git-svn.perlon commit builtin-merge: avoid run_command_v_opt() for recursive and subtree (18668f5)
   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        $url = defined $_commit_url ? $_commit_url : $gs->full_url;
 425        my $last_rev = $_revision if defined $_revision;
 426        if ($url) {
 427                print "Committing to $url ...\n";
 428        }
 429        unless ($gs) {
 430                die "Unable to determine upstream SVN information from ",
 431                    "$head history.\nPerhaps the repository is empty.";
 432        }
 433        my ($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 cmd_info {
 807        my $path = canonicalize_path(defined($_[0]) ? $_[0] : ".");
 808        if (exists $_[1]) {
 809                die "Too many arguments specified\n";
 810        }
 811
 812        my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
 813
 814        if (!$file_type && !$diff_status) {
 815                print STDERR "$path:  (Not a versioned resource)\n\n";
 816                return;
 817        }
 818
 819        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 820        unless ($gs) {
 821                die "Unable to determine upstream SVN information from ",
 822                    "working tree history\n";
 823        }
 824
 825        # canonicalize_path() will return "" to make libsvn 1.5.x happy,
 826        $path = "." if $path eq "";
 827
 828        my $full_url = $url . ($path eq "." ? "" : "/$path");
 829
 830        if ($_url) {
 831                print $full_url, "\n";
 832                return;
 833        }
 834
 835        my $result = "Path: $path\n";
 836        $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
 837        $result .= "URL: " . $full_url . "\n";
 838
 839        eval {
 840                my $repos_root = $gs->repos_root;
 841                Git::SVN::remove_username($repos_root);
 842                $result .= "Repository Root: $repos_root\n";
 843        };
 844        if ($@) {
 845                $result .= "Repository Root: (offline)\n";
 846        }
 847        $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A";
 848        $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
 849
 850        $result .= "Node Kind: " .
 851                   ($file_type eq "dir" ? "directory" : "file") . "\n";
 852
 853        my $schedule = $diff_status eq "A"
 854                       ? "add"
 855                       : ($diff_status eq "D" ? "delete" : "normal");
 856        $result .= "Schedule: $schedule\n";
 857
 858        if ($diff_status eq "A") {
 859                print $result, "\n";
 860                return;
 861        }
 862
 863        my ($lc_author, $lc_rev, $lc_date_utc);
 864        my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $path);
 865        my $log = command_output_pipe(@args);
 866        my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
 867        while (<$log>) {
 868                if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
 869                        $lc_author = $1;
 870                        $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
 871                } elsif (/^${esc_color}    (git-svn-id:.+)$/o) {
 872                        (undef, $lc_rev, undef) = ::extract_metadata($1);
 873                }
 874        }
 875        close $log;
 876
 877        Git::SVN::Log::set_local_timezone();
 878
 879        $result .= "Last Changed Author: $lc_author\n";
 880        $result .= "Last Changed Rev: $lc_rev\n";
 881        $result .= "Last Changed Date: " .
 882                   Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
 883
 884        if ($file_type ne "dir") {
 885                my $text_last_updated_date =
 886                    ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
 887                $result .=
 888                    "Text Last Updated: " .
 889                    Git::SVN::Log::format_svn_date($text_last_updated_date) .
 890                    "\n";
 891                my $checksum;
 892                if ($diff_status eq "D") {
 893                        my ($fh, $ctx) =
 894                            command_output_pipe(qw(cat-file blob), "HEAD:$path");
 895                        if ($file_type eq "link") {
 896                                my $file_name = <$fh>;
 897                                $checksum = md5sum("link $file_name");
 898                        } else {
 899                                $checksum = md5sum($fh);
 900                        }
 901                        command_close_pipe($fh, $ctx);
 902                } elsif ($file_type eq "link") {
 903                        my $file_name =
 904                            command(qw(cat-file blob), "HEAD:$path");
 905                        $checksum =
 906                            md5sum("link " . $file_name);
 907                } else {
 908                        open FILE, "<", $path or die $!;
 909                        $checksum = md5sum(\*FILE);
 910                        close FILE or die $!;
 911                }
 912                $result .= "Checksum: " . $checksum . "\n";
 913        }
 914
 915        print $result, "\n";
 916}
 917
 918########################### utility functions #########################
 919
 920sub rebase_cmd {
 921        my @cmd = qw/rebase/;
 922        push @cmd, '-v' if $_verbose;
 923        push @cmd, qw/--merge/ if $_merge;
 924        push @cmd, "--strategy=$_strategy" if $_strategy;
 925        @cmd;
 926}
 927
 928sub post_fetch_checkout {
 929        return if $_no_checkout;
 930        my $gs = $Git::SVN::_head or return;
 931        return if verify_ref('refs/heads/master^0');
 932
 933        my $valid_head = verify_ref('HEAD^0');
 934        command_noisy(qw(update-ref refs/heads/master), $gs->refname);
 935        return if ($valid_head || !verify_ref('HEAD^0'));
 936
 937        return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
 938        my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
 939        return if -f $index;
 940
 941        return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
 942        return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
 943        command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
 944        print STDERR "Checked out HEAD:\n  ",
 945                     $gs->full_url, " r", $gs->last_rev, "\n";
 946}
 947
 948sub complete_svn_url {
 949        my ($url, $path) = @_;
 950        $path =~ s#/+$##;
 951        if ($path !~ m#^[a-z\+]+://#) {
 952                if (!defined $url || $url !~ m#^[a-z\+]+://#) {
 953                        fatal("E: '$path' is not a complete URL ",
 954                              "and a separate URL is not specified");
 955                }
 956                return ($url, $path);
 957        }
 958        return ($path, '');
 959}
 960
 961sub complete_url_ls_init {
 962        my ($ra, $repo_path, $switch, $pfx) = @_;
 963        unless ($repo_path) {
 964                print STDERR "W: $switch not specified\n";
 965                return;
 966        }
 967        $repo_path =~ s#/+$##;
 968        if ($repo_path =~ m#^[a-z\+]+://#) {
 969                $ra = Git::SVN::Ra->new($repo_path);
 970                $repo_path = '';
 971        } else {
 972                $repo_path =~ s#^/+##;
 973                unless ($ra) {
 974                        fatal("E: '$repo_path' is not a complete URL ",
 975                              "and a separate URL is not specified");
 976                }
 977        }
 978        my $url = $ra->{url};
 979        my $gs = Git::SVN->init($url, undef, undef, undef, 1);
 980        my $k = "svn-remote.$gs->{repo_id}.url";
 981        my $orig_url = eval { command_oneline(qw/config --get/, $k) };
 982        if ($orig_url && ($orig_url ne $gs->{url})) {
 983                die "$k already set: $orig_url\n",
 984                    "wanted to set to: $gs->{url}\n";
 985        }
 986        command_oneline('config', $k, $gs->{url}) unless $orig_url;
 987        my $remote_path = "$ra->{svn_path}/$repo_path";
 988        $remote_path =~ s#/+#/#g;
 989        $remote_path =~ s#^/##g;
 990        $remote_path .= "/*" if $remote_path !~ /\*/;
 991        my ($n) = ($switch =~ /^--(\w+)/);
 992        if (length $pfx && $pfx !~ m#/$#) {
 993                die "--prefix='$pfx' must have a trailing slash '/'\n";
 994        }
 995        command_noisy('config',
 996                      "svn-remote.$gs->{repo_id}.$n",
 997                      "$remote_path:refs/remotes/$pfx*" .
 998                        ('/*' x (($remote_path =~ tr/*/*/) - 1)) );
 999}
1000
1001sub verify_ref {
1002        my ($ref) = @_;
1003        eval { command_oneline([ 'rev-parse', '--verify', $ref ],
1004                               { STDERR => 0 }); };
1005}
1006
1007sub get_tree_from_treeish {
1008        my ($treeish) = @_;
1009        # $treeish can be a symbolic ref, too:
1010        my $type = command_oneline(qw/cat-file -t/, $treeish);
1011        my $expected;
1012        while ($type eq 'tag') {
1013                ($treeish, $type) = command(qw/cat-file tag/, $treeish);
1014        }
1015        if ($type eq 'commit') {
1016                $expected = (grep /^tree /, command(qw/cat-file commit/,
1017                                                    $treeish))[0];
1018                ($expected) = ($expected =~ /^tree ($sha1)$/o);
1019                die "Unable to get tree from $treeish\n" unless $expected;
1020        } elsif ($type eq 'tree') {
1021                $expected = $treeish;
1022        } else {
1023                die "$treeish is a $type, expected tree, tag or commit\n";
1024        }
1025        return $expected;
1026}
1027
1028sub get_commit_entry {
1029        my ($treeish) = shift;
1030        my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
1031        my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
1032        my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
1033        open my $log_fh, '>', $commit_editmsg or croak $!;
1034
1035        my $type = command_oneline(qw/cat-file -t/, $treeish);
1036        if ($type eq 'commit' || $type eq 'tag') {
1037                my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1038                                                         $type, $treeish);
1039                my $in_msg = 0;
1040                my $author;
1041                my $saw_from = 0;
1042                my $msgbuf = "";
1043                while (<$msg_fh>) {
1044                        if (!$in_msg) {
1045                                $in_msg = 1 if (/^\s*$/);
1046                                $author = $1 if (/^author (.*>)/);
1047                        } elsif (/^git-svn-id: /) {
1048                                # skip this for now, we regenerate the
1049                                # correct one on re-fetch anyways
1050                                # TODO: set *:merge properties or like...
1051                        } else {
1052                                if (/^From:/ || /^Signed-off-by:/) {
1053                                        $saw_from = 1;
1054                                }
1055                                $msgbuf .= $_;
1056                        }
1057                }
1058                $msgbuf =~ s/\s+$//s;
1059                if ($Git::SVN::_add_author_from && defined($author)
1060                    && !$saw_from) {
1061                        $msgbuf .= "\n\nFrom: $author";
1062                }
1063                print $log_fh $msgbuf or croak $!;
1064                command_close_pipe($msg_fh, $ctx);
1065        }
1066        close $log_fh or croak $!;
1067
1068        if ($_edit || ($type eq 'tree')) {
1069                my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
1070                # TODO: strip out spaces, comments, like git-commit.sh
1071                system($editor, $commit_editmsg);
1072        }
1073        rename $commit_editmsg, $commit_msg or croak $!;
1074        open $log_fh, '<', $commit_msg or croak $!;
1075        { local $/; chomp($log_entry{log} = <$log_fh>); }
1076        close $log_fh or croak $!;
1077        unlink $commit_msg;
1078        \%log_entry;
1079}
1080
1081sub s_to_file {
1082        my ($str, $file, $mode) = @_;
1083        open my $fd,'>',$file or croak $!;
1084        print $fd $str,"\n" or croak $!;
1085        close $fd or croak $!;
1086        chmod ($mode &~ umask, $file) if (defined $mode);
1087}
1088
1089sub file_to_s {
1090        my $file = shift;
1091        open my $fd,'<',$file or croak "$!: file: $file\n";
1092        local $/;
1093        my $ret = <$fd>;
1094        close $fd or croak $!;
1095        $ret =~ s/\s*$//s;
1096        return $ret;
1097}
1098
1099# '<svn username> = real-name <email address>' mapping based on git-svnimport:
1100sub load_authors {
1101        open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1102        my $log = $cmd eq 'log';
1103        while (<$authors>) {
1104                chomp;
1105                next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1106                my ($user, $name, $email) = ($1, $2, $3);
1107                if ($log) {
1108                        $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1109                } else {
1110                        $users{$user} = [$name, $email];
1111                }
1112        }
1113        close $authors or croak $!;
1114}
1115
1116# convert GetOpt::Long specs for use by git-config
1117sub read_repo_config {
1118        return unless -d $ENV{GIT_DIR};
1119        my $opts = shift;
1120        my @config_only;
1121        foreach my $o (keys %$opts) {
1122                # if we have mixedCase and a long option-only, then
1123                # it's a config-only variable that we don't need for
1124                # the command-line.
1125                push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1126                my $v = $opts->{$o};
1127                my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1128                $key =~ s/-//g;
1129                my $arg = 'git-config';
1130                $arg .= ' --int' if ($o =~ /[:=]i$/);
1131                $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1132                if (ref $v eq 'ARRAY') {
1133                        chomp(my @tmp = `$arg --get-all svn.$key`);
1134                        @$v = @tmp if @tmp;
1135                } else {
1136                        chomp(my $tmp = `$arg --get svn.$key`);
1137                        if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1138                                $$v = $tmp;
1139                        }
1140                }
1141        }
1142        delete @$opts{@config_only} if @config_only;
1143}
1144
1145sub extract_metadata {
1146        my $id = shift or return (undef, undef, undef);
1147        my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1148                                                        \s([a-f\d\-]+)$/x);
1149        if (!defined $rev || !$uuid || !$url) {
1150                # some of the original repositories I made had
1151                # identifiers like this:
1152                ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/);
1153        }
1154        return ($url, $rev, $uuid);
1155}
1156
1157sub cmt_metadata {
1158        return extract_metadata((grep(/^git-svn-id: /,
1159                command(qw/cat-file commit/, shift)))[-1]);
1160}
1161
1162sub working_head_info {
1163        my ($head, $refs) = @_;
1164        my @args = ('log', '--no-color', '--first-parent', '--pretty=medium');
1165        my ($fh, $ctx) = command_output_pipe(@args, $head);
1166        my $hash;
1167        my %max;
1168        while (<$fh>) {
1169                if ( m{^commit ($::sha1)$} ) {
1170                        unshift @$refs, $hash if $hash and $refs;
1171                        $hash = $1;
1172                        next;
1173                }
1174                next unless s{^\s*(git-svn-id:)}{$1};
1175                my ($url, $rev, $uuid) = extract_metadata($_);
1176                if (defined $url && defined $rev) {
1177                        next if $max{$url} and $max{$url} < $rev;
1178                        if (my $gs = Git::SVN->find_by_url($url)) {
1179                                my $c = $gs->rev_map_get($rev, $uuid);
1180                                if ($c && $c eq $hash) {
1181                                        close $fh; # break the pipe
1182                                        return ($url, $rev, $uuid, $gs);
1183                                } else {
1184                                        $max{$url} ||= $gs->rev_map_max;
1185                                }
1186                        }
1187                }
1188        }
1189        command_close_pipe($fh, $ctx);
1190        (undef, undef, undef, undef);
1191}
1192
1193sub read_commit_parents {
1194        my ($parents, $c) = @_;
1195        chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1196        $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1197        @{$parents->{$c}} = split(/ /, $p);
1198}
1199
1200sub linearize_history {
1201        my ($gs, $refs) = @_;
1202        my %parents;
1203        foreach my $c (@$refs) {
1204                read_commit_parents(\%parents, $c);
1205        }
1206
1207        my @linear_refs;
1208        my %skip = ();
1209        my $last_svn_commit = $gs->last_commit;
1210        foreach my $c (reverse @$refs) {
1211                next if $c eq $last_svn_commit;
1212                last if $skip{$c};
1213
1214                unshift @linear_refs, $c;
1215                $skip{$c} = 1;
1216
1217                # we only want the first parent to diff against for linear
1218                # history, we save the rest to inject when we finalize the
1219                # svn commit
1220                my $fp_a = verify_ref("$c~1");
1221                my $fp_b = shift @{$parents{$c}} if $parents{$c};
1222                if (!$fp_a || !$fp_b) {
1223                        die "Commit $c\n",
1224                            "has no parent commit, and therefore ",
1225                            "nothing to diff against.\n",
1226                            "You should be working from a repository ",
1227                            "originally created by git-svn\n";
1228                }
1229                if ($fp_a ne $fp_b) {
1230                        die "$c~1 = $fp_a, however parsing commit $c ",
1231                            "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1232                }
1233
1234                foreach my $p (@{$parents{$c}}) {
1235                        $skip{$p} = 1;
1236                }
1237        }
1238        (\@linear_refs, \%parents);
1239}
1240
1241sub find_file_type_and_diff_status {
1242        my ($path) = @_;
1243        return ('dir', '') if $path eq '';
1244
1245        my $diff_output =
1246            command_oneline(qw(diff --cached --name-status --), $path) || "";
1247        my $diff_status = (split(' ', $diff_output))[0] || "";
1248
1249        my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1250
1251        return (undef, undef) if !$diff_status && !$ls_tree;
1252
1253        if ($diff_status eq "A") {
1254                return ("link", $diff_status) if -l $path;
1255                return ("dir", $diff_status) if -d $path;
1256                return ("file", $diff_status);
1257        }
1258
1259        my $mode = (split(' ', $ls_tree))[0] || "";
1260
1261        return ("link", $diff_status) if $mode eq "120000";
1262        return ("dir", $diff_status) if $mode eq "040000";
1263        return ("file", $diff_status);
1264}
1265
1266sub md5sum {
1267        my $arg = shift;
1268        my $ref = ref $arg;
1269        my $md5 = Digest::MD5->new();
1270        if ($ref eq 'GLOB' || $ref eq 'IO::File' || $ref eq 'File::Temp') {
1271                $md5->addfile($arg) or croak $!;
1272        } elsif ($ref eq 'SCALAR') {
1273                $md5->add($$arg) or croak $!;
1274        } elsif (!$ref) {
1275                $md5->add($arg) or croak $!;
1276        } else {
1277                ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
1278        }
1279        return $md5->hexdigest();
1280}
1281
1282package Git::SVN;
1283use strict;
1284use warnings;
1285use Fcntl qw/:DEFAULT :seek/;
1286use constant rev_map_fmt => 'NH40';
1287use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1288            $_repack $_repack_flags $_use_svm_props $_head
1289            $_use_svnsync_props $no_reuse_existing $_minimize_url
1290            $_use_log_author $_add_author_from/;
1291use Carp qw/croak/;
1292use File::Path qw/mkpath/;
1293use File::Copy qw/copy/;
1294use IPC::Open3;
1295
1296my ($_gc_nr, $_gc_period);
1297
1298# properties that we do not log:
1299my %SKIP_PROP;
1300BEGIN {
1301        %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1302                                        svn:special svn:executable
1303                                        svn:entry:committed-rev
1304                                        svn:entry:last-author
1305                                        svn:entry:uuid
1306                                        svn:entry:committed-date/;
1307
1308        # some options are read globally, but can be overridden locally
1309        # per [svn-remote "..."] section.  Command-line options will *NOT*
1310        # override options set in an [svn-remote "..."] section
1311        no strict 'refs';
1312        for my $option (qw/follow_parent no_metadata use_svm_props
1313                           use_svnsync_props/) {
1314                my $key = $option;
1315                $key =~ tr/_//d;
1316                my $prop = "-$option";
1317                *$option = sub {
1318                        my ($self) = @_;
1319                        return $self->{$prop} if exists $self->{$prop};
1320                        my $k = "svn-remote.$self->{repo_id}.$key";
1321                        eval { command_oneline(qw/config --get/, $k) };
1322                        if ($@) {
1323                                $self->{$prop} = ${"Git::SVN::_$option"};
1324                        } else {
1325                                my $v = command_oneline(qw/config --bool/,$k);
1326                                $self->{$prop} = $v eq 'false' ? 0 : 1;
1327                        }
1328                        return $self->{$prop};
1329                }
1330        }
1331}
1332
1333
1334my (%LOCKFILES, %INDEX_FILES);
1335END {
1336        unlink keys %LOCKFILES if %LOCKFILES;
1337        unlink keys %INDEX_FILES if %INDEX_FILES;
1338}
1339
1340sub resolve_local_globs {
1341        my ($url, $fetch, $glob_spec) = @_;
1342        return unless defined $glob_spec;
1343        my $ref = $glob_spec->{ref};
1344        my $path = $glob_spec->{path};
1345        foreach (command(qw#for-each-ref --format=%(refname) refs/remotes#)) {
1346                next unless m#^refs/remotes/$ref->{regex}$#;
1347                my $p = $1;
1348                my $pathname = desanitize_refname($path->full_path($p));
1349                my $refname = desanitize_refname($ref->full_path($p));
1350                if (my $existing = $fetch->{$pathname}) {
1351                        if ($existing ne $refname) {
1352                                die "Refspec conflict:\n",
1353                                    "existing: refs/remotes/$existing\n",
1354                                    " globbed: refs/remotes/$refname\n";
1355                        }
1356                        my $u = (::cmt_metadata("refs/remotes/$refname"))[0];
1357                        $u =~ s!^\Q$url\E(/|$)!! or die
1358                          "refs/remotes/$refname: '$url' not found in '$u'\n";
1359                        if ($pathname ne $u) {
1360                                warn "W: Refspec glob conflict ",
1361                                     "(ref: refs/remotes/$refname):\n",
1362                                     "expected path: $pathname\n",
1363                                     "    real path: $u\n",
1364                                     "Continuing ahead with $u\n";
1365                                next;
1366                        }
1367                } else {
1368                        $fetch->{$pathname} = $refname;
1369                }
1370        }
1371}
1372
1373sub parse_revision_argument {
1374        my ($base, $head) = @_;
1375        if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1376                return ($base, $head);
1377        }
1378        return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1379        return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1380        return ($head, $head) if ($::_revision eq 'HEAD');
1381        return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1382        return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1383        die "revision argument: $::_revision not understood by git-svn\n";
1384}
1385
1386sub fetch_all {
1387        my ($repo_id, $remotes) = @_;
1388        if (ref $repo_id) {
1389                my $gs = $repo_id;
1390                $repo_id = undef;
1391                $repo_id = $gs->{repo_id};
1392        }
1393        $remotes ||= read_all_remotes();
1394        my $remote = $remotes->{$repo_id} or
1395                     die "[svn-remote \"$repo_id\"] unknown\n";
1396        my $fetch = $remote->{fetch};
1397        my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1398        my (@gs, @globs);
1399        my $ra = Git::SVN::Ra->new($url);
1400        my $uuid = $ra->get_uuid;
1401        my $head = $ra->get_latest_revnum;
1402        my $base = defined $fetch ? $head : 0;
1403
1404        # read the max revs for wildcard expansion (branches/*, tags/*)
1405        foreach my $t (qw/branches tags/) {
1406                defined $remote->{$t} or next;
1407                push @globs, $remote->{$t};
1408                my $max_rev = eval { tmp_config(qw/--int --get/,
1409                                         "svn-remote.$repo_id.${t}-maxRev") };
1410                if (defined $max_rev && ($max_rev < $base)) {
1411                        $base = $max_rev;
1412                } elsif (!defined $max_rev) {
1413                        $base = 0;
1414                }
1415        }
1416
1417        if ($fetch) {
1418                foreach my $p (sort keys %$fetch) {
1419                        my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1420                        my $lr = $gs->rev_map_max;
1421                        if (defined $lr) {
1422                                $base = $lr if ($lr < $base);
1423                        }
1424                        push @gs, $gs;
1425                }
1426        }
1427
1428        ($base, $head) = parse_revision_argument($base, $head);
1429        $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1430}
1431
1432sub read_all_remotes {
1433        my $r = {};
1434        my $use_svm_props = eval { command_oneline(qw/config --bool
1435            svn.useSvmProps/) };
1436        $use_svm_props = $use_svm_props eq 'true' if $use_svm_props;
1437        foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1438                if (m!^(.+)\.fetch=\s*(.*)\s*:\s*(.+)\s*$!) {
1439                        my ($remote, $local_ref, $_remote_ref) = ($1, $2, $3);
1440                        die("svn-remote.$remote: remote ref '$_remote_ref' "
1441                            . "must start with 'refs/remotes/'\n")
1442                                unless $_remote_ref =~ m{^refs/remotes/(.+)};
1443                        my $remote_ref = $1;
1444                        $local_ref =~ s{^/}{};
1445                        $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1446                        $r->{$remote}->{svm} = {} if $use_svm_props;
1447                } elsif (m!^(.+)\.usesvmprops=\s*(.*)\s*$!) {
1448                        $r->{$1}->{svm} = {};
1449                } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1450                        $r->{$1}->{url} = $2;
1451                } elsif (m!^(.+)\.(branches|tags)=
1452                           (.*):refs/remotes/(.+)\s*$/!x) {
1453                        my ($p, $g) = ($3, $4);
1454                        my $rs = $r->{$1}->{$2} = {
1455                                          t => $2,
1456                                          remote => $1,
1457                                          path => Git::SVN::GlobSpec->new($p),
1458                                          ref => Git::SVN::GlobSpec->new($g) };
1459                        if (length($rs->{ref}->{right}) != 0) {
1460                                die "The '*' glob character must be the last ",
1461                                    "character of '$g'\n";
1462                        }
1463                }
1464        }
1465
1466        map {
1467                if (defined $r->{$_}->{svm}) {
1468                        my $svm;
1469                        eval {
1470                                my $section = "svn-remote.$_";
1471                                $svm = {
1472                                        source => tmp_config('--get',
1473                                            "$section.svm-source"),
1474                                        replace => tmp_config('--get',
1475                                            "$section.svm-replace"),
1476                                }
1477                        };
1478                        $r->{$_}->{svm} = $svm;
1479                }
1480        } keys %$r;
1481
1482        $r;
1483}
1484
1485sub init_vars {
1486        $_gc_nr = $_gc_period = 1000;
1487        if (defined $_repack || defined $_repack_flags) {
1488               warn "Repack options are obsolete; they have no effect.\n";
1489        }
1490}
1491
1492sub verify_remotes_sanity {
1493        return unless -d $ENV{GIT_DIR};
1494        my %seen;
1495        foreach (command(qw/config -l/)) {
1496                if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1497                        if ($seen{$1}) {
1498                                die "Remote ref refs/remote/$1 is tracked by",
1499                                    "\n  \"$_\"\nand\n  \"$seen{$1}\"\n",
1500                                    "Please resolve this ambiguity in ",
1501                                    "your git configuration file before ",
1502                                    "continuing\n";
1503                        }
1504                        $seen{$1} = $_;
1505                }
1506        }
1507}
1508
1509sub find_existing_remote {
1510        my ($url, $remotes) = @_;
1511        return undef if $no_reuse_existing;
1512        my $existing;
1513        foreach my $repo_id (keys %$remotes) {
1514                my $u = $remotes->{$repo_id}->{url} or next;
1515                next if $u ne $url;
1516                $existing = $repo_id;
1517                last;
1518        }
1519        $existing;
1520}
1521
1522sub init_remote_config {
1523        my ($self, $url, $no_write) = @_;
1524        $url =~ s!/+$!!; # strip trailing slash
1525        my $r = read_all_remotes();
1526        my $existing = find_existing_remote($url, $r);
1527        if ($existing) {
1528                unless ($no_write) {
1529                        print STDERR "Using existing ",
1530                                     "[svn-remote \"$existing\"]\n";
1531                }
1532                $self->{repo_id} = $existing;
1533        } elsif ($_minimize_url) {
1534                my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1535                $existing = find_existing_remote($min_url, $r);
1536                if ($existing) {
1537                        unless ($no_write) {
1538                                print STDERR "Using existing ",
1539                                             "[svn-remote \"$existing\"]\n";
1540                        }
1541                        $self->{repo_id} = $existing;
1542                }
1543                if ($min_url ne $url) {
1544                        unless ($no_write) {
1545                                print STDERR "Using higher level of URL: ",
1546                                             "$url => $min_url\n";
1547                        }
1548                        my $old_path = $self->{path};
1549                        $self->{path} = $url;
1550                        $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1551                        if (length $old_path) {
1552                                $self->{path} .= "/$old_path";
1553                        }
1554                        $url = $min_url;
1555                }
1556        }
1557        my $orig_url;
1558        if (!$existing) {
1559                # verify that we aren't overwriting anything:
1560                $orig_url = eval {
1561                        command_oneline('config', '--get',
1562                                        "svn-remote.$self->{repo_id}.url")
1563                };
1564                if ($orig_url && ($orig_url ne $url)) {
1565                        die "svn-remote.$self->{repo_id}.url already set: ",
1566                            "$orig_url\nwanted to set to: $url\n";
1567                }
1568        }
1569        my ($xrepo_id, $xpath) = find_ref($self->refname);
1570        if (defined $xpath) {
1571                die "svn-remote.$xrepo_id.fetch already set to track ",
1572                    "$xpath:refs/remotes/", $self->refname, "\n";
1573        }
1574        unless ($no_write) {
1575                command_noisy('config',
1576                              "svn-remote.$self->{repo_id}.url", $url);
1577                $self->{path} =~ s{^/}{};
1578                command_noisy('config', '--add',
1579                              "svn-remote.$self->{repo_id}.fetch",
1580                              "$self->{path}:".$self->refname);
1581        }
1582        $self->{url} = $url;
1583}
1584
1585sub find_by_url { # repos_root and, path are optional
1586        my ($class, $full_url, $repos_root, $path) = @_;
1587
1588        return undef unless defined $full_url;
1589        remove_username($full_url);
1590        remove_username($repos_root) if defined $repos_root;
1591        my $remotes = read_all_remotes();
1592        if (defined $full_url && defined $repos_root && !defined $path) {
1593                $path = $full_url;
1594                $path =~ s#^\Q$repos_root\E(?:/|$)##;
1595        }
1596        foreach my $repo_id (keys %$remotes) {
1597                my $u = $remotes->{$repo_id}->{url} or next;
1598                remove_username($u);
1599                next if defined $repos_root && $repos_root ne $u;
1600
1601                my $fetch = $remotes->{$repo_id}->{fetch} || {};
1602                foreach (qw/branches tags/) {
1603                        resolve_local_globs($u, $fetch,
1604                                            $remotes->{$repo_id}->{$_});
1605                }
1606                my $p = $path;
1607                my $rwr = rewrite_root({repo_id => $repo_id});
1608                my $svm = $remotes->{$repo_id}->{svm}
1609                        if defined $remotes->{$repo_id}->{svm};
1610                unless (defined $p) {
1611                        $p = $full_url;
1612                        my $z = $u;
1613                        my $prefix = '';
1614                        if ($rwr) {
1615                                $z = $rwr;
1616                        } elsif (defined $svm) {
1617                                $z = $svm->{source};
1618                                $prefix = $svm->{replace};
1619                                $prefix =~ s#^\Q$u\E(?:/|$)##;
1620                                $prefix =~ s#/$##;
1621                        }
1622                        $p =~ s#^\Q$z\E(?:/|$)#$prefix# or next;
1623                }
1624                foreach my $f (keys %$fetch) {
1625                        next if $f ne $p;
1626                        return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1627                }
1628        }
1629        undef;
1630}
1631
1632sub init {
1633        my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1634        my $self = _new($class, $repo_id, $ref_id, $path);
1635        if (defined $url) {
1636                $self->init_remote_config($url, $no_write);
1637        }
1638        $self;
1639}
1640
1641sub find_ref {
1642        my ($ref_id) = @_;
1643        foreach (command(qw/config -l/)) {
1644                next unless m!^svn-remote\.(.+)\.fetch=
1645                              \s*(.*)\s*:\s*refs/remotes/(.+)\s*$!x;
1646                my ($repo_id, $path, $ref) = ($1, $2, $3);
1647                if ($ref eq $ref_id) {
1648                        $path = '' if ($path =~ m#^\./?#);
1649                        return ($repo_id, $path);
1650                }
1651        }
1652        (undef, undef, undef);
1653}
1654
1655sub new {
1656        my ($class, $ref_id, $repo_id, $path) = @_;
1657        if (defined $ref_id && !defined $repo_id && !defined $path) {
1658                ($repo_id, $path) = find_ref($ref_id);
1659                if (!defined $repo_id) {
1660                        die "Could not find a \"svn-remote.*.fetch\" key ",
1661                            "in the repository configuration matching: ",
1662                            "refs/remotes/$ref_id\n";
1663                }
1664        }
1665        my $self = _new($class, $repo_id, $ref_id, $path);
1666        if (!defined $self->{path} || !length $self->{path}) {
1667                my $fetch = command_oneline('config', '--get',
1668                                            "svn-remote.$repo_id.fetch",
1669                                            ":refs/remotes/$ref_id\$") or
1670                     die "Failed to read \"svn-remote.$repo_id.fetch\" ",
1671                         "\":refs/remotes/$ref_id\$\" in config\n";
1672                ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
1673        }
1674        $self->{url} = command_oneline('config', '--get',
1675                                       "svn-remote.$repo_id.url") or
1676                  die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
1677        $self->rebuild;
1678        $self;
1679}
1680
1681sub refname {
1682        my ($refname) = "refs/remotes/$_[0]->{ref_id}" ;
1683
1684        # It cannot end with a slash /, we'll throw up on this because
1685        # SVN can't have directories with a slash in their name, either:
1686        if ($refname =~ m{/$}) {
1687                die "ref: '$refname' ends with a trailing slash, this is ",
1688                    "not permitted by git nor Subversion\n";
1689        }
1690
1691        # It cannot have ASCII control character space, tilde ~, caret ^,
1692        # colon :, question-mark ?, asterisk *, space, or open bracket [
1693        # anywhere.
1694        #
1695        # Additionally, % must be escaped because it is used for escaping
1696        # and we want our escaped refname to be reversible
1697        $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
1698
1699        # no slash-separated component can begin with a dot .
1700        # /.* becomes /%2E*
1701        $refname =~ s{/\.}{/%2E}g;
1702
1703        # It cannot have two consecutive dots .. anywhere
1704        # .. becomes %2E%2E
1705        $refname =~ s{\.\.}{%2E%2E}g;
1706
1707        return $refname;
1708}
1709
1710sub desanitize_refname {
1711        my ($refname) = @_;
1712        $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
1713        return $refname;
1714}
1715
1716sub svm_uuid {
1717        my ($self) = @_;
1718        return $self->{svm}->{uuid} if $self->svm;
1719        $self->ra;
1720        unless ($self->{svm}) {
1721                die "SVM UUID not cached, and reading remotely failed\n";
1722        }
1723        $self->{svm}->{uuid};
1724}
1725
1726sub svm {
1727        my ($self) = @_;
1728        return $self->{svm} if $self->{svm};
1729        my $svm;
1730        # see if we have it in our config, first:
1731        eval {
1732                my $section = "svn-remote.$self->{repo_id}";
1733                $svm = {
1734                  source => tmp_config('--get', "$section.svm-source"),
1735                  uuid => tmp_config('--get', "$section.svm-uuid"),
1736                  replace => tmp_config('--get', "$section.svm-replace"),
1737                }
1738        };
1739        if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
1740                $self->{svm} = $svm;
1741        }
1742        $self->{svm};
1743}
1744
1745sub _set_svm_vars {
1746        my ($self, $ra) = @_;
1747        return $ra if $self->svm;
1748
1749        my @err = ( "useSvmProps set, but failed to read SVM properties\n",
1750                    "(svm:source, svm:uuid) ",
1751                    "from the following URLs:\n" );
1752        sub read_svm_props {
1753                my ($self, $ra, $path, $r) = @_;
1754                my $props = ($ra->get_dir($path, $r))[2];
1755                my $src = $props->{'svm:source'};
1756                my $uuid = $props->{'svm:uuid'};
1757                return undef if (!$src || !$uuid);
1758
1759                chomp($src, $uuid);
1760
1761                $uuid =~ m{^[0-9a-f\-]{30,}$}
1762                    or die "doesn't look right - svm:uuid is '$uuid'\n";
1763
1764                # the '!' is used to mark the repos_root!/relative/path
1765                $src =~ s{/?!/?}{/};
1766                $src =~ s{/+$}{}; # no trailing slashes please
1767                # username is of no interest
1768                $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
1769
1770                my $replace = $ra->{url};
1771                $replace .= "/$path" if length $path;
1772
1773                my $section = "svn-remote.$self->{repo_id}";
1774                tmp_config("$section.svm-source", $src);
1775                tmp_config("$section.svm-replace", $replace);
1776                tmp_config("$section.svm-uuid", $uuid);
1777                $self->{svm} = {
1778                        source => $src,
1779                        uuid => $uuid,
1780                        replace => $replace
1781                };
1782        }
1783
1784        my $r = $ra->get_latest_revnum;
1785        my $path = $self->{path};
1786        my %tried;
1787        while (length $path) {
1788                unless ($tried{"$self->{url}/$path"}) {
1789                        return $ra if $self->read_svm_props($ra, $path, $r);
1790                        $tried{"$self->{url}/$path"} = 1;
1791                }
1792                $path =~ s#/?[^/]+$##;
1793        }
1794        die "Path: '$path' should be ''\n" if $path ne '';
1795        return $ra if $self->read_svm_props($ra, $path, $r);
1796        $tried{"$self->{url}/$path"} = 1;
1797
1798        if ($ra->{repos_root} eq $self->{url}) {
1799                die @err, (map { "  $_\n" } keys %tried), "\n";
1800        }
1801
1802        # nope, make sure we're connected to the repository root:
1803        my $ok;
1804        my @tried_b;
1805        $path = $ra->{svn_path};
1806        $ra = Git::SVN::Ra->new($ra->{repos_root});
1807        while (length $path) {
1808                unless ($tried{"$ra->{url}/$path"}) {
1809                        $ok = $self->read_svm_props($ra, $path, $r);
1810                        last if $ok;
1811                        $tried{"$ra->{url}/$path"} = 1;
1812                }
1813                $path =~ s#/?[^/]+$##;
1814        }
1815        die "Path: '$path' should be ''\n" if $path ne '';
1816        $ok ||= $self->read_svm_props($ra, $path, $r);
1817        $tried{"$ra->{url}/$path"} = 1;
1818        if (!$ok) {
1819                die @err, (map { "  $_\n" } keys %tried), "\n";
1820        }
1821        Git::SVN::Ra->new($self->{url});
1822}
1823
1824sub svnsync {
1825        my ($self) = @_;
1826        return $self->{svnsync} if $self->{svnsync};
1827
1828        if ($self->no_metadata) {
1829                die "Can't have both 'noMetadata' and ",
1830                    "'useSvnsyncProps' options set!\n";
1831        }
1832        if ($self->rewrite_root) {
1833                die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
1834                    "options set!\n";
1835        }
1836
1837        my $svnsync;
1838        # see if we have it in our config, first:
1839        eval {
1840                my $section = "svn-remote.$self->{repo_id}";
1841
1842                my $url = tmp_config('--get', "$section.svnsync-url");
1843                ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1844                   die "doesn't look right - svn:sync-from-url is '$url'\n";
1845
1846                my $uuid = tmp_config('--get', "$section.svnsync-uuid");
1847                ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1848                   die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1849
1850                $svnsync = { url => $url, uuid => $uuid }
1851        };
1852        if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
1853                return $self->{svnsync} = $svnsync;
1854        }
1855
1856        my $err = "useSvnsyncProps set, but failed to read " .
1857                  "svnsync property: svn:sync-from-";
1858        my $rp = $self->ra->rev_proplist(0);
1859
1860        my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
1861        ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1862                   die "doesn't look right - svn:sync-from-url is '$url'\n";
1863
1864        my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
1865        ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1866                   die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1867
1868        my $section = "svn-remote.$self->{repo_id}";
1869        tmp_config('--add', "$section.svnsync-uuid", $uuid);
1870        tmp_config('--add', "$section.svnsync-url", $url);
1871        return $self->{svnsync} = { url => $url, uuid => $uuid };
1872}
1873
1874# this allows us to memoize our SVN::Ra UUID locally and avoid a
1875# remote lookup (useful for 'git svn log').
1876sub ra_uuid {
1877        my ($self) = @_;
1878        unless ($self->{ra_uuid}) {
1879                my $key = "svn-remote.$self->{repo_id}.uuid";
1880                my $uuid = eval { tmp_config('--get', $key) };
1881                if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/) {
1882                        $self->{ra_uuid} = $uuid;
1883                } else {
1884                        die "ra_uuid called without URL\n" unless $self->{url};
1885                        $self->{ra_uuid} = $self->ra->get_uuid;
1886                        tmp_config('--add', $key, $self->{ra_uuid});
1887                }
1888        }
1889        $self->{ra_uuid};
1890}
1891
1892sub _set_repos_root {
1893        my ($self, $repos_root) = @_;
1894        my $k = "svn-remote.$self->{repo_id}.reposRoot";
1895        $repos_root ||= $self->ra->{repos_root};
1896        tmp_config($k, $repos_root);
1897        $repos_root;
1898}
1899
1900sub repos_root {
1901        my ($self) = @_;
1902        my $k = "svn-remote.$self->{repo_id}.reposRoot";
1903        eval { tmp_config('--get', $k) } || $self->_set_repos_root;
1904}
1905
1906sub ra {
1907        my ($self) = shift;
1908        my $ra = Git::SVN::Ra->new($self->{url});
1909        $self->_set_repos_root($ra->{repos_root});
1910        if ($self->use_svm_props && !$self->{svm}) {
1911                if ($self->no_metadata) {
1912                        die "Can't have both 'noMetadata' and ",
1913                            "'useSvmProps' options set!\n";
1914                } elsif ($self->use_svnsync_props) {
1915                        die "Can't have both 'useSvnsyncProps' and ",
1916                            "'useSvmProps' options set!\n";
1917                }
1918                $ra = $self->_set_svm_vars($ra);
1919                $self->{-want_revprops} = 1;
1920        }
1921        $ra;
1922}
1923
1924sub rel_path {
1925        my ($self) = @_;
1926        my $repos_root = $self->ra->{repos_root};
1927        return $self->{path} if ($self->{url} eq $repos_root);
1928        my $url = $self->{url} .
1929                  (length $self->{path} ? "/$self->{path}" : $self->{path});
1930        $url =~ s!^\Q$repos_root\E(?:/+|$)!!g;
1931        $url;
1932}
1933
1934# prop_walk(PATH, REV, SUB)
1935# -------------------------
1936# Recursively traverse PATH at revision REV and invoke SUB for each
1937# directory that contains a SVN property.  SUB will be invoked as
1938# follows:  &SUB(gs, path, props);  where `gs' is this instance of
1939# Git::SVN, `path' the path to the directory where the properties
1940# `props' were found.  The `path' will be relative to point of checkout,
1941# that is, if url://repo/trunk is the current Git branch, and that
1942# directory contains a sub-directory `d', SUB will be invoked with `/d/'
1943# as `path' (note the trailing `/').
1944sub prop_walk {
1945        my ($self, $path, $rev, $sub) = @_;
1946
1947        $path =~ s#^/##;
1948        my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
1949        $path =~ s#^/*#/#g;
1950        my $p = $path;
1951        # Strip the irrelevant part of the path.
1952        $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
1953        # Ensure the path is terminated by a `/'.
1954        $p =~ s#/*$#/#;
1955
1956        # The properties contain all the internal SVN stuff nobody
1957        # (usually) cares about.
1958        my $interesting_props = 0;
1959        foreach (keys %{$props}) {
1960                # If it doesn't start with `svn:', it must be a
1961                # user-defined property.
1962                ++$interesting_props and next if $_ !~ /^svn:/;
1963                # FIXME: Fragile, if SVN adds new public properties,
1964                # this needs to be updated.
1965                ++$interesting_props if /^svn:(?:ignore|keywords|executable
1966                                                 |eol-style|mime-type
1967                                                 |externals|needs-lock)$/x;
1968        }
1969        &$sub($self, $p, $props) if $interesting_props;
1970
1971        foreach (sort keys %$dirent) {
1972                next if $dirent->{$_}->{kind} != $SVN::Node::dir;
1973                $self->prop_walk($self->{path} . $p . $_, $rev, $sub);
1974        }
1975}
1976
1977sub last_rev { ($_[0]->last_rev_commit)[0] }
1978sub last_commit { ($_[0]->last_rev_commit)[1] }
1979
1980# returns the newest SVN revision number and newest commit SHA1
1981sub last_rev_commit {
1982        my ($self) = @_;
1983        if (defined $self->{last_rev} && defined $self->{last_commit}) {
1984                return ($self->{last_rev}, $self->{last_commit});
1985        }
1986        my $c = ::verify_ref($self->refname.'^0');
1987        if ($c && !$self->use_svm_props && !$self->no_metadata) {
1988                my $rev = (::cmt_metadata($c))[1];
1989                if (defined $rev) {
1990                        ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
1991                        return ($rev, $c);
1992                }
1993        }
1994        my $map_path = $self->map_path;
1995        unless (-e $map_path) {
1996                ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
1997                return (undef, undef);
1998        }
1999        my ($rev, $commit) = $self->rev_map_max(1);
2000        ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
2001        return ($rev, $commit);
2002}
2003
2004sub get_fetch_range {
2005        my ($self, $min, $max) = @_;
2006        $max ||= $self->ra->get_latest_revnum;
2007        $min ||= $self->rev_map_max;
2008        (++$min, $max);
2009}
2010
2011sub tmp_config {
2012        my (@args) = @_;
2013        my $old_def_config = "$ENV{GIT_DIR}/svn/config";
2014        my $config = "$ENV{GIT_DIR}/svn/.metadata";
2015        if (! -f $config && -f $old_def_config) {
2016                rename $old_def_config, $config or
2017                       die "Failed rename $old_def_config => $config: $!\n";
2018        }
2019        my $old_config = $ENV{GIT_CONFIG};
2020        $ENV{GIT_CONFIG} = $config;
2021        $@ = undef;
2022        my @ret = eval {
2023                unless (-f $config) {
2024                        mkfile($config);
2025                        open my $fh, '>', $config or
2026                            die "Can't open $config: $!\n";
2027                        print $fh "; This file is used internally by ",
2028                                  "git-svn\n" or die
2029                                  "Couldn't write to $config: $!\n";
2030                        print $fh "; You should not have to edit it\n" or
2031                              die "Couldn't write to $config: $!\n";
2032                        close $fh or die "Couldn't close $config: $!\n";
2033                }
2034                command('config', @args);
2035        };
2036        my $err = $@;
2037        if (defined $old_config) {
2038                $ENV{GIT_CONFIG} = $old_config;
2039        } else {
2040                delete $ENV{GIT_CONFIG};
2041        }
2042        die $err if $err;
2043        wantarray ? @ret : $ret[0];
2044}
2045
2046sub tmp_index_do {
2047        my ($self, $sub) = @_;
2048        my $old_index = $ENV{GIT_INDEX_FILE};
2049        $ENV{GIT_INDEX_FILE} = $self->{index};
2050        $@ = undef;
2051        my @ret = eval {
2052                my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
2053                mkpath([$dir]) unless -d $dir;
2054                &$sub;
2055        };
2056        my $err = $@;
2057        if (defined $old_index) {
2058                $ENV{GIT_INDEX_FILE} = $old_index;
2059        } else {
2060                delete $ENV{GIT_INDEX_FILE};
2061        }
2062        die $err if $err;
2063        wantarray ? @ret : $ret[0];
2064}
2065
2066sub assert_index_clean {
2067        my ($self, $treeish) = @_;
2068
2069        $self->tmp_index_do(sub {
2070                command_noisy('read-tree', $treeish) unless -e $self->{index};
2071                my $x = command_oneline('write-tree');
2072                my ($y) = (command(qw/cat-file commit/, $treeish) =~
2073                           /^tree ($::sha1)/mo);
2074                return if $y eq $x;
2075
2076                warn "Index mismatch: $y != $x\nrereading $treeish\n";
2077                unlink $self->{index} or die "unlink $self->{index}: $!\n";
2078                command_noisy('read-tree', $treeish);
2079                $x = command_oneline('write-tree');
2080                if ($y ne $x) {
2081                        ::fatal "trees ($treeish) $y != $x\n",
2082                                "Something is seriously wrong...";
2083                }
2084        });
2085}
2086
2087sub get_commit_parents {
2088        my ($self, $log_entry) = @_;
2089        my (%seen, @ret, @tmp);
2090        # legacy support for 'set-tree'; this is only used by set_tree_cb:
2091        if (my $ip = $self->{inject_parents}) {
2092                if (my $commit = delete $ip->{$log_entry->{revision}}) {
2093                        push @tmp, $commit;
2094                }
2095        }
2096        if (my $cur = ::verify_ref($self->refname.'^0')) {
2097                push @tmp, $cur;
2098        }
2099        if (my $ipd = $self->{inject_parents_dcommit}) {
2100                if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2101                        push @tmp, @$commit;
2102                }
2103        }
2104        push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2105        while (my $p = shift @tmp) {
2106                next if $seen{$p};
2107                $seen{$p} = 1;
2108                push @ret, $p;
2109                # MAXPARENT is defined to 16 in commit-tree.c:
2110                last if @ret >= 16;
2111        }
2112        if (@tmp) {
2113                die "r$log_entry->{revision}: No room for parents:\n\t",
2114                    join("\n\t", @tmp), "\n";
2115        }
2116        @ret;
2117}
2118
2119sub rewrite_root {
2120        my ($self) = @_;
2121        return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2122        my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2123        my $rwr = eval { command_oneline(qw/config --get/, $k) };
2124        if ($rwr) {
2125                $rwr =~ s#/+$##;
2126                if ($rwr !~ m#^[a-z\+]+://#) {
2127                        die "$rwr is not a valid URL (key: $k)\n";
2128                }
2129        }
2130        $self->{-rewrite_root} = $rwr;
2131}
2132
2133sub metadata_url {
2134        my ($self) = @_;
2135        ($self->rewrite_root || $self->{url}) .
2136           (length $self->{path} ? '/' . $self->{path} : '');
2137}
2138
2139sub full_url {
2140        my ($self) = @_;
2141        $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2142}
2143
2144
2145sub set_commit_header_env {
2146        my ($log_entry) = @_;
2147        my %env;
2148        foreach my $ned (qw/NAME EMAIL DATE/) {
2149                foreach my $ac (qw/AUTHOR COMMITTER/) {
2150                        $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2151                }
2152        }
2153
2154        $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2155        $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2156        $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2157
2158        $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2159                                                ? $log_entry->{commit_name}
2160                                                : $log_entry->{name};
2161        $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2162                                                ? $log_entry->{commit_email}
2163                                                : $log_entry->{email};
2164        \%env;
2165}
2166
2167sub restore_commit_header_env {
2168        my ($env) = @_;
2169        foreach my $ned (qw/NAME EMAIL DATE/) {
2170                foreach my $ac (qw/AUTHOR COMMITTER/) {
2171                        my $k = "GIT_${ac}_${ned}";
2172                        if (defined $env->{$k}) {
2173                                $ENV{$k} = $env->{$k};
2174                        } else {
2175                                delete $ENV{$k};
2176                        }
2177                }
2178        }
2179}
2180
2181sub gc {
2182        command_noisy('gc', '--auto');
2183};
2184
2185sub do_git_commit {
2186        my ($self, $log_entry) = @_;
2187        my $lr = $self->last_rev;
2188        if (defined $lr && $lr >= $log_entry->{revision}) {
2189                die "Last fetched revision of ", $self->refname,
2190                    " was r$lr, but we are about to fetch: ",
2191                    "r$log_entry->{revision}!\n";
2192        }
2193        if (my $c = $self->rev_map_get($log_entry->{revision})) {
2194                croak "$log_entry->{revision} = $c already exists! ",
2195                      "Why are we refetching it?\n";
2196        }
2197        my $old_env = set_commit_header_env($log_entry);
2198        my $tree = $log_entry->{tree};
2199        if (!defined $tree) {
2200                $tree = $self->tmp_index_do(sub {
2201                                            command_oneline('write-tree') });
2202        }
2203        die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2204
2205        my @exec = ('git-commit-tree', $tree);
2206        foreach ($self->get_commit_parents($log_entry)) {
2207                push @exec, '-p', $_;
2208        }
2209        defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2210                                                                   or croak $!;
2211        print $msg_fh $log_entry->{log} or croak $!;
2212        restore_commit_header_env($old_env);
2213        unless ($self->no_metadata) {
2214                print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2215                              or croak $!;
2216        }
2217        $msg_fh->flush == 0 or croak $!;
2218        close $msg_fh or croak $!;
2219        chomp(my $commit = do { local $/; <$out_fh> });
2220        close $out_fh or croak $!;
2221        waitpid $pid, 0;
2222        croak $? if $?;
2223        if ($commit !~ /^$::sha1$/o) {
2224                die "Failed to commit, invalid sha1: $commit\n";
2225        }
2226
2227        $self->rev_map_set($log_entry->{revision}, $commit, 1);
2228
2229        $self->{last_rev} = $log_entry->{revision};
2230        $self->{last_commit} = $commit;
2231        print "r$log_entry->{revision}";
2232        if (defined $log_entry->{svm_revision}) {
2233                 print " (\@$log_entry->{svm_revision})";
2234                 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2235                                   0, $self->svm_uuid);
2236        }
2237        print " = $commit ($self->{ref_id})\n";
2238        if (--$_gc_nr == 0) {
2239                $_gc_nr = $_gc_period;
2240                gc();
2241        }
2242        return $commit;
2243}
2244
2245sub match_paths {
2246        my ($self, $paths, $r) = @_;
2247        return 1 if $self->{path} eq '';
2248        if (my $path = $paths->{"/$self->{path}"}) {
2249                return ($path->{action} eq 'D') ? 0 : 1;
2250        }
2251        $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
2252        if (grep /$self->{path_regex}/, keys %$paths) {
2253                return 1;
2254        }
2255        my $c = '';
2256        foreach (split m#/#, $self->{path}) {
2257                $c .= "/$_";
2258                next unless ($paths->{$c} &&
2259                             ($paths->{$c}->{action} =~ /^[AR]$/));
2260                if ($self->ra->check_path($self->{path}, $r) ==
2261                    $SVN::Node::dir) {
2262                        return 1;
2263                }
2264        }
2265        return 0;
2266}
2267
2268sub find_parent_branch {
2269        my ($self, $paths, $rev) = @_;
2270        return undef unless $self->follow_parent;
2271        unless (defined $paths) {
2272                my $err_handler = $SVN::Error::handler;
2273                $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2274                $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1, sub {
2275                                   $paths =
2276                                      Git::SVN::Ra::dup_changed_paths($_[0]) });
2277                $SVN::Error::handler = $err_handler;
2278        }
2279        return undef unless defined $paths;
2280
2281        # look for a parent from another branch:
2282        my @b_path_components = split m#/#, $self->rel_path;
2283        my @a_path_components;
2284        my $i;
2285        while (@b_path_components) {
2286                $i = $paths->{'/'.join('/', @b_path_components)};
2287                last if $i && defined $i->{copyfrom_path};
2288                unshift(@a_path_components, pop(@b_path_components));
2289        }
2290        return undef unless defined $i && defined $i->{copyfrom_path};
2291        my $branch_from = $i->{copyfrom_path};
2292        if (@a_path_components) {
2293                print STDERR "branch_from: $branch_from => ";
2294                $branch_from .= '/'.join('/', @a_path_components);
2295                print STDERR $branch_from, "\n";
2296        }
2297        my $r = $i->{copyfrom_rev};
2298        my $repos_root = $self->ra->{repos_root};
2299        my $url = $self->ra->{url};
2300        my $new_url = $repos_root . $branch_from;
2301        print STDERR  "Found possible branch point: ",
2302                      "$new_url => ", $self->full_url, ", $r\n";
2303        $branch_from =~ s#^/##;
2304        my $gs = Git::SVN->find_by_url($new_url, $repos_root, $branch_from);
2305        unless ($gs) {
2306                my $ref_id = $self->{ref_id};
2307                $ref_id =~ s/\@\d+$//;
2308                $ref_id .= "\@$r";
2309                # just grow a tail if we're not unique enough :x
2310                $ref_id .= '-' while find_ref($ref_id);
2311                print STDERR "Initializing parent: $ref_id\n";
2312                my ($u, $p, $repo_id) = ($new_url, '', $ref_id);
2313                if ($u =~ s#^\Q$url\E(/|$)##) {
2314                        $p = $u;
2315                        $u = $url;
2316                        $repo_id = $self->{repo_id};
2317                }
2318                $gs = Git::SVN->init($u, $p, $repo_id, $ref_id, 1);
2319        }
2320        my ($r0, $parent) = $gs->find_rev_before($r, 1);
2321        if (!defined $r0 || !defined $parent) {
2322                my ($base, $head) = parse_revision_argument(0, $r);
2323                if ($base <= $r) {
2324                        $gs->fetch($base, $r);
2325                }
2326                ($r0, $parent) = $gs->last_rev_commit;
2327        }
2328        if (defined $r0 && defined $parent) {
2329                print STDERR "Found branch parent: ($self->{ref_id}) $parent\n";
2330                my $ed;
2331                if ($self->ra->can_do_switch) {
2332                        $self->assert_index_clean($parent);
2333                        print STDERR "Following parent with do_switch\n";
2334                        # do_switch works with svn/trunk >= r22312, but that
2335                        # is not included with SVN 1.4.3 (the latest version
2336                        # at the moment), so we can't rely on it
2337                        $self->{last_commit} = $parent;
2338                        $ed = SVN::Git::Fetcher->new($self);
2339                        $gs->ra->gs_do_switch($r0, $rev, $gs,
2340                                              $self->full_url, $ed)
2341                          or die "SVN connection failed somewhere...\n";
2342                } elsif ($self->ra->trees_match($new_url, $r0,
2343                                                $self->full_url, $rev)) {
2344                        print STDERR "Trees match:\n",
2345                                     "  $new_url\@$r0\n",
2346                                     "  ${\$self->full_url}\@$rev\n",
2347                                     "Following parent with no changes\n";
2348                        $self->tmp_index_do(sub {
2349                            command_noisy('read-tree', $parent);
2350                        });
2351                        $self->{last_commit} = $parent;
2352                } else {
2353                        print STDERR "Following parent with do_update\n";
2354                        $ed = SVN::Git::Fetcher->new($self);
2355                        $self->ra->gs_do_update($rev, $rev, $self, $ed)
2356                          or die "SVN connection failed somewhere...\n";
2357                }
2358                print STDERR "Successfully followed parent\n";
2359                return $self->make_log_entry($rev, [$parent], $ed);
2360        }
2361        return undef;
2362}
2363
2364sub do_fetch {
2365        my ($self, $paths, $rev) = @_;
2366        my $ed;
2367        my ($last_rev, @parents);
2368        if (my $lc = $self->last_commit) {
2369                # we can have a branch that was deleted, then re-added
2370                # under the same name but copied from another path, in
2371                # which case we'll have multiple parents (we don't
2372                # want to break the original ref, nor lose copypath info):
2373                if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2374                        push @{$log_entry->{parents}}, $lc;
2375                        return $log_entry;
2376                }
2377                $ed = SVN::Git::Fetcher->new($self);
2378                $last_rev = $self->{last_rev};
2379                $ed->{c} = $lc;
2380                @parents = ($lc);
2381        } else {
2382                $last_rev = $rev;
2383                if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2384                        return $log_entry;
2385                }
2386                $ed = SVN::Git::Fetcher->new($self);
2387        }
2388        unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2389                die "SVN connection failed somewhere...\n";
2390        }
2391        $self->make_log_entry($rev, \@parents, $ed);
2392}
2393
2394sub get_untracked {
2395        my ($self, $ed) = @_;
2396        my @out;
2397        my $h = $ed->{empty};
2398        foreach (sort keys %$h) {
2399                my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2400                push @out, "  $act: " . uri_encode($_);
2401                warn "W: $act: $_\n";
2402        }
2403        foreach my $t (qw/dir_prop file_prop/) {
2404                $h = $ed->{$t} or next;
2405                foreach my $path (sort keys %$h) {
2406                        my $ppath = $path eq '' ? '.' : $path;
2407                        foreach my $prop (sort keys %{$h->{$path}}) {
2408                                next if $SKIP_PROP{$prop};
2409                                my $v = $h->{$path}->{$prop};
2410                                my $t_ppath_prop = "$t: " .
2411                                                    uri_encode($ppath) . ' ' .
2412                                                    uri_encode($prop);
2413                                if (defined $v) {
2414                                        push @out, "  +$t_ppath_prop " .
2415                                                   uri_encode($v);
2416                                } else {
2417                                        push @out, "  -$t_ppath_prop";
2418                                }
2419                        }
2420                }
2421        }
2422        foreach my $t (qw/absent_file absent_directory/) {
2423                $h = $ed->{$t} or next;
2424                foreach my $parent (sort keys %$h) {
2425                        foreach my $path (sort @{$h->{$parent}}) {
2426                                push @out, "  $t: " .
2427                                           uri_encode("$parent/$path");
2428                                warn "W: $t: $parent/$path ",
2429                                     "Insufficient permissions?\n";
2430                        }
2431                }
2432        }
2433        \@out;
2434}
2435
2436sub parse_svn_date {
2437        my $date = shift || return '+0000 1970-01-01 00:00:00';
2438        my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2439                                            (\d\d)\:(\d\d)\:(\d\d).\d+Z$/x) or
2440                                         croak "Unable to parse date: $date\n";
2441        "+0000 $Y-$m-$d $H:$M:$S";
2442}
2443
2444sub check_author {
2445        my ($author) = @_;
2446        if (!defined $author || length $author == 0) {
2447                $author = '(no author)';
2448        } elsif (defined $::_authors && ! defined $::users{$author}) {
2449                die "Author: $author not defined in $::_authors file\n";
2450        }
2451        $author;
2452}
2453
2454sub make_log_entry {
2455        my ($self, $rev, $parents, $ed) = @_;
2456        my $untracked = $self->get_untracked($ed);
2457
2458        open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
2459        print $un "r$rev\n" or croak $!;
2460        print $un $_, "\n" foreach @$untracked;
2461        my %log_entry = ( parents => $parents || [], revision => $rev,
2462                          log => '');
2463
2464        my $headrev;
2465        my $logged = delete $self->{logged_rev_props};
2466        if (!$logged || $self->{-want_revprops}) {
2467                my $rp = $self->ra->rev_proplist($rev);
2468                foreach (sort keys %$rp) {
2469                        my $v = $rp->{$_};
2470                        if (/^svn:(author|date|log)$/) {
2471                                $log_entry{$1} = $v;
2472                        } elsif ($_ eq 'svm:headrev') {
2473                                $headrev = $v;
2474                        } else {
2475                                print $un "  rev_prop: ", uri_encode($_), ' ',
2476                                          uri_encode($v), "\n";
2477                        }
2478                }
2479        } else {
2480                map { $log_entry{$_} = $logged->{$_} } keys %$logged;
2481        }
2482        close $un or croak $!;
2483
2484        $log_entry{date} = parse_svn_date($log_entry{date});
2485        $log_entry{log} .= "\n";
2486        my $author = $log_entry{author} = check_author($log_entry{author});
2487        my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
2488                                                       : ($author, undef);
2489
2490        my ($commit_name, $commit_email) = ($name, $email);
2491        if ($_use_log_author) {
2492                my $name_field;
2493                if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
2494                        $name_field = $1;
2495                } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
2496                        $name_field = $1;
2497                }
2498                if (!defined $name_field) {
2499                        if (!defined $email) {
2500                                $email = $name;
2501                        }
2502                } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
2503                        ($name, $email) = ($1, $2);
2504                } elsif ($name_field =~ /(.*)@/) {
2505                        ($name, $email) = ($1, $name_field);
2506                } else {
2507                        ($name, $email) = ($name_field, $name_field);
2508                }
2509        }
2510        if (defined $headrev && $self->use_svm_props) {
2511                if ($self->rewrite_root) {
2512                        die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
2513                            "options set!\n";
2514                }
2515                my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$};
2516                # we don't want "SVM: initializing mirror for junk" ...
2517                return undef if $r == 0;
2518                my $svm = $self->svm;
2519                if ($uuid ne $svm->{uuid}) {
2520                        die "UUID mismatch on SVM path:\n",
2521                            "expected: $svm->{uuid}\n",
2522                            "     got: $uuid\n";
2523                }
2524                my $full_url = $self->full_url;
2525                $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
2526                             die "Failed to replace '$svm->{replace}' with ",
2527                                 "'$svm->{source}' in $full_url\n";
2528                # throw away username for storing in records
2529                remove_username($full_url);
2530                $log_entry{metadata} = "$full_url\@$r $uuid";
2531                $log_entry{svm_revision} = $r;
2532                $email ||= "$author\@$uuid";
2533                $commit_email ||= "$author\@$uuid";
2534        } elsif ($self->use_svnsync_props) {
2535                my $full_url = $self->svnsync->{url};
2536                $full_url .= "/$self->{path}" if length $self->{path};
2537                remove_username($full_url);
2538                my $uuid = $self->svnsync->{uuid};
2539                $log_entry{metadata} = "$full_url\@$rev $uuid";
2540                $email ||= "$author\@$uuid";
2541                $commit_email ||= "$author\@$uuid";
2542        } else {
2543                my $url = $self->metadata_url;
2544                remove_username($url);
2545                $log_entry{metadata} = "$url\@$rev " .
2546                                       $self->ra->get_uuid;
2547                $email ||= "$author\@" . $self->ra->get_uuid;
2548                $commit_email ||= "$author\@" . $self->ra->get_uuid;
2549        }
2550        $log_entry{name} = $name;
2551        $log_entry{email} = $email;
2552        $log_entry{commit_name} = $commit_name;
2553        $log_entry{commit_email} = $commit_email;
2554        \%log_entry;
2555}
2556
2557sub fetch {
2558        my ($self, $min_rev, $max_rev, @parents) = @_;
2559        my ($last_rev, $last_commit) = $self->last_rev_commit;
2560        my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
2561        $self->ra->gs_fetch_loop_common($base, $head, [$self]);
2562}
2563
2564sub set_tree_cb {
2565        my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
2566        $self->{inject_parents} = { $rev => $tree };
2567        $self->fetch(undef, undef);
2568}
2569
2570sub set_tree {
2571        my ($self, $tree) = (shift, shift);
2572        my $log_entry = ::get_commit_entry($tree);
2573        unless ($self->{last_rev}) {
2574                fatal("Must have an existing revision to commit");
2575        }
2576        my %ed_opts = ( r => $self->{last_rev},
2577                        log => $log_entry->{log},
2578                        ra => $self->ra,
2579                        tree_a => $self->{last_commit},
2580                        tree_b => $tree,
2581                        editor_cb => sub {
2582                               $self->set_tree_cb($log_entry, $tree, @_) },
2583                        svn_path => $self->{path} );
2584        if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
2585                print "No changes\nr$self->{last_rev} = $tree\n";
2586        }
2587}
2588
2589sub rebuild_from_rev_db {
2590        my ($self, $path) = @_;
2591        my $r = -1;
2592        open my $fh, '<', $path or croak "open: $!";
2593        binmode $fh or croak "binmode: $!";
2594        while (<$fh>) {
2595                length($_) == 41 or croak "inconsistent size in ($_) != 41";
2596                chomp($_);
2597                ++$r;
2598                next if $_ eq ('0' x 40);
2599                $self->rev_map_set($r, $_);
2600                print "r$r = $_\n";
2601        }
2602        close $fh or croak "close: $!";
2603        unlink $path or croak "unlink: $!";
2604}
2605
2606sub rebuild {
2607        my ($self) = @_;
2608        my $map_path = $self->map_path;
2609        return if (-e $map_path && ! -z $map_path);
2610        return unless ::verify_ref($self->refname.'^0');
2611        if ($self->use_svm_props || $self->no_metadata) {
2612                my $rev_db = $self->rev_db_path;
2613                $self->rebuild_from_rev_db($rev_db);
2614                if ($self->use_svm_props) {
2615                        my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
2616                        $self->rebuild_from_rev_db($svm_rev_db);
2617                }
2618                $self->unlink_rev_db_symlink;
2619                return;
2620        }
2621        print "Rebuilding $map_path ...\n";
2622        my ($log, $ctx) =
2623            command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
2624                                $self->refname, '--');
2625        my $metadata_url = $self->metadata_url;
2626        remove_username($metadata_url);
2627        my $svn_uuid = $self->ra_uuid;
2628        my $c;
2629        while (<$log>) {
2630                if ( m{^commit ($::sha1)$} ) {
2631                        $c = $1;
2632                        next;
2633                }
2634                next unless s{^\s*(git-svn-id:)}{$1};
2635                my ($url, $rev, $uuid) = ::extract_metadata($_);
2636                remove_username($url);
2637
2638                # ignore merges (from set-tree)
2639                next if (!defined $rev || !$uuid);
2640
2641                # if we merged or otherwise started elsewhere, this is
2642                # how we break out of it
2643                if (($uuid ne $svn_uuid) ||
2644                    ($metadata_url && $url && ($url ne $metadata_url))) {
2645                        next;
2646                }
2647
2648                $self->rev_map_set($rev, $c);
2649                print "r$rev = $c\n";
2650        }
2651        command_close_pipe($log, $ctx);
2652        print "Done rebuilding $map_path\n";
2653        my $rev_db_path = $self->rev_db_path;
2654        if (-f $self->rev_db_path) {
2655                unlink $self->rev_db_path or croak "unlink: $!";
2656        }
2657        $self->unlink_rev_db_symlink;
2658}
2659
2660# rev_map:
2661# Tie::File seems to be prone to offset errors if revisions get sparse,
2662# it's not that fast, either.  Tie::File is also not in Perl 5.6.  So
2663# one of my favorite modules is out :<  Next up would be one of the DBM
2664# modules, but I'm not sure which is most portable...
2665#
2666# This is the replacement for the rev_db format, which was too big
2667# and inefficient for large repositories with a lot of sparse history
2668# (mainly tags)
2669#
2670# The format is this:
2671#   - 24 bytes for every record,
2672#     * 4 bytes for the integer representing an SVN revision number
2673#     * 20 bytes representing the sha1 of a git commit
2674#   - No empty padding records like the old format
2675#     (except the last record, which can be overwritten)
2676#   - new records are written append-only since SVN revision numbers
2677#     increase monotonically
2678#   - lookups on SVN revision number are done via a binary search
2679#   - Piping the file to xxd -c24 is a good way of dumping it for
2680#     viewing or editing (piped back through xxd -r), should the need
2681#     ever arise.
2682#   - The last record can be padding revision with an all-zero sha1
2683#     This is used to optimize fetch performance when using multiple
2684#     "fetch" directives in .git/config
2685#
2686# These files are disposable unless noMetadata or useSvmProps is set
2687
2688sub _rev_map_set {
2689        my ($fh, $rev, $commit) = @_;
2690
2691        binmode $fh or croak "binmode: $!";
2692        my $size = (stat($fh))[7];
2693        ($size % 24) == 0 or croak "inconsistent size: $size";
2694
2695        my $wr_offset = 0;
2696        if ($size > 0) {
2697                sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2698                my $read = sysread($fh, my $buf, 24) or croak "read: $!";
2699                $read == 24 or croak "read only $read bytes (!= 24)";
2700                my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
2701                if ($last_commit eq ('0' x40)) {
2702                        if ($size >= 48) {
2703                                sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2704                                $read = sysread($fh, $buf, 24) or
2705                                    croak "read: $!";
2706                                $read == 24 or
2707                                    croak "read only $read bytes (!= 24)";
2708                                ($last_rev, $last_commit) =
2709                                    unpack(rev_map_fmt, $buf);
2710                                if ($last_commit eq ('0' x40)) {
2711                                        croak "inconsistent .rev_map\n";
2712                                }
2713                        }
2714                        if ($last_rev >= $rev) {
2715                                croak "last_rev is higher!: $last_rev >= $rev";
2716                        }
2717                        $wr_offset = -24;
2718                }
2719        }
2720        sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
2721        syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
2722          croak "write: $!";
2723}
2724
2725sub mkfile {
2726        my ($path) = @_;
2727        unless (-e $path) {
2728                my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
2729                mkpath([$dir]) unless -d $dir;
2730                open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
2731                close $fh or die "Couldn't close (create) $path: $!\n";
2732        }
2733}
2734
2735sub rev_map_set {
2736        my ($self, $rev, $commit, $update_ref, $uuid) = @_;
2737        length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
2738        my $db = $self->map_path($uuid);
2739        my $db_lock = "$db.lock";
2740        my $sig;
2741        if ($update_ref) {
2742                $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2743                            $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
2744        }
2745        mkfile($db);
2746
2747        $LOCKFILES{$db_lock} = 1;
2748        my $sync;
2749        # both of these options make our .rev_db file very, very important
2750        # and we can't afford to lose it because rebuild() won't work
2751        if ($self->use_svm_props || $self->no_metadata) {
2752                $sync = 1;
2753                copy($db, $db_lock) or die "rev_map_set(@_): ",
2754                                           "Failed to copy: ",
2755                                           "$db => $db_lock ($!)\n";
2756        } else {
2757                rename $db, $db_lock or die "rev_map_set(@_): ",
2758                                            "Failed to rename: ",
2759                                            "$db => $db_lock ($!)\n";
2760        }
2761
2762        sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
2763             or croak "Couldn't open $db_lock: $!\n";
2764        _rev_map_set($fh, $rev, $commit);
2765        if ($sync) {
2766                $fh->flush or die "Couldn't flush $db_lock: $!\n";
2767                $fh->sync or die "Couldn't sync $db_lock: $!\n";
2768        }
2769        close $fh or croak $!;
2770        if ($update_ref) {
2771                $_head = $self;
2772                command_noisy('update-ref', '-m', "r$rev",
2773                              $self->refname, $commit);
2774        }
2775        rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
2776                                    "$db_lock => $db ($!)\n";
2777        delete $LOCKFILES{$db_lock};
2778        if ($update_ref) {
2779                $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2780                            $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
2781                kill $sig, $$ if defined $sig;
2782        }
2783}
2784
2785# If want_commit, this will return an array of (rev, commit) where
2786# commit _must_ be a valid commit in the archive.
2787# Otherwise, it'll return the max revision (whether or not the
2788# commit is valid or just a 0x40 placeholder).
2789sub rev_map_max {
2790        my ($self, $want_commit) = @_;
2791        $self->rebuild;
2792        my $map_path = $self->map_path;
2793        stat $map_path or return $want_commit ? (0, undef) : 0;
2794        sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2795        binmode $fh or croak "binmode: $!";
2796        my $size = (stat($fh))[7];
2797        ($size % 24) == 0 or croak "inconsistent size: $size";
2798
2799        if ($size == 0) {
2800                close $fh or croak "close: $!";
2801                return $want_commit ? (0, undef) : 0;
2802        }
2803
2804        sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2805        sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2806        my ($r, $c) = unpack(rev_map_fmt, $buf);
2807        if ($want_commit && $c eq ('0' x40)) {
2808                if ($size < 48) {
2809                        return $want_commit ? (0, undef) : 0;
2810                }
2811                sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2812                sysread($fh, $buf, 24) == 24 or croak "read: $!";
2813                ($r, $c) = unpack(rev_map_fmt, $buf);
2814                if ($c eq ('0'x40)) {
2815                        croak "Penultimate record is all-zeroes in $map_path";
2816                }
2817        }
2818        close $fh or croak "close: $!";
2819        $want_commit ? ($r, $c) : $r;
2820}
2821
2822sub rev_map_get {
2823        my ($self, $rev, $uuid) = @_;
2824        my $map_path = $self->map_path($uuid);
2825        return undef unless -e $map_path;
2826
2827        sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2828        binmode $fh or croak "binmode: $!";
2829        my $size = (stat($fh))[7];
2830        ($size % 24) == 0 or croak "inconsistent size: $size";
2831
2832        if ($size == 0) {
2833                close $fh or croak "close: $fh";
2834                return undef;
2835        }
2836
2837        my ($l, $u) = (0, $size - 24);
2838        my ($r, $c, $buf);
2839
2840        while ($l <= $u) {
2841                my $i = int(($l/24 + $u/24) / 2) * 24;
2842                sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
2843                sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2844                my ($r, $c) = unpack('NH40', $buf);
2845
2846                if ($r < $rev) {
2847                        $l = $i + 24;
2848                } elsif ($r > $rev) {
2849                        $u = $i - 24;
2850                } else { # $r == $rev
2851                        close($fh) or croak "close: $!";
2852                        return $c eq ('0' x 40) ? undef : $c;
2853                }
2854        }
2855        close($fh) or croak "close: $!";
2856        undef;
2857}
2858
2859# Finds the first svn revision that exists on (if $eq_ok is true) or
2860# before $rev for the current branch.  It will not search any lower
2861# than $min_rev.  Returns the git commit hash and svn revision number
2862# if found, else (undef, undef).
2863sub find_rev_before {
2864        my ($self, $rev, $eq_ok, $min_rev) = @_;
2865        --$rev unless $eq_ok;
2866        $min_rev ||= 1;
2867        while ($rev >= $min_rev) {
2868                if (my $c = $self->rev_map_get($rev)) {
2869                        return ($rev, $c);
2870                }
2871                --$rev;
2872        }
2873        return (undef, undef);
2874}
2875
2876# Finds the first svn revision that exists on (if $eq_ok is true) or
2877# after $rev for the current branch.  It will not search any higher
2878# than $max_rev.  Returns the git commit hash and svn revision number
2879# if found, else (undef, undef).
2880sub find_rev_after {
2881        my ($self, $rev, $eq_ok, $max_rev) = @_;
2882        ++$rev unless $eq_ok;
2883        $max_rev ||= $self->rev_map_max;
2884        while ($rev <= $max_rev) {
2885                if (my $c = $self->rev_map_get($rev)) {
2886                        return ($rev, $c);
2887                }
2888                ++$rev;
2889        }
2890        return (undef, undef);
2891}
2892
2893sub _new {
2894        my ($class, $repo_id, $ref_id, $path) = @_;
2895        unless (defined $repo_id && length $repo_id) {
2896                $repo_id = $Git::SVN::default_repo_id;
2897        }
2898        unless (defined $ref_id && length $ref_id) {
2899                $_[2] = $ref_id = $Git::SVN::default_ref_id;
2900        }
2901        $_[1] = $repo_id;
2902        my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
2903        $_[3] = $path = '' unless (defined $path);
2904        mkpath(["$ENV{GIT_DIR}/svn"]);
2905        bless {
2906                ref_id => $ref_id, dir => $dir, index => "$dir/index",
2907                path => $path, config => "$ENV{GIT_DIR}/svn/config",
2908                map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
2909}
2910
2911# for read-only access of old .rev_db formats
2912sub unlink_rev_db_symlink {
2913        my ($self) = @_;
2914        my $link = $self->rev_db_path;
2915        $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
2916        if (-l $link) {
2917                unlink $link or croak "unlink: $link failed!";
2918        }
2919}
2920
2921sub rev_db_path {
2922        my ($self, $uuid) = @_;
2923        my $db_path = $self->map_path($uuid);
2924        $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
2925            or croak "map_path: $db_path does not contain '/.rev_map.' !";
2926        $db_path;
2927}
2928
2929# the new replacement for .rev_db
2930sub map_path {
2931        my ($self, $uuid) = @_;
2932        $uuid ||= $self->ra_uuid;
2933        "$self->{map_root}.$uuid";
2934}
2935
2936sub uri_encode {
2937        my ($f) = @_;
2938        $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
2939        $f
2940}
2941
2942sub remove_username {
2943        $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
2944}
2945
2946package Git::SVN::Prompt;
2947use strict;
2948use warnings;
2949require SVN::Core;
2950use vars qw/$_no_auth_cache $_username/;
2951
2952sub simple {
2953        my ($cred, $realm, $default_username, $may_save, $pool) = @_;
2954        $may_save = undef if $_no_auth_cache;
2955        $default_username = $_username if defined $_username;
2956        if (defined $default_username && length $default_username) {
2957                if (defined $realm && length $realm) {
2958                        print STDERR "Authentication realm: $realm\n";
2959                        STDERR->flush;
2960                }
2961                $cred->username($default_username);
2962        } else {
2963                username($cred, $realm, $may_save, $pool);
2964        }
2965        $cred->password(_read_password("Password for '" .
2966                                       $cred->username . "': ", $realm));
2967        $cred->may_save($may_save);
2968        $SVN::_Core::SVN_NO_ERROR;
2969}
2970
2971sub ssl_server_trust {
2972        my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
2973        $may_save = undef if $_no_auth_cache;
2974        print STDERR "Error validating server certificate for '$realm':\n";
2975        {
2976                no warnings 'once';
2977                # All variables SVN::Auth::SSL::* are used only once,
2978                # so we're shutting up Perl warnings about this.
2979                if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
2980                        print STDERR " - The certificate is not issued ",
2981                            "by a trusted authority. Use the\n",
2982                            "   fingerprint to validate ",
2983                            "the certificate manually!\n";
2984                }
2985                if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
2986                        print STDERR " - The certificate hostname ",
2987                            "does not match.\n";
2988                }
2989                if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
2990                        print STDERR " - The certificate is not yet valid.\n";
2991                }
2992                if ($failures & $SVN::Auth::SSL::EXPIRED) {
2993                        print STDERR " - The certificate has expired.\n";
2994                }
2995                if ($failures & $SVN::Auth::SSL::OTHER) {
2996                        print STDERR " - The certificate has ",
2997                            "an unknown error.\n";
2998                }
2999        } # no warnings 'once'
3000        printf STDERR
3001                "Certificate information:\n".
3002                " - Hostname: %s\n".
3003                " - Valid: from %s until %s\n".
3004                " - Issuer: %s\n".
3005                " - Fingerprint: %s\n",
3006                map $cert_info->$_, qw(hostname valid_from valid_until
3007                                       issuer_dname fingerprint);
3008        my $choice;
3009prompt:
3010        print STDERR $may_save ?
3011              "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
3012              "(R)eject or accept (t)emporarily? ";
3013        STDERR->flush;
3014        $choice = lc(substr(<STDIN> || 'R', 0, 1));
3015        if ($choice =~ /^t$/i) {
3016                $cred->may_save(undef);
3017        } elsif ($choice =~ /^r$/i) {
3018                return -1;
3019        } elsif ($may_save && $choice =~ /^p$/i) {
3020                $cred->may_save($may_save);
3021        } else {
3022                goto prompt;
3023        }
3024        $cred->accepted_failures($failures);
3025        $SVN::_Core::SVN_NO_ERROR;
3026}
3027
3028sub ssl_client_cert {
3029        my ($cred, $realm, $may_save, $pool) = @_;
3030        $may_save = undef if $_no_auth_cache;
3031        print STDERR "Client certificate filename: ";
3032        STDERR->flush;
3033        chomp(my $filename = <STDIN>);
3034        $cred->cert_file($filename);
3035        $cred->may_save($may_save);
3036        $SVN::_Core::SVN_NO_ERROR;
3037}
3038
3039sub ssl_client_cert_pw {
3040        my ($cred, $realm, $may_save, $pool) = @_;
3041        $may_save = undef if $_no_auth_cache;
3042        $cred->password(_read_password("Password: ", $realm));
3043        $cred->may_save($may_save);
3044        $SVN::_Core::SVN_NO_ERROR;
3045}
3046
3047sub username {
3048        my ($cred, $realm, $may_save, $pool) = @_;
3049        $may_save = undef if $_no_auth_cache;
3050        if (defined $realm && length $realm) {
3051                print STDERR "Authentication realm: $realm\n";
3052        }
3053        my $username;
3054        if (defined $_username) {
3055                $username = $_username;
3056        } else {
3057                print STDERR "Username: ";
3058                STDERR->flush;
3059                chomp($username = <STDIN>);
3060        }
3061        $cred->username($username);
3062        $cred->may_save($may_save);
3063        $SVN::_Core::SVN_NO_ERROR;
3064}
3065
3066sub _read_password {
3067        my ($prompt, $realm) = @_;
3068        print STDERR $prompt;
3069        STDERR->flush;
3070        require Term::ReadKey;
3071        Term::ReadKey::ReadMode('noecho');
3072        my $password = '';
3073        while (defined(my $key = Term::ReadKey::ReadKey(0))) {
3074                last if $key =~ /[\012\015]/; # \n\r
3075                $password .= $key;
3076        }
3077        Term::ReadKey::ReadMode('restore');
3078        print STDERR "\n";
3079        STDERR->flush;
3080        $password;
3081}
3082
3083package SVN::Git::Fetcher;
3084use vars qw/@ISA/;
3085use strict;
3086use warnings;
3087use Carp qw/croak/;
3088use File::Temp qw/tempfile/;
3089use IO::File qw//;
3090
3091# file baton members: path, mode_a, mode_b, pool, fh, blob, base
3092sub new {
3093        my ($class, $git_svn) = @_;
3094        my $self = SVN::Delta::Editor->new;
3095        bless $self, $class;
3096        $self->{c} = $git_svn->{last_commit} if exists $git_svn->{last_commit};
3097        $self->{empty} = {};
3098        $self->{dir_prop} = {};
3099        $self->{file_prop} = {};
3100        $self->{absent_dir} = {};
3101        $self->{absent_file} = {};
3102        $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
3103        $self;
3104}
3105
3106sub set_path_strip {
3107        my ($self, $path) = @_;
3108        $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
3109}
3110
3111sub open_root {
3112        { path => '' };
3113}
3114
3115sub open_directory {
3116        my ($self, $path, $pb, $rev) = @_;
3117        { path => $path };
3118}
3119
3120sub git_path {
3121        my ($self, $path) = @_;
3122        if ($self->{path_strip}) {
3123                $path =~ s!$self->{path_strip}!! or
3124                  die "Failed to strip path '$path' ($self->{path_strip})\n";
3125        }
3126        $path;
3127}
3128
3129sub delete_entry {
3130        my ($self, $path, $rev, $pb) = @_;
3131
3132        my $gpath = $self->git_path($path);
3133        return undef if ($gpath eq '');
3134
3135        # remove entire directories.
3136        if (command('ls-tree', $self->{c}, '--', $gpath) =~ /^040000 tree/) {
3137                my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3138                                                     -r --name-only -z/,
3139                                                     $self->{c}, '--', $gpath);
3140                local $/ = "\0";
3141                while (<$ls>) {
3142                        chomp;
3143                        $self->{gii}->remove($_);
3144                        print "\tD\t$_\n" unless $::_q;
3145                }
3146                print "\tD\t$gpath/\n" unless $::_q;
3147                command_close_pipe($ls, $ctx);
3148                $self->{empty}->{$path} = 0
3149        } else {
3150                $self->{gii}->remove($gpath);
3151                print "\tD\t$gpath\n" unless $::_q;
3152        }
3153        undef;
3154}
3155
3156sub open_file {
3157        my ($self, $path, $pb, $rev) = @_;
3158        my $gpath = $self->git_path($path);
3159        my ($mode, $blob) = (command('ls-tree', $self->{c}, '--', $gpath)
3160                             =~ /^(\d{6}) blob ([a-f\d]{40})\t/);
3161        unless (defined $mode && defined $blob) {
3162                die "$path was not found in commit $self->{c} (r$rev)\n";
3163        }
3164        { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3165          pool => SVN::Pool->new, action => 'M' };
3166}
3167
3168sub add_file {
3169        my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3170        my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3171        delete $self->{empty}->{$dir};
3172        { path => $path, mode_a => 100644, mode_b => 100644,
3173          pool => SVN::Pool->new, action => 'A' };
3174}
3175
3176sub add_directory {
3177        my ($self, $path, $cp_path, $cp_rev) = @_;
3178        my $gpath = $self->git_path($path);
3179        if ($gpath eq '') {
3180                my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3181                                                     -r --name-only -z/,
3182                                                     $self->{c});
3183                local $/ = "\0";
3184                while (<$ls>) {
3185                        chomp;
3186                        $self->{gii}->remove($_);
3187                        print "\tD\t$_\n" unless $::_q;
3188                }
3189                command_close_pipe($ls, $ctx);
3190                $self->{empty}->{$path} = 0;
3191        }
3192        my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3193        delete $self->{empty}->{$dir};
3194        $self->{empty}->{$path} = 1;
3195        { path => $path };
3196}
3197
3198sub change_dir_prop {
3199        my ($self, $db, $prop, $value) = @_;
3200        $self->{dir_prop}->{$db->{path}} ||= {};
3201        $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3202        undef;
3203}
3204
3205sub absent_directory {
3206        my ($self, $path, $pb) = @_;
3207        $self->{absent_dir}->{$pb->{path}} ||= [];
3208        push @{$self->{absent_dir}->{$pb->{path}}}, $path;
3209        undef;
3210}
3211
3212sub absent_file {
3213        my ($self, $path, $pb) = @_;
3214        $self->{absent_file}->{$pb->{path}} ||= [];
3215        push @{$self->{absent_file}->{$pb->{path}}}, $path;
3216        undef;
3217}
3218
3219sub change_file_prop {
3220        my ($self, $fb, $prop, $value) = @_;
3221        if ($prop eq 'svn:executable') {
3222                if ($fb->{mode_b} != 120000) {
3223                        $fb->{mode_b} = defined $value ? 100755 : 100644;
3224                }
3225        } elsif ($prop eq 'svn:special') {
3226                $fb->{mode_b} = defined $value ? 120000 : 100644;
3227        } else {
3228                $self->{file_prop}->{$fb->{path}} ||= {};
3229                $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
3230        }
3231        undef;
3232}
3233
3234sub apply_textdelta {
3235        my ($self, $fb, $exp) = @_;
3236        my $fh = Git::temp_acquire('svn_delta');
3237        # $fh gets auto-closed() by SVN::TxDelta::apply(),
3238        # (but $base does not,) so dup() it for reading in close_file
3239        open my $dup, '<&', $fh or croak $!;
3240        my $base = Git::temp_acquire('git_blob');
3241        if ($fb->{blob}) {
3242                print $base 'link ' if ($fb->{mode_a} == 120000);
3243                my $size = $::_repository->cat_blob($fb->{blob}, $base);
3244                die "Failed to read object $fb->{blob}" if ($size < 0);
3245
3246                if (defined $exp) {
3247                        seek $base, 0, 0 or croak $!;
3248                        my $got = ::md5sum($base);
3249                        die "Checksum mismatch: $fb->{path} $fb->{blob}\n",
3250                            "expected: $exp\n",
3251                            "     got: $got\n" if ($got ne $exp);
3252                }
3253        }
3254        seek $base, 0, 0 or croak $!;
3255        $fb->{fh} = $fh;
3256        $fb->{base} = $base;
3257        [ SVN::TxDelta::apply($base, $dup, undef, $fb->{path}, $fb->{pool}) ];
3258}
3259
3260sub close_file {
3261        my ($self, $fb, $exp) = @_;
3262        my $hash;
3263        my $path = $self->git_path($fb->{path});
3264        if (my $fh = $fb->{fh}) {
3265                if (defined $exp) {
3266                        seek($fh, 0, 0) or croak $!;
3267                        my $got = ::md5sum($fh);
3268                        if ($got ne $exp) {
3269                                die "Checksum mismatch: $path\n",
3270                                    "expected: $exp\n    got: $got\n";
3271                        }
3272                }
3273                if ($fb->{mode_b} == 120000) {
3274                        sysseek($fh, 0, 0) or croak $!;
3275                        sysread($fh, my $buf, 5) == 5 or croak $!;
3276
3277                        unless ($buf eq 'link ') {
3278                                warn "$path has mode 120000",
3279                                                " but is not a link\n";
3280                        } else {
3281                                my $tmp_fh = Git::temp_acquire('svn_hash');
3282                                my $res;
3283                                while ($res = sysread($fh, my $str, 1024)) {
3284                                        my $out = syswrite($tmp_fh, $str, $res);
3285                                        defined($out) && $out == $res
3286                                                or croak("write ",
3287                                                        $tmp_fh->filename,
3288                                                        ": $!\n");
3289                                }
3290                                defined $res or croak $!;
3291
3292                                ($fh, $tmp_fh) = ($tmp_fh, $fh);
3293                                Git::temp_release($tmp_fh, 1);
3294                        }
3295                }
3296
3297                $hash = $::_repository->hash_and_insert_object(
3298                                $fh->filename);
3299                $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
3300
3301                Git::temp_release($fb->{base}, 1);
3302                Git::temp_release($fh, 1);
3303        } else {
3304                $hash = $fb->{blob} or die "no blob information\n";
3305        }
3306        $fb->{pool}->clear;
3307        $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
3308        print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
3309        undef;
3310}
3311
3312sub abort_edit {
3313        my $self = shift;
3314        $self->{nr} = $self->{gii}->{nr};
3315        delete $self->{gii};
3316        $self->SUPER::abort_edit(@_);
3317}
3318
3319sub close_edit {
3320        my $self = shift;
3321        $self->{git_commit_ok} = 1;
3322        $self->{nr} = $self->{gii}->{nr};
3323        delete $self->{gii};
3324        $self->SUPER::close_edit(@_);
3325}
3326
3327package SVN::Git::Editor;
3328use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
3329use strict;
3330use warnings;
3331use Carp qw/croak/;
3332use IO::File;
3333
3334sub new {
3335        my ($class, $opts) = @_;
3336        foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
3337                die "$_ required!\n" unless (defined $opts->{$_});
3338        }
3339
3340        my $pool = SVN::Pool->new;
3341        my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
3342        my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
3343                                     $opts->{r}, $mods);
3344
3345        # $opts->{ra} functions should not be used after this:
3346        my @ce  = $opts->{ra}->get_commit_editor($opts->{log},
3347                                                $opts->{editor_cb}, $pool);
3348        my $self = SVN::Delta::Editor->new(@ce, $pool);
3349        bless $self, $class;
3350        foreach (qw/svn_path r tree_a tree_b/) {
3351                $self->{$_} = $opts->{$_};
3352        }
3353        $self->{url} = $opts->{ra}->{url};
3354        $self->{mods} = $mods;
3355        $self->{types} = $types;
3356        $self->{pool} = $pool;
3357        $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
3358        $self->{rm} = { };
3359        $self->{path_prefix} = length $self->{svn_path} ?
3360                               "$self->{svn_path}/" : '';
3361        $self->{config} = $opts->{config};
3362        return $self;
3363}
3364
3365sub generate_diff {
3366        my ($tree_a, $tree_b) = @_;
3367        my @diff_tree = qw(diff-tree -z -r);
3368        if ($_cp_similarity) {
3369                push @diff_tree, "-C$_cp_similarity";
3370        } else {
3371                push @diff_tree, '-C';
3372        }
3373        push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
3374        push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
3375        push @diff_tree, $tree_a, $tree_b;
3376        my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
3377        local $/ = "\0";
3378        my $state = 'meta';
3379        my @mods;
3380        while (<$diff_fh>) {
3381                chomp $_; # this gets rid of the trailing "\0"
3382                if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
3383                                        $::sha1\s($::sha1)\s
3384                                        ([MTCRAD])\d*$/xo) {
3385                        push @mods, {   mode_a => $1, mode_b => $2,
3386                                        sha1_b => $3, chg => $4 };
3387                        if ($4 =~ /^(?:C|R)$/) {
3388                                $state = 'file_a';
3389                        } else {
3390                                $state = 'file_b';
3391                        }
3392                } elsif ($state eq 'file_a') {
3393                        my $x = $mods[$#mods] or croak "Empty array\n";
3394                        if ($x->{chg} !~ /^(?:C|R)$/) {
3395                                croak "Error parsing $_, $x->{chg}\n";
3396                        }
3397                        $x->{file_a} = $_;
3398                        $state = 'file_b';
3399                } elsif ($state eq 'file_b') {
3400                        my $x = $mods[$#mods] or croak "Empty array\n";
3401                        if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
3402                                croak "Error parsing $_, $x->{chg}\n";
3403                        }
3404                        if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
3405                                croak "Error parsing $_, $x->{chg}\n";
3406                        }
3407                        $x->{file_b} = $_;
3408                        $state = 'meta';
3409                } else {
3410                        croak "Error parsing $_\n";
3411                }
3412        }
3413        command_close_pipe($diff_fh, $ctx);
3414        \@mods;
3415}
3416
3417sub check_diff_paths {
3418        my ($ra, $pfx, $rev, $mods) = @_;
3419        my %types;
3420        $pfx .= '/' if length $pfx;
3421
3422        sub type_diff_paths {
3423                my ($ra, $types, $path, $rev) = @_;
3424                my @p = split m#/+#, $path;
3425                my $c = shift @p;
3426                unless (defined $types->{$c}) {
3427                        $types->{$c} = $ra->check_path($c, $rev);
3428                }
3429                while (@p) {
3430                        $c .= '/' . shift @p;
3431                        next if defined $types->{$c};
3432                        $types->{$c} = $ra->check_path($c, $rev);
3433                }
3434        }
3435
3436        foreach my $m (@$mods) {
3437                foreach my $f (qw/file_a file_b/) {
3438                        next unless defined $m->{$f};
3439                        my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
3440                        if (length $pfx.$dir && ! defined $types{$dir}) {
3441                                type_diff_paths($ra, \%types, $pfx.$dir, $rev);
3442                        }
3443                }
3444        }
3445        \%types;
3446}
3447
3448sub split_path {
3449        return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
3450}
3451
3452sub repo_path {
3453        my ($self, $path) = @_;
3454        $self->{path_prefix}.(defined $path ? $path : '');
3455}
3456
3457sub url_path {
3458        my ($self, $path) = @_;
3459        if ($self->{url} =~ m#^https?://#) {
3460                $path =~ s/([^a-zA-Z0-9_.-])/uc sprintf("%%%02x",ord($1))/eg;
3461        }
3462        $self->{url} . '/' . $self->repo_path($path);
3463}
3464
3465sub rmdirs {
3466        my ($self) = @_;
3467        my $rm = $self->{rm};
3468        delete $rm->{''}; # we never delete the url we're tracking
3469        return unless %$rm;
3470
3471        foreach (keys %$rm) {
3472                my @d = split m#/#, $_;
3473                my $c = shift @d;
3474                $rm->{$c} = 1;
3475                while (@d) {
3476                        $c .= '/' . shift @d;
3477                        $rm->{$c} = 1;
3478                }
3479        }
3480        delete $rm->{$self->{svn_path}};
3481        delete $rm->{''}; # we never delete the url we're tracking
3482        return unless %$rm;
3483
3484        my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
3485                                             $self->{tree_b});
3486        local $/ = "\0";
3487        while (<$fh>) {
3488                chomp;
3489                my @dn = split m#/#, $_;
3490                while (pop @dn) {
3491                        delete $rm->{join '/', @dn};
3492                }
3493                unless (%$rm) {
3494                        close $fh;
3495                        return;
3496                }
3497        }
3498        command_close_pipe($fh, $ctx);
3499
3500        my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
3501        foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
3502                $self->close_directory($bat->{$d}, $p);
3503                my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
3504                print "\tD+\t$d/\n" unless $::_q;
3505                $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
3506                delete $bat->{$d};
3507        }
3508}
3509
3510sub open_or_add_dir {
3511        my ($self, $full_path, $baton) = @_;
3512        my $t = $self->{types}->{$full_path};
3513        if (!defined $t) {
3514                die "$full_path not known in r$self->{r} or we have a bug!\n";
3515        }
3516        {
3517                no warnings 'once';
3518                # SVN::Node::none and SVN::Node::file are used only once,
3519                # so we're shutting up Perl's warnings about them.
3520                if ($t == $SVN::Node::none) {
3521                        return $self->add_directory($full_path, $baton,
3522                            undef, -1, $self->{pool});
3523                } elsif ($t == $SVN::Node::dir) {
3524                        return $self->open_directory($full_path, $baton,
3525                            $self->{r}, $self->{pool});
3526                } # no warnings 'once'
3527                print STDERR "$full_path already exists in repository at ",
3528                    "r$self->{r} and it is not a directory (",
3529                    ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
3530        } # no warnings 'once'
3531        exit 1;
3532}
3533
3534sub ensure_path {
3535        my ($self, $path) = @_;
3536        my $bat = $self->{bat};
3537        my $repo_path = $self->repo_path($path);
3538        return $bat->{''} unless (length $repo_path);
3539        my @p = split m#/+#, $repo_path;
3540        my $c = shift @p;
3541        $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
3542        while (@p) {
3543                my $c0 = $c;
3544                $c .= '/' . shift @p;
3545                $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
3546        }
3547        return $bat->{$c};
3548}
3549
3550# Subroutine to convert a globbing pattern to a regular expression.
3551# From perl cookbook.
3552sub glob2pat {
3553        my $globstr = shift;
3554        my %patmap = ('*' => '.*', '?' => '.', '[' => '[', ']' => ']');
3555        $globstr =~ s{(.)} { $patmap{$1} || "\Q$1" }ge;
3556        return '^' . $globstr . '$';
3557}
3558
3559sub check_autoprop {
3560        my ($self, $pattern, $properties, $file, $fbat) = @_;
3561        # Convert the globbing pattern to a regular expression.
3562        my $regex = glob2pat($pattern);
3563        # Check if the pattern matches the file name.
3564        if($file =~ m/($regex)/) {
3565                # Parse the list of properties to set.
3566                my @props = split(/;/, $properties);
3567                foreach my $prop (@props) {
3568                        # Parse 'name=value' syntax and set the property.
3569                        if ($prop =~ /([^=]+)=(.*)/) {
3570                                my ($n,$v) = ($1,$2);
3571                                for ($n, $v) {
3572                                        s/^\s+//; s/\s+$//;
3573                                }
3574                                $self->change_file_prop($fbat, $n, $v);
3575                        }
3576                }
3577        }
3578}
3579
3580sub apply_autoprops {
3581        my ($self, $file, $fbat) = @_;
3582        my $conf_t = ${$self->{config}}{'config'};
3583        no warnings 'once';
3584        # Check [miscellany]/enable-auto-props in svn configuration.
3585        if (SVN::_Core::svn_config_get_bool(
3586                $conf_t,
3587                $SVN::_Core::SVN_CONFIG_SECTION_MISCELLANY,
3588                $SVN::_Core::SVN_CONFIG_OPTION_ENABLE_AUTO_PROPS,
3589                0)) {
3590                # Auto-props are enabled.  Enumerate them to look for matches.
3591                my $callback = sub {
3592                        $self->check_autoprop($_[0], $_[1], $file, $fbat);
3593                };
3594                SVN::_Core::svn_config_enumerate(
3595                        $conf_t,
3596                        $SVN::_Core::SVN_CONFIG_SECTION_AUTO_PROPS,
3597                        $callback);
3598        }
3599}
3600
3601sub A {
3602        my ($self, $m) = @_;
3603        my ($dir, $file) = split_path($m->{file_b});
3604        my $pbat = $self->ensure_path($dir);
3605        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3606                                        undef, -1);
3607        print "\tA\t$m->{file_b}\n" unless $::_q;
3608        $self->apply_autoprops($file, $fbat);
3609        $self->chg_file($fbat, $m);
3610        $self->close_file($fbat,undef,$self->{pool});
3611}
3612
3613sub C {
3614        my ($self, $m) = @_;
3615        my ($dir, $file) = split_path($m->{file_b});
3616        my $pbat = $self->ensure_path($dir);
3617        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3618                                $self->url_path($m->{file_a}), $self->{r});
3619        print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3620        $self->chg_file($fbat, $m);
3621        $self->close_file($fbat,undef,$self->{pool});
3622}
3623
3624sub delete_entry {
3625        my ($self, $path, $pbat) = @_;
3626        my $rpath = $self->repo_path($path);
3627        my ($dir, $file) = split_path($rpath);
3628        $self->{rm}->{$dir} = 1;
3629        $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
3630}
3631
3632sub R {
3633        my ($self, $m) = @_;
3634        my ($dir, $file) = split_path($m->{file_b});
3635        my $pbat = $self->ensure_path($dir);
3636        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3637                                $self->url_path($m->{file_a}), $self->{r});
3638        print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3639        $self->chg_file($fbat, $m);
3640        $self->close_file($fbat,undef,$self->{pool});
3641
3642        ($dir, $file) = split_path($m->{file_a});
3643        $pbat = $self->ensure_path($dir);
3644        $self->delete_entry($m->{file_a}, $pbat);
3645}
3646
3647sub M {
3648        my ($self, $m) = @_;
3649        my ($dir, $file) = split_path($m->{file_b});
3650        my $pbat = $self->ensure_path($dir);
3651        my $fbat = $self->open_file($self->repo_path($m->{file_b}),
3652                                $pbat,$self->{r},$self->{pool});
3653        print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
3654        $self->chg_file($fbat, $m);
3655        $self->close_file($fbat,undef,$self->{pool});
3656}
3657
3658sub T { shift->M(@_) }
3659
3660sub change_file_prop {
3661        my ($self, $fbat, $pname, $pval) = @_;
3662        $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
3663}
3664
3665sub chg_file {
3666        my ($self, $fbat, $m) = @_;
3667        if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
3668                $self->change_file_prop($fbat,'svn:executable','*');
3669        } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
3670                $self->change_file_prop($fbat,'svn:executable',undef);
3671        }
3672        my $fh = Git::temp_acquire('git_blob');
3673        if ($m->{mode_b} =~ /^120/) {
3674                print $fh 'link ' or croak $!;
3675                $self->change_file_prop($fbat,'svn:special','*');
3676        } elsif ($m->{mode_a} =~ /^120/ && $m->{mode_b} !~ /^120/) {
3677                $self->change_file_prop($fbat,'svn:special',undef);
3678        }
3679        my $size = $::_repository->cat_blob($m->{sha1_b}, $fh);
3680        croak "Failed to read object $m->{sha1_b}" if ($size < 0);
3681        $fh->flush == 0 or croak $!;
3682        seek $fh, 0, 0 or croak $!;
3683
3684        my $exp = ::md5sum($fh);
3685        seek $fh, 0, 0 or croak $!;
3686
3687        my $pool = SVN::Pool->new;
3688        my $atd = $self->apply_textdelta($fbat, undef, $pool);
3689        my $got = SVN::TxDelta::send_stream($fh, @$atd, $pool);
3690        die "Checksum mismatch\nexpected: $exp\ngot: $got\n" if ($got ne $exp);
3691        Git::temp_release($fh, 1);
3692        $pool->clear;
3693}
3694
3695sub D {
3696        my ($self, $m) = @_;
3697        my ($dir, $file) = split_path($m->{file_b});
3698        my $pbat = $self->ensure_path($dir);
3699        print "\tD\t$m->{file_b}\n" unless $::_q;
3700        $self->delete_entry($m->{file_b}, $pbat);
3701}
3702
3703sub close_edit {
3704        my ($self) = @_;
3705        my ($p,$bat) = ($self->{pool}, $self->{bat});
3706        foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
3707                next if $_ eq '';
3708                $self->close_directory($bat->{$_}, $p);
3709        }
3710        $self->close_directory($bat->{''}, $p);
3711        $self->SUPER::close_edit($p);
3712        $p->clear;
3713}
3714
3715sub abort_edit {
3716        my ($self) = @_;
3717        $self->SUPER::abort_edit($self->{pool});
3718}
3719
3720sub DESTROY {
3721        my $self = shift;
3722        $self->SUPER::DESTROY(@_);
3723        $self->{pool}->clear;
3724}
3725
3726# this drives the editor
3727sub apply_diff {
3728        my ($self) = @_;
3729        my $mods = $self->{mods};
3730        my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
3731        foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
3732                my $f = $m->{chg};
3733                if (defined $o{$f}) {
3734                        $self->$f($m);
3735                } else {
3736                        fatal("Invalid change type: $f");
3737                }
3738        }
3739        $self->rmdirs if $_rmdir;
3740        if (@$mods == 0) {
3741                $self->abort_edit;
3742        } else {
3743                $self->close_edit;
3744        }
3745        return scalar @$mods;
3746}
3747
3748package Git::SVN::Ra;
3749use vars qw/@ISA $config_dir $_log_window_size/;
3750use strict;
3751use warnings;
3752my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
3753
3754BEGIN {
3755        # enforce temporary pool usage for some simple functions
3756        no strict 'refs';
3757        for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root/) {
3758                my $SUPER = "SUPER::$f";
3759                *$f = sub {
3760                        my $self = shift;
3761                        my $pool = SVN::Pool->new;
3762                        my @ret = $self->$SUPER(@_,$pool);
3763                        $pool->clear;
3764                        wantarray ? @ret : $ret[0];
3765                };
3766        }
3767}
3768
3769sub _auth_providers () {
3770        [
3771          SVN::Client::get_simple_provider(),
3772          SVN::Client::get_ssl_server_trust_file_provider(),
3773          SVN::Client::get_simple_prompt_provider(
3774            \&Git::SVN::Prompt::simple, 2),
3775          SVN::Client::get_ssl_client_cert_file_provider(),
3776          SVN::Client::get_ssl_client_cert_prompt_provider(
3777            \&Git::SVN::Prompt::ssl_client_cert, 2),
3778          SVN::Client::get_ssl_client_cert_pw_file_provider(),
3779          SVN::Client::get_ssl_client_cert_pw_prompt_provider(
3780            \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
3781          SVN::Client::get_username_provider(),
3782          SVN::Client::get_ssl_server_trust_prompt_provider(
3783            \&Git::SVN::Prompt::ssl_server_trust),
3784          SVN::Client::get_username_prompt_provider(
3785            \&Git::SVN::Prompt::username, 2)
3786        ]
3787}
3788
3789sub escape_uri_only {
3790        my ($uri) = @_;
3791        my @tmp;
3792        foreach (split m{/}, $uri) {
3793                s/([^\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
3794                push @tmp, $_;
3795        }
3796        join('/', @tmp);
3797}
3798
3799sub escape_url {
3800        my ($url) = @_;
3801        if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
3802                my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
3803                $url = "$scheme://$domain$uri";
3804        }
3805        $url;
3806}
3807
3808sub new {
3809        my ($class, $url) = @_;
3810        $url =~ s!/+$!!;
3811        return $RA if ($RA && $RA->{url} eq $url);
3812
3813        SVN::_Core::svn_config_ensure($config_dir, undef);
3814        my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
3815        my $config = SVN::Core::config_get_config($config_dir);
3816        $RA = undef;
3817        my $dont_store_passwords = 1;
3818        my $conf_t = ${$config}{'config'};
3819        {
3820                no warnings 'once';
3821                # The usage of $SVN::_Core::SVN_CONFIG_* variables
3822                # produces warnings that variables are used only once.
3823                # I had not found the better way to shut them up, so
3824                # the warnings of type 'once' are disabled in this block.
3825                if (SVN::_Core::svn_config_get_bool($conf_t,
3826                    $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3827                    $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
3828                    1) == 0) {
3829                        SVN::_Core::svn_auth_set_parameter($baton,
3830                            $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
3831                            bless (\$dont_store_passwords, "_p_void"));
3832                }
3833                if (SVN::_Core::svn_config_get_bool($conf_t,
3834                    $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3835                    $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
3836                    1) == 0) {
3837                        $Git::SVN::Prompt::_no_auth_cache = 1;
3838                }
3839        } # no warnings 'once'
3840        my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
3841                              config => $config,
3842                              pool => SVN::Pool->new,
3843                              auth_provider_callbacks => $callbacks);
3844        $self->{url} = $url;
3845        $self->{svn_path} = $url;
3846        $self->{repos_root} = $self->get_repos_root;
3847        $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
3848        $self->{cache} = { check_path => { r => 0, data => {} },
3849                           get_dir => { r => 0, data => {} } };
3850        $RA = bless $self, $class;
3851}
3852
3853sub check_path {
3854        my ($self, $path, $r) = @_;
3855        my $cache = $self->{cache}->{check_path};
3856        if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
3857                return $cache->{data}->{$path};
3858        }
3859        my $pool = SVN::Pool->new;
3860        my $t = $self->SUPER::check_path($path, $r, $pool);
3861        $pool->clear;
3862        if ($r != $cache->{r}) {
3863                %{$cache->{data}} = ();
3864                $cache->{r} = $r;
3865        }
3866        $cache->{data}->{$path} = $t;
3867}
3868
3869sub get_dir {
3870        my ($self, $dir, $r) = @_;
3871        my $cache = $self->{cache}->{get_dir};
3872        if ($r == $cache->{r}) {
3873                if (my $x = $cache->{data}->{$dir}) {
3874                        return wantarray ? @$x : $x->[0];
3875                }
3876        }
3877        my $pool = SVN::Pool->new;
3878        my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
3879        my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
3880        $pool->clear;
3881        if ($r != $cache->{r}) {
3882                %{$cache->{data}} = ();
3883                $cache->{r} = $r;
3884        }
3885        $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
3886        wantarray ? (\%dirents, $r, $props) : \%dirents;
3887}
3888
3889sub DESTROY {
3890        # do not call the real DESTROY since we store ourselves in $RA
3891}
3892
3893sub get_log {
3894        my ($self, @args) = @_;
3895        my $pool = SVN::Pool->new;
3896        splice(@args, 3, 1) if ($SVN::Core::VERSION le '1.2.0');
3897        my $ret = $self->SUPER::get_log(@args, $pool);
3898        $pool->clear;
3899        $ret;
3900}
3901
3902sub trees_match {
3903        my ($self, $url1, $rev1, $url2, $rev2) = @_;
3904        my $ctx = SVN::Client->new(auth => _auth_providers);
3905        my $out = IO::File->new_tmpfile;
3906
3907        # older SVN (1.1.x) doesn't take $pool as the last parameter for
3908        # $ctx->diff(), so we'll create a default one
3909        my $pool = SVN::Pool->new_default_sub;
3910
3911        $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
3912        $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
3913        $out->flush;
3914        my $ret = (($out->stat)[7] == 0);
3915        close $out or croak $!;
3916
3917        $ret;
3918}
3919
3920sub get_commit_editor {
3921        my ($self, $log, $cb, $pool) = @_;
3922        my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
3923        $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
3924}
3925
3926sub gs_do_update {
3927        my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
3928        my $new = ($rev_a == $rev_b);
3929        my $path = $gs->{path};
3930
3931        if ($new && -e $gs->{index}) {
3932                unlink $gs->{index} or die
3933                  "Couldn't unlink index: $gs->{index}: $!\n";
3934        }
3935        my $pool = SVN::Pool->new;
3936        $editor->set_path_strip($path);
3937        my (@pc) = split m#/#, $path;
3938        my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
3939                                        1, $editor, $pool);
3940        my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3941
3942        # Since we can't rely on svn_ra_reparent being available, we'll
3943        # just have to do some magic with set_path to make it so
3944        # we only want a partial path.
3945        my $sp = '';
3946        my $final = join('/', @pc);
3947        while (@pc) {
3948                $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
3949                $sp .= '/' if length $sp;
3950                $sp .= shift @pc;
3951        }
3952        die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
3953
3954        $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
3955
3956        $reporter->finish_report($pool);
3957        $pool->clear;
3958        $editor->{git_commit_ok};
3959}
3960
3961# this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
3962# svn_ra_reparent didn't work before 1.4)
3963sub gs_do_switch {
3964        my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
3965        my $path = $gs->{path};
3966        my $pool = SVN::Pool->new;
3967
3968        my $full_url = $self->{url};
3969        my $old_url = $full_url;
3970        $full_url .= '/' . escape_uri_only($path) if length $path;
3971        my ($ra, $reparented);
3972        if ($old_url ne $full_url) {
3973                if ($old_url !~ m#^svn(\+ssh)?://#) {
3974                        SVN::_Ra::svn_ra_reparent($self->{session}, $full_url,
3975                                                  $pool);
3976                        $self->{url} = $full_url;
3977                        $reparented = 1;
3978                } else {
3979                        $_[0] = undef;
3980                        $self = undef;
3981                        $RA = undef;
3982                        $ra = Git::SVN::Ra->new($full_url);
3983                        $ra_invalid = 1;
3984                }
3985        }
3986        $ra ||= $self;
3987        my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
3988        my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3989        $reporter->set_path('', $rev_a, 0, @lock, $pool);
3990        $reporter->finish_report($pool);
3991
3992        if ($reparented) {
3993                SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
3994                $self->{url} = $old_url;
3995        }
3996
3997        $pool->clear;
3998        $editor->{git_commit_ok};
3999}
4000
4001sub longest_common_path {
4002        my ($gsv, $globs) = @_;
4003        my %common;
4004        my $common_max = scalar @$gsv;
4005
4006        foreach my $gs (@$gsv) {
4007                my @tmp = split m#/#, $gs->{path};
4008                my $p = '';
4009                foreach (@tmp) {
4010                        $p .= length($p) ? "/$_" : $_;
4011                        $common{$p} ||= 0;
4012                        $common{$p}++;
4013                }
4014        }
4015        $globs ||= [];
4016        $common_max += scalar @$globs;
4017        foreach my $glob (@$globs) {
4018                my @tmp = split m#/#, $glob->{path}->{left};
4019                my $p = '';
4020                foreach (@tmp) {
4021                        $p .= length($p) ? "/$_" : $_;
4022                        $common{$p} ||= 0;
4023                        $common{$p}++;
4024                }
4025        }
4026
4027        my $longest_path = '';
4028        foreach (sort {length $b <=> length $a} keys %common) {
4029                if ($common{$_} == $common_max) {
4030                        $longest_path = $_;
4031                        last;
4032                }
4033        }
4034        $longest_path;
4035}
4036
4037sub gs_fetch_loop_common {
4038        my ($self, $base, $head, $gsv, $globs) = @_;
4039        return if ($base > $head);
4040        my $inc = $_log_window_size;
4041        my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
4042        my $longest_path = longest_common_path($gsv, $globs);
4043        my $ra_url = $self->{url};
4044        while (1) {
4045                my %revs;
4046                my $err;
4047                my $err_handler = $SVN::Error::handler;
4048                $SVN::Error::handler = sub {
4049                        ($err) = @_;
4050                        skip_unknown_revs($err);
4051                };
4052                sub _cb {
4053                        my ($paths, $r, $author, $date, $log) = @_;
4054                        [ dup_changed_paths($paths),
4055                          { author => $author, date => $date, log => $log } ];
4056                }
4057                $self->get_log([$longest_path], $min, $max, 0, 1, 1,
4058                               sub { $revs{$_[1]} = _cb(@_) });
4059                if ($err && $max >= $head) {
4060                        print STDERR "Path '$longest_path' ",
4061                                     "was probably deleted:\n",
4062                                     $err->expanded_message,
4063                                     "\nWill attempt to follow ",
4064                                     "revisions r$min .. r$max ",
4065                                     "committed before the deletion\n";
4066                        my $hi = $max;
4067                        while (--$hi >= $min) {
4068                                my $ok;
4069                                $self->get_log([$longest_path], $min, $hi,
4070                                               0, 1, 1, sub {
4071                                               $ok ||= $_[1];
4072                                               $revs{$_[1]} = _cb(@_) });
4073                                if ($ok) {
4074                                        print STDERR "r$min .. r$ok OK\n";
4075                                        last;
4076                                }
4077                        }
4078                }
4079                $SVN::Error::handler = $err_handler;
4080
4081                my %exists = map { $_->{path} => $_ } @$gsv;
4082                foreach my $r (sort {$a <=> $b} keys %revs) {
4083                        my ($paths, $logged) = @{$revs{$r}};
4084
4085                        foreach my $gs ($self->match_globs(\%exists, $paths,
4086                                                           $globs, $r)) {
4087                                if ($gs->rev_map_max >= $r) {
4088                                        next;
4089                                }
4090                                next unless $gs->match_paths($paths, $r);
4091                                $gs->{logged_rev_props} = $logged;
4092                                if (my $last_commit = $gs->last_commit) {
4093                                        $gs->assert_index_clean($last_commit);
4094                                }
4095                                my $log_entry = $gs->do_fetch($paths, $r);
4096                                if ($log_entry) {
4097                                        $gs->do_git_commit($log_entry);
4098                                }
4099                                $INDEX_FILES{$gs->{index}} = 1;
4100                        }
4101                        foreach my $g (@$globs) {
4102                                my $k = "svn-remote.$g->{remote}." .
4103                                        "$g->{t}-maxRev";
4104                                Git::SVN::tmp_config($k, $r);
4105                        }
4106                        if ($ra_invalid) {
4107                                $_[0] = undef;
4108                                $self = undef;
4109                                $RA = undef;
4110                                $self = Git::SVN::Ra->new($ra_url);
4111                                $ra_invalid = undef;
4112                        }
4113                }
4114                # pre-fill the .rev_db since it'll eventually get filled in
4115                # with '0' x40 if something new gets committed
4116                foreach my $gs (@$gsv) {
4117                        next if $gs->rev_map_max >= $max;
4118                        next if defined $gs->rev_map_get($max);
4119                        $gs->rev_map_set($max, 0 x40);
4120                }
4121                foreach my $g (@$globs) {
4122                        my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
4123                        Git::SVN::tmp_config($k, $max);
4124                }
4125                last if $max >= $head;
4126                $min = $max + 1;
4127                $max += $inc;
4128                $max = $head if ($max > $head);
4129        }
4130        Git::SVN::gc();
4131}
4132
4133sub get_dir_globbed {
4134        my ($self, $left, $depth, $r) = @_;
4135
4136        my @x = eval { $self->get_dir($left, $r) };
4137        return unless scalar @x == 3;
4138        my $dirents = $x[0];
4139        my @finalents;
4140        foreach my $de (keys %$dirents) {
4141                next if $dirents->{$de}->{kind} != $SVN::Node::dir;
4142                if ($depth > 1) {
4143                        my @args = ("$left/$de", $depth - 1, $r);
4144                        foreach my $dir ($self->get_dir_globbed(@args)) {
4145                                push @finalents, "$de/$dir";
4146                        }
4147                } else {
4148                        push @finalents, $de;
4149                }
4150        }
4151        @finalents;
4152}
4153
4154sub match_globs {
4155        my ($self, $exists, $paths, $globs, $r) = @_;
4156
4157        sub get_dir_check {
4158                my ($self, $exists, $g, $r) = @_;
4159
4160                my @dirs = $self->get_dir_globbed($g->{path}->{left},
4161                                                  $g->{path}->{depth},
4162                                                  $r);
4163
4164                foreach my $de (@dirs) {
4165                        my $p = $g->{path}->full_path($de);
4166                        next if $exists->{$p};
4167                        next if (length $g->{path}->{right} &&
4168                                 ($self->check_path($p, $r) !=
4169                                  $SVN::Node::dir));
4170                        $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
4171                                         $g->{ref}->full_path($de), 1);
4172                }
4173        }
4174        foreach my $g (@$globs) {
4175                if (my $path = $paths->{"/$g->{path}->{left}"}) {
4176                        if ($path->{action} =~ /^[AR]$/) {
4177                                get_dir_check($self, $exists, $g, $r);
4178                        }
4179                }
4180                foreach (keys %$paths) {
4181                        if (/$g->{path}->{left_regex}/ &&
4182                            !/$g->{path}->{regex}/) {
4183                                next if $paths->{$_}->{action} !~ /^[AR]$/;
4184                                get_dir_check($self, $exists, $g, $r);
4185                        }
4186                        next unless /$g->{path}->{regex}/;
4187                        my $p = $1;
4188                        my $pathname = $g->{path}->full_path($p);
4189                        next if $exists->{$pathname};
4190                        next if ($self->check_path($pathname, $r) !=
4191                                 $SVN::Node::dir);
4192                        $exists->{$pathname} = Git::SVN->init(
4193                                              $self->{url}, $pathname, undef,
4194                                              $g->{ref}->full_path($p), 1);
4195                }
4196                my $c = '';
4197                foreach (split m#/#, $g->{path}->{left}) {
4198                        $c .= "/$_";
4199                        next unless ($paths->{$c} &&
4200                                     ($paths->{$c}->{action} =~ /^[AR]$/));
4201                        get_dir_check($self, $exists, $g, $r);
4202                }
4203        }
4204        values %$exists;
4205}
4206
4207sub minimize_url {
4208        my ($self) = @_;
4209        return $self->{url} if ($self->{url} eq $self->{repos_root});
4210        my $url = $self->{repos_root};
4211        my @components = split(m!/!, $self->{svn_path});
4212        my $c = '';
4213        do {
4214                $url .= "/$c" if length $c;
4215                eval { (ref $self)->new($url)->get_latest_revnum };
4216        } while ($@ && ($c = shift @components));
4217        $url;
4218}
4219
4220sub can_do_switch {
4221        my $self = shift;
4222        unless (defined $can_do_switch) {
4223                my $pool = SVN::Pool->new;
4224                my $rep = eval {
4225                        $self->do_switch(1, '', 0, $self->{url},
4226                                         SVN::Delta::Editor->new, $pool);
4227                };
4228                if ($@) {
4229                        $can_do_switch = 0;
4230                } else {
4231                        $rep->abort_report($pool);
4232                        $can_do_switch = 1;
4233                }
4234                $pool->clear;
4235        }
4236        $can_do_switch;
4237}
4238
4239sub skip_unknown_revs {
4240        my ($err) = @_;
4241        my $errno = $err->apr_err();
4242        # Maybe the branch we're tracking didn't
4243        # exist when the repo started, so it's
4244        # not an error if it doesn't, just continue
4245        #
4246        # Wonderfully consistent library, eh?
4247        # 160013 - svn:// and file://
4248        # 175002 - http(s)://
4249        # 175007 - http(s):// (this repo required authorization, too...)
4250        #   More codes may be discovered later...
4251        if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
4252                my $err_key = $err->expanded_message;
4253                # revision numbers change every time, filter them out
4254                $err_key =~ s/\d+/\0/g;
4255                $err_key = "$errno\0$err_key";
4256                unless ($ignored_err{$err_key}) {
4257                        warn "W: Ignoring error from SVN, path probably ",
4258                             "does not exist: ($errno): ",
4259                             $err->expanded_message,"\n";
4260                        warn "W: Do not be alarmed at the above message ",
4261                             "git-svn is just searching aggressively for ",
4262                             "old history.\n",
4263                             "This may take a while on large repositories\n";
4264                        $ignored_err{$err_key} = 1;
4265                }
4266                return;
4267        }
4268        die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
4269}
4270
4271# svn_log_changed_path_t objects passed to get_log are likely to be
4272# overwritten even if only the refs are copied to an external variable,
4273# so we should dup the structures in their entirety.  Using an externally
4274# passed pool (instead of our temporary and quickly cleared pool in
4275# Git::SVN::Ra) does not help matters at all...
4276sub dup_changed_paths {
4277        my ($paths) = @_;
4278        return undef unless $paths;
4279        my %ret;
4280        foreach my $p (keys %$paths) {
4281                my $i = $paths->{$p};
4282                my %s = map { $_ => $i->$_ }
4283                              qw/copyfrom_path copyfrom_rev action/;
4284                $ret{$p} = \%s;
4285        }
4286        \%ret;
4287}
4288
4289package Git::SVN::Log;
4290use strict;
4291use warnings;
4292use POSIX qw/strftime/;
4293use constant commit_log_separator => ('-' x 72) . "\n";
4294use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
4295            %rusers $show_commit $incremental/;
4296my $l_fmt;
4297
4298sub cmt_showable {
4299        my ($c) = @_;
4300        return 1 if defined $c->{r};
4301
4302        # big commit message got truncated by the 16k pretty buffer in rev-list
4303        if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
4304                                $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
4305                @{$c->{l}} = ();
4306                my @log = command(qw/cat-file commit/, $c->{c});
4307
4308                # shift off the headers
4309                shift @log while ($log[0] ne '');
4310                shift @log;
4311
4312                # TODO: make $c->{l} not have a trailing newline in the future
4313                @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
4314
4315                (undef, $c->{r}, undef) = ::extract_metadata(
4316                                (grep(/^git-svn-id: /, @log))[-1]);
4317        }
4318        return defined $c->{r};
4319}
4320
4321sub log_use_color {
4322        return $color || Git->repository->get_colorbool('color.diff');
4323}
4324
4325sub git_svn_log_cmd {
4326        my ($r_min, $r_max, @args) = @_;
4327        my $head = 'HEAD';
4328        my (@files, @log_opts);
4329        foreach my $x (@args) {
4330                if ($x eq '--' || @files) {
4331                        push @files, $x;
4332                } else {
4333                        if (::verify_ref("$x^0")) {
4334                                $head = $x;
4335                        } else {
4336                                push @log_opts, $x;
4337                        }
4338                }
4339        }
4340
4341        my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
4342        $gs ||= Git::SVN->_new;
4343        my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
4344                   $gs->refname);
4345        push @cmd, '-r' unless $non_recursive;
4346        push @cmd, qw/--raw --name-status/ if $verbose;
4347        push @cmd, '--color' if log_use_color();
4348        push @cmd, @log_opts;
4349        if (defined $r_max && $r_max == $r_min) {
4350                push @cmd, '--max-count=1';
4351                if (my $c = $gs->rev_map_get($r_max)) {
4352                        push @cmd, $c;
4353                }
4354        } elsif (defined $r_max) {
4355                if ($r_max < $r_min) {
4356                        ($r_min, $r_max) = ($r_max, $r_min);
4357                }
4358                my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
4359                my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
4360                # If there are no commits in the range, both $c_max and $c_min
4361                # will be undefined.  If there is at least 1 commit in the
4362                # range, both will be defined.
4363                return () if !defined $c_min || !defined $c_max;
4364                if ($c_min eq $c_max) {
4365                        push @cmd, '--max-count=1', $c_min;
4366                } else {
4367                        push @cmd, '--boundary', "$c_min..$c_max";
4368                }
4369        }
4370        return (@cmd, @files);
4371}
4372
4373# adapted from pager.c
4374sub config_pager {
4375        $pager ||= $ENV{GIT_PAGER} || $ENV{PAGER};
4376        if (!defined $pager) {
4377                $pager = 'less';
4378        } elsif (length $pager == 0 || $pager eq 'cat') {
4379                $pager = undef;
4380        }
4381        $ENV{GIT_PAGER_IN_USE} = defined($pager);
4382}
4383
4384sub run_pager {
4385        return unless -t *STDOUT && defined $pager;
4386        pipe my $rfd, my $wfd or return;
4387        defined(my $pid = fork) or ::fatal "Can't fork: $!";
4388        if (!$pid) {
4389                open STDOUT, '>&', $wfd or
4390                                     ::fatal "Can't redirect to stdout: $!";
4391                return;
4392        }
4393        open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
4394        $ENV{LESS} ||= 'FRSX';
4395        exec $pager or ::fatal "Can't run pager: $! ($pager)";
4396}
4397
4398sub format_svn_date {
4399        return strftime("%Y-%m-%d %H:%M:%S %z (%a, %d %b %Y)", localtime(shift));
4400}
4401
4402sub parse_git_date {
4403        my ($t, $tz) = @_;
4404        # Date::Parse isn't in the standard Perl distro :(
4405        if ($tz =~ s/^\+//) {
4406                $t += tz_to_s_offset($tz);
4407        } elsif ($tz =~ s/^\-//) {
4408                $t -= tz_to_s_offset($tz);
4409        }
4410        return $t;
4411}
4412
4413sub set_local_timezone {
4414        if (defined $TZ) {
4415                $ENV{TZ} = $TZ;
4416        } else {
4417                delete $ENV{TZ};
4418        }
4419}
4420
4421sub tz_to_s_offset {
4422        my ($tz) = @_;
4423        $tz =~ s/(\d\d)$//;
4424        return ($1 * 60) + ($tz * 3600);
4425}
4426
4427sub get_author_info {
4428        my ($dest, $author, $t, $tz) = @_;
4429        $author =~ s/(?:^\s*|\s*$)//g;
4430        $dest->{a_raw} = $author;
4431        my $au;
4432        if ($::_authors) {
4433                $au = $rusers{$author} || undef;
4434        }
4435        if (!$au) {
4436                ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
4437        }
4438        $dest->{t} = $t;
4439        $dest->{tz} = $tz;
4440        $dest->{a} = $au;
4441        $dest->{t_utc} = parse_git_date($t, $tz);
4442}
4443
4444sub process_commit {
4445        my ($c, $r_min, $r_max, $defer) = @_;
4446        if (defined $r_min && defined $r_max) {
4447                if ($r_min == $c->{r} && $r_min == $r_max) {
4448                        show_commit($c);
4449                        return 0;
4450                }
4451                return 1 if $r_min == $r_max;
4452                if ($r_min < $r_max) {
4453                        # we need to reverse the print order
4454                        return 0 if (defined $limit && --$limit < 0);
4455                        push @$defer, $c;
4456                        return 1;
4457                }
4458                if ($r_min != $r_max) {
4459                        return 1 if ($r_min < $c->{r});
4460                        return 1 if ($r_max > $c->{r});
4461                }
4462        }
4463        return 0 if (defined $limit && --$limit < 0);
4464        show_commit($c);
4465        return 1;
4466}
4467
4468sub show_commit {
4469        my $c = shift;
4470        if ($oneline) {
4471                my $x = "\n";
4472                if (my $l = $c->{l}) {
4473                        while ($l->[0] =~ /^\s*$/) { shift @$l }
4474                        $x = $l->[0];
4475                }
4476                $l_fmt ||= 'A' . length($c->{r});
4477                print 'r',pack($l_fmt, $c->{r}),' | ';
4478                print "$c->{c} | " if $show_commit;
4479                print $x;
4480        } else {
4481                show_commit_normal($c);
4482        }
4483}
4484
4485sub show_commit_changed_paths {
4486        my ($c) = @_;
4487        return unless $c->{changed};
4488        print "Changed paths:\n", @{$c->{changed}};
4489}
4490
4491sub show_commit_normal {
4492        my ($c) = @_;
4493        print commit_log_separator, "r$c->{r} | ";
4494        print "$c->{c} | " if $show_commit;
4495        print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
4496        my $nr_line = 0;
4497
4498        if (my $l = $c->{l}) {
4499                while ($l->[$#$l] eq "\n" && $#$l > 0
4500                                          && $l->[($#$l - 1)] eq "\n") {
4501                        pop @$l;
4502                }
4503                $nr_line = scalar @$l;
4504                if (!$nr_line) {
4505                        print "1 line\n\n\n";
4506                } else {
4507                        if ($nr_line == 1) {
4508                                $nr_line = '1 line';
4509                        } else {
4510                                $nr_line .= ' lines';
4511                        }
4512                        print $nr_line, "\n";
4513                        show_commit_changed_paths($c);
4514                        print "\n";
4515                        print $_ foreach @$l;
4516                }
4517        } else {
4518                print "1 line\n";
4519                show_commit_changed_paths($c);
4520                print "\n";
4521
4522        }
4523        foreach my $x (qw/raw stat diff/) {
4524                if ($c->{$x}) {
4525                        print "\n";
4526                        print $_ foreach @{$c->{$x}}
4527                }
4528        }
4529}
4530
4531sub cmd_show_log {
4532        my (@args) = @_;
4533        my ($r_min, $r_max);
4534        my $r_last = -1; # prevent dupes
4535        set_local_timezone();
4536        if (defined $::_revision) {
4537                if ($::_revision =~ /^(\d+):(\d+)$/) {
4538                        ($r_min, $r_max) = ($1, $2);
4539                } elsif ($::_revision =~ /^\d+$/) {
4540                        $r_min = $r_max = $::_revision;
4541                } else {
4542                        ::fatal "-r$::_revision is not supported, use ",
4543                                "standard 'git log' arguments instead";
4544                }
4545        }
4546
4547        config_pager();
4548        @args = git_svn_log_cmd($r_min, $r_max, @args);
4549        if (!@args) {
4550                print commit_log_separator unless $incremental || $oneline;
4551                return;
4552        }
4553        my $log = command_output_pipe(@args);
4554        run_pager();
4555        my (@k, $c, $d, $stat);
4556        my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
4557        while (<$log>) {
4558                if (/^${esc_color}commit -?($::sha1_short)/o) {
4559                        my $cmt = $1;
4560                        if ($c && cmt_showable($c) && $c->{r} != $r_last) {
4561                                $r_last = $c->{r};
4562                                process_commit($c, $r_min, $r_max, \@k) or
4563                                                                goto out;
4564                        }
4565                        $d = undef;
4566                        $c = { c => $cmt };
4567                } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
4568                        get_author_info($c, $1, $2, $3);
4569                } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
4570                        # ignore
4571                } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
4572                        push @{$c->{raw}}, $_;
4573                } elsif (/^${esc_color}[ACRMDT]\t/) {
4574                        # we could add $SVN->{svn_path} here, but that requires
4575                        # remote access at the moment (repo_path_split)...
4576                        s#^(${esc_color})([ACRMDT])\t#$1   $2 #o;
4577                        push @{$c->{changed}}, $_;
4578                } elsif (/^${esc_color}diff /o) {
4579                        $d = 1;
4580                        push @{$c->{diff}}, $_;
4581                } elsif ($d) {
4582                        push @{$c->{diff}}, $_;
4583                } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
4584                          $esc_color*[\+\-]*$esc_color$/x) {
4585                        $stat = 1;
4586                        push @{$c->{stat}}, $_;
4587                } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
4588                        push @{$c->{stat}}, $_;
4589                        $stat = undef;
4590                } elsif (/^${esc_color}    (git-svn-id:.+)$/o) {
4591                        ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
4592                } elsif (s/^${esc_color}    //o) {
4593                        push @{$c->{l}}, $_;
4594                }
4595        }
4596        if ($c && defined $c->{r} && $c->{r} != $r_last) {
4597                $r_last = $c->{r};
4598                process_commit($c, $r_min, $r_max, \@k);
4599        }
4600        if (@k) {
4601                ($r_min, $r_max) = ($r_max, $r_min);
4602                process_commit($_, $r_min, $r_max) foreach reverse @k;
4603        }
4604out:
4605        close $log;
4606        print commit_log_separator unless $incremental || $oneline;
4607}
4608
4609sub cmd_blame {
4610        my $path = pop;
4611
4612        config_pager();
4613        run_pager();
4614
4615        my ($fh, $ctx, $rev);
4616
4617        if ($_git_format) {
4618                ($fh, $ctx) = command_output_pipe('blame', @_, $path);
4619                while (my $line = <$fh>) {
4620                        if ($line =~ /^\^?([[:xdigit:]]+)\s/) {
4621                                # Uncommitted edits show up as a rev ID of
4622                                # all zeros, which we can't look up with
4623                                # cmt_metadata
4624                                if ($1 !~ /^0+$/) {
4625                                        (undef, $rev, undef) =
4626                                                ::cmt_metadata($1);
4627                                        $rev = '0' if (!$rev);
4628                                } else {
4629                                        $rev = '0';
4630                                }
4631                                $rev = sprintf('%-10s', $rev);
4632                                $line =~ s/^\^?[[:xdigit:]]+(\s)/$rev$1/;
4633                        }
4634                        print $line;
4635                }
4636        } else {
4637                ($fh, $ctx) = command_output_pipe('blame', '-p', @_, 'HEAD',
4638                                                  '--', $path);
4639                my ($sha1);
4640                my %authors;
4641                while (my $line = <$fh>) {
4642                        if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
4643                                $sha1 = $1;
4644                                (undef, $rev, undef) = ::cmt_metadata($1);
4645                                $rev = '0' if (!$rev);
4646                        }
4647                        elsif ($line =~ /^author (.*)/) {
4648                                $authors{$rev} = $1;
4649                                $authors{$rev} =~ s/\s/_/g;
4650                        }
4651                        elsif ($line =~ /^\t(.*)$/) {
4652                                printf("%6s %10s %s\n", $rev, $authors{$rev}, $1);
4653                        }
4654                }
4655        }
4656        command_close_pipe($fh, $ctx);
4657}
4658
4659package Git::SVN::Migration;
4660# these version numbers do NOT correspond to actual version numbers
4661# of git nor git-svn.  They are just relative.
4662#
4663# v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
4664#
4665# v1 layout: .git/$id/info/url, refs/remotes/$id
4666#
4667# v2 layout: .git/svn/$id/info/url, refs/remotes/$id
4668#
4669# v3 layout: .git/svn/$id, refs/remotes/$id
4670#            - info/url may remain for backwards compatibility
4671#            - this is what we migrate up to this layout automatically,
4672#            - this will be used by git svn init on single branches
4673# v3.1 layout (auto migrated):
4674#            - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
4675#              for backwards compatibility
4676#
4677# v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
4678#            - this is only created for newly multi-init-ed
4679#              repositories.  Similar in spirit to the
4680#              --use-separate-remotes option in git-clone (now default)
4681#            - we do not automatically migrate to this (following
4682#              the example set by core git)
4683#
4684# v5 layout: .rev_db.$UUID => .rev_map.$UUID
4685#            - newer, more-efficient format that uses 24-bytes per record
4686#              with no filler space.
4687#            - use xxd -c24 < .rev_map.$UUID to view and debug
4688#            - This is a one-way migration, repositories updated to the
4689#              new format will not be able to use old git-svn without
4690#              rebuilding the .rev_db.  Rebuilding the rev_db is not
4691#              possible if noMetadata or useSvmProps are set; but should
4692#              be no problem for users that use the (sensible) defaults.
4693use strict;
4694use warnings;
4695use Carp qw/croak/;
4696use File::Path qw/mkpath/;
4697use File::Basename qw/dirname basename/;
4698use vars qw/$_minimize/;
4699
4700sub migrate_from_v0 {
4701        my $git_dir = $ENV{GIT_DIR};
4702        return undef unless -d $git_dir;
4703        my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4704        my $migrated = 0;
4705        while (<$fh>) {
4706                chomp;
4707                my ($id, $orig_ref) = ($_, $_);
4708                next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
4709                next unless -f "$git_dir/$id/info/url";
4710                my $new_ref = "refs/remotes/$id";
4711                if (::verify_ref("$new_ref^0")) {
4712                        print STDERR "W: $orig_ref is probably an old ",
4713                                     "branch used by an ancient version of ",
4714                                     "git-svn.\n",
4715                                     "However, $new_ref also exists.\n",
4716                                     "We will not be able ",
4717                                     "to use this branch until this ",
4718                                     "ambiguity is resolved.\n";
4719                        next;
4720                }
4721                print STDERR "Migrating from v0 layout...\n" if !$migrated;
4722                print STDERR "Renaming ref: $orig_ref => $new_ref\n";
4723                command_noisy('update-ref', $new_ref, $orig_ref);
4724                command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
4725                $migrated++;
4726        }
4727        command_close_pipe($fh, $ctx);
4728        print STDERR "Done migrating from v0 layout...\n" if $migrated;
4729        $migrated;
4730}
4731
4732sub migrate_from_v1 {
4733        my $git_dir = $ENV{GIT_DIR};
4734        my $migrated = 0;
4735        return $migrated unless -d $git_dir;
4736        my $svn_dir = "$git_dir/svn";
4737
4738        # just in case somebody used 'svn' as their $id at some point...
4739        return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
4740
4741        print STDERR "Migrating from a git-svn v1 layout...\n";
4742        mkpath([$svn_dir]);
4743        print STDERR "Data from a previous version of git-svn exists, but\n\t",
4744                     "$svn_dir\n\t(required for this version ",
4745                     "($::VERSION) of git-svn) does not exist.\n";
4746        my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4747        while (<$fh>) {
4748                my $x = $_;
4749                next unless $x =~ s#^refs/remotes/##;
4750                chomp $x;
4751                next unless -f "$git_dir/$x/info/url";
4752                my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
4753                next unless $u;
4754                my $dn = dirname("$git_dir/svn/$x");
4755                mkpath([$dn]) unless -d $dn;
4756                if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
4757                        mkpath(["$git_dir/svn/svn"]);
4758                        print STDERR " - $git_dir/$x/info => ",
4759                                        "$git_dir/svn/$x/info\n";
4760                        rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
4761                               croak "$!: $x";
4762                        # don't worry too much about these, they probably
4763                        # don't exist with repos this old (save for index,
4764                        # and we can easily regenerate that)
4765                        foreach my $f (qw/unhandled.log index .rev_db/) {
4766                                rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
4767                        }
4768                } else {
4769                        print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
4770                        rename "$git_dir/$x", "$git_dir/svn/$x" or
4771                               croak "$!: $x";
4772                }
4773                $migrated++;
4774        }
4775        command_close_pipe($fh, $ctx);
4776        print STDERR "Done migrating from a git-svn v1 layout\n";
4777        $migrated;
4778}
4779
4780sub read_old_urls {
4781        my ($l_map, $pfx, $path) = @_;
4782        my @dir;
4783        foreach (<$path/*>) {
4784                if (-r "$_/info/url") {
4785                        $pfx .= '/' if $pfx && $pfx !~ m!/$!;
4786                        my $ref_id = $pfx . basename $_;
4787                        my $url = ::file_to_s("$_/info/url");
4788                        $l_map->{$ref_id} = $url;
4789                } elsif (-d $_) {
4790                        push @dir, $_;
4791                }
4792        }
4793        foreach (@dir) {
4794                my $x = $_;
4795                $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
4796                read_old_urls($l_map, $x, $_);
4797        }
4798}
4799
4800sub migrate_from_v2 {
4801        my @cfg = command(qw/config -l/);
4802        return if grep /^svn-remote\..+\.url=/, @cfg;
4803        my %l_map;
4804        read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
4805        my $migrated = 0;
4806
4807        foreach my $ref_id (sort keys %l_map) {
4808                eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
4809                if ($@) {
4810                        Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
4811                }
4812                $migrated++;
4813        }
4814        $migrated;
4815}
4816
4817sub minimize_connections {
4818        my $r = Git::SVN::read_all_remotes();
4819        my $new_urls = {};
4820        my $root_repos = {};
4821        foreach my $repo_id (keys %$r) {
4822                my $url = $r->{$repo_id}->{url} or next;
4823                my $fetch = $r->{$repo_id}->{fetch} or next;
4824                my $ra = Git::SVN::Ra->new($url);
4825
4826                # skip existing cases where we already connect to the root
4827                if (($ra->{url} eq $ra->{repos_root}) ||
4828                    ($ra->{repos_root} eq $repo_id)) {
4829                        $root_repos->{$ra->{url}} = $repo_id;
4830                        next;
4831                }
4832
4833                my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
4834                my $root_path = $ra->{url};
4835                $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
4836                foreach my $path (keys %$fetch) {
4837                        my $ref_id = $fetch->{$path};
4838                        my $gs = Git::SVN->new($ref_id, $repo_id, $path);
4839
4840                        # make sure we can read when connecting to
4841                        # a higher level of a repository
4842                        my ($last_rev, undef) = $gs->last_rev_commit;
4843                        if (!defined $last_rev) {
4844                                $last_rev = eval {
4845                                        $root_ra->get_latest_revnum;
4846                                };
4847                                next if $@;
4848                        }
4849                        my $new = $root_path;
4850                        $new .= length $path ? "/$path" : '';
4851                        eval {
4852                                $root_ra->get_log([$new], $last_rev, $last_rev,
4853                                                  0, 0, 1, sub { });
4854                        };
4855                        next if $@;
4856                        $new_urls->{$ra->{repos_root}}->{$new} =
4857                                { ref_id => $ref_id,
4858                                  old_repo_id => $repo_id,
4859                                  old_path => $path };
4860                }
4861        }
4862
4863        my @emptied;
4864        foreach my $url (keys %$new_urls) {
4865                # see if we can re-use an existing [svn-remote "repo_id"]
4866                # instead of creating a(n ugly) new section:
4867                my $repo_id = $root_repos->{$url} || $url;
4868
4869                my $fetch = $new_urls->{$url};
4870                foreach my $path (keys %$fetch) {
4871                        my $x = $fetch->{$path};
4872                        Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
4873                        my $pfx = "svn-remote.$x->{old_repo_id}";
4874
4875                        my $old_fetch = quotemeta("$x->{old_path}:".
4876                                                  "refs/remotes/$x->{ref_id}");
4877                        command_noisy(qw/config --unset/,
4878                                      "$pfx.fetch", '^'. $old_fetch . '$');
4879                        delete $r->{$x->{old_repo_id}}->
4880                               {fetch}->{$x->{old_path}};
4881                        if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
4882                                command_noisy(qw/config --unset/,
4883                                              "$pfx.url");
4884                                push @emptied, $x->{old_repo_id}
4885                        }
4886                }
4887        }
4888        if (@emptied) {
4889                my $file = $ENV{GIT_CONFIG} || $ENV{GIT_CONFIG_LOCAL} ||
4890                           "$ENV{GIT_DIR}/config";
4891                print STDERR <<EOF;
4892The following [svn-remote] sections in your config file ($file) are empty
4893and can be safely removed:
4894EOF
4895                print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
4896        }
4897}
4898
4899sub migration_check {
4900        migrate_from_v0();
4901        migrate_from_v1();
4902        migrate_from_v2();
4903        minimize_connections() if $_minimize;
4904}
4905
4906package Git::IndexInfo;
4907use strict;
4908use warnings;
4909use Git qw/command_input_pipe command_close_pipe/;
4910
4911sub new {
4912        my ($class) = @_;
4913        my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
4914        bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
4915}
4916
4917sub remove {
4918        my ($self, $path) = @_;
4919        if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
4920                return ++$self->{nr};
4921        }
4922        undef;
4923}
4924
4925sub update {
4926        my ($self, $mode, $hash, $path) = @_;
4927        if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
4928                return ++$self->{nr};
4929        }
4930        undef;
4931}
4932
4933sub DESTROY {
4934        my ($self) = @_;
4935        command_close_pipe($self->{gui}, $self->{ctx});
4936}
4937
4938package Git::SVN::GlobSpec;
4939use strict;
4940use warnings;
4941
4942sub new {
4943        my ($class, $glob) = @_;
4944        my $re = $glob;
4945        $re =~ s!/+$!!g; # no need for trailing slashes
4946        $re =~ m!^([^*]*)(\*(?:/\*)*)([^*]*)$!;
4947        my $temp = $re;
4948        my ($left, $right) = ($1, $3);
4949        $re = $2;
4950        my $depth = $re =~ tr/*/*/;
4951        if ($depth != $temp =~ tr/*/*/) {
4952                die "Only one set of wildcard directories " .
4953                        "(e.g. '*' or '*/*/*') is supported: '$glob'\n";
4954        }
4955        if ($depth == 0) {
4956                die "One '*' is needed for glob: '$glob'\n";
4957        }
4958        $re =~ s!\*!\[^/\]*!g;
4959        $re = quotemeta($left) . "($re)" . quotemeta($right);
4960        if (length $left && !($left =~ s!/+$!!g)) {
4961                die "Missing trailing '/' on left side of: '$glob' ($left)\n";
4962        }
4963        if (length $right && !($right =~ s!^/+!!g)) {
4964                die "Missing leading '/' on right side of: '$glob' ($right)\n";
4965        }
4966        my $left_re = qr/^\/\Q$left\E(\/|$)/;
4967        bless { left => $left, right => $right, left_regex => $left_re,
4968                regex => qr/$re/, glob => $glob, depth => $depth }, $class;
4969}
4970
4971sub full_path {
4972        my ($self, $path) = @_;
4973        return (length $self->{left} ? "$self->{left}/" : '') .
4974               $path . (length $self->{right} ? "/$self->{right}" : '');
4975}
4976
4977__END__
4978
4979Data structures:
4980
4981
4982$remotes = { # returned by read_all_remotes()
4983        'svn' => {
4984                # svn-remote.svn.url=https://svn.musicpd.org
4985                url => 'https://svn.musicpd.org',
4986                # svn-remote.svn.fetch=mpd/trunk:trunk
4987                fetch => {
4988                        'mpd/trunk' => 'trunk',
4989                },
4990                # svn-remote.svn.tags=mpd/tags/*:tags/*
4991                tags => {
4992                        path => {
4993                                left => 'mpd/tags',
4994                                right => '',
4995                                regex => qr!mpd/tags/([^/]+)$!,
4996                                glob => 'tags/*',
4997                        },
4998                        ref => {
4999                                left => 'tags',
5000                                right => '',
5001                                regex => qr!tags/([^/]+)$!,
5002                                glob => 'tags/*',
5003                        },
5004                }
5005        }
5006};
5007
5008$log_entry hashref as returned by libsvn_log_entry()
5009{
5010        log => 'whitespace-formatted log entry
5011',                                              # trailing newline is preserved
5012        revision => '8',                        # integer
5013        date => '2004-02-24T17:01:44.108345Z',  # commit date
5014        author => 'committer name'
5015};
5016
5017
5018# this is generated by generate_diff();
5019@mods = array of diff-index line hashes, each element represents one line
5020        of diff-index output
5021
5022diff-index line ($m hash)
5023{
5024        mode_a => first column of diff-index output, no leading ':',
5025        mode_b => second column of diff-index output,
5026        sha1_b => sha1sum of the final blob,
5027        chg => change type [MCRADT],
5028        file_a => original file name of a file (iff chg is 'C' or 'R')
5029        file_b => new/current file name of a file (any chg)
5030}
5031;
5032
5033# retval of read_url_paths{,_all}();
5034$l_map = {
5035        # repository root url
5036        'https://svn.musicpd.org' => {
5037                # repository path               # GIT_SVN_ID
5038                'mpd/trunk'             =>      'trunk',
5039                'mpd/tags/0.11.5'       =>      'tags/0.11.5',
5040        },
5041}
5042
5043Notes:
5044        I don't trust the each() function on unless I created %hash myself
5045        because the internal iterator may not have started at base.