
 815
 816sub escape_url {
 817        my ($url) = @_;
 818        if ($url =~ m#^([^:]+)://([^/]*)(.*)$#) {
 819                my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
 820                $url = "$scheme://$domain$uri";
 821        }
 822        $url;
 823}
 824
 825sub cmd_info {
 826        my $path = canonicalize_path(defined($_[0]) ? $_[0] : ".");
 827        my $fullpath = canonicalize_path($cmd_dir_prefix . $path);
 828        if (exists $_[1]) {
 829                die "Too many arguments specified\n";
 830        }
 831
 832        my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
 833
 834        if (!$file_type && !$diff_status) {
 835                print STDERR "svn: '$path' is not under version control\n";
 836                exit 1;
 837        }
 838
 839        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 840        unless ($gs) {
 841                die "Unable to determine upstream SVN information from ",
 842                    "working tree history\n";
 843        }
 844
 845        # canonicalize_path() will return "" to make libsvn 1.5.x happy,
 846        $path = "." if $path eq "";
 847
 848        my $full_url = $url . ($fullpath eq "" ? "" : "/$fullpath");
 849
 850        if ($_url) {
 851                print escape_url($full_url), "\n";
 852                return;
 853        }
 854
 855        my $result = "Path: $path\n";
 856        $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
 857        $result .= "URL: " . escape_url($full_url) . "\n";
 858
 859        eval {
 860                my $repos_root = $gs->repos_root;
 861                Git::SVN::remove_username($repos_root);
 862                $result .= "Repository Root: " . escape_url($repos_root) . "\n";
 863        };
 864        if ($@) {
 865                $result .= "Repository Root: (offline)\n";
 866        }
 867        $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A";
 868        $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
 869
 870        $result .= "Node Kind: " .
 871                   ($file_type eq "dir" ? "directory" : "file") . "\n";
 872
 873        my $schedule = $diff_status eq "A"
 874                       ? "add"
 875                       : ($diff_status eq "D" ? "delete" : "normal");
 876        $result .= "Schedule: $schedule\n";
 877
 878        if ($diff_status eq "A") {
 879                print $result, "\n";
 880                return;
 881        }
 882
 883        my ($lc_author, $lc_rev, $lc_date_utc);
 884        my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $fullpath);
 885        my $log = command_output_pipe(@args);
 886        my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
 887        while (<$log>) {
 888                if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
 889                        $lc_author = $1;
 890                        $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
 891                } elsif (/^${esc_color}    (git-svn-id:.+)$/o) {
 892                        (undef, $lc_rev, undef) = ::extract_metadata($1);
 893                }
 894        }
 895        close $log;
 896
 897        Git::SVN::Log::set_local_timezone();
 898
 899        $result .= "Last Changed Author: $lc_author\n";
 900        $result .= "Last Changed Rev: $lc_rev\n";
 901        $result .= "Last Changed Date: " .
 902                   Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
 903
 904        if ($file_type ne "dir") {
 905                my $text_last_updated_date =
 906                    ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
 907                $result .=
 908                    "Text Last Updated: " .
 909                    Git::SVN::Log::format_svn_date($text_last_updated_date) .
 910                    "\n";
 911                my $checksum;
 912                if ($diff_status eq "D") {
 913                        my ($fh, $ctx) =
 914                            command_output_pipe(qw(cat-file blob), "HEAD:$path");
 915                        if ($file_type eq "link") {
 916                                my $file_name = <$fh>;
 917                                $checksum = md5sum("link $file_name");
 918                        } else {
 919                                $checksum = md5sum($fh);
 920                        }
 921                        command_close_pipe($fh, $ctx);
 922                } elsif ($file_type eq "link") {
 923                        my $file_name =
 924                            command(qw(cat-file blob), "HEAD:$path");
 925                        $checksum =
 926                            md5sum("link " . $file_name);
 927                } else {
 928                        open FILE, "<", $path or die $!;
 929                        $checksum = md5sum(\*FILE);
 930                        close FILE or die $!;
 931                }
 932                $result .= "Checksum: " . $checksum . "\n";
 933        }
 934
 935        print $result, "\n";
 936}
 937
 938########################### utility functions #########################
 939
 940sub rebase_cmd {
 941        my @cmd = qw/rebase/;
 942        push @cmd, '-v' if $_verbose;
 943        push @cmd, qw/--merge/ if $_merge;
 944        push @cmd, "--strategy=$_strategy" if $_strategy;
 945        @cmd;
 946}
 947
 948sub post_fetch_checkout {
 949        return if $_no_checkout;
 950        my $gs = $Git::SVN::_head or return;
 951        return if verify_ref('refs/heads/master^0');
 952
 953        my $valid_head = verify_ref('HEAD^0');
 954        command_noisy(qw(update-ref refs/heads/master), $gs->refname);
 955        return if ($valid_head || !verify_ref('HEAD^0'));
 956
 957        return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
 958        my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
 959        return if -f $index;
 960
 961        return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
 962        return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
 963        command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
 964        print STDERR "Checked out HEAD:\n  ",
 965                     $gs->full_url, " r", $gs->last_rev, "\n";
 966}
 967
 968sub complete_svn_url {
 969        my ($url, $path) = @_;
 970        $path =~ s#/+$##;
 971        if ($path !~ m#^[a-z\+]+://#) {
 972                if (!defined $url || $url !~ m#^[a-z\+]+://#) {
 973                        fatal("E: '$path' is not a complete URL ",
 974                              "and a separate URL is not specified");
 975                }
 976                return ($url, $path);
 977        }
 978        return ($path, '');
 979}
 980
 981sub complete_url_ls_init {
 982        my ($ra, $repo_path, $switch, $pfx) = @_;
 983        unless ($repo_path) {
 984                print STDERR "W: $switch not specified\n";
 985                return;
 986        }
 987        $repo_path =~ s#/+$##;
 988        if ($repo_path =~ m#^[a-z\+]+://#) {
 989                $ra = Git::SVN::Ra->new($repo_path);
 990                $repo_path = '';
 991        } else {
 992                $repo_path =~ s#^/+##;
 993                unless ($ra) {
 994                        fatal("E: '$repo_path' is not a complete URL ",
 995                              "and a separate URL is not specified");
 996                }
 997        }
 998        my $url = $ra->{url};
 999        my $gs = Git::SVN->init($url, undef, undef, undef, 1);
1000        my $k = "svn-remote.$gs->{repo_id}.url";
1001        my $orig_url = eval { command_oneline(qw/config --get/, $k) };
1002        if ($orig_url && ($orig_url ne $gs->{url})) {
1003                die "$k already set: $orig_url\n",
1004                    "wanted to set to: $gs->{url}\n";
1005        }
1006        command_oneline('config', $k, $gs->{url}) unless $orig_url;
1007        my $remote_path = "$ra->{svn_path}/$repo_path";
1008        $remote_path =~ s#/+#/#g;
1009        $remote_path =~ s#^/##g;
1010        $remote_path .= "/*" if $remote_path !~ /\*/;
1011        my ($n) = ($switch =~ /^--(\w+)/);
1012        if (length $pfx && $pfx !~ m#/$#) {
1013                die "--prefix='$pfx' must have a trailing slash '/'\n";
1014        }
1015        command_noisy('config',
1016                      "svn-remote.$gs->{repo_id}.$n",
1017                      "$remote_path:refs/remotes/$pfx*" .
1018                        ('/*' x (($remote_path =~ tr/*/*/) - 1)) );
1019}
1020
1021sub verify_ref {
1022        my ($ref) = @_;
1023        eval { command_oneline([ 'rev-parse', '--verify', $ref ],
1024                               { STDERR => 0 }); };
1025}
1026
1027sub get_tree_from_treeish {
1028        my ($treeish) = @_;
1029        # $treeish can be a symbolic ref, too:
1030        my $type = command_oneline(qw/cat-file -t/, $treeish);
1031        my $expected;
1032        while ($type eq 'tag') {
1033                ($treeish, $type) = command(qw/cat-file tag/, $treeish);
1034        }
1035        if ($type eq 'commit') {
1036                $expected = (grep /^tree /, command(qw/cat-file commit/,
1037                                                    $treeish))[0];
1038                ($expected) = ($expected =~ /^tree ($sha1)$/o);
1039                die "Unable to get tree from $treeish\n" unless $expected;
1040        } elsif ($type eq 'tree') {
1041                $expected = $treeish;
1042        } else {
1043                die "$treeish is a $type, expected tree, tag or commit\n";
1044        }
1045        return $expected;
1046}
1047
1048sub get_commit_entry {
1049        my ($treeish) = shift;
1050        my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
1051        my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
1052        my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
1053        open my $log_fh, '>', $commit_editmsg or croak $!;
1054
1055        my $type = command_oneline(qw/cat-file -t/, $treeish);
1056        if ($type eq 'commit' || $type eq 'tag') {
1057                my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1058                                                         $type, $treeish);
1059                my $in_msg = 0;
1060                my $author;
1061                my $saw_from = 0;
1062                my $msgbuf = "";
1063                while (<$msg_fh>) {
1064                        if (!$in_msg) {
1065                                $in_msg = 1 if (/^\s*$/);
1066                                $author = $1 if (/^author (.*>)/);
1067                        } elsif (/^git-svn-id: /) {
1068                                # skip this for now, we regenerate the
1069                                # correct one on re-fetch anyways
1070                                # TODO: set *:merge properties or like...
1071                        } else {
1072                                if (/^From:/ || /^Signed-off-by:/) {
1073                                        $saw_from = 1;
1074                                }
1075                                $msgbuf .= $_;
1076                        }
1077                }
1078                $msgbuf =~ s/\s+$//s;
1079                if ($Git::SVN::_add_author_from && defined($author)
1080                    && !$saw_from) {
1081                        $msgbuf .= "\n\nFrom: $author";
1082                }
1083                print $log_fh $msgbuf or croak $!;
1084                command_close_pipe($msg_fh, $ctx);
1085        }
1086        close $log_fh or croak $!;
1087
1088        if ($_edit || ($type eq 'tree')) {
1089                my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
1090                # TODO: strip out spaces, comments, like git-commit.sh
1091                system($editor, $commit_editmsg);
1092        }
1093        rename $commit_editmsg, $commit_msg or croak $!;
1094        open $log_fh, '<', $commit_msg or croak $!;
1095        { local $/; chomp($log_entry{log} = <$log_fh>); }
1096        close $log_fh or croak $!;
1097        unlink $commit_msg;
1098        \%log_entry;
1099}
1100
1101sub s_to_file {
1102        my ($str, $file, $mode) = @_;
1103        open my $fd,'>',$file or croak $!;
1104        print $fd $str,"\n" or croak $!;
1105        close $fd or croak $!;
1106        chmod ($mode &~ umask, $file) if (defined $mode);
1107}
1108
1109sub file_to_s {
1110        my $file = shift;
1111        open my $fd,'<',$file or croak "$!: file: $file\n";
1112        local $/;
1113        my $ret = <$fd>;
1114        close $fd or croak $!;
1115        $ret =~ s/\s*$//s;
1116        return $ret;
1117}
1118
1119# '<svn username> = real-name <email address>' mapping based on git-svnimport:
1120sub load_authors {
1121        open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1122        my $log = $cmd eq 'log';
1123        while (<$authors>) {
1124                chomp;
1125                next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1126                my ($user, $name, $email) = ($1, $2, $3);
1127                if ($log) {
1128                        $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1129                } else {
1130                        $users{$user} = [$name, $email];
1131                }
1132        }
1133        close $authors or croak $!;
1134}
1135
1136# convert GetOpt::Long specs for use by git-config
1137sub read_repo_config {
1138        return unless -d $ENV{GIT_DIR};
1139        my $opts = shift;
1140        my @config_only;
1141        foreach my $o (keys %$opts) {
1142                # if we have mixedCase and a long option-only, then
1143                # it's a config-only variable that we don't need for
1144                # the command-line.
1145                push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1146                my $v = $opts->{$o};
1147                my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1148                $key =~ s/-//g;
1149                my $arg = 'git-config';
1150                $arg .= ' --int' if ($o =~ /[:=]i$/);
1151                $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1152                if (ref $v eq 'ARRAY') {
1153                        chomp(my @tmp = `$arg --get-all svn.$key`);
1154                        @$v = @tmp if @tmp;
1155                } else {
1156                        chomp(my $tmp = `$arg --get svn.$key`);
1157                        if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1158                                $$v = $tmp;
1159                        }
1160                }
1161        }
1162        delete @$opts{@config_only} if @config_only;
1163}
1164
1165sub extract_metadata {
1166        my $id = shift or return (undef, undef, undef);
1167        my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1168                                                        \s([a-f\d\-]+)$/x);
1169        if (!defined $rev || !$uuid || !$url) {
1170                # some of the original repositories I made had
1171                # identifiers like this:
1172                ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/);
1173        }
1174        return ($url, $rev, $uuid);
1175}
1176
1177sub cmt_metadata {
1178        return extract_metadata((grep(/^git-svn-id: /,
1179                command(qw/cat-file commit/, shift)))[-1]);
1180}
1181
1182sub working_head_info {
1183        my ($head, $refs) = @_;
1184        my @args = ('log', '--no-color', '--first-parent', '--pretty=medium');
1185        my ($fh, $ctx) = command_output_pipe(@args, $head);
1186        my $hash;
1187        my %max;
1188        while (<$fh>) {
1189                if ( m{^commit ($::sha1)$} ) {
1190                        unshift @$refs, $hash if $hash and $refs;
1191                        $hash = $1;
1192                        next;
1193                }
1194                next unless s{^\s*(git-svn-id:)}{$1};
1195                my ($url, $rev, $uuid) = extract_metadata($_);
1196                if (defined $url && defined $rev) {
1197                        next if $max{$url} and $max{$url} < $rev;
1198                        if (my $gs = Git::SVN->find_by_url($url)) {
1199                                my $c = $gs->rev_map_get($rev, $uuid);
1200                                if ($c && $c eq $hash) {
1201                                        close $fh; # break the pipe
1202                                        return ($url, $rev, $uuid, $gs);
1203                                } else {
1204                                        $max{$url} ||= $gs->rev_map_max;
1205                                }
1206                        }
1207                }
1208        }
1209        command_close_pipe($fh, $ctx);
1210        (undef, undef, undef, undef);
1211}
1212
1213sub read_commit_parents {
1214        my ($parents, $c) = @_;
1215        chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1216        $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1217        @{$parents->{$c}} = split(/ /, $p);
1218}
1219
1220sub linearize_history {
1221        my ($gs, $refs) = @_;
1222        my %parents;
1223        foreach my $c (@$refs) {
1224                read_commit_parents(\%parents, $c);
1225        }
1226
1227        my @linear_refs;
1228        my %skip = ();
1229        my $last_svn_commit = $gs->last_commit;
1230        foreach my $c (reverse @$refs) {
1231                next if $c eq $last_svn_commit;
1232                last if $skip{$c};
1233
1234                unshift @linear_refs, $c;
1235                $skip{$c} = 1;
1236
1237                # we only want the first parent to diff against for linear
1238                # history, we save the rest to inject when we finalize the
1239                # svn commit
1240                my $fp_a = verify_ref("$c~1");
1241                my $fp_b = shift @{$parents{$c}} if $parents{$c};
1242                if (!$fp_a || !$fp_b) {
1243                        die "Commit $c\n",
1244                            "has no parent commit, and therefore ",
1245                            "nothing to diff against.\n",
1246                            "You should be working from a repository ",
1247                            "originally created by git-svn\n";
1248                }
1249                if ($fp_a ne $fp_b) {
1250                        die "$c~1 = $fp_a, however parsing commit $c ",
1251                            "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1252                }
1253
1254                foreach my $p (@{$parents{$c}}) {
1255                        $skip{$p} = 1;
1256                }
1257        }
1258        (\@linear_refs, \%parents);
1259}
1260
1261sub find_file_type_and_diff_status {
1262        my ($path) = @_;
1263        return ('dir', '') if $path eq '';
1264
1265        my $diff_output =
1266            command_oneline(qw(diff --cached --name-status --), $path) || "";
1267        my $diff_status = (split(' ', $diff_output))[0] || "";
1268
1269        my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1270
1271        return (undef, undef) if !$diff_status && !$ls_tree;
1272
1273        if ($diff_status eq "A") {
1274                return ("link", $diff_status) if -l $path;
1275                return ("dir", $diff_status) if -d $path;
1276                return ("file", $diff_status);
1277        }
1278
1279        my $mode = (split(' ', $ls_tree))[0] || "";
1280
1281        return ("link", $diff_status) if $mode eq "120000";
1282        return ("dir", $diff_status) if $mode eq "040000";
1283        return ("file", $diff_status);
1284}
1285
1286sub md5sum {
1287        my $arg = shift;
1288        my $ref = ref $arg;
1289        my $md5 = Digest::MD5->new();
1290        if ($ref eq 'GLOB' || $ref eq 'IO::File' || $ref eq 'File::Temp') {
1291                $md5->addfile($arg) or croak $!;
1292        } elsif ($ref eq 'SCALAR') {
1293                $md5->add($$arg) or croak $!;
1294        } elsif (!$ref) {
1295                $md5->add($arg) or croak $!;
1296        } else {
1297                ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
1298        }
1299        return $md5->hexdigest();
1300}
1301
1302package Git::SVN;
1303use strict;
1304use warnings;
1305use Fcntl qw/:DEFAULT :seek/;
1306use constant rev_map_fmt => 'NH40';
1307use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1308            $_repack $_repack_flags $_use_svm_props $_head
1309            $_use_svnsync_props $no_reuse_existing $_minimize_url
1310            $_use_log_author $_add_author_from/;
1311use Carp qw/croak/;
1312use File::Path qw/mkpath/;
1313use File::Copy qw/copy/;
1314use IPC::Open3;
1315
1316my ($_gc_nr, $_gc_period);
1317
1318# properties that we do not log:
1319my %SKIP_PROP;
1320BEGIN {
1321        %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1322                                        svn:special svn:executable
1323                                        svn:entry:committed-rev
1324                                        svn:entry:last-author
1325                                        svn:entry:uuid
1326                                        svn:entry:committed-date/;
1327
1328        # some options are read globally, but can be overridden locally
1329        # per [svn-remote "..."] section.  Command-line options will *NOT*
1330        # override options set in an [svn-remote "..."] section
1331        no strict 'refs';
1332        for my $option (qw/follow_parent no_metadata use_svm_props
1333                           use_svnsync_props/) {
1334                my $key = $option;
1335                $key =~ tr/_//d;
1336                my $prop = "-$option";
1337                *$option = sub {
1338                        my ($self) = @_;
1339                        return $self->{$prop} if exists $self->{$prop};
1340                        my $k = "svn-remote.$self->{repo_id}.$key";
1341                        eval { command_oneline(qw/config --get/, $k) };
1342                        if ($@) {
1343                                $self->{$prop} = ${"Git::SVN::_$option"};
1344                        } else {
1345                                my $v = command_oneline(qw/config --bool/,$k);
1346                                $self->{$prop} = $v eq 'false' ? 0 : 1;
1347                        }
1348                        return $self->{$prop};
1349                }
1350        }
1351}
1352
1353
1354my (%LOCKFILES, %INDEX_FILES);
1355END {
1356        unlink keys %LOCKFILES if %LOCKFILES;
1357        unlink keys %INDEX_FILES if %INDEX_FILES;
1358}
1359
1360sub resolve_local_globs {
1361        my ($url, $fetch, $glob_spec) = @_;
1362        return unless defined $glob_spec;
1363        my $ref = $glob_spec->{ref};
1364        my $path = $glob_spec->{path};
1365        foreach (command(qw#for-each-ref --format=%(refname) refs/remotes#)) {
1366                next unless m#^refs/remotes/$ref->{regex}$#;
1367                my $p = $1;
1368                my $pathname = desanitize_refname($path->full_path($p));
1369                my $refname = desanitize_refname($ref->full_path($p));
1370                if (my $existing = $fetch->{$pathname}) {
1371                        if ($existing ne $refname) {
1372                                die "Refspec conflict:\n",
1373                                    "existing: refs/remotes/$existing\n",
1374                                    " globbed: refs/remotes/$refname\n";
1375                        }
1376                        my $u = (::cmt_metadata("refs/remotes/$refname"))[0];
1377                        $u =~ s!^\Q$url\E(/|$)!! or die
1378                          "refs/remotes/$refname: '$url' not found in '$u'\n";
1379                        if ($pathname ne $u) {
1380                                warn "W: Refspec glob conflict ",
1381                                     "(ref: refs/remotes/$refname):\n",
1382                                     "expected path: $pathname\n",
1383                                     "    real path: $u\n",
1384                                     "Continuing ahead with $u\n";
1385                                next;
1386                        }
1387                } else {
1388                        $fetch->{$pathname} = $refname;
1389                }
1390        }
1391}
1392
1393sub parse_revision_argument {
1394        my ($base, $head) = @_;
1395        if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1396                return ($base, $head);
1397        }
1398        return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1399        return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1400        return ($head, $head) if ($::_revision eq 'HEAD');
1401        return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1402        return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1403        die "revision argument: $::_revision not understood by git-svn\n";
1404}
1405
1406sub fetch_all {
1407        my ($repo_id, $remotes) = @_;
1408        if (ref $repo_id) {
1409                my $gs = $repo_id;
1410                $repo_id = undef;
1411                $repo_id = $gs->{repo_id};
1412        }
1413        $remotes ||= read_all_remotes();
1414        my $remote = $remotes->{$repo_id} or
1415                     die "[svn-remote \"$repo_id\"] unknown\n";
1416        my $fetch = $remote->{fetch};
1417        my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1418        my (@gs, @globs);
1419        my $ra = Git::SVN::Ra->new($url);
1420        my $uuid = $ra->get_uuid;
1421        my $head = $ra->get_latest_revnum;
1422        my $base = defined $fetch ? $head : 0;
1423
1424        # read the max revs for wildcard expansion (branches/*, tags/*)
1425        foreach my $t (qw/branches tags/) {
1426                defined $remote->{$t} or next;
1427                push @globs, $remote->{$t};
1428                my $max_rev = eval { tmp_config(qw/--int --get/,
1429                                         "svn-remote.$repo_id.${t}-maxRev") };
1430                if (defined $max_rev && ($max_rev < $base)) {
1431                        $base = $max_rev;
1432                } elsif (!defined $max_rev) {
1433                        $base = 0;
1434                }
1435        }
1436
1437        if ($fetch) {
1438                foreach my $p (sort keys %$fetch) {
1439                        my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1440                        my $lr = $gs->rev_map_max;
1441                        if (defined $lr) {
1442                                $base = $lr if ($lr < $base);
1443                        }
1444                        push @gs, $gs;
1445                }
1446        }
1447
1448        ($base, $head) = parse_revision_argument($base, $head);
1449        $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1450}
1451
1452sub read_all_remotes {
1453        my $r = {};
1454        my $use_svm_props = eval { command_oneline(qw/config --bool
1455            svn.useSvmProps/) };
1456        $use_svm_props = $use_svm_props eq 'true' if $use_svm_props;
1457        foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1458                if (m!^(.+)\.fetch=\s*(.*)\s*:\s*(.+)\s*$!) {
1459                        my ($remote, $local_ref, $_remote_ref) = ($1, $2, $3);
1460                        die("svn-remote.$remote: remote ref '$_remote_ref' "
1461                            . "must start with 'refs/remotes/'\n")
1462                                unless $_remote_ref =~ m{^refs/remotes/(.+)};
1463                        my $remote_ref = $1;
1464                        $local_ref =~ s{^/}{};
1465                        $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1466                        $r->{$remote}->{svm} = {} if $use_svm_props;
1467                } elsif (m!^(.+)\.usesvmprops=\s*(.*)\s*$!) {
1468                        $r->{$1}->{svm} = {};
1469                } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1470                        $r->{$1}->{url} = $2;
1471                } elsif (m!^(.+)\.(branches|tags)=
1472                           (.*):refs/remotes/(.+)\s*$/!x) {
1473                        my ($p, $g) = ($3, $4);
1474                        my $rs = $r->{$1}->{$2} = {
1475                                          t => $2,
1476                                          remote => $1,
1477                                          path => Git::SVN::GlobSpec->new($p),
1478                                          ref => Git::SVN::GlobSpec->new($g) };
1479                        if (length($rs->{ref}->{right}) != 0) {
1480                                die "The '*' glob character must be the last ",
1481                                    "character of '$g'\n";
1482                        }
1483                }
1484        }
1485
1486        map {
1487                if (defined $r->{$_}->{svm}) {
1488                        my $svm;
1489                        eval {
1490                                my $section = "svn-remote.$_";
1491                                $svm = {
1492                                        source => tmp_config('--get',
1493                                            "$section.svm-source"),
1494                                        replace => tmp_config('--get',
1495                                            "$section.svm-replace"),
1496                                }
1497                        };
1498                        $r->{$_}->{svm} = $svm;
1499                }
1500        } keys %$r;
1501
1502        $r;
1503}
1504
1505sub init_vars {
1506        $_gc_nr = $_gc_period = 1000;
1507        if (defined $_repack || defined $_repack_flags) {
1508               warn "Repack options are obsolete; they have no effect.\n";
1509        }
1510}
1511
1512sub verify_remotes_sanity {
1513        return unless -d $ENV{GIT_DIR};
1514        my %seen;
1515        foreach (command(qw/config -l/)) {
1516                if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1517                        if ($seen{$1}) {
1518                                die "Remote ref refs/remote/$1 is tracked by",
1519                                    "\n  \"$_\"\nand\n  \"$seen{$1}\"\n",
1520                                    "Please resolve this ambiguity in ",
1521                                    "your git configuration file before ",
1522                                    "continuing\n";
1523                        }
1524                        $seen{$1} = $_;
1525                }
1526        }
1527}
1528
1529sub find_existing_remote {
1530        my ($url, $remotes) = @_;
1531        return undef if $no_reuse_existing;
1532        my $existing;
1533        foreach my $repo_id (keys %$remotes) {
1534                my $u = $remotes->{$repo_id}->{url} or next;
1535                next if $u ne $url;
1536                $existing = $repo_id;
1537                last;
1538        }
1539        $existing;
1540}
1541
1542sub init_remote_config {
1543        my ($self, $url, $no_write) = @_;
1544        $url =~ s!/+$!!; # strip trailing slash
1545        my $r = read_all_remotes();
1546        my $existing = find_existing_remote($url, $r);
1547        if ($existing) {
1548                unless ($no_write) {
1549                        print STDERR "Using existing ",
1550                                     "[svn-remote \"$existing\"]\n";
1551                }
1552                $self->{repo_id} = $existing;
1553        } elsif ($_minimize_url) {
1554                my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1555                $existing = find_existing_remote($min_url, $r);
1556                if ($existing) {
1557                        unless ($no_write) {
1558                                print STDERR "Using existing ",
1559                                             "[svn-remote \"$existing\"]\n";
1560                        }
1561                        $self->{repo_id} = $existing;
1562                }
1563                if ($min_url ne $url) {
1564                        unless ($no_write) {
1565                                print STDERR "Using higher level of URL: ",
1566                                             "$url => $min_url\n";
1567                        }
1568                        my $old_path = $self->{path};
1569                        $self->{path} = $url;
1570                        $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1571                        if (length $old_path) {
1572                                $self->{path} .= "/$old_path";
1573                        }
1574                        $url = $min_url;
1575                }
1576        }
1577        my $orig_url;
1578        if (!$existing) {
1579                # verify that we aren't overwriting anything:
1580                $orig_url = eval {
1581                        command_oneline('config', '--get',
1582                                        "svn-remote.$self->{repo_id}.url")
1583                };
1584                if ($orig_url && ($orig_url ne $url)) {
1585                        die "svn-remote.$self->{repo_id}.url already set: ",
1586                            "$orig_url\nwanted to set to: $url\n";
1587                }
1588        }
1589        my ($xrepo_id, $xpath) = find_ref($self->refname);
1590        if (defined $xpath) {
1591                die "svn-remote.$xrepo_id.fetch already set to track ",
1592                    "$xpath:refs/remotes/", $self->refname, "\n";
1593        }
1594        unless ($no_write) {
1595                command_noisy('config',
1596                              "svn-remote.$self->{repo_id}.url", $url);
1597                $self->{path} =~ s{^/}{};
1598                command_noisy('config', '--add',
1599                              "svn-remote.$self->{repo_id}.fetch",
1600                              "$self->{path}:".$self->refname);
1601        }
1602        $self->{url} = $url;
1603}
1604
1605sub find_by_url { # repos_root and, path are optional
1606        my ($class, $full_url, $repos_root, $path) = @_;
1607
1608        return undef unless defined $full_url;
1609        remove_username($full_url);
1610        remove_username($repos_root) if defined $repos_root;
1611        my $remotes = read_all_remotes();
1612        if (defined $full_url && defined $repos_root && !defined $path) {
1613                $path = $full_url;
1614                $path =~ s#^\Q$repos_root\E(?:/|$)##;
1615        }
1616        foreach my $repo_id (keys %$remotes) {
1617                my $u = $remotes->{$repo_id}->{url} or next;
1618                remove_username($u);
1619                next if defined $repos_root && $repos_root ne $u;
1620
1621                my $fetch = $remotes->{$repo_id}->{fetch} || {};
1622                foreach (qw/branches tags/) {
1623                        resolve_local_globs($u, $fetch,
1624                                            $remotes->{$repo_id}->{$_});
1625                }
1626                my $p = $path;
1627                my $rwr = rewrite_root({repo_id => $repo_id});
1628                my $svm = $remotes->{$repo_id}->{svm}
1629                        if defined $remotes->{$repo_id}->{svm};
1630                unless (defined $p) {
1631                        $p = $full_url;
1632                        my $z = $u;
1633                        my $prefix = '';
1634                        if ($rwr) {
1635                                $z = $rwr;
1636                        } elsif (defined $svm) {
1637                                $z = $svm->{source};
1638                                $prefix = $svm->{replace};
1639                                $prefix =~ s#^\Q$u\E(?:/|$)##;
1640                                $prefix =~ s#/$##;
1641                        }
1642                        $p =~ s#^\Q$z\E(?:/|$)#$prefix# or next;
1643                }
1644                foreach my $f (keys %$fetch) {
1645                        next if $f ne $p;
1646                        return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1647                }
1648        }
1649        undef;
1650}
1651
1652sub init {
1653        my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1654        my $self = _new($class, $repo_id, $ref_id, $path);
1655        if (defined $url) {
1656                $self->init_remote_config($url, $no_write);
1657        }
1658        $self;
1659}
1660
1661sub find_ref {
1662        my ($ref_id) = @_;
1663        foreach (command(qw/config -l/)) {
1664                next unless m!^svn-remote\.(.+)\.fetch=
1665                              \s*(.*)\s*:\s*refs/remotes/(.+)\s*$!x;
1666                my ($repo_id, $path, $ref) = ($1, $2, $3);
1667                if ($ref eq $ref_id) {
1668                        $path = '' if ($path =~ m#^\./?#);
1669                        return ($repo_id, $path);
1670                }
1671        }
1672        (undef, undef, undef);
1673}
1674
1675sub new {
1676        my ($class, $ref_id, $repo_id, $path) = @_;
1677        if (defined $ref_id && !defined $repo_id && !defined $path) {
1678                ($repo_id, $path) = find_ref($ref_id);
1679                if (!defined $repo_id) {
1680                        die "Could not find a \"svn-remote.*.fetch\" key ",
1681                            "in the repository configuration matching: ",
1682                            "refs/remotes/$ref_id\n";
1683                }
1684        }
1685        my $self = _new($class, $repo_id, $ref_id, $path);
1686        if (!defined $self->{path} || !length $self->{path}) {
1687                my $fetch = command_oneline('config', '--get',
1688                                            "svn-remote.$repo_id.fetch",
1689                                            ":refs/remotes/$ref_id\$") or
1690                     die "Failed to read \"svn-remote.$repo_id.fetch\" ",
1691                         "\":refs/remotes/$ref_id\$\" in config\n";
1692                ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
1693        }
1694        $self->{url} = command_oneline('config', '--get',
1695                                       "svn-remote.$repo_id.url") or
1696                  die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
1697        $self->rebuild;
1698        $self;
1699}
1700
1701sub refname {
1702        my ($refname) = "refs/remotes/$_[0]->{ref_id}" ;
1703
1704        # It cannot end with a slash /, we'll throw up on this because
1705        # SVN can't have directories with a slash in their name, either:
1706        if ($refname =~ m{/$}) {
1707                die "ref: '$refname' ends with a trailing slash, this is ",
1708                    "not permitted by git nor Subversion\n";
1709        }
1710
1711        # It cannot have ASCII control character space, tilde ~, caret ^,
1712        # colon :, question-mark ?, asterisk *, space, or open bracket [
1713        # anywhere.
1714        #
1715        # Additionally, % must be escaped because it is used for escaping
1716        # and we want our escaped refname to be reversible
1717        $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
1718
1719        # no slash-separated component can begin with a dot .
1720        # /.* becomes /%2E*
1721        $refname =~ s{/\.}{/%2E}g;
1722
1723        # It cannot have two consecutive dots .. anywhere
1724        # .. becomes %2E%2E
1725        $refname =~ s{\.\.}{%2E%2E}g;
1726
1727        return $refname;
1728}
1729
1730sub desanitize_refname {
1731        my ($refname) = @_;
1732        $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
1733        return $refname;
1734}
1735
1736sub svm_uuid {
1737        my ($self) = @_;
1738        return $self->{svm}->{uuid} if $self->svm;
1739        $self->ra;
1740        unless ($self->{svm}) {
1741                die "SVM UUID not cached, and reading remotely failed\n";
1742        }
1743        $self->{svm}->{uuid};
1744}
1745
1746sub svm {
1747        my ($self) = @_;
1748        return $self->{svm} if $self->{svm};
1749        my $svm;
1750        # see if we have it in our config, first:
1751        eval {
1752                my $section = "svn-remote.$self->{repo_id}";
1753                $svm = {
1754                  source => tmp_config('--get', "$section.svm-source"),
1755                  uuid => tmp_config('--get', "$section.svm-uuid"),
1756                  replace => tmp_config('--get', "$section.svm-replace"),
1757                }
1758        };
1759        if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
1760                $self->{svm} = $svm;
1761        }
1762        $self->{svm};
1763}
1764
1765sub _set_svm_vars {
1766        my ($self, $ra) = @_;
1767        return $ra if $self->svm;
1768
1769        my @err = ( "useSvmProps set, but failed to read SVM properties\n",
1770                    "(svm:source, svm:uuid) ",
1771                    "from the following URLs:\n" );
1772        sub read_svm_props {
1773                my ($self, $ra, $path, $r) = @_;
1774                my $props = ($ra->get_dir($path, $r))[2];
1775                my $src = $props->{'svm:source'};
1776                my $uuid = $props->{'svm:uuid'};
1777                return undef if (!$src || !$uuid);
1778
1779                chomp($src, $uuid);
1780
1781                $uuid =~ m{^[0-9a-f\-]{30,}$}
1782                    or die "doesn't look right - svm:uuid is '$uuid'\n";
1783
1784                # the '!' is used to mark the repos_root!/relative/path
1785                $src =~ s{/?!/?}{/};
1786                $src =~ s{/+$}{}; # no trailing slashes please
1787                # username is of no interest
1788                $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
1789
1790                my $replace = $ra->{url};
1791                $replace .= "/$path" if length $path;
1792
1793                my $section = "svn-remote.$self->{repo_id}";
1794                tmp_config("$section.svm-source", $src);
1795                tmp_config("$section.svm-replace", $replace);
1796                tmp_config("$section.svm-uuid", $uuid);
1797                $self->{svm} = {
1798                        source => $src,
1799                        uuid => $uuid,
1800                        replace => $replace
1801                };
1802        }
1803
1804        my $r = $ra->get_latest_revnum;
1805        my $path = $self->{path};
1806        my %tried;
1807        while (length $path) {
1808                unless ($tried{"$self->{url}/$path"}) {
1809                        return $ra if $self->read_svm_props($ra, $path, $r);
1810                        $tried{"$self->{url}/$path"} = 1;
1811                }
1812                $path =~ s#/?[^/]+$##;
1813        }
1814        die "Path: '$path' should be ''\n" if $path ne '';
1815        return $ra if $self->read_svm_props($ra, $path, $r);
1816        $tried{"$self->{url}/$path"} = 1;
1817
1818        if ($ra->{repos_root} eq $self->{url}) {
1819                die @err, (map { "  $_\n" } keys %tried), "\n";
1820        }
1821
1822        # nope, make sure we're connected to the repository root:
1823        my $ok;
1824        my @tried_b;
1825        $path = $ra->{svn_path};
1826        $ra = Git::SVN::Ra->new($ra->{repos_root});
1827        while (length $path) {
1828                unless ($tried{"$ra->{url}/$path"}) {
1829                        $ok = $self->read_svm_props($ra, $path, $r);
1830                        last if $ok;
1831                        $tried{"$ra->{url}/$path"} = 1;
1832                }
1833                $path =~ s#/?[^/]+$##;
1834        }
1835        die "Path: '$path' should be ''\n" if $path ne '';
1836        $ok ||= $self->read_svm_props($ra, $path, $r);
1837        $tried{"$ra->{url}/$path"} = 1;
1838        if (!$ok) {
1839                die @err, (map { "  $_\n" } keys %tried), "\n";
1840        }
1841        Git::SVN::Ra->new($self->{url});
1842}
1843
1844sub svnsync {
1845        my ($self) = @_;
1846        return $self->{svnsync} if $self->{svnsync};
1847
1848        if ($self->no_metadata) {
1849                die "Can't have both 'noMetadata' and ",
1850                    "'useSvnsyncProps' options set!\n";
1851        }
1852        if ($self->rewrite_root) {
1853                die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
1854                    "options set!\n";
1855        }
1856
1857        my $svnsync;
1858        # see if we have it in our config, first:
1859        eval {
1860                my $section = "svn-remote.$self->{repo_id}";
1861
1862                my $url = tmp_config('--get', "$section.svnsync-url");
1863                ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1864                   die "doesn't look right - svn:sync-from-url is '$url'\n";
1865
1866                my $uuid = tmp_config('--get', "$section.svnsync-uuid");
1867                ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1868                   die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1869
1870                $svnsync = { url => $url, uuid => $uuid }
1871        };
1872        if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
1873                return $self->{svnsync} = $svnsync;
1874        }
1875
1876        my $err = "useSvnsyncProps set, but failed to read " .
1877                  "svnsync property: svn:sync-from-";
1878        my $rp = $self->ra->rev_proplist(0);
1879
1880        my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
1881        ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1882                   die "doesn't look right - svn:sync-from-url is '$url'\n";
1883
1884        my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
1885        ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1886                   die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1887
1888        my $section = "svn-remote.$self->{repo_id}";
1889        tmp_config('--add', "$section.svnsync-uuid", $uuid);
1890        tmp_config('--add', "$section.svnsync-url", $url);
1891        return $self->{svnsync} = { url => $url, uuid => $uuid };
1892}
1893
1894# this allows us to memoize our SVN::Ra UUID locally and avoid a
1895# remote lookup (useful for 'git svn log').
1896sub ra_uuid {
1897        my ($self) = @_;
1898        unless ($self->{ra_uuid}) {
1899                my $key = "svn-remote.$self->{repo_id}.uuid";
1900                my $uuid = eval { tmp_config('--get', $key) };
1901                if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/) {
1902                        $self->{ra_uuid} = $uuid;
1903                } else {
1904                        die "ra_uuid called without URL\n" unless $self->{url};
1905                        $self->{ra_uuid} = $self->ra->get_uuid;
1906                        tmp_config('--add', $key, $self->{ra_uuid});
1907                }
1908        }
1909        $self->{ra_uuid};
1910}
1911
1912sub _set_repos_root {
1913        my ($self, $repos_root) = @_;
1914        my $k = "svn-remote.$self->{repo_id}.reposRoot";
1915        $repos_root ||= $self->ra->{repos_root};
1916        tmp_config($k, $repos_root);
1917        $repos_root;
1918}
1919
1920sub repos_root {
1921        my ($self) = @_;
1922        my $k = "svn-remote.$self->{repo_id}.reposRoot";
1923        eval { tmp_config('--get', $k) } || $self->_set_repos_root;
1924}
1925
1926sub ra {
1927        my ($self) = shift;
1928        my $ra = Git::SVN::Ra->new($self->{url});
1929        $self->_set_repos_root($ra->{repos_root});
1930        if ($self->use_svm_props && !$self->{svm}) {
1931                if ($self->no_metadata) {
1932                        die "Can't have both 'noMetadata' and ",
1933                            "'useSvmProps' options set!\n";
1934                } elsif ($self->use_svnsync_props) {
1935                        die "Can't have both 'useSvnsyncProps' and ",
1936                            "'useSvmProps' options set!\n";
1937                }
1938                $ra = $self->_set_svm_vars($ra);
1939                $self->{-want_revprops} = 1;
1940        }
1941        $ra;
1942}
1943
1944sub rel_path {
1945        my ($self) = @_;
1946        my $repos_root = $self->ra->{repos_root};
1947        return $self->{path} if ($self->{url} eq $repos_root);
1948        my $url = $self->{url} .
1949                  (length $self->{path} ? "/$self->{path}" : $self->{path});
1950        $url =~ s!^\Q$repos_root\E(?:/+|$)!!g;
1951        $url;
1952}
1953
1954# prop_walk(PATH, REV, SUB)
1955# -------------------------
1956# Recursively traverse PATH at revision REV and invoke SUB for each
1957# directory that contains a SVN property.  SUB will be invoked as
1958# follows:  &SUB(gs, path, props);  where `gs' is this instance of
1959# Git::SVN, `path' the path to the directory where the properties
1960# `props' were found.  The `path' will be relative to point of checkout,
1961# that is, if url://repo/trunk is the current Git branch, and that
1962# directory contains a sub-directory `d', SUB will be invoked with `/d/'
1963# as `path' (note the trailing `/').
1964sub prop_walk {
1965        my ($self, $path, $rev, $sub) = @_;
1966
1967        $path =~ s#^/##;
1968        my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
1969        $path =~ s#^/*#/#g;
1970        my $p = $path;
1971        # Strip the irrelevant part of the path.
1972        $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
1973        # Ensure the path is terminated by a `/'.
1974        $p =~ s#/*$#/#;
1975
1976        # The properties contain all the internal SVN stuff nobody
1977        # (usually) cares about.
1978        my $interesting_props = 0;
1979        foreach (keys %{$props}) {
1980                # If it doesn't start with `svn:', it must be a
1981                # user-defined property.
1982                ++$interesting_props and next if $_ !~ /^svn:/;
1983                # FIXME: Fragile, if SVN adds new public properties,
1984                # this needs to be updated.
1985                ++$interesting_props if /^svn:(?:ignore|keywords|executable
1986                                                 |eol-style|mime-type
1987                                                 |externals|needs-lock)$/x;
1988        }
1989        &$sub($self, $p, $props) if $interesting_props;
1990
1991        foreach (sort keys %$dirent) {
1992                next if $dirent->{$_}->{kind} != $SVN::Node::dir;
1993                $self->prop_walk($self->{path} . $p . $_, $rev, $sub);
1994        }
1995}
1996
1997sub last_rev { ($_[0]->last_rev_commit)[0] }
1998sub last_commit { ($_[0]->last_rev_commit)[1] }
1999
2000# returns the newest SVN revision number and newest commit SHA1
2001sub last_rev_commit {
2002        my ($self) = @_;
2003        if (defined $self->{last_rev} && defined $self->{last_commit}) {
2004                return ($self->{last_rev}, $self->{last_commit});
2005        }
2006        my $c = ::verify_ref($self->refname.'^0');
2007        if ($c && !$self->use_svm_props && !$self->no_metadata) {
2008                my $rev = (::cmt_metadata($c))[1];
2009                if (defined $rev) {
2010                        ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
2011                        return ($rev, $c);
2012                }
2013        }
2014        my $map_path = $self->map_path;
2015        unless (-e $map_path) {
2016                ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
2017                return (undef, undef);
2018        }
2019        my ($rev, $commit) = $self->rev_map_max(1);
2020        ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
2021        return ($rev, $commit);
2022}
2023
2024sub get_fetch_range {
2025        my ($self, $min, $max) = @_;
2026        $max ||= $self->ra->get_latest_revnum;
2027        $min ||= $self->rev_map_max;
2028        (++$min, $max);
2029}
2030
2031sub tmp_config {
2032        my (@args) = @_;
2033        my $old_def_config = "$ENV{GIT_DIR}/svn/config";
2034        my $config = "$ENV{GIT_DIR}/svn/.metadata";
2035        if (! -f $config && -f $old_def_config) {
2036                rename $old_def_config, $config or
2037                       die "Failed rename $old_def_config => $config: $!\n";
2038        }
2039        my $old_config = $ENV{GIT_CONFIG};
2040        $ENV{GIT_CONFIG} = $config;
2041        $@ = undef;
2042        my @ret = eval {
2043                unless (-f $config) {
2044                        mkfile($config);
2045                        open my $fh, '>', $config or
2046                            die "Can't open $config: $!\n";
2047                        print $fh "; This file is used internally by ",
2048                                  "git-svn\n" or die
2049                                  "Couldn't write to $config: $!\n";
2050                        print $fh "; You should not have to edit it\n" or
2051                              die "Couldn't write to $config: $!\n";
2052                        close $fh or die "Couldn't close $config: $!\n";
2053                }
2054                command('config', @args);
2055        };
2056        my $err = $@;
2057        if (defined $old_config) {
2058                $ENV{GIT_CONFIG} = $old_config;
2059        } else {
2060                delete $ENV{GIT_CONFIG};
2061        }
2062        die $err if $err;
2063        wantarray ? @ret : $ret[0];
2064}
2065
2066sub tmp_index_do {
2067        my ($self, $sub) = @_;
2068        my $old_index = $ENV{GIT_INDEX_FILE};
2069        $ENV{GIT_INDEX_FILE} = $self->{index};
2070        $@ = undef;
2071        my @ret = eval {
2072                my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
2073                mkpath([$dir]) unless -d $dir;
2074                &$sub;
2075        };
2076        my $err = $@;
2077        if (defined $old_index) {
2078                $ENV{GIT_INDEX_FILE} = $old_index;
2079        } else {
2080                delete $ENV{GIT_INDEX_FILE};
2081        }
2082        die $err if $err;
2083        wantarray ? @ret : $ret[0];
2084}
2085
2086sub assert_index_clean {
2087        my ($self, $treeish) = @_;
2088
2089        $self->tmp_index_do(sub {
2090                command_noisy('read-tree', $treeish) unless -e $self->{index};
2091                my $x = command_oneline('write-tree');
2092                my ($y) = (command(qw/cat-file commit/, $treeish) =~
2093                           /^tree ($::sha1)/mo);
2094                return if $y eq $x;
2095
2096                warn "Index mismatch: $y != $x\nrereading $treeish\n";
2097                unlink $self->{index} or die "unlink $self->{index}: $!\n";
2098                command_noisy('read-tree', $treeish);
2099                $x = command_oneline('write-tree');
2100                if ($y ne $x) {
2101                        ::fatal "trees ($treeish) $y != $x\n",
2102                                "Something is seriously wrong...";
2103                }
2104        });
2105}
2106
2107sub get_commit_parents {
2108        my ($self, $log_entry) = @_;
2109        my (%seen, @ret, @tmp);
2110        # legacy support for 'set-tree'; this is only used by set_tree_cb:
2111        if (my $ip = $self->{inject_parents}) {
2112                if (my $commit = delete $ip->{$log_entry->{revision}}) {
2113                        push @tmp, $commit;
2114                }
2115        }
2116        if (my $cur = ::verify_ref($self->refname.'^0')) {
2117                push @tmp, $cur;
2118        }
2119        if (my $ipd = $self->{inject_parents_dcommit}) {
2120                if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2121                        push @tmp, @$commit;
2122                }
2123        }
2124        push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2125        while (my $p = shift @tmp) {
2126                next if $seen{$p};
2127                $seen{$p} = 1;
2128                push @ret, $p;
2129                # MAXPARENT is defined to 16 in commit-tree.c:
2130                last if @ret >= 16;
2131        }
2132        if (@tmp) {
2133                die "r$log_entry->{revision}: No room for parents:\n\t",
2134                    join("\n\t", @tmp), "\n";
2135        }
2136        @ret;
2137}
2138
2139sub rewrite_root {
2140        my ($self) = @_;
2141        return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2142        my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2143        my $rwr = eval { command_oneline(qw/config --get/, $k) };
2144        if ($rwr) {
2145                $rwr =~ s#/+$##;
2146                if ($rwr !~ m#^[a-z\+]+://#) {
2147                        die "$rwr is not a valid URL (key: $k)\n";
2148                }
2149        }
2150        $self->{-rewrite_root} = $rwr;
2151}
2152
2153sub metadata_url {
2154        my ($self) = @_;
2155        ($self->rewrite_root || $self->{url}) .
2156           (length $self->{path} ? '/' . $self->{path} : '');
2157}
2158
2159sub full_url {
2160        my ($self) = @_;
2161        $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2162}
2163
2164
2165sub set_commit_header_env {
2166        my ($log_entry) = @_;
2167        my %env;
2168        foreach my $ned (qw/NAME EMAIL DATE/) {
2169                foreach my $ac (qw/AUTHOR COMMITTER/) {
2170                        $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2171                }
2172        }
2173
2174        $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2175        $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2176        $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2177
2178        $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2179                                                ? $log_entry->{commit_name}
2180                                                : $log_entry->{name};
2181        $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2182                                                ? $log_entry->{commit_email}
2183                                                : $log_entry->{email};
2184        \%env;
2185}
2186
2187sub restore_commit_header_env {
2188        my ($env) = @_;
2189        foreach my $ned (qw/NAME EMAIL DATE/) {
2190                foreach my $ac (qw/AUTHOR COMMITTER/) {
2191                        my $k = "GIT_${ac}_${ned}";
2192                        if (defined $env->{$k}) {
2193                                $ENV{$k} = $env->{$k};
2194                        } else {
2195                                delete $ENV{$k};
2196                        }
2197                }
2198        }
2199}
2200
2201sub gc {
2202        command_noisy('gc', '--auto');
2203};
2204
2205sub do_git_commit {
2206        my ($self, $log_entry) = @_;
2207        my $lr = $self->last_rev;
2208        if (defined $lr && $lr >= $log_entry->{revision}) {
2209                die "Last fetched revision of ", $self->refname,
2210                    " was r$lr, but we are about to fetch: ",
2211                    "r$log_entry->{revision}!\n";
2212        }
2213        if (my $c = $self->rev_map_get($log_entry->{revision})) {
2214                croak "$log_entry->{revision} = $c already exists! ",
2215                      "Why are we refetching it?\n";
2216        }
2217        my $old_env = set_commit_header_env($log_entry);
2218        my $tree = $log_entry->{tree};
2219        if (!defined $tree) {
2220                $tree = $self->tmp_index_do(sub {
2221                                            command_oneline('write-tree') });
2222        }
2223        die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2224
2225        my @exec = ('git-commit-tree', $tree);
2226        foreach ($self->get_commit_parents($log_entry)) {
2227                push @exec, '-p', $_;
2228        }
2229        defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2230                                                                   or croak $!;
2231        print $msg_fh $log_entry->{log} or croak $!;
2232        restore_commit_header_env($old_env);
2233        unless ($self->no_metadata) {
2234                print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2235                              or croak $!;
2236        }
2237        $msg_fh->flush == 0 or croak $!;
2238        close $msg_fh or croak $!;
2239        chomp(my $commit = do { local $/; <$out_fh> });
2240        close $out_fh or croak $!;
2241        waitpid $pid, 0;
2242        croak $? if $?;
2243        if ($commit !~ /^$::sha1$/o) {
2244                die "Failed to commit, invalid sha1: $commit\n";
2245        }
2246
2247        $self->rev_map_set($log_entry->{revision}, $commit, 1);
2248
2249        $self->{last_rev} = $log_entry->{revision};
2250        $self->{last_commit} = $commit;
2251        print "r$log_entry->{revision}";
2252        if (defined $log_entry->{svm_revision}) {
2253                 print " (\@$log_entry->{svm_revision})";
2254                 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2255                                   0, $self->svm_uuid);
2256        }
2257        print " = $commit ($self->{ref_id})\n";
2258        if (--$_gc_nr == 0) {
2259                $_gc_nr = $_gc_period;
2260                gc();
2261        }
2262        return $commit;
2263}
2264
2265sub match_paths {
2266        my ($self, $paths, $r) = @_;
2267        return 1 if $self->{path} eq '';
2268        if (my $path = $paths->{"/$self->{path}"}) {
2269                return ($path->{action} eq 'D') ? 0 : 1;
2270        }
2271        $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
2272        if (grep /$self->{path_regex}/, keys %$paths) {
2273                return 1;
2274        }
2275        my $c = '';
2276        foreach (split m#/#, $self->{path}) {
2277                $c .= "/$_";
2278                next unless ($paths->{$c} &&
2279                             ($paths->{$c}->{action} =~ /^[AR]$/));
2280                if ($self->ra->check_path($self->{path}, $r) ==
2281                    $SVN::Node::dir) {
2282                        return 1;
2283                }
2284        }
2285        return 0;
2286}
2287
2288sub find_parent_branch {
2289        my ($self, $paths, $rev) = @_;
2290        return undef unless $self->follow_parent;
2291        unless (defined $paths) {
2292                my $err_handler = $SVN::Error::handler;
2293                $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2294                $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1, sub {
2295                                   $paths =
2296                                      Git::SVN::Ra::dup_changed_paths($_[0]) });
2297                $SVN::Error::handler = $err_handler;
2298        }
2299        return undef unless defined $paths;
2300
2301        # look for a parent from another branch:
2302        my @b_path_components = split m#/#, $self->rel_path;
2303        my @a_path_components;
2304        my $i;
2305        while (@b_path_components) {
2306                $i = $paths->{'/'.join('/', @b_path_components)};
2307                last if $i && defined $i->{copyfrom_path};
2308                unshift(@a_path_components, pop(@b_path_components));
2309        }
2310        return undef unless defined $i && defined $i->{copyfrom_path};
2311        my $branch_from = $i->{copyfrom_path};
2312        if (@a_path_components) {
2313                print STDERR "branch_from: $branch_from => ";
2314                $branch_from .= '/'.join('/', @a_path_components);
2315                print STDERR $branch_from, "\n";
2316        }
2317        my $r = $i->{copyfrom_rev};
2318        my $repos_root = $self->ra->{repos_root};
2319        my $url = $self->ra->{url};
2320        my $new_url = $repos_root . $branch_from;
2321        print STDERR  "Found possible branch point: ",
2322                      "$new_url => ", $self->full_url, ", $r\n";
2323        $branch_from =~ s#^/##;
2324        my $gs = Git::SVN->find_by_url($new_url, $repos_root, $branch_from);
2325        unless ($gs) {
2326                my $ref_id = $self->{ref_id};
2327                $ref_id =~ s/\@\d+$//;
2328                $ref_id .= "\@$r";
2329                # just grow a tail if we're not unique enough :x
2330                $ref_id .= '-' while find_ref($ref_id);
2331                print STDERR "Initializing parent: $ref_id\n";
2332                my ($u, $p, $repo_id) = ($new_url, '', $ref_id);
2333                if ($u =~ s#^\Q$url\E(/|$)##) {
2334                        $p = $u;
2335                        $u = $url;
2336                        $repo_id = $self->{repo_id};
2337                }
2338                $gs = Git::SVN->init($u, $p, $repo_id, $ref_id, 1);
2339        }
2340        my ($r0, $parent) = $gs->find_rev_before($r, 1);
2341        if (!defined $r0 || !defined $parent) {
2342                my ($base, $head) = parse_revision_argument(0, $r);
2343                if ($base <= $r) {
2344                        $gs->fetch($base, $r);
2345                }
2346                ($r0, $parent) = $gs->last_rev_commit;
2347        }
2348        if (defined $r0 && defined $parent) {
2349                print STDERR "Found branch parent: ($self->{ref_id}) $parent\n";
2350                my $ed;
2351                if ($self->ra->can_do_switch) {
2352                        $self->assert_index_clean($parent);
2353                        print STDERR "Following parent with do_switch\n";
2354                        # do_switch works with svn/trunk >= r22312, but that
2355                        # is not included with SVN 1.4.3 (the latest version
2356                        # at the moment), so we can't rely on it
2357                        $self->{last_commit} = $parent;
2358                        $ed = SVN::Git::Fetcher->new($self);
2359                        $gs->ra->gs_do_switch($r0, $rev, $gs,
2360                                              $self->full_url, $ed)
2361                          or die "SVN connection failed somewhere...\n";
2362                } elsif ($self->ra->trees_match($new_url, $r0,
2363                                                $self->full_url, $rev)) {
2364                        print STDERR "Trees match:\n",
2365                                     "  $new_url\@$r0\n",
2366                                     "  ${\$self->full_url}\@$rev\n",
2367                                     "Following parent with no changes\n";
2368                        $self->tmp_index_do(sub {
2369                            command_noisy('read-tree', $parent);
2370                        });
2371                        $self->{last_commit} = $parent;
2372                } else {
2373                        print STDERR "Following parent with do_update\n";
2374                        $ed = SVN::Git::Fetcher->new($self);
2375                        $self->ra->gs_do_update($rev, $rev, $self, $ed)
2376                          or die "SVN connection failed somewhere...\n";
2377                }
2378                print STDERR "Successfully followed parent\n";
2379                return $self->make_log_entry($rev, [$parent], $ed);
2380        }
2381        return undef;
2382}
2383
2384sub do_fetch {
2385        my ($self, $paths, $rev) = @_;
2386        my $ed;
2387        my ($last_rev, @parents);
2388        if (my $lc = $self->last_commit) {
2389                # we can have a branch that was deleted, then re-added
2390                # under the same name but copied from another path, in
2391                # which case we'll have multiple parents (we don't
2392                # want to break the original ref, nor lose copypath info):
2393                if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2394                        push @{$log_entry->{parents}}, $lc;
2395                        return $log_entry;
2396                }
2397                $ed = SVN::Git::Fetcher->new($self);
2398                $last_rev = $self->{last_rev};
2399                $ed->{c} = $lc;
2400                @parents = ($lc);
2401        } else {
2402                $last_rev = $rev;
2403                if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2404                        return $log_entry;
2405                }
2406                $ed = SVN::Git::Fetcher->new($self);
2407        }
2408        unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2409                die "SVN connection failed somewhere...\n";
2410        }
2411        $self->make_log_entry($rev, \@parents, $ed);
2412}
2413
2414sub get_untracked {
2415        my ($self, $ed) = @_;
2416        my @out;
2417        my $h = $ed->{empty};
2418        foreach (sort keys %$h) {
2419                my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2420                push @out, "  $act: " . uri_encode($_);
2421                warn "W: $act: $_\n";
2422        }
2423        foreach my $t (qw/dir_prop file_prop/) {
2424                $h = $ed->{$t} or next;
2425                foreach my $path (sort keys %$h) {
2426                        my $ppath = $path eq '' ? '.' : $path;
2427                        foreach my $prop (sort keys %{$h->{$path}}) {
2428                                next if $SKIP_PROP{$prop};
2429                                my $v = $h->{$path}->{$prop};
2430                                my $t_ppath_prop = "$t: " .
2431                                                    uri_encode($ppath) . ' ' .
2432                                                    uri_encode($prop);
2433                                if (defined $v) {
2434                                        push @out, "  +$t_ppath_prop " .
2435                                                   uri_encode($v);
2436                                } else {
2437                                        push @out, "  -$t_ppath_prop";
2438                                }
2439                        }
2440                }
2441        }
2442        foreach my $t (qw/absent_file absent_directory/) {
2443                $h = $ed->{$t} or next;
2444                foreach my $parent (sort keys %$h) {
2445                        foreach my $path (sort @{$h->{$parent}}) {
2446                                push @out, "  $t: " .
2447                                           uri_encode("$parent/$path");
2448                                warn "W: $t: $parent/$path ",
2449                                     "Insufficient permissions?\n";
2450                        }
2451                }
2452        }
2453        \@out;
2454}
2455
2456sub parse_svn_date {
2457        my $date = shift || return '+0000 1970-01-01 00:00:00';
2458        my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2459                                            (\d\d)\:(\d\d)\:(\d\d).\d+Z$/x) or
2460                                         croak "Unable to parse date: $date\n";
2461        "+0000 $Y-$m-$d $H:$M:$S";
2462}
2463
2464sub check_author {
2465        my ($author) = @_;
2466        if (!defined $author || length $author == 0) {
2467                $author = '(no author)';
2468        } elsif (defined $::_authors && ! defined $::users{$author}) {
2469                die "Author: $author not defined in $::_authors file\n";
2470        }
2471        $author;
2472}
2473
2474sub make_log_entry {
2475        my ($self, $rev, $parents, $ed) = @_;
2476        my $untracked = $self->get_untracked($ed);
2477
2478        open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
2479        print $un "r$rev\n" or croak $!;
2480        print $un $_, "\n" foreach @$untracked;
2481        my %log_entry = ( parents => $parents || [], revision => $rev,
2482                          log => '');
2483
2484        my $headrev;
2485        my $logged = delete $self->{logged_rev_props};
2486        if (!$logged || $self->{-want_revprops}) {
2487                my $rp = $self->ra->rev_proplist($rev);
2488                foreach (sort keys %$rp) {
2489                        my $v = $rp->{$_};
2490                        if (/^svn:(author|date|log)$/) {
2491                                $log_entry{$1} = $v;
2492                        } elsif ($_ eq 'svm:headrev') {
2493                                $headrev = $v;
2494                        } else {
2495                                print $un "  rev_prop: ", uri_encode($_), ' ',
2496                                          uri_encode($v), "\n";
2497                        }
2498                }
2499        } else {
2500                map { $log_entry{$_} = $logged->{$_} } keys %$logged;
2501        }
2502        close $un or croak $!;
2503
2504        $log_entry{date} = parse_svn_date($log_entry{date});
2505        $log_entry{log} .= "\n";
2506        my $author = $log_entry{author} = check_author($log_entry{author});
2507        my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
2508                                                       : ($author, undef);
2509
2510        my ($commit_name, $commit_email) = ($name, $email);
2511        if ($_use_log_author) {
2512                my $name_field;
2513                if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
2514                        $name_field = $1;
2515                } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
2516                        $name_field = $1;
2517                }
2518                if (!defined $name_field) {
2519                        if (!defined $email) {
2520                                $email = $name;
2521                        }
2522                } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
2523                        ($name, $email) = ($1, $2);
2524                } elsif ($name_field =~ /(.*)@/) {
2525                        ($name, $email) = ($1, $name_field);
2526                } else {
2527                        ($name, $email) = ($name_field, $name_field);
2528                }
2529        }
2530        if (defined $headrev && $self->use_svm_props) {
2531                if ($self->rewrite_root) {
2532                        die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
2533                            "options set!\n";
2534                }
2535                my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$};
2536                # we don't want "SVM: initializing mirror for junk" ...
2537                return undef if $r == 0;
2538                my $svm = $self->svm;
2539                if ($uuid ne $svm->{uuid}) {
2540                        die "UUID mismatch on SVM path:\n",
2541                            "expected: $svm->{uuid}\n",
2542                            "     got: $uuid\n";
2543                }
2544                my $full_url = $self->full_url;
2545                $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
2546                             die "Failed to replace '$svm->{replace}' with ",
2547                                 "'$svm->{source}' in $full_url\n";
2548                # throw away username for storing in records
2549                remove_username($full_url);
2550                $log_entry{metadata} = "$full_url\@$r $uuid";
2551                $log_entry{svm_revision} = $r;
2552                $email ||= "$author\@$uuid";
2553                $commit_email ||= "$author\@$uuid";
2554        } elsif ($self->use_svnsync_props) {
2555                my $full_url = $self->svnsync->{url};
2556                $full_url .= "/$self->{path}" if length $self->{path};
2557                remove_username($full_url);
2558                my $uuid = $self->svnsync->{uuid};
2559                $log_entry{metadata} = "$full_url\@$rev $uuid";
2560                $email ||= "$author\@$uuid";
2561                $commit_email ||= "$author\@$uuid";
2562        } else {
2563                my $url = $self->metadata_url;
2564                remove_username($url);
2565                $log_entry{metadata} = "$url\@$rev " .
2566                                       $self->ra->get_uuid;
2567                $email ||= "$author\@" . $self->ra->get_uuid;
2568                $commit_email ||= "$author\@" . $self->ra->get_uuid;
2569        }
2570        $log_entry{name} = $name;
2571        $log_entry{email} = $email;
2572        $log_entry{commit_name} = $commit_name;
2573        $log_entry{commit_email} = $commit_email;
2574        \%log_entry;
2575}
2576
2577sub fetch {
2578        my ($self, $min_rev, $max_rev, @parents) = @_;
2579        my ($last_rev, $last_commit) = $self->last_rev_commit;
2580        my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
2581        $self->ra->gs_fetch_loop_common($base, $head, [$self]);
2582}
2583
2584sub set_tree_cb {
2585        my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
2586        $self->{inject_parents} = { $rev => $tree };
2587        $self->fetch(undef, undef);
2588}
2589
2590sub set_tree {
2591        my ($self, $tree) = (shift, shift);
2592        my $log_entry = ::get_commit_entry($tree);
2593        unless ($self->{last_rev}) {
2594                ::fatal("Must have an existing revision to commit");
2595        }
2596        my %ed_opts = ( r => $self->{last_rev},
2597                        log => $log_entry->{log},
2598                        ra => $self->ra,
2599                        tree_a => $self->{last_commit},
2600                        tree_b => $tree,
2601                        editor_cb => sub {
2602                               $self->set_tree_cb($log_entry, $tree, @_) },
2603                        svn_path => $self->{path} );
2604        if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
2605                print "No changes\nr$self->{last_rev} = $tree\n";
2606        }
2607}
2608
2609sub rebuild_from_rev_db {
2610        my ($self, $path) = @_;
2611        my $r = -1;
2612        open my $fh, '<', $path or croak "open: $!";
2613        binmode $fh or croak "binmode: $!";
2614        while (<$fh>) {
2615                length($_) == 41 or croak "inconsistent size in ($_) != 41";
2616                chomp($_);
2617                ++$r;
2618                next if $_ eq ('0' x 40);
2619                $self->rev_map_set($r, $_);
2620                print "r$r = $_\n";
2621        }
2622        close $fh or croak "close: $!";
2623        unlink $path or croak "unlink: $!";
2624}
2625
2626sub rebuild {
2627        my ($self) = @_;
2628        my $map_path = $self->map_path;
2629        my $partial = (-e $map_path && ! -z $map_path);
2630        return unless ::verify_ref($self->refname.'^0');
2631        if (!$partial && ($self->use_svm_props || $self->no_metadata)) {
2632                my $rev_db = $self->rev_db_path;
2633                $self->rebuild_from_rev_db($rev_db);
2634                if ($self->use_svm_props) {
2635                        my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
2636                        $self->rebuild_from_rev_db($svm_rev_db);
2637                }
2638                $self->unlink_rev_db_symlink;
2639                return;
2640        }
2641        print "Rebuilding $map_path ...\n" if (!$partial);
2642        my ($base_rev, $head) = ($partial ? $self->rev_map_max_norebuild(1) :
2643                (undef, undef));
2644        my ($log, $ctx) =
2645            command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
2646                                ($head ? "$head.." : "") . $self->refname,
2647                                '--');
2648        my $metadata_url = $self->metadata_url;
2649        remove_username($metadata_url);
2650        my $svn_uuid = $self->ra_uuid;
2651        my $c;
2652        while (<$log>) {
2653                if ( m{^commit ($::sha1)$} ) {
2654                        $c = $1;
2655                        next;
2656                }
2657                next unless s{^\s*(git-svn-id:)}{$1};
2658                my ($url, $rev, $uuid) = ::extract_metadata($_);
2659                remove_username($url);
2660
2661                # ignore merges (from set-tree)
2662                next if (!defined $rev || !$uuid);
2663
2664                # if we merged or otherwise started elsewhere, this is
2665                # how we break out of it
2666                if (($uuid ne $svn_uuid) ||
2667                    ($metadata_url && $url && ($url ne $metadata_url))) {
2668                        next;
2669                }
2670                if ($partial && $head) {
2671                        print "Partial-rebuilding $map_path ...\n";
2672                        print "Currently at $base_rev = $head\n";
2673                        $head = undef;
2674                }
2675
2676                $self->rev_map_set($rev, $c);
2677                print "r$rev = $c\n";
2678        }
2679        command_close_pipe($log, $ctx);
2680        print "Done rebuilding $map_path\n" if (!$partial || !$head);
2681        my $rev_db_path = $self->rev_db_path;
2682        if (-f $self->rev_db_path) {
2683                unlink $self->rev_db_path or croak "unlink: $!";
2684        }
2685        $self->unlink_rev_db_symlink;
2686}
2687
2688# rev_map:
2689# Tie::File seems to be prone to offset errors if revisions get sparse,
2690# it's not that fast, either.  Tie::File is also not in Perl 5.6.  So
2691# one of my favorite modules is out :<  Next up would be one of the DBM
2692# modules, but I'm not sure which is most portable...
2693#
2694# This is the replacement for the rev_db format, which was too big
2695# and inefficient for large repositories with a lot of sparse history
2696# (mainly tags)
2697#
2698# The format is this:
2699#   - 24 bytes for every record,
2700#     * 4 bytes for the integer representing an SVN revision number
2701#     * 20 bytes representing the sha1 of a git commit
2702#   - No empty padding records like the old format
2703#     (except the last record, which can be overwritten)
2704#   - new records are written append-only since SVN revision numbers
2705#     increase monotonically
2706#   - lookups on SVN revision number are done via a binary search
2707#   - Piping the file to xxd -c24 is a good way of dumping it for
2708#     viewing or editing (piped back through xxd -r), should the need
2709#     ever arise.
2710#   - The last record can be padding revision with an all-zero sha1
2711#     This is used to optimize fetch performance when using multiple
2712#     "fetch" directives in .git/config
2713#
2714# These files are disposable unless noMetadata or useSvmProps is set
2715
2716sub _rev_map_set {
2717        my ($fh, $rev, $commit) = @_;
2718
2719        binmode $fh or croak "binmode: $!";
2720        my $size = (stat($fh))[7];
2721        ($size % 24) == 0 or croak "inconsistent size: $size";
2722
2723        my $wr_offset = 0;
2724        if ($size > 0) {
2725                sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2726                my $read = sysread($fh, my $buf, 24) or croak "read: $!";
2727                $read == 24 or croak "read only $read bytes (!= 24)";
2728                my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
2729                if ($last_commit eq ('0' x40)) {
2730                        if ($size >= 48) {
2731                                sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2732                                $read = sysread($fh, $buf, 24) or
2733                                    croak "read: $!";
2734                                $read == 24 or
2735                                    croak "read only $read bytes (!= 24)";
2736                                ($last_rev, $last_commit) =
2737                                    unpack(rev_map_fmt, $buf);
2738                                if ($last_commit eq ('0' x40)) {
2739                                        croak "inconsistent .rev_map\n";
2740                                }
2741                        }
2742                        if ($last_rev >= $rev) {
2743                                croak "last_rev is higher!: $last_rev >= $rev";
2744                        }
2745                        $wr_offset = -24;
2746                }
2747        }
2748        sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
2749        syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
2750          croak "write: $!";
2751}
2752
2753sub mkfile {
2754        my ($path) = @_;
2755        unless (-e $path) {
2756                my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
2757                mkpath([$dir]) unless -d $dir;
2758                open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
2759                close $fh or die "Couldn't close (create) $path: $!\n";
2760        }
2761}
2762
2763sub rev_map_set {
2764        my ($self, $rev, $commit, $update_ref, $uuid) = @_;
2765        length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
2766        my $db = $self->map_path($uuid);
2767        my $db_lock = "$db.lock";
2768        my $sig;
2769        if ($update_ref) {
2770                $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2771                            $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
2772        }
2773        mkfile($db);
2774
2775        $LOCKFILES{$db_lock} = 1;
2776        my $sync;
2777        # both of these options make our .rev_db file very, very important
2778        # and we can't afford to lose it because rebuild() won't work
2779        if ($self->use_svm_props || $self->no_metadata) {
2780                $sync = 1;
2781                copy($db, $db_lock) or die "rev_map_set(@_): ",
2782                                           "Failed to copy: ",
2783                                           "$db => $db_lock ($!)\n";
2784        } else {
2785                rename $db, $db_lock or die "rev_map_set(@_): ",
2786                                            "Failed to rename: ",
2787                                            "$db => $db_lock ($!)\n";
2788        }
2789
2790        sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
2791             or croak "Couldn't open $db_lock: $!\n";
2792        _rev_map_set($fh, $rev, $commit);
2793        if ($sync) {
2794                $fh->flush or die "Couldn't flush $db_lock: $!\n";
2795                $fh->sync or die "Couldn't sync $db_lock: $!\n";
2796        }
2797        close $fh or croak $!;
2798        if ($update_ref) {
2799                $_head = $self;
2800                command_noisy('update-ref', '-m', "r$rev",
2801                              $self->refname, $commit);
2802        }
2803        rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
2804                                    "$db_lock => $db ($!)\n";
2805        delete $LOCKFILES{$db_lock};
2806        if ($update_ref) {
2807                $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2808                            $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
2809                kill $sig, $$ if defined $sig;
2810        }
2811}
2812
2813# If want_commit, this will return an array of (rev, commit) where
2814# commit _must_ be a valid commit in the archive.
2815# Otherwise, it'll return the max revision (whether or not the
2816# commit is valid or just a 0x40 placeholder).
2817sub rev_map_max {
2818        my ($self, $want_commit) = @_;
2819        $self->rebuild;
2820        my ($r, $c) = $self->rev_map_max_norebuild($want_commit);
2821        $want_commit ? ($r, $c) : $r;
2822}
2823
2824sub rev_map_max_norebuild {
2825        my ($self, $want_commit) = @_;
2826        my $map_path = $self->map_path;
2827        stat $map_path or return $want_commit ? (0, undef) : 0;
2828        sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2829        binmode $fh or croak "binmode: $!";
2830        my $size = (stat($fh))[7];
2831        ($size % 24) == 0 or croak "inconsistent size: $size";
2832
2833        if ($size == 0) {
2834                close $fh or croak "close: $!";
2835                return $want_commit ? (0, undef) : 0;
2836        }
2837
2838        sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2839        sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2840        my ($r, $c) = unpack(rev_map_fmt, $buf);
2841        if ($want_commit && $c eq ('0' x40)) {
2842                if ($size < 48) {
2843                        return $want_commit ? (0, undef) : 0;
2844                }
2845                sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2846                sysread($fh, $buf, 24) == 24 or croak "read: $!";
2847                ($r, $c) = unpack(rev_map_fmt, $buf);
2848                if ($c eq ('0'x40)) {
2849                        croak "Penultimate record is all-zeroes in $map_path";
2850                }
2851        }
2852        close $fh or croak "close: $!";
2853        $want_commit ? ($r, $c) : $r;
2854}
2855
2856sub rev_map_get {
2857        my ($self, $rev, $uuid) = @_;
2858        my $map_path = $self->map_path($uuid);
2859        return undef unless -e $map_path;
2860
2861        sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2862        binmode $fh or croak "binmode: $!";
2863        my $size = (stat($fh))[7];
2864        ($size % 24) == 0 or croak "inconsistent size: $size";
2865
2866        if ($size == 0) {
2867                close $fh or croak "close: $fh";
2868                return undef;
2869        }
2870
2871        my ($l, $u) = (0, $size - 24);
2872        my ($r, $c, $buf);
2873
2874        while ($l <= $u) {
2875                my $i = int(($l/24 + $u/24) / 2) * 24;
2876                sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
2877                sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2878                my ($r, $c) = unpack('NH40', $buf);
2879
2880                if ($r < $rev) {
2881                        $l = $i + 24;
2882                } elsif ($r > $rev) {
2883                        $u = $i - 24;
2884                } else { # $r == $rev
2885                        close($fh) or croak "close: $!";
2886                        return $c eq ('0' x 40) ? undef : $c;
2887                }
2888        }
2889        close($fh) or croak "close: $!";
2890        undef;
2891}
2892
2893# Finds the first svn revision that exists on (if $eq_ok is true) or
2894# before $rev for the current branch.  It will not search any lower
2895# than $min_rev.  Returns the git commit hash and svn revision number
2896# if found, else (undef, undef).
2897sub find_rev_before {
2898        my ($self, $rev, $eq_ok, $min_rev) = @_;
2899        --$rev unless $eq_ok;
2900        $min_rev ||= 1;
2901        while ($rev >= $min_rev) {
2902                if (my $c = $self->rev_map_get($rev)) {
2903                        return ($rev, $c);
2904                }
2905                --$rev;
2906        }
2907        return (undef, undef);
2908}
2909
2910# Finds the first svn revision that exists on (if $eq_ok is true) or
2911# after $rev for the current branch.  It will not search any higher
2912# than $max_rev.  Returns the git commit hash and svn revision number
2913# if found, else (undef, undef).
2914sub find_rev_after {
2915        my ($self, $rev, $eq_ok, $max_rev) = @_;
2916        ++$rev unless $eq_ok;
2917        $max_rev ||= $self->rev_map_max;
2918        while ($rev <= $max_rev) {
2919                if (my $c = $self->rev_map_get($rev)) {
2920                        return ($rev, $c);
2921                }
2922                ++$rev;
2923        }
2924        return (undef, undef);
2925}
2926
2927sub _new {
2928        my ($class, $repo_id, $ref_id, $path) = @_;
2929        unless (defined $repo_id && length $repo_id) {
2930                $repo_id = $Git::SVN::default_repo_id;
2931        }
2932        unless (defined $ref_id && length $ref_id) {
2933                $_[2] = $ref_id = $Git::SVN::default_ref_id;
2934        }
2935        $_[1] = $repo_id;
2936        my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
2937        $_[3] = $path = '' unless (defined $path);
2938        mkpath(["$ENV{GIT_DIR}/svn"]);
2939        bless {
2940                ref_id => $ref_id, dir => $dir, index => "$dir/index",
2941                path => $path, config => "$ENV{GIT_DIR}/svn/config",
2942                map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
2943}
2944
2945# for read-only access of old .rev_db formats
2946sub unlink_rev_db_symlink {
2947        my ($self) = @_;
2948        my $link = $self->rev_db_path;
2949        $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
2950        if (-l $link) {
2951                unlink $link or croak "unlink: $link failed!";
2952        }
2953}
2954
2955sub rev_db_path {
2956        my ($self, $uuid) = @_;
2957        my $db_path = $self->map_path($uuid);
2958        $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
2959            or croak "map_path: $db_path does not contain '/.rev_map.' !";
2960        $db_path;
2961}
2962
2963# the new replacement for .rev_db
2964sub map_path {
2965        my ($self, $uuid) = @_;
2966        $uuid ||= $self->ra_uuid;
2967        "$self->{map_root}.$uuid";
2968}
2969
2970sub uri_encode {
2971        my ($f) = @_;
2972        $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
2973        $f
2974}
2975
2976sub remove_username {
2977        $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
2978}
2979
2980package Git::SVN::Prompt;
2981use strict;
2982use warnings;
2983require SVN::Core;
2984use vars qw/$_no_auth_cache $_username/;
2985
2986sub simple {
2987        my ($cred, $realm, $default_username, $may_save, $pool) = @_;
2988        $may_save = undef if $_no_auth_cache;
2989        $default_username = $_username if defined $_username;
2990        if (defined $default_username && length $default_username) {
2991                if (defined $realm && length $realm) {
2992                        print STDERR "Authentication realm: $realm\n";
2993                        STDERR->flush;
2994                }
2995                $cred->username($default_username);
2996        } else {
2997                username($cred, $realm, $may_save, $pool);
2998        }
2999        $cred->password(_read_password("Password for '" .
3000                                       $cred->username . "': ", $realm));
3001        $cred->may_save($may_save);
3002        $SVN::_Core::SVN_NO_ERROR;
3003}
3004
3005sub ssl_server_trust {
3006        my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
3007        $may_save = undef if $_no_auth_cache;
3008        print STDERR "Error validating server certificate for '$realm':\n";
3009        {
3010                no warnings 'once';
3011                # All variables SVN::Auth::SSL::* are used only once,
3012                # so we're shutting up Perl warnings about this.
3013                if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
3014                        print STDERR " - The certificate is not issued ",
3015                            "by a trusted authority. Use the\n",
3016                            "   fingerprint to validate ",
3017                            "the certificate manually!\n";
3018                }
3019                if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
3020                        print STDERR " - The certificate hostname ",
3021                            "does not match.\n";
3022                }
3023                if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
3024                        print STDERR " - The certificate is not yet valid.\n";
3025                }
3026                if ($failures & $SVN::Auth::SSL::EXPIRED) {
3027                        print STDERR " - The certificate has expired.\n";
3028                }
3029                if ($failures & $SVN::Auth::SSL::OTHER) {
3030                        print STDERR " - The certificate has ",
3031                            "an unknown error.\n";
3032                }
3033        } # no warnings 'once'
3034        printf STDERR
3035                "Certificate information:\n".
3036                " - Hostname: %s\n".
3037                " - Valid: from %s until %s\n".
3038                " - Issuer: %s\n".
3039                " - Fingerprint: %s\n",
3040                map $cert_info->$_, qw(hostname valid_from valid_until
3041                                       issuer_dname fingerprint);
3042        my $choice;
3043prompt:
3044        print STDERR $may_save ?
3045              "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
3046              "(R)eject or accept (t)emporarily? ";
3047        STDERR->flush;
3048        $choice = lc(substr(<STDIN> || 'R', 0, 1));
3049        if ($choice =~ /^t$/i) {
3050                $cred->may_save(undef);
3051        } elsif ($choice =~ /^r$/i) {
3052                return -1;
3053        } elsif ($may_save && $choice =~ /^p$/i) {
3054                $cred->may_save($may_save);
3055        } else {
3056                goto prompt;
3057        }
3058        $cred->accepted_failures($failures);
3059        $SVN::_Core::SVN_NO_ERROR;
3060}
3061
3062sub ssl_client_cert {
3063        my ($cred, $realm, $may_save, $pool) = @_;
3064        $may_save = undef if $_no_auth_cache;
3065        print STDERR "Client certificate filename: ";
3066        STDERR->flush;
3067        chomp(my $filename = <STDIN>);
3068        $cred->cert_file($filename);
3069        $cred->may_save($may_save);
3070        $SVN::_Core::SVN_NO_ERROR;
3071}
3072
3073sub ssl_client_cert_pw {
3074        my ($cred, $realm, $may_save, $pool) = @_;
3075        $may_save = undef if $_no_auth_cache;
3076        $cred->password(_read_password("Password: ", $realm));
3077        $cred->may_save($may_save);
3078        $SVN::_Core::SVN_NO_ERROR;
3079}
3080
3081sub username {
3082        my ($cred, $realm, $may_save, $pool) = @_;
3083        $may_save = undef if $_no_auth_cache;
3084        if (defined $realm && length $realm) {
3085                print STDERR "Authentication realm: $realm\n";
3086        }
3087        my $username;
3088        if (defined $_username) {
3089                $username = $_username;
3090        } else {
3091                print STDERR "Username: ";
3092                STDERR->flush;
3093                chomp($username = <STDIN>);
3094        }
3095        $cred->username($username);
3096        $cred->may_save($may_save);
3097        $SVN::_Core::SVN_NO_ERROR;
3098}
3099
3100sub _read_password {
3101        my ($prompt, $realm) = @_;
3102        print STDERR $prompt;
3103        STDERR->flush;
3104        require Term::ReadKey;
3105        Term::ReadKey::ReadMode('noecho');
3106        my $password = '';
3107        while (defined(my $key = Term::ReadKey::ReadKey(0))) {
3108                last if $key =~ /[\012\015]/; # \n\r
3109                $password .= $key;
3110        }
3111        Term::ReadKey::ReadMode('restore');
3112        print STDERR "\n";
3113        STDERR->flush;
3114        $password;
3115}
3116
3117package SVN::Git::Fetcher;
3118use vars qw/@ISA/;
3119use strict;
3120use warnings;
3121use Carp qw/croak/;
3122use File::Temp qw/tempfile/;
3123use IO::File qw//;
3124
3125# file baton members: path, mode_a, mode_b, pool, fh, blob, base
3126sub new {
3127        my ($class, $git_svn) = @_;
3128        my $self = SVN::Delta::Editor->new;
3129        bless $self, $class;
3130        $self->{c} = $git_svn->{last_commit} if exists $git_svn->{last_commit};
3131        $self->{empty} = {};
3132        $self->{dir_prop} = {};
3133        $self->{file_prop} = {};
3134        $self->{absent_dir} = {};
3135        $self->{absent_file} = {};
3136        $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
3137        $self;
3138}
3139
3140sub set_path_strip {
3141        my ($self, $path) = @_;
3142        $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
3143}
3144
3145sub open_root {
3146        { path => '' };
3147}
3148
3149sub open_directory {
3150        my ($self, $path, $pb, $rev) = @_;
3151        { path => $path };
3152}
3153
3154sub git_path {
3155        my ($self, $path) = @_;
3156        if ($self->{path_strip}) {
3157                $path =~ s!$self->{path_strip}!! or
3158                  die "Failed to strip path '$path' ($self->{path_strip})\n";
3159        }
3160        $path;
3161}
3162
3163sub delete_entry {
3164        my ($self, $path, $rev, $pb) = @_;
3165
3166        my $gpath = $self->git_path($path);
3167        return undef if ($gpath eq '');
3168
3169        # remove entire directories.
3170        if (command('ls-tree', $self->{c}, '--', $gpath) =~ /^040000 tree/) {
3171                my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3172                                                     -r --name-only -z/,
3173                                                     $self->{c}, '--', $gpath);
3174                local $/ = "\0";
3175                while (<$ls>) {
3176                        chomp;
3177                        $self->{gii}->remove($_);
3178                        print "\tD\t$_\n" unless $::_q;
3179                }
3180                print "\tD\t$gpath/\n" unless $::_q;
3181                command_close_pipe($ls, $ctx);
3182                $self->{empty}->{$path} = 0
3183        } else {
3184                $self->{gii}->remove($gpath);
3185                print "\tD\t$gpath\n" unless $::_q;
3186        }
3187        undef;
3188}
3189
3190sub open_file {
3191        my ($self, $path, $pb, $rev) = @_;
3192        my $gpath = $self->git_path($path);
3193        my ($mode, $blob) = (command('ls-tree', $self->{c}, '--', $gpath)
3194                             =~ /^(\d{6}) blob ([a-f\d]{40})\t/);
3195        unless (defined $mode && defined $blob) {
3196                die "$path was not found in commit $self->{c} (r$rev)\n";
3197        }
3198        { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3199          pool => SVN::Pool->new, action => 'M' };
3200}
3201
3202sub add_file {
3203        my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3204        my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3205        delete $self->{empty}->{$dir};
3206        { path => $path, mode_a => 100644, mode_b => 100644,
3207          pool => SVN::Pool->new, action => 'A' };
3208}
3209
3210sub add_directory {
3211        my ($self, $path, $cp_path, $cp_rev) = @_;
3212        my $gpath = $self->git_path($path);
3213        if ($gpath eq '') {
3214                my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3215                                                     -r --name-only -z/,
3216                                                     $self->{c});
3217                local $/ = "\0";
3218                while (<$ls>) {
3219                        chomp;
3220                        $self->{gii}->remove($_);
3221                        print "\tD\t$_\n" unless $::_q;
3222                }
3223                command_close_pipe($ls, $ctx);
3224                $self->{empty}->{$path} = 0;
3225        }
3226        my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3227        delete $self->{empty}->{$dir};
3228        $self->{empty}->{$path} = 1;
3229        { path => $path };
3230}
3231
3232sub change_dir_prop {
3233        my ($self, $db, $prop, $value) = @_;
3234        $self->{dir_prop}->{$db->{path}} ||= {};
3235        $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3236        undef;
3237}
3238
3239sub absent_directory {
3240        my ($self, $path, $pb) = @_;
3241        $self->{absent_dir}->{$pb->{path}} ||= [];
3242        push @{$self->{absent_dir}->{$pb->{path}}}, $path;
3243        undef;
3244}
3245
3246sub absent_file {
3247        my ($self, $path, $pb) = @_;
3248        $self->{absent_file}->{$pb->{path}} ||= [];
3249        push @{$self->{absent_file}->{$pb->{path}}}, $path;
3250        undef;
3251}
3252
3253sub change_file_prop {
3254        my ($self, $fb, $prop, $value) = @_;
3255        if ($prop eq 'svn:executable') {
3256                if ($fb->{mode_b} != 120000) {
3257                        $fb->{mode_b} = defined $value ? 100755 : 100644;
3258                }
3259        } elsif ($prop eq 'svn:special') {
3260                $fb->{mode_b} = defined $value ? 120000 : 100644;
3261        } else {
3262                $self->{file_prop}->{$fb->{path}} ||= {};
3263                $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
3264        }
3265        undef;
3266}
3267
3268sub apply_textdelta {
3269        my ($self, $fb, $exp) = @_;
3270        my $fh = Git::temp_acquire('svn_delta');
3271        # $fh gets auto-closed() by SVN::TxDelta::apply(),
3272        # (but $base does not,) so dup() it for reading in close_file
3273        open my $dup, '<&', $fh or croak $!;
3274        my $base = Git::temp_acquire('git_blob');
3275        if ($fb->{blob}) {
3276                print $base 'link ' if ($fb->{mode_a} == 120000);
3277                my $size = $::_repository->cat_blob($fb->{blob}, $base);
3278                die "Failed to read object $fb->{blob}" if ($size < 0);
3279
3280                if (defined $exp) {
3281                        seek $base, 0, 0 or croak $!;
3282                        my $got = ::md5sum($base);
3283                        die "Checksum mismatch: $fb->{path} $fb->{blob}\n",
3284                            "expected: $exp\n",
3285                            "     got: $got\n" if ($got ne $exp);
3286                }
3287        }
3288        seek $base, 0, 0 or croak $!;
3289        $fb->{fh} = $fh;
3290        $fb->{base} = $base;
3291        [ SVN::TxDelta::apply($base, $dup, undef, $fb->{path}, $fb->{pool}) ];
3292}
3293
3294sub close_file {
3295        my ($self, $fb, $exp) = @_;
3296        my $hash;
3297        my $path = $self->git_path($fb->{path});
3298        if (my $fh = $fb->{fh}) {
3299                if (defined $exp) {
3300                        seek($fh, 0, 0) or croak $!;
3301                        my $got = ::md5sum($fh);
3302                        if ($got ne $exp) {
3303                                die "Checksum mismatch: $path\n",
3304                                    "expected: $exp\n    got: $got\n";
3305                        }
3306                }
3307                if ($fb->{mode_b} == 120000) {
3308                        sysseek($fh, 0, 0) or croak $!;
3309                        sysread($fh, my $buf, 5) == 5 or croak $!;
3310
3311                        unless ($buf eq 'link ') {
3312                                warn "$path has mode 120000",
3313                                                " but is not a link\n";
3314                        } else {
3315                                my $tmp_fh = Git::temp_acquire('svn_hash');
3316                                my $res;
3317                                while ($res = sysread($fh, my $str, 1024)) {
3318                                        my $out = syswrite($tmp_fh, $str, $res);
3319                                        defined($out) && $out == $res
3320                                                or croak("write ",
3321                                                        Git::temp_path($tmp_fh),
3322                                                        ": $!\n");
3323                                }
3324                                defined $res or croak $!;
3325
3326                                ($fh, $tmp_fh) = ($tmp_fh, $fh);
3327                                Git::temp_release($tmp_fh, 1);
3328                        }
3329                }
3330
3331                $hash = $::_repository->hash_and_insert_object(
3332                                Git::temp_path($fh));
3333                $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
3334
3335                Git::temp_release($fb->{base}, 1);
3336                Git::temp_release($fh, 1);
3337        } else {
3338                $hash = $fb->{blob} or die "no blob information\n";
3339        }
3340        $fb->{pool}->clear;
3341        $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
3342        print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
3343        undef;
3344}
3345
3346sub abort_edit {
3347        my $self = shift;
3348        $self->{nr} = $self->{gii}->{nr};
3349        delete $self->{gii};
3350        $self->SUPER::abort_edit(@_);
3351}
3352
3353sub close_edit {
3354        my $self = shift;
3355        $self->{git_commit_ok} = 1;
3356        $self->{nr} = $self->{gii}->{nr};
3357        delete $self->{gii};
3358        $self->SUPER::close_edit(@_);
3359}
3360
3361package SVN::Git::Editor;
3362use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
3363use strict;
3364use warnings;
3365use Carp qw/croak/;
3366use IO::File;
3367
3368sub new {
3369        my ($class, $opts) = @_;
3370        foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
3371                die "$_ required!\n" unless (defined $opts->{$_});
3372        }
3373
3374        my $pool = SVN::Pool->new;
3375        my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
3376        my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
3377                                     $opts->{r}, $mods);
3378
3379        # $opts->{ra} functions should not be used after this:
3380        my @ce  = $opts->{ra}->get_commit_editor($opts->{log},
3381                                                $opts->{editor_cb}, $pool);
3382        my $self = SVN::Delta::Editor->new(@ce, $pool);
3383        bless $self, $class;
3384        foreach (qw/svn_path r tree_a tree_b/) {
3385                $self->{$_} = $opts->{$_};
3386        }
3387        $self->{url} = $opts->{ra}->{url};
3388        $self->{mods} = $mods;
3389        $self->{types} = $types;
3390        $self->{pool} = $pool;
3391        $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
3392        $self->{rm} = { };
3393        $self->{path_prefix} = length $self->{svn_path} ?
3394                               "$self->{svn_path}/" : '';
3395        $self->{config} = $opts->{config};
3396        return $self;
3397}
3398
3399sub generate_diff {
3400        my ($tree_a, $tree_b) = @_;
3401        my @diff_tree = qw(diff-tree -z -r);
3402        if ($_cp_similarity) {
3403                push @diff_tree, "-C$_cp_similarity";
3404        } else {
3405                push @diff_tree, '-C';
3406        }
3407        push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
3408        push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
3409        push @diff_tree, $tree_a, $tree_b;
3410        my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
3411        local $/ = "\0";
3412        my $state = 'meta';
3413        my @mods;
3414        while (<$diff_fh>) {
3415                chomp $_; # this gets rid of the trailing "\0"
3416                if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
3417                                        ($::sha1)\s($::sha1)\s
3418                                        ([MTCRAD])\d*$/xo) {
3419                        push @mods, {   mode_a => $1, mode_b => $2,
3420                                        sha1_a => $3, sha1_b => $4,
3421                                        chg => $5 };
3422                        if ($5 =~ /^(?:C|R)$/) {
3423                                $state = 'file_a';
3424                        } else {
3425                                $state = 'file_b';
3426                        }
3427                } elsif ($state eq 'file_a') {
3428                        my $x = $mods[$#mods] or croak "Empty array\n";
3429                        if ($x->{chg} !~ /^(?:C|R)$/) {
3430                                croak "Error parsing $_, $x->{chg}\n";
3431                        }
3432                        $x->{file_a} = $_;
3433                        $state = 'file_b';
3434                } elsif ($state eq 'file_b') {
3435                        my $x = $mods[$#mods] or croak "Empty array\n";
3436                        if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
3437                                croak "Error parsing $_, $x->{chg}\n";
3438                        }
3439                        if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
3440                                croak "Error parsing $_, $x->{chg}\n";
3441                        }
3442                        $x->{file_b} = $_;
3443                        $state = 'meta';
3444                } else {
3445                        croak "Error parsing $_\n";
3446                }
3447        }
3448        command_close_pipe($diff_fh, $ctx);
3449        \@mods;
3450}
3451
3452sub check_diff_paths {
3453        my ($ra, $pfx, $rev, $mods) = @_;
3454        my %types;
3455        $pfx .= '/' if length $pfx;
3456
3457        sub type_diff_paths {
3458                my ($ra, $types, $path, $rev) = @_;
3459                my @p = split m#/+#, $path;
3460                my $c = shift @p;
3461                unless (defined $types->{$c}) {
3462                        $types->{$c} = $ra->check_path($c, $rev);
3463                }
3464                while (@p) {
3465                        $c .= '/' . shift @p;
3466                        next if defined $types->{$c};
3467                        $types->{$c} = $ra->check_path($c, $rev);
3468                }
3469        }
3470
3471        foreach my $m (@$mods) {
3472                foreach my $f (qw/file_a file_b/) {
3473                        next unless defined $m->{$f};
3474                        my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
3475                        if (length $pfx.$dir && ! defined $types{$dir}) {
3476                                type_diff_paths($ra, \%types, $pfx.$dir, $rev);
3477                        }
3478                }
3479        }
3480        \%types;
3481}
3482
3483sub split_path {
3484        return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
3485}
3486
3487sub repo_path {
3488        my ($self, $path) = @_;
3489        $self->{path_prefix}.(defined $path ? $path : '');
3490}
3491
3492sub url_path {
3493        my ($self, $path) = @_;
3494        if ($self->{url} =~ m#^https?://#) {
3495                $path =~ s/([^a-zA-Z0-9_.-])/uc sprintf("%%%02x",ord($1))/eg;
3496        }
3497        $self->{url} . '/' . $self->repo_path($path);
3498}
3499
3500sub rmdirs {
3501        my ($self) = @_;
3502        my $rm = $self->{rm};
3503        delete $rm->{''}; # we never delete the url we're tracking
3504        return unless %$rm;
3505
3506        foreach (keys %$rm) {
3507                my @d = split m#/#, $_;
3508                my $c = shift @d;
3509                $rm->{$c} = 1;
3510                while (@d) {
3511                        $c .= '/' . shift @d;
3512                        $rm->{$c} = 1;
3513                }
3514        }
3515        delete $rm->{$self->{svn_path}};
3516        delete $rm->{''}; # we never delete the url we're tracking
3517        return unless %$rm;
3518
3519        my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
3520                                             $self->{tree_b});
3521        local $/ = "\0";
3522        while (<$fh>) {
3523                chomp;
3524                my @dn = split m#/#, $_;
3525                while (pop @dn) {
3526                        delete $rm->{join '/', @dn};
3527                }
3528                unless (%$rm) {
3529                        close $fh;
3530                        return;
3531                }
3532        }
3533        command_close_pipe($fh, $ctx);
3534
3535        my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
3536        foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
3537                $self->close_directory($bat->{$d}, $p);
3538                my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
3539                print "\tD+\t$d/\n" unless $::_q;
3540                $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
3541                delete $bat->{$d};
3542        }
3543}
3544
3545sub open_or_add_dir {
3546        my ($self, $full_path, $baton) = @_;
3547        my $t = $self->{types}->{$full_path};
3548        if (!defined $t) {
3549                die "$full_path not known in r$self->{r} or we have a bug!\n";
3550        }
3551        {
3552                no warnings 'once';
3553                # SVN::Node::none and SVN::Node::file are used only once,
3554                # so we're shutting up Perl's warnings about them.
3555                if ($t == $SVN::Node::none) {
3556                        return $self->add_directory($full_path, $baton,
3557                            undef, -1, $self->{pool});
3558                } elsif ($t == $SVN::Node::dir) {
3559                        return $self->open_directory($full_path, $baton,
3560                            $self->{r}, $self->{pool});
3561                } # no warnings 'once'
3562                print STDERR "$full_path already exists in repository at ",
3563                    "r$self->{r} and it is not a directory (",
3564                    ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
3565        } # no warnings 'once'
3566        exit 1;
3567}
3568
3569sub ensure_path {
3570        my ($self, $path) = @_;
3571        my $bat = $self->{bat};
3572        my $repo_path = $self->repo_path($path);
3573        return $bat->{''} unless (length $repo_path);
3574        my @p = split m#/+#, $repo_path;
3575        my $c = shift @p;
3576        $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
3577        while (@p) {
3578                my $c0 = $c;
3579                $c .= '/' . shift @p;
3580                $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
3581        }
3582        return $bat->{$c};
3583}
3584
3585# Subroutine to convert a globbing pattern to a regular expression.
3586# From perl cookbook.
3587sub glob2pat {
3588        my $globstr = shift;
3589        my %patmap = ('*' => '.*', '?' => '.', '[' => '[', ']' => ']');
3590        $globstr =~ s{(.)} { $patmap{$1} || "\Q$1" }ge;
3591        return '^' . $globstr . '$';
3592}
3593
3594sub check_autoprop {
3595        my ($self, $pattern, $properties, $file, $fbat) = @_;
3596        # Convert the globbing pattern to a regular expression.
3597        my $regex = glob2pat($pattern);
3598        # Check if the pattern matches the file name.
3599        if($file =~ m/($regex)/) {
3600                # Parse the list of properties to set.
3601                my @props = split(/;/, $properties);
3602                foreach my $prop (@props) {
3603                        # Parse 'name=value' syntax and set the property.
3604                        if ($prop =~ /([^=]+)=(.*)/) {
3605                                my ($n,$v) = ($1,$2);
3606                                for ($n, $v) {
3607                                        s/^\s+//; s/\s+$//;
3608                                }
3609                                $self->change_file_prop($fbat, $n, $v);
3610                        }
3611                }
3612        }
3613}
3614
3615sub apply_autoprops {
3616        my ($self, $file, $fbat) = @_;
3617        my $conf_t = ${$self->{config}}{'config'};
3618        no warnings 'once';
3619        # Check [miscellany]/enable-auto-props in svn configuration.
3620        if (SVN::_Core::svn_config_get_bool(
3621                $conf_t,
3622                $SVN::_Core::SVN_CONFIG_SECTION_MISCELLANY,
3623                $SVN::_Core::SVN_CONFIG_OPTION_ENABLE_AUTO_PROPS,
3624                0)) {
3625                # Auto-props are enabled.  Enumerate them to look for matches.
3626                my $callback = sub {
3627                        $self->check_autoprop($_[0], $_[1], $file, $fbat);
3628                };
3629                SVN::_Core::svn_config_enumerate(
3630                        $conf_t,
3631                        $SVN::_Core::SVN_CONFIG_SECTION_AUTO_PROPS,
3632                        $callback);
3633        }
3634}
3635
3636sub A {
3637        my ($self, $m) = @_;
3638        my ($dir, $file) = split_path($m->{file_b});
3639        my $pbat = $self->ensure_path($dir);
3640        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3641                                        undef, -1);
3642        print "\tA\t$m->{file_b}\n" unless $::_q;
3643        $self->apply_autoprops($file, $fbat);
3644        $self->chg_file($fbat, $m);
3645        $self->close_file($fbat,undef,$self->{pool});
3646}
3647
3648sub C {
3649        my ($self, $m) = @_;
3650        my ($dir, $file) = split_path($m->{file_b});
3651        my $pbat = $self->ensure_path($dir);
3652        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3653                                $self->url_path($m->{file_a}), $self->{r});
3654        print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3655        $self->chg_file($fbat, $m);
3656        $self->close_file($fbat,undef,$self->{pool});
3657}
3658
3659sub delete_entry {
3660        my ($self, $path, $pbat) = @_;
3661        my $rpath = $self->repo_path($path);
3662        my ($dir, $file) = split_path($rpath);
3663        $self->{rm}->{$dir} = 1;
3664        $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
3665}
3666
3667sub R {
3668        my ($self, $m) = @_;
3669        my ($dir, $file) = split_path($m->{file_b});
3670        my $pbat = $self->ensure_path($dir);
3671        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3672                                $self->url_path($m->{file_a}), $self->{r});
3673        print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3674        $self->apply_autoprops($file, $fbat);
3675        $self->chg_file($fbat, $m);
3676        $self->close_file($fbat,undef,$self->{pool});
3677
3678        ($dir, $file) = split_path($m->{file_a});
3679        $pbat = $self->ensure_path($dir);
3680        $self->delete_entry($m->{file_a}, $pbat);
3681}
3682
3683sub M {
3684        my ($self, $m) = @_;
3685        my ($dir, $file) = split_path($m->{file_b});
3686        my $pbat = $self->ensure_path($dir);
3687        my $fbat = $self->open_file($self->repo_path($m->{file_b}),
3688                                $pbat,$self->{r},$self->{pool});
3689        print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
3690        $self->chg_file($fbat, $m);
3691        $self->close_file($fbat,undef,$self->{pool});
3692}
3693
3694sub T { shift->M(@_) }
3695
3696sub change_file_prop {
3697        my ($self, $fbat, $pname, $pval) = @_;
3698        $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
3699}
3700
3701sub _chg_file_get_blob ($$$$) {
3702        my ($self, $fbat, $m, $which) = @_;
3703        my $fh = Git::temp_acquire("git_blob_$which");
3704        if ($m->{"mode_$which"} =~ /^120/) {
3705                print $fh 'link ' or croak $!;
3706                $self->change_file_prop($fbat,'svn:special','*');
3707        } elsif ($m->{mode_a} =~ /^120/ && $m->{"mode_$which"} !~ /^120/) {
3708                $self->change_file_prop($fbat,'svn:special',undef);
3709        }
3710        my $blob = $m->{"sha1_$which"};
3711        return ($fh,) if ($blob =~ /^0{40}$/);
3712        my $size = $::_repository->cat_blob($blob, $fh);
3713        croak "Failed to read object $blob" if ($size < 0);
3714        $fh->flush == 0 or croak $!;
3715        seek $fh, 0, 0 or croak $!;
3716
3717        my $exp = ::md5sum($fh);
3718        seek $fh, 0, 0 or croak $!;
3719        return ($fh, $exp);
3720}
3721
3722sub chg_file {
3723        my ($self, $fbat, $m) = @_;
3724        if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
3725                $self->change_file_prop($fbat,'svn:executable','*');
3726        } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
3727                $self->change_file_prop($fbat,'svn:executable',undef);
3728        }
3729        my ($fh_a, $exp_a) = _chg_file_get_blob $self, $fbat, $m, 'a';
3730        my ($fh_b, $exp_b) = _chg_file_get_blob $self, $fbat, $m, 'b';
3731        my $pool = SVN::Pool->new;
3732        my $atd = $self->apply_textdelta($fbat, $exp_a, $pool);
3733        if (-s $fh_a) {
3734                my $txstream = SVN::TxDelta::new ($fh_a, $fh_b, $pool);
3735                my $res = SVN::TxDelta::send_txstream($txstream, @$atd, $pool);
3736                if (defined $res) {
3737                        die "Unexpected result from send_txstream: $res\n",
3738                            "(SVN::Core::VERSION: $SVN::Core::VERSION)\n";
3739                }
3740        } else {
3741                my $got = SVN::TxDelta::send_stream($fh_b, @$atd, $pool);
3742                die "Checksum mismatch\nexpected: $exp_b\ngot: $got\n"
3743                    if ($got ne $exp_b);
3744        }
3745        Git::temp_release($fh_b, 1);
3746        Git::temp_release($fh_a, 1);
3747        $pool->clear;
3748}
3749
3750sub D {
3751        my ($self, $m) = @_;
3752        my ($dir, $file) = split_path($m->{file_b});
3753        my $pbat = $self->ensure_path($dir);
3754        print "\tD\t$m->{file_b}\n" unless $::_q;
3755        $self->delete_entry($m->{file_b}, $pbat);
3756}
3757
3758sub close_edit {
3759        my ($self) = @_;
3760        my ($p,$bat) = ($self->{pool}, $self->{bat});
3761        foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
3762                next if $_ eq '';
3763                $self->close_directory($bat->{$_}, $p);
3764        }
3765        $self->close_directory($bat->{''}, $p);
3766        $self->SUPER::close_edit($p);
3767        $p->clear;
3768}
3769
3770sub abort_edit {
3771        my ($self) = @_;
3772        $self->SUPER::abort_edit($self->{pool});
3773}
3774
3775sub DESTROY {
3776        my $self = shift;
3777        $self->SUPER::DESTROY(@_);
3778        $self->{pool}->clear;
3779}
3780
3781# this drives the editor
3782sub apply_diff {
3783        my ($self) = @_;
3784        my $mods = $self->{mods};
3785        my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
3786        foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
3787                my $f = $m->{chg};
3788                if (defined $o{$f}) {
3789                        $self->$f($m);
3790                } else {
3791                        fatal("Invalid change type: $f");
3792                }
3793        }
3794        $self->rmdirs if $_rmdir;
3795        if (@$mods == 0) {
3796                $self->abort_edit;
3797        } else {
3798                $self->close_edit;
3799        }
3800        return scalar @$mods;
3801}
3802
3803package Git::SVN::Ra;
3804use vars qw/@ISA $config_dir $_log_window_size/;
3805use strict;
3806use warnings;
3807my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
3808
3809BEGIN {
3810        # enforce temporary pool usage for some simple functions
3811        no strict 'refs';
3812        for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root/) {
3813                my $SUPER = "SUPER::$f";
3814                *$f = sub {
3815                        my $self = shift;
3816                        my $pool = SVN::Pool->new;
3817                        my @ret = $self->$SUPER(@_,$pool);
3818                        $pool->clear;
3819                        wantarray ? @ret : $ret[0];
3820                };
3821        }
3822}
3823
3824sub _auth_providers () {
3825        [
3826          SVN::Client::get_simple_provider(),
3827          SVN::Client::get_ssl_server_trust_file_provider(),
3828          SVN::Client::get_simple_prompt_provider(
3829            \&Git::SVN::Prompt::simple, 2),
3830          SVN::Client::get_ssl_client_cert_file_provider(),
3831          SVN::Client::get_ssl_client_cert_prompt_provider(
3832            \&Git::SVN::Prompt::ssl_client_cert, 2),
3833          SVN::Client::get_ssl_client_cert_pw_file_provider(),
3834          SVN::Client::get_ssl_client_cert_pw_prompt_provider(
3835            \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
3836          SVN::Client::get_username_provider(),
3837          SVN::Client::get_ssl_server_trust_prompt_provider(
3838            \&Git::SVN::Prompt::ssl_server_trust),
3839          SVN::Client::get_username_prompt_provider(
3840            \&Git::SVN::Prompt::username, 2)
3841        ]
3842}
3843
3844sub escape_uri_only {
3845        my ($uri) = @_;
3846        my @tmp;
3847        foreach (split m{/}, $uri) {
3848                s/([^\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
3849                push @tmp, $_;
3850        }
3851        join('/', @tmp);
3852}
3853
3854sub escape_url {
3855        my ($url) = @_;
3856        if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
3857                my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
3858                $url = "$scheme://$domain$uri";
3859        }
3860        $url;
3861}
3862
3863sub new {
3864        my ($class, $url) = @_;
3865        $url =~ s!/+$!!;
3866        return $RA if ($RA && $RA->{url} eq $url);
3867
3868        SVN::_Core::svn_config_ensure($config_dir, undef);
3869        my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
3870        my $config = SVN::Core::config_get_config($config_dir);
3871        $RA = undef;
3872        my $dont_store_passwords = 1;
3873        my $conf_t = ${$config}{'config'};
3874        {
3875                no warnings 'once';
3876                # The usage of $SVN::_Core::SVN_CONFIG_* variables
3877                # produces warnings that variables are used only once.
3878                # I had not found the better way to shut them up, so
3879                # the warnings of type 'once' are disabled in this block.
3880                if (SVN::_Core::svn_config_get_bool($conf_t,
3881                    $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3882                    $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
3883                    1) == 0) {
3884                        SVN::_Core::svn_auth_set_parameter($baton,
3885                            $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
3886                            bless (\$dont_store_passwords, "_p_void"));
3887                }
3888                if (SVN::_Core::svn_config_get_bool($conf_t,
3889                    $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3890                    $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
3891                    1) == 0) {
3892                        $Git::SVN::Prompt::_no_auth_cache = 1;
3893                }
3894        } # no warnings 'once'
3895        my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
3896                              config => $config,
3897                              pool => SVN::Pool->new,
3898                              auth_provider_callbacks => $callbacks);
3899        $self->{url} = $url;
3900        $self->{svn_path} = $url;
3901        $self->{repos_root} = $self->get_repos_root;
3902        $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
3903        $self->{cache} = { check_path => { r => 0, data => {} },
3904                           get_dir => { r => 0, data => {} } };
3905        $RA = bless $self, $class;
3906}
3907
3908sub check_path {
3909        my ($self, $path, $r) = @_;
3910        my $cache = $self->{cache}->{check_path};
3911        if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
3912                return $cache->{data}->{$path};
3913        }
3914        my $pool = SVN::Pool->new;
3915        my $t = $self->SUPER::check_path($path, $r, $pool);
3916        $pool->clear;
3917        if ($r != $cache->{r}) {
3918                %{$cache->{data}} = ();
3919                $cache->{r} = $r;
3920        }
3921        $cache->{data}->{$path} = $t;
3922}
3923
3924sub get_dir {
3925        my ($self, $dir, $r) = @_;
3926        my $cache = $self->{cache}->{get_dir};
3927        if ($r == $cache->{r}) {
3928                if (my $x = $cache->{data}->{$dir}) {
3929                        return wantarray ? @$x : $x->[0];
3930                }
3931        }
3932        my $pool = SVN::Pool->new;
3933        my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
3934        my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
3935        $pool->clear;
3936        if ($r != $cache->{r}) {
3937                %{$cache->{data}} = ();
3938                $cache->{r} = $r;
3939        }
3940        $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
3941        wantarray ? (\%dirents, $r, $props) : \%dirents;
3942}
3943
3944sub DESTROY {
3945        # do not call the real DESTROY since we store ourselves in $RA
3946}
3947
3948sub get_log {
3949        my ($self, @args) = @_;
3950        my $pool = SVN::Pool->new;
3951        splice(@args, 3, 1) if ($SVN::Core::VERSION le '1.2.0');
3952        my $ret = $self->SUPER::get_log(@args, $pool);
3953        $pool->clear;
3954        $ret;
3955}
3956
3957sub trees_match {
3958        my ($self, $url1, $rev1, $url2, $rev2) = @_;
3959        my $ctx = SVN::Client->new(auth => _auth_providers);
3960        my $out = IO::File->new_tmpfile;
3961
3962        # older SVN (1.1.x) doesn't take $pool as the last parameter for
3963        # $ctx->diff(), so we'll create a default one
3964        my $pool = SVN::Pool->new_default_sub;
3965
3966        $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
3967        $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
3968        $out->flush;
3969        my $ret = (($out->stat)[7] == 0);
3970        close $out or croak $!;
3971
3972        $ret;
3973}
3974
3975sub get_commit_editor {
3976        my ($self, $log, $cb, $pool) = @_;
3977        my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
3978        $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
3979}
3980
3981sub gs_do_update {
3982        my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
3983        my $new = ($rev_a == $rev_b);
3984        my $path = $gs->{path};
3985
3986        if ($new && -e $gs->{index}) {
3987                unlink $gs->{index} or die
3988                  "Couldn't unlink index: $gs->{index}: $!\n";
3989        }
3990        my $pool = SVN::Pool->new;
3991        $editor->set_path_strip($path);
3992        my (@pc) = split m#/#, $path;
3993        my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
3994                                        1, $editor, $pool);
3995        my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3996
3997        # Since we can't rely on svn_ra_reparent being available, we'll
3998        # just have to do some magic with set_path to make it so
3999        # we only want a partial path.
4000        my $sp = '';
4001        my $final = join('/', @pc);
4002        while (@pc) {
4003                $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
4004                $sp .= '/' if length $sp;
4005                $sp .= shift @pc;
4006        }
4007        die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
4008
4009        $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
4010
4011        $reporter->finish_report($pool);
4012        $pool->clear;
4013        $editor->{git_commit_ok};
4014}
4015
4016# this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
4017# svn_ra_reparent didn't work before 1.4)
4018sub gs_do_switch {
4019        my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
4020        my $path = $gs->{path};
4021        my $pool = SVN::Pool->new;
4022
4023        my $full_url = $self->{url};
4024        my $old_url = $full_url;
4025        $full_url .= '/' . escape_uri_only($path) if length $path;
4026        my ($ra, $reparented);
4027
4028        if ($old_url =~ m#^svn(\+ssh)?://#) {
4029                $_[0] = undef;
4030                $self = undef;
4031                $RA = undef;
4032                $ra = Git::SVN::Ra->new($full_url);
4033                $ra_invalid = 1;
4034        } elsif ($old_url ne $full_url) {
4035                SVN::_Ra::svn_ra_reparent($self->{session}, $full_url, $pool);
4036                $self->{url} = $full_url;
4037                $reparented = 1;
4038        }
4039
4040        $ra ||= $self;
4041        $url_b = escape_url($url_b);
4042        my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
4043        my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
4044        $reporter->set_path('', $rev_a, 0, @lock, $pool);
4045        $reporter->finish_report($pool);
4046
4047        if ($reparented) {
4048                SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
4049                $self->{url} = $old_url;
4050        }
4051
4052        $pool->clear;
4053        $editor->{git_commit_ok};
4054}
4055
4056sub longest_common_path {
4057        my ($gsv, $globs) = @_;
4058        my %common;
4059        my $common_max = scalar @$gsv;
4060
4061        foreach my $gs (@$gsv) {
4062                my @tmp = split m#/#, $gs->{path};
4063                my $p = '';
4064                foreach (@tmp) {
4065                        $p .= length($p) ? "/$_" : $_;
4066                        $common{$p} ||= 0;
4067                        $common{$p}++;
4068                }
4069        }
4070        $globs ||= [];
4071        $common_max += scalar @$globs;
4072        foreach my $glob (@$globs) {
4073                my @tmp = split m#/#, $glob->{path}->{left};
4074                my $p = '';
4075                foreach (@tmp) {
4076                        $p .= length($p) ? "/$_" : $_;
4077                        $common{$p} ||= 0;
4078                        $common{$p}++;
4079                }
4080        }
4081
4082        my $longest_path = '';
4083        foreach (sort {length $b <=> length $a} keys %common) {
4084                if ($common{$_} == $common_max) {
4085                        $longest_path = $_;
4086                        last;
4087                }
4088        }
4089        $longest_path;
4090}
4091
4092sub gs_fetch_loop_common {
4093        my ($self, $base, $head, $gsv, $globs) = @_;
4094        return if ($base > $head);
4095        my $inc = $_log_window_size;
4096        my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
4097        my $longest_path = longest_common_path($gsv, $globs);
4098        my $ra_url = $self->{url};
4099        while (1) {
4100                my %revs;
4101                my $err;
4102                my $err_handler = $SVN::Error::handler;
4103                $SVN::Error::handler = sub {
4104                        ($err) = @_;
4105                        skip_unknown_revs($err);
4106                };
4107                sub _cb {
4108                        my ($paths, $r, $author, $date, $log) = @_;
4109                        [ dup_changed_paths($paths),
4110                          { author => $author, date => $date, log => $log } ];
4111                }
4112                $self->get_log([$longest_path], $min, $max, 0, 1, 1,
4113                               sub { $revs{$_[1]} = _cb(@_) });
4114                if ($err && $max >= $head) {
4115                        print STDERR "Path '$longest_path' ",
4116                                     "was probably deleted:\n",
4117                                     $err->expanded_message,
4118                                     "\nWill attempt to follow ",
4119                                     "revisions r$min .. r$max ",
4120                                     "committed before the deletion\n";
4121                        my $hi = $max;
4122                        while (--$hi >= $min) {
4123                                my $ok;
4124                                $self->get_log([$longest_path], $min, $hi,
4125                                               0, 1, 1, sub {
4126                                               $ok ||= $_[1];
4127                                               $revs{$_[1]} = _cb(@_) });
4128                                if ($ok) {
4129                                        print STDERR "r$min .. r$ok OK\n";
4130                                        last;
4131                                }
4132                        }
4133                }
4134                $SVN::Error::handler = $err_handler;
4135
4136                my %exists = map { $_->{path} => $_ } @$gsv;
4137                foreach my $r (sort {$a <=> $b} keys %revs) {
4138                        my ($paths, $logged) = @{$revs{$r}};
4139
4140                        foreach my $gs ($self->match_globs(\%exists, $paths,
4141                                                           $globs, $r)) {
4142                                if ($gs->rev_map_max >= $r) {
4143                                        next;
4144                                }
4145                                next unless $gs->match_paths($paths, $r);
4146                                $gs->{logged_rev_props} = $logged;
4147                                if (my $last_commit = $gs->last_commit) {
4148                                        $gs->assert_index_clean($last_commit);
4149                                }
4150                                my $log_entry = $gs->do_fetch($paths, $r);
4151                                if ($log_entry) {
4152                                        $gs->do_git_commit($log_entry);
4153                                }
4154                                $INDEX_FILES{$gs->{index}} = 1;
4155                        }
4156                        foreach my $g (@$globs) {
4157                                my $k = "svn-remote.$g->{remote}." .
4158                                        "$g->{t}-maxRev";
4159                                Git::SVN::tmp_config($k, $r);
4160                        }
4161                        if ($ra_invalid) {
4162                                $_[0] = undef;
4163                                $self = undef;
4164                                $RA = undef;
4165                                $self = Git::SVN::Ra->new($ra_url);
4166                                $ra_invalid = undef;
4167                        }
4168                }
4169                # pre-fill the .rev_db since it'll eventually get filled in
4170                # with '0' x40 if something new gets committed
4171                foreach my $gs (@$gsv) {
4172                        next if $gs->rev_map_max >= $max;
4173                        next if defined $gs->rev_map_get($max);
4174                        $gs->rev_map_set($max, 0 x40);
4175                }
4176                foreach my $g (@$globs) {
4177                        my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
4178                        Git::SVN::tmp_config($k, $max);
4179                }
4180                last if $max >= $head;
4181                $min = $max + 1;
4182                $max += $inc;
4183                $max = $head if ($max > $head);
4184        }
4185        Git::SVN::gc();
4186}
4187
4188sub get_dir_globbed {
4189        my ($self, $left, $depth, $r) = @_;
4190
4191        my @x = eval { $self->get_dir($left, $r) };
4192        return unless scalar @x == 3;
4193        my $dirents = $x[0];
4194        my @finalents;
4195        foreach my $de (keys %$dirents) {
4196                next if $dirents->{$de}->{kind} != $SVN::Node::dir;
4197                if ($depth > 1) {
4198                        my @args = ("$left/$de", $depth - 1, $r);
4199                        foreach my $dir ($self->get_dir_globbed(@args)) {
4200                                push @finalents, "$de/$dir";
4201                        }
4202                } else {
4203                        push @finalents, $de;
4204                }
4205        }
4206        @finalents;
4207}
4208
4209sub match_globs {
4210        my ($self, $exists, $paths, $globs, $r) = @_;
4211
4212        sub get_dir_check {
4213                my ($self, $exists, $g, $r) = @_;
4214
4215                my @dirs = $self->get_dir_globbed($g->{path}->{left},
4216                                                  $g->{path}->{depth},
4217                                                  $r);
4218
4219                foreach my $de (@dirs) {
4220                        my $p = $g->{path}->full_path($de);
4221                        next if $exists->{$p};
4222                        next if (length $g->{path}->{right} &&
4223                                 ($self->check_path($p, $r) !=
4224                                  $SVN::Node::dir));
4225                        $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
4226                                         $g->{ref}->full_path($de), 1);
4227                }
4228        }
4229        foreach my $g (@$globs) {
4230                if (my $path = $paths->{"/$g->{path}->{left}"}) {
4231                        if ($path->{action} =~ /^[AR]$/) {
4232                                get_dir_check($self, $exists, $g, $r);
4233                        }
4234                }
4235                foreach (keys %$paths) {
4236                        if (/$g->{path}->{left_regex}/ &&
4237                            !/$g->{path}->{regex}/) {
4238                                next if $paths->{$_}->{action} !~ /^[AR]$/;
4239                                get_dir_check($self, $exists, $g, $r);
4240                        }
4241                        next unless /$g->{path}->{regex}/;
4242                        my $p = $1;
4243                        my $pathname = $g->{path}->full_path($p);
4244                        next if $exists->{$pathname};
4245                        next if ($self->check_path($pathname, $r) !=
4246                                 $SVN::Node::dir);
4247                        $exists->{$pathname} = Git::SVN->init(
4248                                              $self->{url}, $pathname, undef,
4249                                              $g->{ref}->full_path($p), 1);
4250                }
4251                my $c = '';
4252                foreach (split m#/#, $g->{path}->{left}) {
4253                        $c .= "/$_";
4254                        next unless ($paths->{$c} &&
4255                                     ($paths->{$c}->{action} =~ /^[AR]$/));
4256                        get_dir_check($self, $exists, $g, $r);
4257                }
4258        }
4259        values %$exists;
4260}
4261
4262sub minimize_url {
4263        my ($self) = @_;
4264        return $self->{url} if ($self->{url} eq $self->{repos_root});
4265        my $url = $self->{repos_root};
4266        my @components = split(m!/!, $self->{svn_path});
4267        my $c = '';
4268        do {
4269                $url .= "/$c" if length $c;
4270                eval { (ref $self)->new($url)->get_latest_revnum };
4271        } while ($@ && ($c = shift @components));
4272        $url;
4273}
4274
4275sub can_do_switch {
4276        my $self = shift;
4277        unless (defined $can_do_switch) {
4278                my $pool = SVN::Pool->new;
4279                my $rep = eval {
4280                        $self->do_switch(1, '', 0, $self->{url},
4281                                         SVN::Delta::Editor->new, $pool);
4282                };
4283                if ($@) {
4284                        $can_do_switch = 0;
4285                } else {
4286                        $rep->abort_report($pool);
4287                        $can_do_switch = 1;
4288                }
4289                $pool->clear;
4290        }
4291        $can_do_switch;
4292}
4293
4294sub skip_unknown_revs {
4295        my ($err) = @_;
4296        my $errno = $err->apr_err();
4297        # Maybe the branch we're tracking didn't
4298        # exist when the repo started, so it's
4299        # not an error if it doesn't, just continue
4300        #
4301        # Wonderfully consistent library, eh?
4302        # 160013 - svn:// and file://
4303        # 175002 - http(s)://
4304        # 175007 - http(s):// (this repo required authorization, too...)
4305        #   More codes may be discovered later...
4306        if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
4307                my $err_key = $err->expanded_message;
4308                # revision numbers change every time, filter them out
4309                $err_key =~ s/\d+/\0/g;
4310                $err_key = "$errno\0$err_key";
4311                unless ($ignored_err{$err_key}) {
4312                        warn "W: Ignoring error from SVN, path probably ",
4313                             "does not exist: ($errno): ",
4314                             $err->expanded_message,"\n";
4315                        warn "W: Do not be alarmed at the above message ",
4316                             "git-svn is just searching aggressively for ",
4317                             "old history.\n",
4318                             "This may take a while on large repositories\n";
4319                        $ignored_err{$err_key} = 1;
4320                }
4321                return;
4322        }
4323        die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
4324}
4325
4326# svn_log_changed_path_t objects passed to get_log are likely to be
4327# overwritten even if only the refs are copied to an external variable,
4328# so we should dup the structures in their entirety.  Using an externally
4329# passed pool (instead of our temporary and quickly cleared pool in
4330# Git::SVN::Ra) does not help matters at all...
4331sub dup_changed_paths {
4332        my ($paths) = @_;
4333        return undef unless $paths;
4334        my %ret;
4335        foreach my $p (keys %$paths) {
4336                my $i = $paths->{$p};
4337                my %s = map { $_ => $i->$_ }
4338                              qw/copyfrom_path copyfrom_rev action/;
4339                $ret{$p} = \%s;
4340        }
4341        \%ret;
4342}
4343
4344package Git::SVN::Log;
4345use strict;
4346use warnings;
4347use POSIX qw/strftime/;
4348use constant commit_log_separator => ('-' x 72) . "\n";
4349use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
4350            %rusers $show_commit $incremental/;
4351my $l_fmt;
4352
4353sub cmt_showable {
4354        my ($c) = @_;
4355        return 1 if defined $c->{r};
4356
4357        # big commit message got truncated by the 16k pretty buffer in rev-list
4358        if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
4359                                $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
4360                @{$c->{l}} = ();
4361                my @log = command(qw/cat-file commit/, $c->{c});
4362
4363                # shift off the headers
4364                shift @log while ($log[0] ne '');
4365                shift @log;
4366
4367                # TODO: make $c->{l} not have a trailing newline in the future
4368                @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
4369
4370                (undef, $c->{r}, undef) = ::extract_metadata(
4371                                (grep(/^git-svn-id: /, @log))[-1]);
4372        }
4373        return defined $c->{r};
4374}
4375
4376sub log_use_color {
4377        return $color || Git->repository->get_colorbool('color.diff');
4378}
4379
4380sub git_svn_log_cmd {
4381        my ($r_min, $r_max, @args) = @_;
4382        my $head = 'HEAD';
4383        my (@files, @log_opts);
4384        foreach my $x (@args) {
4385                if ($x eq '--' || @files) {
4386                        push @files, $x;
4387                } else {
4388                        if (::verify_ref("$x^0")) {
4389                                $head = $x;
4390                        } else {
4391                                push @log_opts, $x;
4392                        }
4393                }
4394        }
4395
4396        my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
4397        $gs ||= Git::SVN->_new;
4398        my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
4399                   $gs->refname);
4400        push @cmd, '-r' unless $non_recursive;
4401        push @cmd, qw/--raw --name-status/ if $verbose;
4402        push @cmd, '--color' if log_use_color();
4403        push @cmd, @log_opts;
4404        if (defined $r_max && $r_max == $r_min) {
4405                push @cmd, '--max-count=1';
4406                if (my $c = $gs->rev_map_get($r_max)) {
4407                        push @cmd, $c;
4408                }
4409        } elsif (defined $r_max) {
4410                if ($r_max < $r_min) {
4411                        ($r_min, $r_max) = ($r_max, $r_min);
4412                }
4413                my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
4414                my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
4415                # If there are no commits in the range, both $c_max and $c_min
4416                # will be undefined.  If there is at least 1 commit in the
4417                # range, both will be defined.
4418                return () if !defined $c_min || !defined $c_max;
4419                if ($c_min eq $c_max) {
4420                        push @cmd, '--max-count=1', $c_min;
4421                } else {
4422                        push @cmd, '--boundary', "$c_min..$c_max";
4423                }
4424        }
4425        return (@cmd, @files);
4426}
4427
4428# adapted from pager.c
4429sub config_pager {
4430        $pager ||= $ENV{GIT_PAGER} || $ENV{PAGER};
4431        if (!defined $pager) {
4432                $pager = 'less';
4433        } elsif (length $pager == 0 || $pager eq 'cat') {
4434                $pager = undef;
4435        }
4436        $ENV{GIT_PAGER_IN_USE} = defined($pager);
4437}
4438
4439sub run_pager {
4440        return unless -t *STDOUT && defined $pager;
4441        pipe my ($rfd, $wfd) or return;
4442        defined(my $pid = fork) or ::fatal "Can't fork: $!";
4443        if (!$pid) {
4444                open STDOUT, '>&', $wfd or
4445                                     ::fatal "Can't redirect to stdout: $!";
4446                return;
4447        }
4448        open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
4449        $ENV{LESS} ||= 'FRSX';
4450        exec $pager or ::fatal "Can't run pager: $! ($pager)";
4451}
4452
4453sub format_svn_date {
4454        return strftime("%Y-%m-%d %H:%M:%S %z (%a, %d %b %Y)", localtime(shift));
4455}
4456
4457sub parse_git_date {
4458        my ($t, $tz) = @_;
4459        # Date::Parse isn't in the standard Perl distro :(
4460        if ($tz =~ s/^\+//) {
4461                $t += tz_to_s_offset($tz);
4462        } elsif ($tz =~ s/^\-//) {
4463                $t -= tz_to_s_offset($tz);
4464        }
4465        return $t;
4466}
4467
4468sub set_local_timezone {
4469        if (defined $TZ) {
4470                $ENV{TZ} = $TZ;
4471        } else {
4472                delete $ENV{TZ};
4473        }
4474}
4475
4476sub tz_to_s_offset {
4477        my ($tz) = @_;
4478        $tz =~ s/(\d\d)$//;
4479        return ($1 * 60) + ($tz * 3600);
4480}
4481
4482sub get_author_info {
4483        my ($dest, $author, $t, $tz) = @_;
4484        $author =~ s/(?:^\s*|\s*$)//g;
4485        $dest->{a_raw} = $author;
4486        my $au;
4487        if ($::_authors) {
4488                $au = $rusers{$author} || undef;
4489        }
4490        if (!$au) {
4491                ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
4492        }
4493        $dest->{t} = $t;
4494        $dest->{tz} = $tz;
4495        $dest->{a} = $au;
4496        $dest->{t_utc} = parse_git_date($t, $tz);
4497}
4498
4499sub process_commit {
4500        my ($c, $r_min, $r_max, $defer) = @_;
4501        if (defined $r_min && defined $r_max) {
4502                if ($r_min == $c->{r} && $r_min == $r_max) {
4503                        show_commit($c);
4504                        return 0;
4505                }
4506                return 1 if $r_min == $r_max;
4507                if ($r_min < $r_max) {
4508                        # we need to reverse the print order
4509                        return 0 if (defined $limit && --$limit < 0);
4510                        push @$defer, $c;
4511                        return 1;
4512                }
4513                if ($r_min != $r_max) {
4514                        return 1 if ($r_min < $c->{r});
4515                        return 1 if ($r_max > $c->{r});
4516                }
4517        }
4518        return 0 if (defined $limit && --$limit < 0);
4519        show_commit($c);
4520        return 1;
4521}
4522
4523sub show_commit {
4524        my $c = shift;
4525        if ($oneline) {
4526                my $x = "\n";
4527                if (my $l = $c->{l}) {
4528                        while ($l->[0] =~ /^\s*$/) { shift @$l }
4529                        $x = $l->[0];
4530                }
4531                $l_fmt ||= 'A' . length($c->{r});
4532                print 'r',pack($l_fmt, $c->{r}),' | ';
4533                print "$c->{c} | " if $show_commit;
4534                print $x;
4535        } else {
4536                show_commit_normal($c);
4537        }
4538}
4539
4540sub show_commit_changed_paths {
4541        my ($c) = @_;
4542        return unless $c->{changed};
4543        print "Changed paths:\n", @{$c->{changed}};
4544}
4545
4546sub show_commit_normal {
4547        my ($c) = @_;
4548        print commit_log_separator, "r$c->{r} | ";
4549        print "$c->{c} | " if $show_commit;
4550        print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
4551        my $nr_line = 0;
4552
4553        if (my $l = $c->{l}) {
4554                while ($l->[$#$l] eq "\n" && $#$l > 0
4555                                          && $l->[($#$l - 1)] eq "\n") {
4556                        pop @$l;
4557                }
4558                $nr_line = scalar @$l;
4559                if (!$nr_line) {
4560                        print "1 line\n\n\n";
4561                } else {
4562                        if ($nr_line == 1) {
4563                                $nr_line = '1 line';
4564                        } else {
4565                                $nr_line .= ' lines';
4566                        }
4567                        print $nr_line, "\n";
4568                        show_commit_changed_paths($c);
4569                        print "\n";
4570                        print $_ foreach @$l;
4571                }
4572        } else {
4573                print "1 line\n";
4574                show_commit_changed_paths($c);
4575                print "\n";
4576
4577        }
4578        foreach my $x (qw/raw stat diff/) {
4579                if ($c->{$x}) {
4580                        print "\n";
4581                        print $_ foreach @{$c->{$x}}
4582                }
4583        }
4584}
4585
4586sub cmd_show_log {
4587        my (@args) = @_;
4588        my ($r_min, $r_max);
4589        my $r_last = -1; # prevent dupes
4590        set_local_timezone();
4591        if (defined $::_revision) {
4592                if ($::_revision =~ /^(\d+):(\d+)$/) {
4593                        ($r_min, $r_max) = ($1, $2);
4594                } elsif ($::_revision =~ /^\d+$/) {
4595                        $r_min = $r_max = $::_revision;
4596                } else {
4597                        ::fatal "-r$::_revision is not supported, use ",
4598                                "standard 'git log' arguments instead";
4599                }
4600        }
4601
4602        config_pager();
4603        @args = git_svn_log_cmd($r_min, $r_max, @args);
4604        if (!@args) {
4605                print commit_log_separator unless $incremental || $oneline;
4606                return;
4607        }
4608        my $log = command_output_pipe(@args);
4609        run_pager();
4610        my (@k, $c, $d, $stat);
4611        my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
4612        while (<$log>) {
4613                if (/^${esc_color}commit -?($::sha1_short)/o) {
4614                        my $cmt = $1;
4615                        if ($c && cmt_showable($c) && $c->{r} != $r_last) {
4616                                $r_last = $c->{r};
4617                                process_commit($c, $r_min, $r_max, \@k) or
4618                                                                goto out;
4619                        }
4620                        $d = undef;
4621                        $c = { c => $cmt };
4622                } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
4623                        get_author_info($c, $1, $2, $3);
4624                } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
4625                        # ignore
4626                } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
4627                        push @{$c->{raw}}, $_;
4628                } elsif (/^${esc_color}[ACRMDT]\t/) {
4629                        # we could add $SVN->{svn_path} here, but that requires
4630                        # remote access at the moment (repo_path_split)...
4631                        s#^(${esc_color})([ACRMDT])\t#$1   $2 #o;
4632                        push @{$c->{changed}}, $_;
4633                } elsif (/^${esc_color}diff /o) {
4634                        $d = 1;
4635                        push @{$c->{diff}}, $_;
4636                } elsif ($d) {
4637                        push @{$c->{diff}}, $_;
4638                } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
4639                          $esc_color*[\+\-]*$esc_color$/x) {
4640                        $stat = 1;
4641                        push @{$c->{stat}}, $_;
4642                } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
4643                        push @{$c->{stat}}, $_;
4644                        $stat = undef;
4645                } elsif (/^${esc_color}    (git-svn-id:.+)$/o) {
4646                        ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
4647                } elsif (s/^${esc_color}    //o) {
4648                        push @{$c->{l}}, $_;
4649                }
4650        }
4651        if ($c && defined $c->{r} && $c->{r} != $r_last) {
4652                $r_last = $c->{r};
4653                process_commit($c, $r_min, $r_max, \@k);
4654        }
4655        if (@k) {
4656                ($r_min, $r_max) = ($r_max, $r_min);
4657                process_commit($_, $r_min, $r_max) foreach reverse @k;
4658        }
4659out:
4660        close $log;
4661        print commit_log_separator unless $incremental || $oneline;
4662}
4663
4664sub cmd_blame {
4665        my $path = pop;
4666
4667        config_pager();
4668        run_pager();
4669
4670        my ($fh, $ctx, $rev);
4671
4672        if ($_git_format) {
4673                ($fh, $ctx) = command_output_pipe('blame', @_, $path);
4674                while (my $line = <$fh>) {
4675                        if ($line =~ /^\^?([[:xdigit:]]+)\s/) {
4676                                # Uncommitted edits show up as a rev ID of
4677                                # all zeros, which we can't look up with
4678                                # cmt_metadata
4679                                if ($1 !~ /^0+$/) {
4680                                        (undef, $rev, undef) =
4681                                                ::cmt_metadata($1);
4682                                        $rev = '0' if (!$rev);
4683                                } else {
4684                                        $rev = '0';
4685                                }
4686                                $rev = sprintf('%-10s', $rev);
4687                                $line =~ s/^\^?[[:xdigit:]]+(\s)/$rev$1/;
4688                        }
4689                        print $line;
4690                }
4691        } else {
4692                ($fh, $ctx) = command_output_pipe('blame', '-p', @_, 'HEAD',
4693                                                  '--', $path);
4694                my ($sha1);
4695                my %authors;
4696                while (my $line = <$fh>) {
4697                        if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
4698                                $sha1 = $1;
4699                                (undef, $rev, undef) = ::cmt_metadata($1);
4700                                $rev = '0' if (!$rev);
4701                        }
4702                        elsif ($line =~ /^author (.*)/) {
4703                                $authors{$rev} = $1;
4704                                $authors{$rev} =~ s/\s/_/g;
4705                        }
4706                        elsif ($line =~ /^\t(.*)$/) {
4707                                printf("%6s %10s %s\n", $rev, $authors{$rev}, $1);
4708                        }
4709                }
4710        }
4711        command_close_pipe($fh, $ctx);
4712}
4713
4714package Git::SVN::Migration;
4715# these version numbers do NOT correspond to actual version numbers
4716# of git nor git-svn.  They are just relative.
4717#
4718# v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
4719#
4720# v1 layout: .git/$id/info/url, refs/remotes/$id
4721#
4722# v2 layout: .git/svn/$id/info/url, refs/remotes/$id
4723#
4724# v3 layout: .git/svn/$id, refs/remotes/$id
4725#            - info/url may remain for backwards compatibility
4726#            - this is what we migrate up to this layout automatically,
4727#            - this will be used by git svn init on single branches
4728# v3.1 layout (auto migrated):
4729#            - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
4730#              for backwards compatibility
4731#
4732# v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
4733#            - this is only created for newly multi-init-ed
4734#              repositories.  Similar in spirit to the
4735#              --use-separate-remotes option in git-clone (now default)
4736#            - we do not automatically migrate to this (following
4737#              the example set by core git)
4738#
4739# v5 layout: .rev_db.$UUID => .rev_map.$UUID
4740#            - newer, more-efficient format that uses 24-bytes per record
4741#              with no filler space.
4742#            - use xxd -c24 < .rev_map.$UUID to view and debug
4743#            - This is a one-way migration, repositories updated to the
4744#              new format will not be able to use old git-svn without
4745#              rebuilding the .rev_db.  Rebuilding the rev_db is not
4746#              possible if noMetadata or useSvmProps are set; but should
4747#              be no problem for users that use the (sensible) defaults.
4748use strict;
4749use warnings;
4750use Carp qw/croak/;
4751use File::Path qw/mkpath/;
4752use File::Basename qw/dirname basename/;
4753use vars qw/$_minimize/;
4754
4755sub migrate_from_v0 {
4756        my $git_dir = $ENV{GIT_DIR};
4757        return undef unless -d $git_dir;
4758        my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4759        my $migrated = 0;
4760        while (<$fh>) {
4761                chomp;
4762                my ($id, $orig_ref) = ($_, $_);
4763                next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
4764                next unless -f "$git_dir/$id/info/url";
4765                my $new_ref = "refs/remotes/$id";
4766                if (::verify_ref("$new_ref^0")) {
4767                        print STDERR "W: $orig_ref is probably an old ",
4768                                     "branch used by an ancient version of ",
4769                                     "git-svn.\n",
4770                                     "However, $new_ref also exists.\n",
4771                                     "We will not be able ",
4772                                     "to use this branch until this ",
4773                                     "ambiguity is resolved.\n";
4774                        next;
4775                }
4776                print STDERR "Migrating from v0 layout...\n" if !$migrated;
4777                print STDERR "Renaming ref: $orig_ref => $new_ref\n";
4778                command_noisy('update-ref', $new_ref, $orig_ref);
4779                command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
4780                $migrated++;
4781        }
4782        command_close_pipe($fh, $ctx);
4783        print STDERR "Done migrating from v0 layout...\n" if $migrated;
4784        $migrated;
4785}
4786
4787sub migrate_from_v1 {
4788        my $git_dir = $ENV{GIT_DIR};
4789        my $migrated = 0;
4790        return $migrated unless -d $git_dir;
4791        my $svn_dir = "$git_dir/svn";
4792
4793        # just in case somebody used 'svn' as their $id at some point...
4794        return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
4795
4796        print STDERR "Migrating from a git-svn v1 layout...\n";
4797        mkpath([$svn_dir]);
4798        print STDERR "Data from a previous version of git-svn exists, but\n\t",
4799                     "$svn_dir\n\t(required for this version ",
4800                     "($::VERSION) of git-svn) does not exist.\n";
4801        my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4802        while (<$fh>) {
4803                my $x = $_;
4804                next unless $x =~ s#^refs/remotes/##;
4805                chomp $x;
4806                next unless -f "$git_dir/$x/info/url";
4807                my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
4808                next unless $u;
4809                my $dn = dirname("$git_dir/svn/$x");
4810                mkpath([$dn]) unless -d $dn;
4811                if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
4812                        mkpath(["$git_dir/svn/svn"]);
4813                        print STDERR " - $git_dir/$x/info => ",
4814                                        "$git_dir/svn/$x/info\n";
4815                        rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
4816                               croak "$!: $x";
4817                        # don't worry too much about these, they probably
4818                        # don't exist with repos this old (save for index,
4819                        # and we can easily regenerate that)
4820                        foreach my $f (qw/unhandled.log index .rev_db/) {
4821                                rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
4822                        }
4823                } else {
4824                        print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
4825                        rename "$git_dir/$x", "$git_dir/svn/$x" or
4826                               croak "$!: $x";
4827                }
4828                $migrated++;
4829        }
4830        command_close_pipe($fh, $ctx);
4831        print STDERR "Done migrating from a git-svn v1 layout\n";
4832        $migrated;
4833}
4834
4835sub read_old_urls {
4836        my ($l_map, $pfx, $path) = @_;
4837        my @dir;
4838        foreach (<$path/*>) {
4839                if (-r "$_/info/url") {
4840                        $pfx .= '/' if $pfx && $pfx !~ m!/$!;
4841                        my $ref_id = $pfx . basename $_;
4842                        my $url = ::file_to_s("$_/info/url");
4843                        $l_map->{$ref_id} = $url;
4844                } elsif (-d $_) {
4845                        push @dir, $_;
4846                }
4847        }
4848        foreach (@dir) {
4849                my $x = $_;
4850                $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
4851                read_old_urls($l_map, $x, $_);
4852        }
4853}
4854
4855sub migrate_from_v2 {
4856        my @cfg = command(qw/config -l/);
4857        return if grep /^svn-remote\..+\.url=/, @cfg;
4858        my %l_map;
4859        read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
4860        my $migrated = 0;
4861
4862        foreach my $ref_id (sort keys %l_map) {
4863                eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
4864                if ($@) {
4865                        Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
4866                }
4867                $migrated++;
4868        }
4869        $migrated;
4870}
4871
4872sub minimize_connections {
4873        my $r = Git::SVN::read_all_remotes();
4874        my $new_urls = {};
4875        my $root_repos = {};
4876        foreach my $repo_id (keys %$r) {
4877                my $url = $r->{$repo_id}->{url} or next;
4878                my $fetch = $r->{$repo_id}->{fetch} or next;
4879                my $ra = Git::SVN::Ra->new($url);
4880
4881                # skip existing cases where we already connect to the root
4882                if (($ra->{url} eq $ra->{repos_root}) ||
4883                    ($ra->{repos_root} eq $repo_id)) {
4884                        $root_repos->{$ra->{url}} = $repo_id;
4885                        next;
4886                }
4887
4888                my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
4889                my $root_path = $ra->{url};
4890                $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
4891                foreach my $path (keys %$fetch) {
4892                        my $ref_id = $fetch->{$path};
4893                        my $gs = Git::SVN->new($ref_id, $repo_id, $path);
4894
4895                        # make sure we can read when connecting to
4896                        # a higher level of a repository
4897                        my ($last_rev, undef) = $gs->last_rev_commit;
4898                        if (!defined $last_rev) {
4899                                $last_rev = eval {
4900                                        $root_ra->get_latest_revnum;
4901                                };
4902                                next if $@;
4903                        }
4904                        my $new = $root_path;
4905                        $new .= length $path ? "/$path" : '';
4906                        eval {
4907                                $root_ra->get_log([$new], $last_rev, $last_rev,
4908                                                  0, 0, 1, sub { });
4909                        };
4910                        next if $@;
4911                        $new_urls->{$ra->{repos_root}}->{$new} =
4912                                { ref_id => $ref_id,
4913                                  old_repo_id => $repo_id,
4914                                  old_path => $path };
4915                }
4916        }
4917
4918        my @emptied;
4919        foreach my $url (keys %$new_urls) {
4920                # see if we can re-use an existing [svn-remote "repo_id"]
4921                # instead of creating a(n ugly) new section:
4922                my $repo_id = $root_repos->{$url} || $url;
4923
4924                my $fetch = $new_urls->{$url};
4925                foreach my $path (keys %$fetch) {
4926                        my $x = $fetch->{$path};
4927                        Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
4928                        my $pfx = "svn-remote.$x->{old_repo_id}";
4929
4930                        my $old_fetch = quotemeta("$x->{old_path}:".
4931                                                  "refs/remotes/$x->{ref_id}");
4932                        command_noisy(qw/config --unset/,
4933                                      "$pfx.fetch", '^'. $old_fetch . '$');
4934                        delete $r->{$x->{old_repo_id}}->
4935                               {fetch}->{$x->{old_path}};
4936                        if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
4937                                command_noisy(qw/config --unset/,
4938                                              "$pfx.url");
4939                                push @emptied, $x->{old_repo_id}
4940                        }
4941                }
4942        }
4943        if (@emptied) {
4944                my $file = $ENV{GIT_CONFIG} || $ENV{GIT_CONFIG_LOCAL} ||
4945                           "$ENV{GIT_DIR}/config";
4946                print STDERR <<EOF;
4947The following [svn-remote] sections in your config file ($file) are empty
4948and can be safely removed:
4949EOF
4950                print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
4951        }
4952}
4953
4954sub migration_check {
4955        migrate_from_v0();
4956        migrate_from_v1();
4957        migrate_from_v2();
4958        minimize_connections() if $_minimize;
4959}
4960
4961package Git::IndexInfo;
4962use strict;
4963use warnings;
4964use Git qw/command_input_pipe command_close_pipe/;
4965
4966sub new {
4967        my ($class) = @_;
4968        my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
4969        bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
4970}
4971
4972sub remove {
4973        my ($self, $path) = @_;
4974        if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
4975                return ++$self->{nr};
4976        }
4977        undef;
4978}
4979
4980sub update {
4981        my ($self, $mode, $hash, $path) = @_;
4982        if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
4983                return ++$self->{nr};
4984        }
4985        undef;
4986}
4987
4988sub DESTROY {
4989        my ($self) = @_;
4990        command_close_pipe($self->{gui}, $self->{ctx});
4991}
4992
4993package Git::SVN::GlobSpec;
4994use strict;
4995use warnings;
4996
4997sub new {
4998        my ($class, $glob) = @_;
4999        my $re = $glob;
5000        $re =~ s!/+$!!g; # no need for trailing slashes
5001        $re =~ m!^([^*]*)(\*(?:/\*)*)([^*]*)$!;
5002        my $temp = $re;
5003        my ($left, $right) = ($1, $3);
5004        $re = $2;
5005        my $depth = $re =~ tr/*/*/;
5006        if ($depth != $temp =~ tr/*/*/) {
5007                die "Only one set of wildcard directories " .
5008                        "(e.g. '*' or '*/*/*') is supported: '$glob'\n";
5009        }
5010        if ($depth == 0) {
5011                die "One '*' is needed for glob: '$glob'\n";
5012        }
5013        $re =~ s!\*!\[^/\]*!g;
5014        $re = quotemeta($left) . "($re)" . quotemeta($right);
5015        if (length $left && !($left =~ s!/+$!!g)) {
5016                die "Missing trailing '/' on left side of: '$glob' ($left)\n";
5017        }
5018        if (length $right && !($right =~ s!^/+!!g)) {
5019                die "Missing leading '/' on right side of: '$glob' ($right)\n";
5020        }
5021        my $left_re = qr/^\/\Q$left\E(\/|$)/;
5022        bless { left => $left, right => $right, left_regex => $left_re,
5023                regex => qr/$re/, glob => $glob, depth => $depth }, $class;
5024}
5025
5026sub full_path {
5027        my ($self, $path) = @_;
5028        return (length $self->{left} ? "$self->{left}/" : '') .
5029               $path . (length $self->{right} ? "/$self->{right}" : '');
5030}
5031
5032__END__
5033
5034Data structures:
5035
5036
5037$remotes = { # returned by read_all_remotes()
5038        'svn' => {
5039                # svn-remote.svn.url=https://svn.musicpd.org
5040                url => 'https://svn.musicpd.org',
5041                # svn-remote.svn.fetch=mpd/trunk:trunk
5042                fetch => {
5043                        'mpd/trunk' => 'trunk',
5044                },
5045                # svn-remote.svn.tags=mpd/tags/*:tags/*
5046                tags => {
5047                        path => {
5048                                left => 'mpd/tags',
5049                                right => '',
5050                                regex => qr!mpd/tags/([^/]+)$!,
5051                                glob => 'tags/*',
5052                        },
5053                        ref => {
5054                                left => 'tags',
5055                                right => '',
5056                                regex => qr!tags/([^/]+)$!,
5057                                glob => 'tags/*',
5058                        },
5059                }
5060        }
5061};
5062
5063$log_entry hashref as returned by libsvn_log_entry()
5064{
5065        log => 'whitespace-formatted log entry
5066',                                              # trailing newline is preserved
5067        revision => '8',                        # integer
5068        date => '2004-02-24T17:01:44.108345Z',  # commit date
5069        author => 'committer name'
5070};
5071
5072
5073# this is generated by generate_diff();
5074@mods = array of diff-index line hashes, each element represents one line
5075        of diff-index output
5076
5077diff-index line ($m hash)
5078{
5079        mode_a => first column of diff-index output, no leading ':',
5080        mode_b => second column of diff-index output,
5081        sha1_b => sha1sum of the final blob,
5082        chg => change type [MCRADT],
5083        file_a => original file name of a file (iff chg is 'C' or 'R')
5084        file_b => new/current file name of a file (any chg)
5085}
5086;
5087
5088# retval of read_url_paths{,_all}();
5089$l_map = {
5090        # repository root url
5091        'https://svn.musicpd.org' => {
5092                # repository path               # GIT_SVN_ID
5093                'mpd/trunk'             =>      'trunk',
5094                'mpd/tags/0.11.5'       =>      'tags/0.11.5',
5095        },
5096}
5097
5098Notes:
5099        I don't trust the each() function on unless I created %hash myself
5100        because the internal iterator may not have started at base.