git-svn.perlon commit apply --whitespace=fix: fix handling of blank lines at the eof (ef2035c)
   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        {
2322                my ($base, $head);
2323                if (!defined $r0 || !defined $parent) {
2324                        ($base, $head) = parse_revision_argument(0, $r);
2325                } else {
2326                        if ($r0 < $r) {
2327                                $gs->ra->get_log([$gs->{path}], $r0 + 1, $r, 1,
2328                                        0, 1, sub { $base = $_[1] - 1 });
2329                        }
2330                }
2331                if (defined $base && $base <= $r) {
2332                        $gs->fetch($base, $r);
2333                }
2334                ($r0, $parent) = $gs->find_rev_before($r, 1);
2335        }
2336        if (defined $r0 && defined $parent) {
2337                print STDERR "Found branch parent: ($self->{ref_id}) $parent\n";
2338                my $ed;
2339                if ($self->ra->can_do_switch) {
2340                        $self->assert_index_clean($parent);
2341                        print STDERR "Following parent with do_switch\n";
2342                        # do_switch works with svn/trunk >= r22312, but that
2343                        # is not included with SVN 1.4.3 (the latest version
2344                        # at the moment), so we can't rely on it
2345                        $self->{last_commit} = $parent;
2346                        $ed = SVN::Git::Fetcher->new($self);
2347                        $gs->ra->gs_do_switch($r0, $rev, $gs,
2348                                              $self->full_url, $ed)
2349                          or die "SVN connection failed somewhere...\n";
2350                } elsif ($self->ra->trees_match($new_url, $r0,
2351                                                $self->full_url, $rev)) {
2352                        print STDERR "Trees match:\n",
2353                                     "  $new_url\@$r0\n",
2354                                     "  ${\$self->full_url}\@$rev\n",
2355                                     "Following parent with no changes\n";
2356                        $self->tmp_index_do(sub {
2357                            command_noisy('read-tree', $parent);
2358                        });
2359                        $self->{last_commit} = $parent;
2360                } else {
2361                        print STDERR "Following parent with do_update\n";
2362                        $ed = SVN::Git::Fetcher->new($self);
2363                        $self->ra->gs_do_update($rev, $rev, $self, $ed)
2364                          or die "SVN connection failed somewhere...\n";
2365                }
2366                print STDERR "Successfully followed parent\n";
2367                return $self->make_log_entry($rev, [$parent], $ed);
2368        }
2369        return undef;
2370}
2371
2372sub do_fetch {
2373        my ($self, $paths, $rev) = @_;
2374        my $ed;
2375        my ($last_rev, @parents);
2376        if (my $lc = $self->last_commit) {
2377                # we can have a branch that was deleted, then re-added
2378                # under the same name but copied from another path, in
2379                # which case we'll have multiple parents (we don't
2380                # want to break the original ref, nor lose copypath info):
2381                if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2382                        push @{$log_entry->{parents}}, $lc;
2383                        return $log_entry;
2384                }
2385                $ed = SVN::Git::Fetcher->new($self);
2386                $last_rev = $self->{last_rev};
2387                $ed->{c} = $lc;
2388                @parents = ($lc);
2389        } else {
2390                $last_rev = $rev;
2391                if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2392                        return $log_entry;
2393                }
2394                $ed = SVN::Git::Fetcher->new($self);
2395        }
2396        unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2397                die "SVN connection failed somewhere...\n";
2398        }
2399        $self->make_log_entry($rev, \@parents, $ed);
2400}
2401
2402sub get_untracked {
2403        my ($self, $ed) = @_;
2404        my @out;
2405        my $h = $ed->{empty};
2406        foreach (sort keys %$h) {
2407                my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2408                push @out, "  $act: " . uri_encode($_);
2409                warn "W: $act: $_\n";
2410        }
2411        foreach my $t (qw/dir_prop file_prop/) {
2412                $h = $ed->{$t} or next;
2413                foreach my $path (sort keys %$h) {
2414                        my $ppath = $path eq '' ? '.' : $path;
2415                        foreach my $prop (sort keys %{$h->{$path}}) {
2416                                next if $SKIP_PROP{$prop};
2417                                my $v = $h->{$path}->{$prop};
2418                                my $t_ppath_prop = "$t: " .
2419                                                    uri_encode($ppath) . ' ' .
2420                                                    uri_encode($prop);
2421                                if (defined $v) {
2422                                        push @out, "  +$t_ppath_prop " .
2423                                                   uri_encode($v);
2424                                } else {
2425                                        push @out, "  -$t_ppath_prop";
2426                                }
2427                        }
2428                }
2429        }
2430        foreach my $t (qw/absent_file absent_directory/) {
2431                $h = $ed->{$t} or next;
2432                foreach my $parent (sort keys %$h) {
2433                        foreach my $path (sort @{$h->{$parent}}) {
2434                                push @out, "  $t: " .
2435                                           uri_encode("$parent/$path");
2436                                warn "W: $t: $parent/$path ",
2437                                     "Insufficient permissions?\n";
2438                        }
2439                }
2440        }
2441        \@out;
2442}
2443
2444sub parse_svn_date {
2445        my $date = shift || return '+0000 1970-01-01 00:00:00';
2446        my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2447                                            (\d\d)\:(\d\d)\:(\d\d).\d+Z$/x) or
2448                                         croak "Unable to parse date: $date\n";
2449        "+0000 $Y-$m-$d $H:$M:$S";
2450}
2451
2452sub check_author {
2453        my ($author) = @_;
2454        if (!defined $author || length $author == 0) {
2455                $author = '(no author)';
2456        } elsif (defined $::_authors && ! defined $::users{$author}) {
2457                die "Author: $author not defined in $::_authors file\n";
2458        }
2459        $author;
2460}
2461
2462sub make_log_entry {
2463        my ($self, $rev, $parents, $ed) = @_;
2464        my $untracked = $self->get_untracked($ed);
2465
2466        open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
2467        print $un "r$rev\n" or croak $!;
2468        print $un $_, "\n" foreach @$untracked;
2469        my %log_entry = ( parents => $parents || [], revision => $rev,
2470                          log => '');
2471
2472        my $headrev;
2473        my $logged = delete $self->{logged_rev_props};
2474        if (!$logged || $self->{-want_revprops}) {
2475                my $rp = $self->ra->rev_proplist($rev);
2476                foreach (sort keys %$rp) {
2477                        my $v = $rp->{$_};
2478                        if (/^svn:(author|date|log)$/) {
2479                                $log_entry{$1} = $v;
2480                        } elsif ($_ eq 'svm:headrev') {
2481                                $headrev = $v;
2482                        } else {
2483                                print $un "  rev_prop: ", uri_encode($_), ' ',
2484                                          uri_encode($v), "\n";
2485                        }
2486                }
2487        } else {
2488                map { $log_entry{$_} = $logged->{$_} } keys %$logged;
2489        }
2490        close $un or croak $!;
2491
2492        $log_entry{date} = parse_svn_date($log_entry{date});
2493        $log_entry{log} .= "\n";
2494        my $author = $log_entry{author} = check_author($log_entry{author});
2495        my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
2496                                                       : ($author, undef);
2497
2498        my ($commit_name, $commit_email) = ($name, $email);
2499        if ($_use_log_author) {
2500                my $name_field;
2501                if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
2502                        $name_field = $1;
2503                } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
2504                        $name_field = $1;
2505                }
2506                if (!defined $name_field) {
2507                        if (!defined $email) {
2508                                $email = $name;
2509                        }
2510                } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
2511                        ($name, $email) = ($1, $2);
2512                } elsif ($name_field =~ /(.*)@/) {
2513                        ($name, $email) = ($1, $name_field);
2514                } else {
2515                        ($name, $email) = ($name_field, $name_field);
2516                }
2517        }
2518        if (defined $headrev && $self->use_svm_props) {
2519                if ($self->rewrite_root) {
2520                        die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
2521                            "options set!\n";
2522                }
2523                my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$};
2524                # we don't want "SVM: initializing mirror for junk" ...
2525                return undef if $r == 0;
2526                my $svm = $self->svm;
2527                if ($uuid ne $svm->{uuid}) {
2528                        die "UUID mismatch on SVM path:\n",
2529                            "expected: $svm->{uuid}\n",
2530                            "     got: $uuid\n";
2531                }
2532                my $full_url = $self->full_url;
2533                $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
2534                             die "Failed to replace '$svm->{replace}' with ",
2535                                 "'$svm->{source}' in $full_url\n";
2536                # throw away username for storing in records
2537                remove_username($full_url);
2538                $log_entry{metadata} = "$full_url\@$r $uuid";
2539                $log_entry{svm_revision} = $r;
2540                $email ||= "$author\@$uuid";
2541                $commit_email ||= "$author\@$uuid";
2542        } elsif ($self->use_svnsync_props) {
2543                my $full_url = $self->svnsync->{url};
2544                $full_url .= "/$self->{path}" if length $self->{path};
2545                remove_username($full_url);
2546                my $uuid = $self->svnsync->{uuid};
2547                $log_entry{metadata} = "$full_url\@$rev $uuid";
2548                $email ||= "$author\@$uuid";
2549                $commit_email ||= "$author\@$uuid";
2550        } else {
2551                my $url = $self->metadata_url;
2552                remove_username($url);
2553                $log_entry{metadata} = "$url\@$rev " .
2554                                       $self->ra->get_uuid;
2555                $email ||= "$author\@" . $self->ra->get_uuid;
2556                $commit_email ||= "$author\@" . $self->ra->get_uuid;
2557        }
2558        $log_entry{name} = $name;
2559        $log_entry{email} = $email;
2560        $log_entry{commit_name} = $commit_name;
2561        $log_entry{commit_email} = $commit_email;
2562        \%log_entry;
2563}
2564
2565sub fetch {
2566        my ($self, $min_rev, $max_rev, @parents) = @_;
2567        my ($last_rev, $last_commit) = $self->last_rev_commit;
2568        my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
2569        $self->ra->gs_fetch_loop_common($base, $head, [$self]);
2570}
2571
2572sub set_tree_cb {
2573        my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
2574        $self->{inject_parents} = { $rev => $tree };
2575        $self->fetch(undef, undef);
2576}
2577
2578sub set_tree {
2579        my ($self, $tree) = (shift, shift);
2580        my $log_entry = ::get_commit_entry($tree);
2581        unless ($self->{last_rev}) {
2582                ::fatal("Must have an existing revision to commit");
2583        }
2584        my %ed_opts = ( r => $self->{last_rev},
2585                        log => $log_entry->{log},
2586                        ra => $self->ra,
2587                        tree_a => $self->{last_commit},
2588                        tree_b => $tree,
2589                        editor_cb => sub {
2590                               $self->set_tree_cb($log_entry, $tree, @_) },
2591                        svn_path => $self->{path} );
2592        if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
2593                print "No changes\nr$self->{last_rev} = $tree\n";
2594        }
2595}
2596
2597sub rebuild_from_rev_db {
2598        my ($self, $path) = @_;
2599        my $r = -1;
2600        open my $fh, '<', $path or croak "open: $!";
2601        binmode $fh or croak "binmode: $!";
2602        while (<$fh>) {
2603                length($_) == 41 or croak "inconsistent size in ($_) != 41";
2604                chomp($_);
2605                ++$r;
2606                next if $_ eq ('0' x 40);
2607                $self->rev_map_set($r, $_);
2608                print "r$r = $_\n";
2609        }
2610        close $fh or croak "close: $!";
2611        unlink $path or croak "unlink: $!";
2612}
2613
2614sub rebuild {
2615        my ($self) = @_;
2616        my $map_path = $self->map_path;
2617        return if (-e $map_path && ! -z $map_path);
2618        return unless ::verify_ref($self->refname.'^0');
2619        if ($self->use_svm_props || $self->no_metadata) {
2620                my $rev_db = $self->rev_db_path;
2621                $self->rebuild_from_rev_db($rev_db);
2622                if ($self->use_svm_props) {
2623                        my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
2624                        $self->rebuild_from_rev_db($svm_rev_db);
2625                }
2626                $self->unlink_rev_db_symlink;
2627                return;
2628        }
2629        print "Rebuilding $map_path ...\n";
2630        my ($log, $ctx) =
2631            command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
2632                                $self->refname, '--');
2633        my $metadata_url = $self->metadata_url;
2634        remove_username($metadata_url);
2635        my $svn_uuid = $self->ra_uuid;
2636        my $c;
2637        while (<$log>) {
2638                if ( m{^commit ($::sha1)$} ) {
2639                        $c = $1;
2640                        next;
2641                }
2642                next unless s{^\s*(git-svn-id:)}{$1};
2643                my ($url, $rev, $uuid) = ::extract_metadata($_);
2644                remove_username($url);
2645
2646                # ignore merges (from set-tree)
2647                next if (!defined $rev || !$uuid);
2648
2649                # if we merged or otherwise started elsewhere, this is
2650                # how we break out of it
2651                if (($uuid ne $svn_uuid) ||
2652                    ($metadata_url && $url && ($url ne $metadata_url))) {
2653                        next;
2654                }
2655
2656                $self->rev_map_set($rev, $c);
2657                print "r$rev = $c\n";
2658        }
2659        command_close_pipe($log, $ctx);
2660        print "Done rebuilding $map_path\n";
2661        my $rev_db_path = $self->rev_db_path;
2662        if (-f $self->rev_db_path) {
2663                unlink $self->rev_db_path or croak "unlink: $!";
2664        }
2665        $self->unlink_rev_db_symlink;
2666}
2667
2668# rev_map:
2669# Tie::File seems to be prone to offset errors if revisions get sparse,
2670# it's not that fast, either.  Tie::File is also not in Perl 5.6.  So
2671# one of my favorite modules is out :<  Next up would be one of the DBM
2672# modules, but I'm not sure which is most portable...
2673#
2674# This is the replacement for the rev_db format, which was too big
2675# and inefficient for large repositories with a lot of sparse history
2676# (mainly tags)
2677#
2678# The format is this:
2679#   - 24 bytes for every record,
2680#     * 4 bytes for the integer representing an SVN revision number
2681#     * 20 bytes representing the sha1 of a git commit
2682#   - No empty padding records like the old format
2683#     (except the last record, which can be overwritten)
2684#   - new records are written append-only since SVN revision numbers
2685#     increase monotonically
2686#   - lookups on SVN revision number are done via a binary search
2687#   - Piping the file to xxd -c24 is a good way of dumping it for
2688#     viewing or editing (piped back through xxd -r), should the need
2689#     ever arise.
2690#   - The last record can be padding revision with an all-zero sha1
2691#     This is used to optimize fetch performance when using multiple
2692#     "fetch" directives in .git/config
2693#
2694# These files are disposable unless noMetadata or useSvmProps is set
2695
2696sub _rev_map_set {
2697        my ($fh, $rev, $commit) = @_;
2698
2699        binmode $fh or croak "binmode: $!";
2700        my $size = (stat($fh))[7];
2701        ($size % 24) == 0 or croak "inconsistent size: $size";
2702
2703        my $wr_offset = 0;
2704        if ($size > 0) {
2705                sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2706                my $read = sysread($fh, my $buf, 24) or croak "read: $!";
2707                $read == 24 or croak "read only $read bytes (!= 24)";
2708                my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
2709                if ($last_commit eq ('0' x40)) {
2710                        if ($size >= 48) {
2711                                sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2712                                $read = sysread($fh, $buf, 24) or
2713                                    croak "read: $!";
2714                                $read == 24 or
2715                                    croak "read only $read bytes (!= 24)";
2716                                ($last_rev, $last_commit) =
2717                                    unpack(rev_map_fmt, $buf);
2718                                if ($last_commit eq ('0' x40)) {
2719                                        croak "inconsistent .rev_map\n";
2720                                }
2721                        }
2722                        if ($last_rev >= $rev) {
2723                                croak "last_rev is higher!: $last_rev >= $rev";
2724                        }
2725                        $wr_offset = -24;
2726                }
2727        }
2728        sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
2729        syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
2730          croak "write: $!";
2731}
2732
2733sub mkfile {
2734        my ($path) = @_;
2735        unless (-e $path) {
2736                my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
2737                mkpath([$dir]) unless -d $dir;
2738                open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
2739                close $fh or die "Couldn't close (create) $path: $!\n";
2740        }
2741}
2742
2743sub rev_map_set {
2744        my ($self, $rev, $commit, $update_ref, $uuid) = @_;
2745        length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
2746        my $db = $self->map_path($uuid);
2747        my $db_lock = "$db.lock";
2748        my $sig;
2749        if ($update_ref) {
2750                $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2751                            $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
2752        }
2753        mkfile($db);
2754
2755        $LOCKFILES{$db_lock} = 1;
2756        my $sync;
2757        # both of these options make our .rev_db file very, very important
2758        # and we can't afford to lose it because rebuild() won't work
2759        if ($self->use_svm_props || $self->no_metadata) {
2760                $sync = 1;
2761                copy($db, $db_lock) or die "rev_map_set(@_): ",
2762                                           "Failed to copy: ",
2763                                           "$db => $db_lock ($!)\n";
2764        } else {
2765                rename $db, $db_lock or die "rev_map_set(@_): ",
2766                                            "Failed to rename: ",
2767                                            "$db => $db_lock ($!)\n";
2768        }
2769
2770        sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
2771             or croak "Couldn't open $db_lock: $!\n";
2772        _rev_map_set($fh, $rev, $commit);
2773        if ($sync) {
2774                $fh->flush or die "Couldn't flush $db_lock: $!\n";
2775                $fh->sync or die "Couldn't sync $db_lock: $!\n";
2776        }
2777        close $fh or croak $!;
2778        if ($update_ref) {
2779                $_head = $self;
2780                command_noisy('update-ref', '-m', "r$rev",
2781                              $self->refname, $commit);
2782        }
2783        rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
2784                                    "$db_lock => $db ($!)\n";
2785        delete $LOCKFILES{$db_lock};
2786        if ($update_ref) {
2787                $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2788                            $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
2789                kill $sig, $$ if defined $sig;
2790        }
2791}
2792
2793# If want_commit, this will return an array of (rev, commit) where
2794# commit _must_ be a valid commit in the archive.
2795# Otherwise, it'll return the max revision (whether or not the
2796# commit is valid or just a 0x40 placeholder).
2797sub rev_map_max {
2798        my ($self, $want_commit) = @_;
2799        $self->rebuild;
2800        my $map_path = $self->map_path;
2801        stat $map_path or return $want_commit ? (0, undef) : 0;
2802        sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2803        binmode $fh or croak "binmode: $!";
2804        my $size = (stat($fh))[7];
2805        ($size % 24) == 0 or croak "inconsistent size: $size";
2806
2807        if ($size == 0) {
2808                close $fh or croak "close: $!";
2809                return $want_commit ? (0, undef) : 0;
2810        }
2811
2812        sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2813        sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2814        my ($r, $c) = unpack(rev_map_fmt, $buf);
2815        if ($want_commit && $c eq ('0' x40)) {
2816                if ($size < 48) {
2817                        return $want_commit ? (0, undef) : 0;
2818                }
2819                sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2820                sysread($fh, $buf, 24) == 24 or croak "read: $!";
2821                ($r, $c) = unpack(rev_map_fmt, $buf);
2822                if ($c eq ('0'x40)) {
2823                        croak "Penultimate record is all-zeroes in $map_path";
2824                }
2825        }
2826        close $fh or croak "close: $!";
2827        $want_commit ? ($r, $c) : $r;
2828}
2829
2830sub rev_map_get {
2831        my ($self, $rev, $uuid) = @_;
2832        my $map_path = $self->map_path($uuid);
2833        return undef unless -e $map_path;
2834
2835        sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2836        binmode $fh or croak "binmode: $!";
2837        my $size = (stat($fh))[7];
2838        ($size % 24) == 0 or croak "inconsistent size: $size";
2839
2840        if ($size == 0) {
2841                close $fh or croak "close: $fh";
2842                return undef;
2843        }
2844
2845        my ($l, $u) = (0, $size - 24);
2846        my ($r, $c, $buf);
2847
2848        while ($l <= $u) {
2849                my $i = int(($l/24 + $u/24) / 2) * 24;
2850                sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
2851                sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2852                my ($r, $c) = unpack('NH40', $buf);
2853
2854                if ($r < $rev) {
2855                        $l = $i + 24;
2856                } elsif ($r > $rev) {
2857                        $u = $i - 24;
2858                } else { # $r == $rev
2859                        close($fh) or croak "close: $!";
2860                        return $c eq ('0' x 40) ? undef : $c;
2861                }
2862        }
2863        close($fh) or croak "close: $!";
2864        undef;
2865}
2866
2867# Finds the first svn revision that exists on (if $eq_ok is true) or
2868# before $rev for the current branch.  It will not search any lower
2869# than $min_rev.  Returns the git commit hash and svn revision number
2870# if found, else (undef, undef).
2871sub find_rev_before {
2872        my ($self, $rev, $eq_ok, $min_rev) = @_;
2873        --$rev unless $eq_ok;
2874        $min_rev ||= 1;
2875        while ($rev >= $min_rev) {
2876                if (my $c = $self->rev_map_get($rev)) {
2877                        return ($rev, $c);
2878                }
2879                --$rev;
2880        }
2881        return (undef, undef);
2882}
2883
2884# Finds the first svn revision that exists on (if $eq_ok is true) or
2885# after $rev for the current branch.  It will not search any higher
2886# than $max_rev.  Returns the git commit hash and svn revision number
2887# if found, else (undef, undef).
2888sub find_rev_after {
2889        my ($self, $rev, $eq_ok, $max_rev) = @_;
2890        ++$rev unless $eq_ok;
2891        $max_rev ||= $self->rev_map_max;
2892        while ($rev <= $max_rev) {
2893                if (my $c = $self->rev_map_get($rev)) {
2894                        return ($rev, $c);
2895                }
2896                ++$rev;
2897        }
2898        return (undef, undef);
2899}
2900
2901sub _new {
2902        my ($class, $repo_id, $ref_id, $path) = @_;
2903        unless (defined $repo_id && length $repo_id) {
2904                $repo_id = $Git::SVN::default_repo_id;
2905        }
2906        unless (defined $ref_id && length $ref_id) {
2907                $_[2] = $ref_id = $Git::SVN::default_ref_id;
2908        }
2909        $_[1] = $repo_id;
2910        my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
2911        $_[3] = $path = '' unless (defined $path);
2912        mkpath(["$ENV{GIT_DIR}/svn"]);
2913        bless {
2914                ref_id => $ref_id, dir => $dir, index => "$dir/index",
2915                path => $path, config => "$ENV{GIT_DIR}/svn/config",
2916                map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
2917}
2918
2919# for read-only access of old .rev_db formats
2920sub unlink_rev_db_symlink {
2921        my ($self) = @_;
2922        my $link = $self->rev_db_path;
2923        $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
2924        if (-l $link) {
2925                unlink $link or croak "unlink: $link failed!";
2926        }
2927}
2928
2929sub rev_db_path {
2930        my ($self, $uuid) = @_;
2931        my $db_path = $self->map_path($uuid);
2932        $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
2933            or croak "map_path: $db_path does not contain '/.rev_map.' !";
2934        $db_path;
2935}
2936
2937# the new replacement for .rev_db
2938sub map_path {
2939        my ($self, $uuid) = @_;
2940        $uuid ||= $self->ra_uuid;
2941        "$self->{map_root}.$uuid";
2942}
2943
2944sub uri_encode {
2945        my ($f) = @_;
2946        $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
2947        $f
2948}
2949
2950sub remove_username {
2951        $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
2952}
2953
2954package Git::SVN::Prompt;
2955use strict;
2956use warnings;
2957require SVN::Core;
2958use vars qw/$_no_auth_cache $_username/;
2959
2960sub simple {
2961        my ($cred, $realm, $default_username, $may_save, $pool) = @_;
2962        $may_save = undef if $_no_auth_cache;
2963        $default_username = $_username if defined $_username;
2964        if (defined $default_username && length $default_username) {
2965                if (defined $realm && length $realm) {
2966                        print STDERR "Authentication realm: $realm\n";
2967                        STDERR->flush;
2968                }
2969                $cred->username($default_username);
2970        } else {
2971                username($cred, $realm, $may_save, $pool);
2972        }
2973        $cred->password(_read_password("Password for '" .
2974                                       $cred->username . "': ", $realm));
2975        $cred->may_save($may_save);
2976        $SVN::_Core::SVN_NO_ERROR;
2977}
2978
2979sub ssl_server_trust {
2980        my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
2981        $may_save = undef if $_no_auth_cache;
2982        print STDERR "Error validating server certificate for '$realm':\n";
2983        {
2984                no warnings 'once';
2985                # All variables SVN::Auth::SSL::* are used only once,
2986                # so we're shutting up Perl warnings about this.
2987                if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
2988                        print STDERR " - The certificate is not issued ",
2989                            "by a trusted authority. Use the\n",
2990                            "   fingerprint to validate ",
2991                            "the certificate manually!\n";
2992                }
2993                if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
2994                        print STDERR " - The certificate hostname ",
2995                            "does not match.\n";
2996                }
2997                if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
2998                        print STDERR " - The certificate is not yet valid.\n";
2999                }
3000                if ($failures & $SVN::Auth::SSL::EXPIRED) {
3001                        print STDERR " - The certificate has expired.\n";
3002                }
3003                if ($failures & $SVN::Auth::SSL::OTHER) {
3004                        print STDERR " - The certificate has ",
3005                            "an unknown error.\n";
3006                }
3007        } # no warnings 'once'
3008        printf STDERR
3009                "Certificate information:\n".
3010                " - Hostname: %s\n".
3011                " - Valid: from %s until %s\n".
3012                " - Issuer: %s\n".
3013                " - Fingerprint: %s\n",
3014                map $cert_info->$_, qw(hostname valid_from valid_until
3015                                       issuer_dname fingerprint);
3016        my $choice;
3017prompt:
3018        print STDERR $may_save ?
3019              "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
3020              "(R)eject or accept (t)emporarily? ";
3021        STDERR->flush;
3022        $choice = lc(substr(<STDIN> || 'R', 0, 1));
3023        if ($choice =~ /^t$/i) {
3024                $cred->may_save(undef);
3025        } elsif ($choice =~ /^r$/i) {
3026                return -1;
3027        } elsif ($may_save && $choice =~ /^p$/i) {
3028                $cred->may_save($may_save);
3029        } else {
3030                goto prompt;
3031        }
3032        $cred->accepted_failures($failures);
3033        $SVN::_Core::SVN_NO_ERROR;
3034}
3035
3036sub ssl_client_cert {
3037        my ($cred, $realm, $may_save, $pool) = @_;
3038        $may_save = undef if $_no_auth_cache;
3039        print STDERR "Client certificate filename: ";
3040        STDERR->flush;
3041        chomp(my $filename = <STDIN>);
3042        $cred->cert_file($filename);
3043        $cred->may_save($may_save);
3044        $SVN::_Core::SVN_NO_ERROR;
3045}
3046
3047sub ssl_client_cert_pw {
3048        my ($cred, $realm, $may_save, $pool) = @_;
3049        $may_save = undef if $_no_auth_cache;
3050        $cred->password(_read_password("Password: ", $realm));
3051        $cred->may_save($may_save);
3052        $SVN::_Core::SVN_NO_ERROR;
3053}
3054
3055sub username {
3056        my ($cred, $realm, $may_save, $pool) = @_;
3057        $may_save = undef if $_no_auth_cache;
3058        if (defined $realm && length $realm) {
3059                print STDERR "Authentication realm: $realm\n";
3060        }
3061        my $username;
3062        if (defined $_username) {
3063                $username = $_username;
3064        } else {
3065                print STDERR "Username: ";
3066                STDERR->flush;
3067                chomp($username = <STDIN>);
3068        }
3069        $cred->username($username);
3070        $cred->may_save($may_save);
3071        $SVN::_Core::SVN_NO_ERROR;
3072}
3073
3074sub _read_password {
3075        my ($prompt, $realm) = @_;
3076        print STDERR $prompt;
3077        STDERR->flush;
3078        require Term::ReadKey;
3079        Term::ReadKey::ReadMode('noecho');
3080        my $password = '';
3081        while (defined(my $key = Term::ReadKey::ReadKey(0))) {
3082                last if $key =~ /[\012\015]/; # \n\r
3083                $password .= $key;
3084        }
3085        Term::ReadKey::ReadMode('restore');
3086        print STDERR "\n";
3087        STDERR->flush;
3088        $password;
3089}
3090
3091package SVN::Git::Fetcher;
3092use vars qw/@ISA/;
3093use strict;
3094use warnings;
3095use Carp qw/croak/;
3096use File::Temp qw/tempfile/;
3097use IO::File qw//;
3098
3099# file baton members: path, mode_a, mode_b, pool, fh, blob, base
3100sub new {
3101        my ($class, $git_svn) = @_;
3102        my $self = SVN::Delta::Editor->new;
3103        bless $self, $class;
3104        $self->{c} = $git_svn->{last_commit} if exists $git_svn->{last_commit};
3105        $self->{empty} = {};
3106        $self->{dir_prop} = {};
3107        $self->{file_prop} = {};
3108        $self->{absent_dir} = {};
3109        $self->{absent_file} = {};
3110        $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
3111        $self;
3112}
3113
3114sub set_path_strip {
3115        my ($self, $path) = @_;
3116        $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
3117}
3118
3119sub open_root {
3120        { path => '' };
3121}
3122
3123sub open_directory {
3124        my ($self, $path, $pb, $rev) = @_;
3125        { path => $path };
3126}
3127
3128sub git_path {
3129        my ($self, $path) = @_;
3130        if ($self->{path_strip}) {
3131                $path =~ s!$self->{path_strip}!! or
3132                  die "Failed to strip path '$path' ($self->{path_strip})\n";
3133        }
3134        $path;
3135}
3136
3137sub delete_entry {
3138        my ($self, $path, $rev, $pb) = @_;
3139
3140        my $gpath = $self->git_path($path);
3141        return undef if ($gpath eq '');
3142
3143        # remove entire directories.
3144        if (command('ls-tree', $self->{c}, '--', $gpath) =~ /^040000 tree/) {
3145                my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3146                                                     -r --name-only -z/,
3147                                                     $self->{c}, '--', $gpath);
3148                local $/ = "\0";
3149                while (<$ls>) {
3150                        chomp;
3151                        $self->{gii}->remove($_);
3152                        print "\tD\t$_\n" unless $::_q;
3153                }
3154                print "\tD\t$gpath/\n" unless $::_q;
3155                command_close_pipe($ls, $ctx);
3156                $self->{empty}->{$path} = 0
3157        } else {
3158                $self->{gii}->remove($gpath);
3159                print "\tD\t$gpath\n" unless $::_q;
3160        }
3161        undef;
3162}
3163
3164sub open_file {
3165        my ($self, $path, $pb, $rev) = @_;
3166        my $gpath = $self->git_path($path);
3167        my ($mode, $blob) = (command('ls-tree', $self->{c}, '--', $gpath)
3168                             =~ /^(\d{6}) blob ([a-f\d]{40})\t/);
3169        unless (defined $mode && defined $blob) {
3170                die "$path was not found in commit $self->{c} (r$rev)\n";
3171        }
3172        { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3173          pool => SVN::Pool->new, action => 'M' };
3174}
3175
3176sub add_file {
3177        my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3178        my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3179        delete $self->{empty}->{$dir};
3180        { path => $path, mode_a => 100644, mode_b => 100644,
3181          pool => SVN::Pool->new, action => 'A' };
3182}
3183
3184sub add_directory {
3185        my ($self, $path, $cp_path, $cp_rev) = @_;
3186        my $gpath = $self->git_path($path);
3187        if ($gpath eq '') {
3188                my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3189                                                     -r --name-only -z/,
3190                                                     $self->{c});
3191                local $/ = "\0";
3192                while (<$ls>) {
3193                        chomp;
3194                        $self->{gii}->remove($_);
3195                        print "\tD\t$_\n" unless $::_q;
3196                }
3197                command_close_pipe($ls, $ctx);
3198                $self->{empty}->{$path} = 0;
3199        }
3200        my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3201        delete $self->{empty}->{$dir};
3202        $self->{empty}->{$path} = 1;
3203        { path => $path };
3204}
3205
3206sub change_dir_prop {
3207        my ($self, $db, $prop, $value) = @_;
3208        $self->{dir_prop}->{$db->{path}} ||= {};
3209        $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3210        undef;
3211}
3212
3213sub absent_directory {
3214        my ($self, $path, $pb) = @_;
3215        $self->{absent_dir}->{$pb->{path}} ||= [];
3216        push @{$self->{absent_dir}->{$pb->{path}}}, $path;
3217        undef;
3218}
3219
3220sub absent_file {
3221        my ($self, $path, $pb) = @_;
3222        $self->{absent_file}->{$pb->{path}} ||= [];
3223        push @{$self->{absent_file}->{$pb->{path}}}, $path;
3224        undef;
3225}
3226
3227sub change_file_prop {
3228        my ($self, $fb, $prop, $value) = @_;
3229        if ($prop eq 'svn:executable') {
3230                if ($fb->{mode_b} != 120000) {
3231                        $fb->{mode_b} = defined $value ? 100755 : 100644;
3232                }
3233        } elsif ($prop eq 'svn:special') {
3234                $fb->{mode_b} = defined $value ? 120000 : 100644;
3235        } else {
3236                $self->{file_prop}->{$fb->{path}} ||= {};
3237                $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
3238        }
3239        undef;
3240}
3241
3242sub apply_textdelta {
3243        my ($self, $fb, $exp) = @_;
3244        my $fh = Git::temp_acquire('svn_delta');
3245        # $fh gets auto-closed() by SVN::TxDelta::apply(),
3246        # (but $base does not,) so dup() it for reading in close_file
3247        open my $dup, '<&', $fh or croak $!;
3248        my $base = Git::temp_acquire('git_blob');
3249        if ($fb->{blob}) {
3250                print $base 'link ' if ($fb->{mode_a} == 120000);
3251                my $size = $::_repository->cat_blob($fb->{blob}, $base);
3252                die "Failed to read object $fb->{blob}" if ($size < 0);
3253
3254                if (defined $exp) {
3255                        seek $base, 0, 0 or croak $!;
3256                        my $got = ::md5sum($base);
3257                        die "Checksum mismatch: $fb->{path} $fb->{blob}\n",
3258                            "expected: $exp\n",
3259                            "     got: $got\n" if ($got ne $exp);
3260                }
3261        }
3262        seek $base, 0, 0 or croak $!;
3263        $fb->{fh} = $fh;
3264        $fb->{base} = $base;
3265        [ SVN::TxDelta::apply($base, $dup, undef, $fb->{path}, $fb->{pool}) ];
3266}
3267
3268sub close_file {
3269        my ($self, $fb, $exp) = @_;
3270        my $hash;
3271        my $path = $self->git_path($fb->{path});
3272        if (my $fh = $fb->{fh}) {
3273                if (defined $exp) {
3274                        seek($fh, 0, 0) or croak $!;
3275                        my $got = ::md5sum($fh);
3276                        if ($got ne $exp) {
3277                                die "Checksum mismatch: $path\n",
3278                                    "expected: $exp\n    got: $got\n";
3279                        }
3280                }
3281                if ($fb->{mode_b} == 120000) {
3282                        sysseek($fh, 0, 0) or croak $!;
3283                        sysread($fh, my $buf, 5) == 5 or croak $!;
3284
3285                        unless ($buf eq 'link ') {
3286                                warn "$path has mode 120000",
3287                                                " but is not a link\n";
3288                        } else {
3289                                my $tmp_fh = Git::temp_acquire('svn_hash');
3290                                my $res;
3291                                while ($res = sysread($fh, my $str, 1024)) {
3292                                        my $out = syswrite($tmp_fh, $str, $res);
3293                                        defined($out) && $out == $res
3294                                                or croak("write ",
3295                                                        Git::temp_path($tmp_fh),
3296                                                        ": $!\n");
3297                                }
3298                                defined $res or croak $!;
3299
3300                                ($fh, $tmp_fh) = ($tmp_fh, $fh);
3301                                Git::temp_release($tmp_fh, 1);
3302                        }
3303                }
3304
3305                $hash = $::_repository->hash_and_insert_object(
3306                                Git::temp_path($fh));
3307                $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
3308
3309                Git::temp_release($fb->{base}, 1);
3310                Git::temp_release($fh, 1);
3311        } else {
3312                $hash = $fb->{blob} or die "no blob information\n";
3313        }
3314        $fb->{pool}->clear;
3315        $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
3316        print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
3317        undef;
3318}
3319
3320sub abort_edit {
3321        my $self = shift;
3322        $self->{nr} = $self->{gii}->{nr};
3323        delete $self->{gii};
3324        $self->SUPER::abort_edit(@_);
3325}
3326
3327sub close_edit {
3328        my $self = shift;
3329        $self->{git_commit_ok} = 1;
3330        $self->{nr} = $self->{gii}->{nr};
3331        delete $self->{gii};
3332        $self->SUPER::close_edit(@_);
3333}
3334
3335package SVN::Git::Editor;
3336use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
3337use strict;
3338use warnings;
3339use Carp qw/croak/;
3340use IO::File;
3341
3342sub new {
3343        my ($class, $opts) = @_;
3344        foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
3345                die "$_ required!\n" unless (defined $opts->{$_});
3346        }
3347
3348        my $pool = SVN::Pool->new;
3349        my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
3350        my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
3351                                     $opts->{r}, $mods);
3352
3353        # $opts->{ra} functions should not be used after this:
3354        my @ce  = $opts->{ra}->get_commit_editor($opts->{log},
3355                                                $opts->{editor_cb}, $pool);
3356        my $self = SVN::Delta::Editor->new(@ce, $pool);
3357        bless $self, $class;
3358        foreach (qw/svn_path r tree_a tree_b/) {
3359                $self->{$_} = $opts->{$_};
3360        }
3361        $self->{url} = $opts->{ra}->{url};
3362        $self->{mods} = $mods;
3363        $self->{types} = $types;
3364        $self->{pool} = $pool;
3365        $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
3366        $self->{rm} = { };
3367        $self->{path_prefix} = length $self->{svn_path} ?
3368                               "$self->{svn_path}/" : '';
3369        $self->{config} = $opts->{config};
3370        return $self;
3371}
3372
3373sub generate_diff {
3374        my ($tree_a, $tree_b) = @_;
3375        my @diff_tree = qw(diff-tree -z -r);
3376        if ($_cp_similarity) {
3377                push @diff_tree, "-C$_cp_similarity";
3378        } else {
3379                push @diff_tree, '-C';
3380        }
3381        push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
3382        push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
3383        push @diff_tree, $tree_a, $tree_b;
3384        my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
3385        local $/ = "\0";
3386        my $state = 'meta';
3387        my @mods;
3388        while (<$diff_fh>) {
3389                chomp $_; # this gets rid of the trailing "\0"
3390                if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
3391                                        $::sha1\s($::sha1)\s
3392                                        ([MTCRAD])\d*$/xo) {
3393                        push @mods, {   mode_a => $1, mode_b => $2,
3394                                        sha1_b => $3, chg => $4 };
3395                        if ($4 =~ /^(?:C|R)$/) {
3396                                $state = 'file_a';
3397                        } else {
3398                                $state = 'file_b';
3399                        }
3400                } elsif ($state eq 'file_a') {
3401                        my $x = $mods[$#mods] or croak "Empty array\n";
3402                        if ($x->{chg} !~ /^(?:C|R)$/) {
3403                                croak "Error parsing $_, $x->{chg}\n";
3404                        }
3405                        $x->{file_a} = $_;
3406                        $state = 'file_b';
3407                } elsif ($state eq 'file_b') {
3408                        my $x = $mods[$#mods] or croak "Empty array\n";
3409                        if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
3410                                croak "Error parsing $_, $x->{chg}\n";
3411                        }
3412                        if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
3413                                croak "Error parsing $_, $x->{chg}\n";
3414                        }
3415                        $x->{file_b} = $_;
3416                        $state = 'meta';
3417                } else {
3418                        croak "Error parsing $_\n";
3419                }
3420        }
3421        command_close_pipe($diff_fh, $ctx);
3422        \@mods;
3423}
3424
3425sub check_diff_paths {
3426        my ($ra, $pfx, $rev, $mods) = @_;
3427        my %types;
3428        $pfx .= '/' if length $pfx;
3429
3430        sub type_diff_paths {
3431                my ($ra, $types, $path, $rev) = @_;
3432                my @p = split m#/+#, $path;
3433                my $c = shift @p;
3434                unless (defined $types->{$c}) {
3435                        $types->{$c} = $ra->check_path($c, $rev);
3436                }
3437                while (@p) {
3438                        $c .= '/' . shift @p;
3439                        next if defined $types->{$c};
3440                        $types->{$c} = $ra->check_path($c, $rev);
3441                }
3442        }
3443
3444        foreach my $m (@$mods) {
3445                foreach my $f (qw/file_a file_b/) {
3446                        next unless defined $m->{$f};
3447                        my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
3448                        if (length $pfx.$dir && ! defined $types{$dir}) {
3449                                type_diff_paths($ra, \%types, $pfx.$dir, $rev);
3450                        }
3451                }
3452        }
3453        \%types;
3454}
3455
3456sub split_path {
3457        return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
3458}
3459
3460sub repo_path {
3461        my ($self, $path) = @_;
3462        $self->{path_prefix}.(defined $path ? $path : '');
3463}
3464
3465sub url_path {
3466        my ($self, $path) = @_;
3467        if ($self->{url} =~ m#^https?://#) {
3468                $path =~ s/([^a-zA-Z0-9_.-])/uc sprintf("%%%02x",ord($1))/eg;
3469        }
3470        $self->{url} . '/' . $self->repo_path($path);
3471}
3472
3473sub rmdirs {
3474        my ($self) = @_;
3475        my $rm = $self->{rm};
3476        delete $rm->{''}; # we never delete the url we're tracking
3477        return unless %$rm;
3478
3479        foreach (keys %$rm) {
3480                my @d = split m#/#, $_;
3481                my $c = shift @d;
3482                $rm->{$c} = 1;
3483                while (@d) {
3484                        $c .= '/' . shift @d;
3485                        $rm->{$c} = 1;
3486                }
3487        }
3488        delete $rm->{$self->{svn_path}};
3489        delete $rm->{''}; # we never delete the url we're tracking
3490        return unless %$rm;
3491
3492        my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
3493                                             $self->{tree_b});
3494        local $/ = "\0";
3495        while (<$fh>) {
3496                chomp;
3497                my @dn = split m#/#, $_;
3498                while (pop @dn) {
3499                        delete $rm->{join '/', @dn};
3500                }
3501                unless (%$rm) {
3502                        close $fh;
3503                        return;
3504                }
3505        }
3506        command_close_pipe($fh, $ctx);
3507
3508        my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
3509        foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
3510                $self->close_directory($bat->{$d}, $p);
3511                my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
3512                print "\tD+\t$d/\n" unless $::_q;
3513                $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
3514                delete $bat->{$d};
3515        }
3516}
3517
3518sub open_or_add_dir {
3519        my ($self, $full_path, $baton) = @_;
3520        my $t = $self->{types}->{$full_path};
3521        if (!defined $t) {
3522                die "$full_path not known in r$self->{r} or we have a bug!\n";
3523        }
3524        {
3525                no warnings 'once';
3526                # SVN::Node::none and SVN::Node::file are used only once,
3527                # so we're shutting up Perl's warnings about them.
3528                if ($t == $SVN::Node::none) {
3529                        return $self->add_directory($full_path, $baton,
3530                            undef, -1, $self->{pool});
3531                } elsif ($t == $SVN::Node::dir) {
3532                        return $self->open_directory($full_path, $baton,
3533                            $self->{r}, $self->{pool});
3534                } # no warnings 'once'
3535                print STDERR "$full_path already exists in repository at ",
3536                    "r$self->{r} and it is not a directory (",
3537                    ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
3538        } # no warnings 'once'
3539        exit 1;
3540}
3541
3542sub ensure_path {
3543        my ($self, $path) = @_;
3544        my $bat = $self->{bat};
3545        my $repo_path = $self->repo_path($path);
3546        return $bat->{''} unless (length $repo_path);
3547        my @p = split m#/+#, $repo_path;
3548        my $c = shift @p;
3549        $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
3550        while (@p) {
3551                my $c0 = $c;
3552                $c .= '/' . shift @p;
3553                $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
3554        }
3555        return $bat->{$c};
3556}
3557
3558# Subroutine to convert a globbing pattern to a regular expression.
3559# From perl cookbook.
3560sub glob2pat {
3561        my $globstr = shift;
3562        my %patmap = ('*' => '.*', '?' => '.', '[' => '[', ']' => ']');
3563        $globstr =~ s{(.)} { $patmap{$1} || "\Q$1" }ge;
3564        return '^' . $globstr . '$';
3565}
3566
3567sub check_autoprop {
3568        my ($self, $pattern, $properties, $file, $fbat) = @_;
3569        # Convert the globbing pattern to a regular expression.
3570        my $regex = glob2pat($pattern);
3571        # Check if the pattern matches the file name.
3572        if($file =~ m/($regex)/) {
3573                # Parse the list of properties to set.
3574                my @props = split(/;/, $properties);
3575                foreach my $prop (@props) {
3576                        # Parse 'name=value' syntax and set the property.
3577                        if ($prop =~ /([^=]+)=(.*)/) {
3578                                my ($n,$v) = ($1,$2);
3579                                for ($n, $v) {
3580                                        s/^\s+//; s/\s+$//;
3581                                }
3582                                $self->change_file_prop($fbat, $n, $v);
3583                        }
3584                }
3585        }
3586}
3587
3588sub apply_autoprops {
3589        my ($self, $file, $fbat) = @_;
3590        my $conf_t = ${$self->{config}}{'config'};
3591        no warnings 'once';
3592        # Check [miscellany]/enable-auto-props in svn configuration.
3593        if (SVN::_Core::svn_config_get_bool(
3594                $conf_t,
3595                $SVN::_Core::SVN_CONFIG_SECTION_MISCELLANY,
3596                $SVN::_Core::SVN_CONFIG_OPTION_ENABLE_AUTO_PROPS,
3597                0)) {
3598                # Auto-props are enabled.  Enumerate them to look for matches.
3599                my $callback = sub {
3600                        $self->check_autoprop($_[0], $_[1], $file, $fbat);
3601                };
3602                SVN::_Core::svn_config_enumerate(
3603                        $conf_t,
3604                        $SVN::_Core::SVN_CONFIG_SECTION_AUTO_PROPS,
3605                        $callback);
3606        }
3607}
3608
3609sub A {
3610        my ($self, $m) = @_;
3611        my ($dir, $file) = split_path($m->{file_b});
3612        my $pbat = $self->ensure_path($dir);
3613        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3614                                        undef, -1);
3615        print "\tA\t$m->{file_b}\n" unless $::_q;
3616        $self->apply_autoprops($file, $fbat);
3617        $self->chg_file($fbat, $m);
3618        $self->close_file($fbat,undef,$self->{pool});
3619}
3620
3621sub C {
3622        my ($self, $m) = @_;
3623        my ($dir, $file) = split_path($m->{file_b});
3624        my $pbat = $self->ensure_path($dir);
3625        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3626                                $self->url_path($m->{file_a}), $self->{r});
3627        print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3628        $self->chg_file($fbat, $m);
3629        $self->close_file($fbat,undef,$self->{pool});
3630}
3631
3632sub delete_entry {
3633        my ($self, $path, $pbat) = @_;
3634        my $rpath = $self->repo_path($path);
3635        my ($dir, $file) = split_path($rpath);
3636        $self->{rm}->{$dir} = 1;
3637        $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
3638}
3639
3640sub R {
3641        my ($self, $m) = @_;
3642        my ($dir, $file) = split_path($m->{file_b});
3643        my $pbat = $self->ensure_path($dir);
3644        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3645                                $self->url_path($m->{file_a}), $self->{r});
3646        print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3647        $self->chg_file($fbat, $m);
3648        $self->close_file($fbat,undef,$self->{pool});
3649
3650        ($dir, $file) = split_path($m->{file_a});
3651        $pbat = $self->ensure_path($dir);
3652        $self->delete_entry($m->{file_a}, $pbat);
3653}
3654
3655sub M {
3656        my ($self, $m) = @_;
3657        my ($dir, $file) = split_path($m->{file_b});
3658        my $pbat = $self->ensure_path($dir);
3659        my $fbat = $self->open_file($self->repo_path($m->{file_b}),
3660                                $pbat,$self->{r},$self->{pool});
3661        print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
3662        $self->chg_file($fbat, $m);
3663        $self->close_file($fbat,undef,$self->{pool});
3664}
3665
3666sub T { shift->M(@_) }
3667
3668sub change_file_prop {
3669        my ($self, $fbat, $pname, $pval) = @_;
3670        $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
3671}
3672
3673sub chg_file {
3674        my ($self, $fbat, $m) = @_;
3675        if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
3676                $self->change_file_prop($fbat,'svn:executable','*');
3677        } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
3678                $self->change_file_prop($fbat,'svn:executable',undef);
3679        }
3680        my $fh = Git::temp_acquire('git_blob');
3681        if ($m->{mode_b} =~ /^120/) {
3682                print $fh 'link ' or croak $!;
3683                $self->change_file_prop($fbat,'svn:special','*');
3684        } elsif ($m->{mode_a} =~ /^120/ && $m->{mode_b} !~ /^120/) {
3685                $self->change_file_prop($fbat,'svn:special',undef);
3686        }
3687        my $size = $::_repository->cat_blob($m->{sha1_b}, $fh);
3688        croak "Failed to read object $m->{sha1_b}" if ($size < 0);
3689        $fh->flush == 0 or croak $!;
3690        seek $fh, 0, 0 or croak $!;
3691
3692        my $exp = ::md5sum($fh);
3693        seek $fh, 0, 0 or croak $!;
3694
3695        my $pool = SVN::Pool->new;
3696        my $atd = $self->apply_textdelta($fbat, undef, $pool);
3697        my $got = SVN::TxDelta::send_stream($fh, @$atd, $pool);
3698        die "Checksum mismatch\nexpected: $exp\ngot: $got\n" if ($got ne $exp);
3699        Git::temp_release($fh, 1);
3700        $pool->clear;
3701}
3702
3703sub D {
3704        my ($self, $m) = @_;
3705        my ($dir, $file) = split_path($m->{file_b});
3706        my $pbat = $self->ensure_path($dir);
3707        print "\tD\t$m->{file_b}\n" unless $::_q;
3708        $self->delete_entry($m->{file_b}, $pbat);
3709}
3710
3711sub close_edit {
3712        my ($self) = @_;
3713        my ($p,$bat) = ($self->{pool}, $self->{bat});
3714        foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
3715                next if $_ eq '';
3716                $self->close_directory($bat->{$_}, $p);
3717        }
3718        $self->close_directory($bat->{''}, $p);
3719        $self->SUPER::close_edit($p);
3720        $p->clear;
3721}
3722
3723sub abort_edit {
3724        my ($self) = @_;
3725        $self->SUPER::abort_edit($self->{pool});
3726}
3727
3728sub DESTROY {
3729        my $self = shift;
3730        $self->SUPER::DESTROY(@_);
3731        $self->{pool}->clear;
3732}
3733
3734# this drives the editor
3735sub apply_diff {
3736        my ($self) = @_;
3737        my $mods = $self->{mods};
3738        my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
3739        foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
3740                my $f = $m->{chg};
3741                if (defined $o{$f}) {
3742                        $self->$f($m);
3743                } else {
3744                        fatal("Invalid change type: $f");
3745                }
3746        }
3747        $self->rmdirs if $_rmdir;
3748        if (@$mods == 0) {
3749                $self->abort_edit;
3750        } else {
3751                $self->close_edit;
3752        }
3753        return scalar @$mods;
3754}
3755
3756package Git::SVN::Ra;
3757use vars qw/@ISA $config_dir $_log_window_size/;
3758use strict;
3759use warnings;
3760my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
3761
3762BEGIN {
3763        # enforce temporary pool usage for some simple functions
3764        no strict 'refs';
3765        for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root/) {
3766                my $SUPER = "SUPER::$f";
3767                *$f = sub {
3768                        my $self = shift;
3769                        my $pool = SVN::Pool->new;
3770                        my @ret = $self->$SUPER(@_,$pool);
3771                        $pool->clear;
3772                        wantarray ? @ret : $ret[0];
3773                };
3774        }
3775}
3776
3777sub _auth_providers () {
3778        [
3779          SVN::Client::get_simple_provider(),
3780          SVN::Client::get_ssl_server_trust_file_provider(),
3781          SVN::Client::get_simple_prompt_provider(
3782            \&Git::SVN::Prompt::simple, 2),
3783          SVN::Client::get_ssl_client_cert_file_provider(),
3784          SVN::Client::get_ssl_client_cert_prompt_provider(
3785            \&Git::SVN::Prompt::ssl_client_cert, 2),
3786          SVN::Client::get_ssl_client_cert_pw_file_provider(),
3787          SVN::Client::get_ssl_client_cert_pw_prompt_provider(
3788            \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
3789          SVN::Client::get_username_provider(),
3790          SVN::Client::get_ssl_server_trust_prompt_provider(
3791            \&Git::SVN::Prompt::ssl_server_trust),
3792          SVN::Client::get_username_prompt_provider(
3793            \&Git::SVN::Prompt::username, 2)
3794        ]
3795}
3796
3797sub escape_uri_only {
3798        my ($uri) = @_;
3799        my @tmp;
3800        foreach (split m{/}, $uri) {
3801                s/([^\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
3802                push @tmp, $_;
3803        }
3804        join('/', @tmp);
3805}
3806
3807sub escape_url {
3808        my ($url) = @_;
3809        if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
3810                my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
3811                $url = "$scheme://$domain$uri";
3812        }
3813        $url;
3814}
3815
3816sub new {
3817        my ($class, $url) = @_;
3818        $url =~ s!/+$!!;
3819        return $RA if ($RA && $RA->{url} eq $url);
3820
3821        SVN::_Core::svn_config_ensure($config_dir, undef);
3822        my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
3823        my $config = SVN::Core::config_get_config($config_dir);
3824        $RA = undef;
3825        my $dont_store_passwords = 1;
3826        my $conf_t = ${$config}{'config'};
3827        {
3828                no warnings 'once';
3829                # The usage of $SVN::_Core::SVN_CONFIG_* variables
3830                # produces warnings that variables are used only once.
3831                # I had not found the better way to shut them up, so
3832                # the warnings of type 'once' are disabled in this block.
3833                if (SVN::_Core::svn_config_get_bool($conf_t,
3834                    $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3835                    $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
3836                    1) == 0) {
3837                        SVN::_Core::svn_auth_set_parameter($baton,
3838                            $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
3839                            bless (\$dont_store_passwords, "_p_void"));
3840                }
3841                if (SVN::_Core::svn_config_get_bool($conf_t,
3842                    $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3843                    $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
3844                    1) == 0) {
3845                        $Git::SVN::Prompt::_no_auth_cache = 1;
3846                }
3847        } # no warnings 'once'
3848        my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
3849                              config => $config,
3850                              pool => SVN::Pool->new,
3851                              auth_provider_callbacks => $callbacks);
3852        $self->{url} = $url;
3853        $self->{svn_path} = $url;
3854        $self->{repos_root} = $self->get_repos_root;
3855        $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
3856        $self->{cache} = { check_path => { r => 0, data => {} },
3857                           get_dir => { r => 0, data => {} } };
3858        $RA = bless $self, $class;
3859}
3860
3861sub check_path {
3862        my ($self, $path, $r) = @_;
3863        my $cache = $self->{cache}->{check_path};
3864        if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
3865                return $cache->{data}->{$path};
3866        }
3867        my $pool = SVN::Pool->new;
3868        my $t = $self->SUPER::check_path($path, $r, $pool);
3869        $pool->clear;
3870        if ($r != $cache->{r}) {
3871                %{$cache->{data}} = ();
3872                $cache->{r} = $r;
3873        }
3874        $cache->{data}->{$path} = $t;
3875}
3876
3877sub get_dir {
3878        my ($self, $dir, $r) = @_;
3879        my $cache = $self->{cache}->{get_dir};
3880        if ($r == $cache->{r}) {
3881                if (my $x = $cache->{data}->{$dir}) {
3882                        return wantarray ? @$x : $x->[0];
3883                }
3884        }
3885        my $pool = SVN::Pool->new;
3886        my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
3887        my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
3888        $pool->clear;
3889        if ($r != $cache->{r}) {
3890                %{$cache->{data}} = ();
3891                $cache->{r} = $r;
3892        }
3893        $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
3894        wantarray ? (\%dirents, $r, $props) : \%dirents;
3895}
3896
3897sub DESTROY {
3898        # do not call the real DESTROY since we store ourselves in $RA
3899}
3900
3901sub get_log {
3902        my ($self, @args) = @_;
3903        my $pool = SVN::Pool->new;
3904        splice(@args, 3, 1) if ($SVN::Core::VERSION le '1.2.0');
3905        my $ret = $self->SUPER::get_log(@args, $pool);
3906        $pool->clear;
3907        $ret;
3908}
3909
3910sub trees_match {
3911        my ($self, $url1, $rev1, $url2, $rev2) = @_;
3912        my $ctx = SVN::Client->new(auth => _auth_providers);
3913        my $out = IO::File->new_tmpfile;
3914
3915        # older SVN (1.1.x) doesn't take $pool as the last parameter for
3916        # $ctx->diff(), so we'll create a default one
3917        my $pool = SVN::Pool->new_default_sub;
3918
3919        $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
3920        $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
3921        $out->flush;
3922        my $ret = (($out->stat)[7] == 0);
3923        close $out or croak $!;
3924
3925        $ret;
3926}
3927
3928sub get_commit_editor {
3929        my ($self, $log, $cb, $pool) = @_;
3930        my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
3931        $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
3932}
3933
3934sub gs_do_update {
3935        my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
3936        my $new = ($rev_a == $rev_b);
3937        my $path = $gs->{path};
3938
3939        if ($new && -e $gs->{index}) {
3940                unlink $gs->{index} or die
3941                  "Couldn't unlink index: $gs->{index}: $!\n";
3942        }
3943        my $pool = SVN::Pool->new;
3944        $editor->set_path_strip($path);
3945        my (@pc) = split m#/#, $path;
3946        my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
3947                                        1, $editor, $pool);
3948        my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3949
3950        # Since we can't rely on svn_ra_reparent being available, we'll
3951        # just have to do some magic with set_path to make it so
3952        # we only want a partial path.
3953        my $sp = '';
3954        my $final = join('/', @pc);
3955        while (@pc) {
3956                $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
3957                $sp .= '/' if length $sp;
3958                $sp .= shift @pc;
3959        }
3960        die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
3961
3962        $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
3963
3964        $reporter->finish_report($pool);
3965        $pool->clear;
3966        $editor->{git_commit_ok};
3967}
3968
3969# this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
3970# svn_ra_reparent didn't work before 1.4)
3971sub gs_do_switch {
3972        my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
3973        my $path = $gs->{path};
3974        my $pool = SVN::Pool->new;
3975
3976        my $full_url = $self->{url};
3977        my $old_url = $full_url;
3978        $full_url .= '/' . escape_uri_only($path) if length $path;
3979        my ($ra, $reparented);
3980
3981        if ($old_url =~ m#^svn(\+ssh)?://#) {
3982                $_[0] = undef;
3983                $self = undef;
3984                $RA = undef;
3985                $ra = Git::SVN::Ra->new($full_url);
3986                $ra_invalid = 1;
3987        } elsif ($old_url ne $full_url) {
3988                SVN::_Ra::svn_ra_reparent($self->{session}, $full_url, $pool);
3989                $self->{url} = $full_url;
3990                $reparented = 1;
3991        }
3992
3993        $ra ||= $self;
3994        $url_b = escape_url($url_b);
3995        my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
3996        my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3997        $reporter->set_path('', $rev_a, 0, @lock, $pool);
3998        $reporter->finish_report($pool);
3999
4000        if ($reparented) {
4001                SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
4002                $self->{url} = $old_url;
4003        }
4004
4005        $pool->clear;
4006        $editor->{git_commit_ok};
4007}
4008
4009sub longest_common_path {
4010        my ($gsv, $globs) = @_;
4011        my %common;
4012        my $common_max = scalar @$gsv;
4013
4014        foreach my $gs (@$gsv) {
4015                my @tmp = split m#/#, $gs->{path};
4016                my $p = '';
4017                foreach (@tmp) {
4018                        $p .= length($p) ? "/$_" : $_;
4019                        $common{$p} ||= 0;
4020                        $common{$p}++;
4021                }
4022        }
4023        $globs ||= [];
4024        $common_max += scalar @$globs;
4025        foreach my $glob (@$globs) {
4026                my @tmp = split m#/#, $glob->{path}->{left};
4027                my $p = '';
4028                foreach (@tmp) {
4029                        $p .= length($p) ? "/$_" : $_;
4030                        $common{$p} ||= 0;
4031                        $common{$p}++;
4032                }
4033        }
4034
4035        my $longest_path = '';
4036        foreach (sort {length $b <=> length $a} keys %common) {
4037                if ($common{$_} == $common_max) {
4038                        $longest_path = $_;
4039                        last;
4040                }
4041        }
4042        $longest_path;
4043}
4044
4045sub gs_fetch_loop_common {
4046        my ($self, $base, $head, $gsv, $globs) = @_;
4047        return if ($base > $head);
4048        my $inc = $_log_window_size;
4049        my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
4050        my $longest_path = longest_common_path($gsv, $globs);
4051        my $ra_url = $self->{url};
4052        while (1) {
4053                my %revs;
4054                my $err;
4055                my $err_handler = $SVN::Error::handler;
4056                $SVN::Error::handler = sub {
4057                        ($err) = @_;
4058                        skip_unknown_revs($err);
4059                };
4060                sub _cb {
4061                        my ($paths, $r, $author, $date, $log) = @_;
4062                        [ dup_changed_paths($paths),
4063                          { author => $author, date => $date, log => $log } ];
4064                }
4065                $self->get_log([$longest_path], $min, $max, 0, 1, 1,
4066                               sub { $revs{$_[1]} = _cb(@_) });
4067                if ($err && $max >= $head) {
4068                        print STDERR "Path '$longest_path' ",
4069                                     "was probably deleted:\n",
4070                                     $err->expanded_message,
4071                                     "\nWill attempt to follow ",
4072                                     "revisions r$min .. r$max ",
4073                                     "committed before the deletion\n";
4074                        my $hi = $max;
4075                        while (--$hi >= $min) {
4076                                my $ok;
4077                                $self->get_log([$longest_path], $min, $hi,
4078                                               0, 1, 1, sub {
4079                                               $ok ||= $_[1];
4080                                               $revs{$_[1]} = _cb(@_) });
4081                                if ($ok) {
4082                                        print STDERR "r$min .. r$ok OK\n";
4083                                        last;
4084                                }
4085                        }
4086                }
4087                $SVN::Error::handler = $err_handler;
4088
4089                my %exists = map { $_->{path} => $_ } @$gsv;
4090                foreach my $r (sort {$a <=> $b} keys %revs) {
4091                        my ($paths, $logged) = @{$revs{$r}};
4092
4093                        foreach my $gs ($self->match_globs(\%exists, $paths,
4094                                                           $globs, $r)) {
4095                                if ($gs->rev_map_max >= $r) {
4096                                        next;
4097                                }
4098                                next unless $gs->match_paths($paths, $r);
4099                                $gs->{logged_rev_props} = $logged;
4100                                if (my $last_commit = $gs->last_commit) {
4101                                        $gs->assert_index_clean($last_commit);
4102                                }
4103                                my $log_entry = $gs->do_fetch($paths, $r);
4104                                if ($log_entry) {
4105                                        $gs->do_git_commit($log_entry);
4106                                }
4107                                $INDEX_FILES{$gs->{index}} = 1;
4108                        }
4109                        foreach my $g (@$globs) {
4110                                my $k = "svn-remote.$g->{remote}." .
4111                                        "$g->{t}-maxRev";
4112                                Git::SVN::tmp_config($k, $r);
4113                        }
4114                        if ($ra_invalid) {
4115                                $_[0] = undef;
4116                                $self = undef;
4117                                $RA = undef;
4118                                $self = Git::SVN::Ra->new($ra_url);
4119                                $ra_invalid = undef;
4120                        }
4121                }
4122                # pre-fill the .rev_db since it'll eventually get filled in
4123                # with '0' x40 if something new gets committed
4124                foreach my $gs (@$gsv) {
4125                        next if $gs->rev_map_max >= $max;
4126                        next if defined $gs->rev_map_get($max);
4127                        $gs->rev_map_set($max, 0 x40);
4128                }
4129                foreach my $g (@$globs) {
4130                        my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
4131                        Git::SVN::tmp_config($k, $max);
4132                }
4133                last if $max >= $head;
4134                $min = $max + 1;
4135                $max += $inc;
4136                $max = $head if ($max > $head);
4137        }
4138        Git::SVN::gc();
4139}
4140
4141sub get_dir_globbed {
4142        my ($self, $left, $depth, $r) = @_;
4143
4144        my @x = eval { $self->get_dir($left, $r) };
4145        return unless scalar @x == 3;
4146        my $dirents = $x[0];
4147        my @finalents;
4148        foreach my $de (keys %$dirents) {
4149                next if $dirents->{$de}->{kind} != $SVN::Node::dir;
4150                if ($depth > 1) {
4151                        my @args = ("$left/$de", $depth - 1, $r);
4152                        foreach my $dir ($self->get_dir_globbed(@args)) {
4153                                push @finalents, "$de/$dir";
4154                        }
4155                } else {
4156                        push @finalents, $de;
4157                }
4158        }
4159        @finalents;
4160}
4161
4162sub match_globs {
4163        my ($self, $exists, $paths, $globs, $r) = @_;
4164
4165        sub get_dir_check {
4166                my ($self, $exists, $g, $r) = @_;
4167
4168                my @dirs = $self->get_dir_globbed($g->{path}->{left},
4169                                                  $g->{path}->{depth},
4170                                                  $r);
4171
4172                foreach my $de (@dirs) {
4173                        my $p = $g->{path}->full_path($de);
4174                        next if $exists->{$p};
4175                        next if (length $g->{path}->{right} &&
4176                                 ($self->check_path($p, $r) !=
4177                                  $SVN::Node::dir));
4178                        $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
4179                                         $g->{ref}->full_path($de), 1);
4180                }
4181        }
4182        foreach my $g (@$globs) {
4183                if (my $path = $paths->{"/$g->{path}->{left}"}) {
4184                        if ($path->{action} =~ /^[AR]$/) {
4185                                get_dir_check($self, $exists, $g, $r);
4186                        }
4187                }
4188                foreach (keys %$paths) {
4189                        if (/$g->{path}->{left_regex}/ &&
4190                            !/$g->{path}->{regex}/) {
4191                                next if $paths->{$_}->{action} !~ /^[AR]$/;
4192                                get_dir_check($self, $exists, $g, $r);
4193                        }
4194                        next unless /$g->{path}->{regex}/;
4195                        my $p = $1;
4196                        my $pathname = $g->{path}->full_path($p);
4197                        next if $exists->{$pathname};
4198                        next if ($self->check_path($pathname, $r) !=
4199                                 $SVN::Node::dir);
4200                        $exists->{$pathname} = Git::SVN->init(
4201                                              $self->{url}, $pathname, undef,
4202                                              $g->{ref}->full_path($p), 1);
4203                }
4204                my $c = '';
4205                foreach (split m#/#, $g->{path}->{left}) {
4206                        $c .= "/$_";
4207                        next unless ($paths->{$c} &&
4208                                     ($paths->{$c}->{action} =~ /^[AR]$/));
4209                        get_dir_check($self, $exists, $g, $r);
4210                }
4211        }
4212        values %$exists;
4213}
4214
4215sub minimize_url {
4216        my ($self) = @_;
4217        return $self->{url} if ($self->{url} eq $self->{repos_root});
4218        my $url = $self->{repos_root};
4219        my @components = split(m!/!, $self->{svn_path});
4220        my $c = '';
4221        do {
4222                $url .= "/$c" if length $c;
4223                eval { (ref $self)->new($url)->get_latest_revnum };
4224        } while ($@ && ($c = shift @components));
4225        $url;
4226}
4227
4228sub can_do_switch {
4229        my $self = shift;
4230        unless (defined $can_do_switch) {
4231                my $pool = SVN::Pool->new;
4232                my $rep = eval {
4233                        $self->do_switch(1, '', 0, $self->{url},
4234                                         SVN::Delta::Editor->new, $pool);
4235                };
4236                if ($@) {
4237                        $can_do_switch = 0;
4238                } else {
4239                        $rep->abort_report($pool);
4240                        $can_do_switch = 1;
4241                }
4242                $pool->clear;
4243        }
4244        $can_do_switch;
4245}
4246
4247sub skip_unknown_revs {
4248        my ($err) = @_;
4249        my $errno = $err->apr_err();
4250        # Maybe the branch we're tracking didn't
4251        # exist when the repo started, so it's
4252        # not an error if it doesn't, just continue
4253        #
4254        # Wonderfully consistent library, eh?
4255        # 160013 - svn:// and file://
4256        # 175002 - http(s)://
4257        # 175007 - http(s):// (this repo required authorization, too...)
4258        #   More codes may be discovered later...
4259        if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
4260                my $err_key = $err->expanded_message;
4261                # revision numbers change every time, filter them out
4262                $err_key =~ s/\d+/\0/g;
4263                $err_key = "$errno\0$err_key";
4264                unless ($ignored_err{$err_key}) {
4265                        warn "W: Ignoring error from SVN, path probably ",
4266                             "does not exist: ($errno): ",
4267                             $err->expanded_message,"\n";
4268                        warn "W: Do not be alarmed at the above message ",
4269                             "git-svn is just searching aggressively for ",
4270                             "old history.\n",
4271                             "This may take a while on large repositories\n";
4272                        $ignored_err{$err_key} = 1;
4273                }
4274                return;
4275        }
4276        die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
4277}
4278
4279# svn_log_changed_path_t objects passed to get_log are likely to be
4280# overwritten even if only the refs are copied to an external variable,
4281# so we should dup the structures in their entirety.  Using an externally
4282# passed pool (instead of our temporary and quickly cleared pool in
4283# Git::SVN::Ra) does not help matters at all...
4284sub dup_changed_paths {
4285        my ($paths) = @_;
4286        return undef unless $paths;
4287        my %ret;
4288        foreach my $p (keys %$paths) {
4289                my $i = $paths->{$p};
4290                my %s = map { $_ => $i->$_ }
4291                              qw/copyfrom_path copyfrom_rev action/;
4292                $ret{$p} = \%s;
4293        }
4294        \%ret;
4295}
4296
4297package Git::SVN::Log;
4298use strict;
4299use warnings;
4300use POSIX qw/strftime/;
4301use constant commit_log_separator => ('-' x 72) . "\n";
4302use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
4303            %rusers $show_commit $incremental/;
4304my $l_fmt;
4305
4306sub cmt_showable {
4307        my ($c) = @_;
4308        return 1 if defined $c->{r};
4309
4310        # big commit message got truncated by the 16k pretty buffer in rev-list
4311        if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
4312                                $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
4313                @{$c->{l}} = ();
4314                my @log = command(qw/cat-file commit/, $c->{c});
4315
4316                # shift off the headers
4317                shift @log while ($log[0] ne '');
4318                shift @log;
4319
4320                # TODO: make $c->{l} not have a trailing newline in the future
4321                @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
4322
4323                (undef, $c->{r}, undef) = ::extract_metadata(
4324                                (grep(/^git-svn-id: /, @log))[-1]);
4325        }
4326        return defined $c->{r};
4327}
4328
4329sub log_use_color {
4330        return $color || Git->repository->get_colorbool('color.diff');
4331}
4332
4333sub git_svn_log_cmd {
4334        my ($r_min, $r_max, @args) = @_;
4335        my $head = 'HEAD';
4336        my (@files, @log_opts);
4337        foreach my $x (@args) {
4338                if ($x eq '--' || @files) {
4339                        push @files, $x;
4340                } else {
4341                        if (::verify_ref("$x^0")) {
4342                                $head = $x;
4343                        } else {
4344                                push @log_opts, $x;
4345                        }
4346                }
4347        }
4348
4349        my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
4350        $gs ||= Git::SVN->_new;
4351        my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
4352                   $gs->refname);
4353        push @cmd, '-r' unless $non_recursive;
4354        push @cmd, qw/--raw --name-status/ if $verbose;
4355        push @cmd, '--color' if log_use_color();
4356        push @cmd, @log_opts;
4357        if (defined $r_max && $r_max == $r_min) {
4358                push @cmd, '--max-count=1';
4359                if (my $c = $gs->rev_map_get($r_max)) {
4360                        push @cmd, $c;
4361                }
4362        } elsif (defined $r_max) {
4363                if ($r_max < $r_min) {
4364                        ($r_min, $r_max) = ($r_max, $r_min);
4365                }
4366                my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
4367                my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
4368                # If there are no commits in the range, both $c_max and $c_min
4369                # will be undefined.  If there is at least 1 commit in the
4370                # range, both will be defined.
4371                return () if !defined $c_min || !defined $c_max;
4372                if ($c_min eq $c_max) {
4373                        push @cmd, '--max-count=1', $c_min;
4374                } else {
4375                        push @cmd, '--boundary', "$c_min..$c_max";
4376                }
4377        }
4378        return (@cmd, @files);
4379}
4380
4381# adapted from pager.c
4382sub config_pager {
4383        $pager ||= $ENV{GIT_PAGER} || $ENV{PAGER};
4384        if (!defined $pager) {
4385                $pager = 'less';
4386        } elsif (length $pager == 0 || $pager eq 'cat') {
4387                $pager = undef;
4388        }
4389        $ENV{GIT_PAGER_IN_USE} = defined($pager);
4390}
4391
4392sub run_pager {
4393        return unless -t *STDOUT && defined $pager;
4394        pipe my ($rfd, $wfd) or return;
4395        defined(my $pid = fork) or ::fatal "Can't fork: $!";
4396        if (!$pid) {
4397                open STDOUT, '>&', $wfd or
4398                                     ::fatal "Can't redirect to stdout: $!";
4399                return;
4400        }
4401        open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
4402        $ENV{LESS} ||= 'FRSX';
4403        exec $pager or ::fatal "Can't run pager: $! ($pager)";
4404}
4405
4406sub format_svn_date {
4407        return strftime("%Y-%m-%d %H:%M:%S %z (%a, %d %b %Y)", localtime(shift));
4408}
4409
4410sub parse_git_date {
4411        my ($t, $tz) = @_;
4412        # Date::Parse isn't in the standard Perl distro :(
4413        if ($tz =~ s/^\+//) {
4414                $t += tz_to_s_offset($tz);
4415        } elsif ($tz =~ s/^\-//) {
4416                $t -= tz_to_s_offset($tz);
4417        }
4418        return $t;
4419}
4420
4421sub set_local_timezone {
4422        if (defined $TZ) {
4423                $ENV{TZ} = $TZ;
4424        } else {
4425                delete $ENV{TZ};
4426        }
4427}
4428
4429sub tz_to_s_offset {
4430        my ($tz) = @_;
4431        $tz =~ s/(\d\d)$//;
4432        return ($1 * 60) + ($tz * 3600);
4433}
4434
4435sub get_author_info {
4436        my ($dest, $author, $t, $tz) = @_;
4437        $author =~ s/(?:^\s*|\s*$)//g;
4438        $dest->{a_raw} = $author;
4439        my $au;
4440        if ($::_authors) {
4441                $au = $rusers{$author} || undef;
4442        }
4443        if (!$au) {
4444                ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
4445        }
4446        $dest->{t} = $t;
4447        $dest->{tz} = $tz;
4448        $dest->{a} = $au;
4449        $dest->{t_utc} = parse_git_date($t, $tz);
4450}
4451
4452sub process_commit {
4453        my ($c, $r_min, $r_max, $defer) = @_;
4454        if (defined $r_min && defined $r_max) {
4455                if ($r_min == $c->{r} && $r_min == $r_max) {
4456                        show_commit($c);
4457                        return 0;
4458                }
4459                return 1 if $r_min == $r_max;
4460                if ($r_min < $r_max) {
4461                        # we need to reverse the print order
4462                        return 0 if (defined $limit && --$limit < 0);
4463                        push @$defer, $c;
4464                        return 1;
4465                }
4466                if ($r_min != $r_max) {
4467                        return 1 if ($r_min < $c->{r});
4468                        return 1 if ($r_max > $c->{r});
4469                }
4470        }
4471        return 0 if (defined $limit && --$limit < 0);
4472        show_commit($c);
4473        return 1;
4474}
4475
4476sub show_commit {
4477        my $c = shift;
4478        if ($oneline) {
4479                my $x = "\n";
4480                if (my $l = $c->{l}) {
4481                        while ($l->[0] =~ /^\s*$/) { shift @$l }
4482                        $x = $l->[0];
4483                }
4484                $l_fmt ||= 'A' . length($c->{r});
4485                print 'r',pack($l_fmt, $c->{r}),' | ';
4486                print "$c->{c} | " if $show_commit;
4487                print $x;
4488        } else {
4489                show_commit_normal($c);
4490        }
4491}
4492
4493sub show_commit_changed_paths {
4494        my ($c) = @_;
4495        return unless $c->{changed};
4496        print "Changed paths:\n", @{$c->{changed}};
4497}
4498
4499sub show_commit_normal {
4500        my ($c) = @_;
4501        print commit_log_separator, "r$c->{r} | ";
4502        print "$c->{c} | " if $show_commit;
4503        print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
4504        my $nr_line = 0;
4505
4506        if (my $l = $c->{l}) {
4507                while ($l->[$#$l] eq "\n" && $#$l > 0
4508                                          && $l->[($#$l - 1)] eq "\n") {
4509                        pop @$l;
4510                }
4511                $nr_line = scalar @$l;
4512                if (!$nr_line) {
4513                        print "1 line\n\n\n";
4514                } else {
4515                        if ($nr_line == 1) {
4516                                $nr_line = '1 line';
4517                        } else {
4518                                $nr_line .= ' lines';
4519                        }
4520                        print $nr_line, "\n";
4521                        show_commit_changed_paths($c);
4522                        print "\n";
4523                        print $_ foreach @$l;
4524                }
4525        } else {
4526                print "1 line\n";
4527                show_commit_changed_paths($c);
4528                print "\n";
4529
4530        }
4531        foreach my $x (qw/raw stat diff/) {
4532                if ($c->{$x}) {
4533                        print "\n";
4534                        print $_ foreach @{$c->{$x}}
4535                }
4536        }
4537}
4538
4539sub cmd_show_log {
4540        my (@args) = @_;
4541        my ($r_min, $r_max);
4542        my $r_last = -1; # prevent dupes
4543        set_local_timezone();
4544        if (defined $::_revision) {
4545                if ($::_revision =~ /^(\d+):(\d+)$/) {
4546                        ($r_min, $r_max) = ($1, $2);
4547                } elsif ($::_revision =~ /^\d+$/) {
4548                        $r_min = $r_max = $::_revision;
4549                } else {
4550                        ::fatal "-r$::_revision is not supported, use ",
4551                                "standard 'git log' arguments instead";
4552                }
4553        }
4554
4555        config_pager();
4556        @args = git_svn_log_cmd($r_min, $r_max, @args);
4557        if (!@args) {
4558                print commit_log_separator unless $incremental || $oneline;
4559                return;
4560        }
4561        my $log = command_output_pipe(@args);
4562        run_pager();
4563        my (@k, $c, $d, $stat);
4564        my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
4565        while (<$log>) {
4566                if (/^${esc_color}commit -?($::sha1_short)/o) {
4567                        my $cmt = $1;
4568                        if ($c && cmt_showable($c) && $c->{r} != $r_last) {
4569                                $r_last = $c->{r};
4570                                process_commit($c, $r_min, $r_max, \@k) or
4571                                                                goto out;
4572                        }
4573                        $d = undef;
4574                        $c = { c => $cmt };
4575                } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
4576                        get_author_info($c, $1, $2, $3);
4577                } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
4578                        # ignore
4579                } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
4580                        push @{$c->{raw}}, $_;
4581                } elsif (/^${esc_color}[ACRMDT]\t/) {
4582                        # we could add $SVN->{svn_path} here, but that requires
4583                        # remote access at the moment (repo_path_split)...
4584                        s#^(${esc_color})([ACRMDT])\t#$1   $2 #o;
4585                        push @{$c->{changed}}, $_;
4586                } elsif (/^${esc_color}diff /o) {
4587                        $d = 1;
4588                        push @{$c->{diff}}, $_;
4589                } elsif ($d) {
4590                        push @{$c->{diff}}, $_;
4591                } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
4592                          $esc_color*[\+\-]*$esc_color$/x) {
4593                        $stat = 1;
4594                        push @{$c->{stat}}, $_;
4595                } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
4596                        push @{$c->{stat}}, $_;
4597                        $stat = undef;
4598                } elsif (/^${esc_color}    (git-svn-id:.+)$/o) {
4599                        ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
4600                } elsif (s/^${esc_color}    //o) {
4601                        push @{$c->{l}}, $_;
4602                }
4603        }
4604        if ($c && defined $c->{r} && $c->{r} != $r_last) {
4605                $r_last = $c->{r};
4606                process_commit($c, $r_min, $r_max, \@k);
4607        }
4608        if (@k) {
4609                ($r_min, $r_max) = ($r_max, $r_min);
4610                process_commit($_, $r_min, $r_max) foreach reverse @k;
4611        }
4612out:
4613        close $log;
4614        print commit_log_separator unless $incremental || $oneline;
4615}
4616
4617sub cmd_blame {
4618        my $path = pop;
4619
4620        config_pager();
4621        run_pager();
4622
4623        my ($fh, $ctx, $rev);
4624
4625        if ($_git_format) {
4626                ($fh, $ctx) = command_output_pipe('blame', @_, $path);
4627                while (my $line = <$fh>) {
4628                        if ($line =~ /^\^?([[:xdigit:]]+)\s/) {
4629                                # Uncommitted edits show up as a rev ID of
4630                                # all zeros, which we can't look up with
4631                                # cmt_metadata
4632                                if ($1 !~ /^0+$/) {
4633                                        (undef, $rev, undef) =
4634                                                ::cmt_metadata($1);
4635                                        $rev = '0' if (!$rev);
4636                                } else {
4637                                        $rev = '0';
4638                                }
4639                                $rev = sprintf('%-10s', $rev);
4640                                $line =~ s/^\^?[[:xdigit:]]+(\s)/$rev$1/;
4641                        }
4642                        print $line;
4643                }
4644        } else {
4645                ($fh, $ctx) = command_output_pipe('blame', '-p', @_, 'HEAD',
4646                                                  '--', $path);
4647                my ($sha1);
4648                my %authors;
4649                while (my $line = <$fh>) {
4650                        if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
4651                                $sha1 = $1;
4652                                (undef, $rev, undef) = ::cmt_metadata($1);
4653                                $rev = '0' if (!$rev);
4654                        }
4655                        elsif ($line =~ /^author (.*)/) {
4656                                $authors{$rev} = $1;
4657                                $authors{$rev} =~ s/\s/_/g;
4658                        }
4659                        elsif ($line =~ /^\t(.*)$/) {
4660                                printf("%6s %10s %s\n", $rev, $authors{$rev}, $1);
4661                        }
4662                }
4663        }
4664        command_close_pipe($fh, $ctx);
4665}
4666
4667package Git::SVN::Migration;
4668# these version numbers do NOT correspond to actual version numbers
4669# of git nor git-svn.  They are just relative.
4670#
4671# v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
4672#
4673# v1 layout: .git/$id/info/url, refs/remotes/$id
4674#
4675# v2 layout: .git/svn/$id/info/url, refs/remotes/$id
4676#
4677# v3 layout: .git/svn/$id, refs/remotes/$id
4678#            - info/url may remain for backwards compatibility
4679#            - this is what we migrate up to this layout automatically,
4680#            - this will be used by git svn init on single branches
4681# v3.1 layout (auto migrated):
4682#            - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
4683#              for backwards compatibility
4684#
4685# v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
4686#            - this is only created for newly multi-init-ed
4687#              repositories.  Similar in spirit to the
4688#              --use-separate-remotes option in git-clone (now default)
4689#            - we do not automatically migrate to this (following
4690#              the example set by core git)
4691#
4692# v5 layout: .rev_db.$UUID => .rev_map.$UUID
4693#            - newer, more-efficient format that uses 24-bytes per record
4694#              with no filler space.
4695#            - use xxd -c24 < .rev_map.$UUID to view and debug
4696#            - This is a one-way migration, repositories updated to the
4697#              new format will not be able to use old git-svn without
4698#              rebuilding the .rev_db.  Rebuilding the rev_db is not
4699#              possible if noMetadata or useSvmProps are set; but should
4700#              be no problem for users that use the (sensible) defaults.
4701use strict;
4702use warnings;
4703use Carp qw/croak/;
4704use File::Path qw/mkpath/;
4705use File::Basename qw/dirname basename/;
4706use vars qw/$_minimize/;
4707
4708sub migrate_from_v0 {
4709        my $git_dir = $ENV{GIT_DIR};
4710        return undef unless -d $git_dir;
4711        my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4712        my $migrated = 0;
4713        while (<$fh>) {
4714                chomp;
4715                my ($id, $orig_ref) = ($_, $_);
4716                next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
4717                next unless -f "$git_dir/$id/info/url";
4718                my $new_ref = "refs/remotes/$id";
4719                if (::verify_ref("$new_ref^0")) {
4720                        print STDERR "W: $orig_ref is probably an old ",
4721                                     "branch used by an ancient version of ",
4722                                     "git-svn.\n",
4723                                     "However, $new_ref also exists.\n",
4724                                     "We will not be able ",
4725                                     "to use this branch until this ",
4726                                     "ambiguity is resolved.\n";
4727                        next;
4728                }
4729                print STDERR "Migrating from v0 layout...\n" if !$migrated;
4730                print STDERR "Renaming ref: $orig_ref => $new_ref\n";
4731                command_noisy('update-ref', $new_ref, $orig_ref);
4732                command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
4733                $migrated++;
4734        }
4735        command_close_pipe($fh, $ctx);
4736        print STDERR "Done migrating from v0 layout...\n" if $migrated;
4737        $migrated;
4738}
4739
4740sub migrate_from_v1 {
4741        my $git_dir = $ENV{GIT_DIR};
4742        my $migrated = 0;
4743        return $migrated unless -d $git_dir;
4744        my $svn_dir = "$git_dir/svn";
4745
4746        # just in case somebody used 'svn' as their $id at some point...
4747        return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
4748
4749        print STDERR "Migrating from a git-svn v1 layout...\n";
4750        mkpath([$svn_dir]);
4751        print STDERR "Data from a previous version of git-svn exists, but\n\t",
4752                     "$svn_dir\n\t(required for this version ",
4753                     "($::VERSION) of git-svn) does not exist.\n";
4754        my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4755        while (<$fh>) {
4756                my $x = $_;
4757                next unless $x =~ s#^refs/remotes/##;
4758                chomp $x;
4759                next unless -f "$git_dir/$x/info/url";
4760                my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
4761                next unless $u;
4762                my $dn = dirname("$git_dir/svn/$x");
4763                mkpath([$dn]) unless -d $dn;
4764                if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
4765                        mkpath(["$git_dir/svn/svn"]);
4766                        print STDERR " - $git_dir/$x/info => ",
4767                                        "$git_dir/svn/$x/info\n";
4768                        rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
4769                               croak "$!: $x";
4770                        # don't worry too much about these, they probably
4771                        # don't exist with repos this old (save for index,
4772                        # and we can easily regenerate that)
4773                        foreach my $f (qw/unhandled.log index .rev_db/) {
4774                                rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
4775                        }
4776                } else {
4777                        print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
4778                        rename "$git_dir/$x", "$git_dir/svn/$x" or
4779                               croak "$!: $x";
4780                }
4781                $migrated++;
4782        }
4783        command_close_pipe($fh, $ctx);
4784        print STDERR "Done migrating from a git-svn v1 layout\n";
4785        $migrated;
4786}
4787
4788sub read_old_urls {
4789        my ($l_map, $pfx, $path) = @_;
4790        my @dir;
4791        foreach (<$path/*>) {
4792                if (-r "$_/info/url") {
4793                        $pfx .= '/' if $pfx && $pfx !~ m!/$!;
4794                        my $ref_id = $pfx . basename $_;
4795                        my $url = ::file_to_s("$_/info/url");
4796                        $l_map->{$ref_id} = $url;
4797                } elsif (-d $_) {
4798                        push @dir, $_;
4799                }
4800        }
4801        foreach (@dir) {
4802                my $x = $_;
4803                $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
4804                read_old_urls($l_map, $x, $_);
4805        }
4806}
4807
4808sub migrate_from_v2 {
4809        my @cfg = command(qw/config -l/);
4810        return if grep /^svn-remote\..+\.url=/, @cfg;
4811        my %l_map;
4812        read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
4813        my $migrated = 0;
4814
4815        foreach my $ref_id (sort keys %l_map) {
4816                eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
4817                if ($@) {
4818                        Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
4819                }
4820                $migrated++;
4821        }
4822        $migrated;
4823}
4824
4825sub minimize_connections {
4826        my $r = Git::SVN::read_all_remotes();
4827        my $new_urls = {};
4828        my $root_repos = {};
4829        foreach my $repo_id (keys %$r) {
4830                my $url = $r->{$repo_id}->{url} or next;
4831                my $fetch = $r->{$repo_id}->{fetch} or next;
4832                my $ra = Git::SVN::Ra->new($url);
4833
4834                # skip existing cases where we already connect to the root
4835                if (($ra->{url} eq $ra->{repos_root}) ||
4836                    ($ra->{repos_root} eq $repo_id)) {
4837                        $root_repos->{$ra->{url}} = $repo_id;
4838                        next;
4839                }
4840
4841                my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
4842                my $root_path = $ra->{url};
4843                $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
4844                foreach my $path (keys %$fetch) {
4845                        my $ref_id = $fetch->{$path};
4846                        my $gs = Git::SVN->new($ref_id, $repo_id, $path);
4847
4848                        # make sure we can read when connecting to
4849                        # a higher level of a repository
4850                        my ($last_rev, undef) = $gs->last_rev_commit;
4851                        if (!defined $last_rev) {
4852                                $last_rev = eval {
4853                                        $root_ra->get_latest_revnum;
4854                                };
4855                                next if $@;
4856                        }
4857                        my $new = $root_path;
4858                        $new .= length $path ? "/$path" : '';
4859                        eval {
4860                                $root_ra->get_log([$new], $last_rev, $last_rev,
4861                                                  0, 0, 1, sub { });
4862                        };
4863                        next if $@;
4864                        $new_urls->{$ra->{repos_root}}->{$new} =
4865                                { ref_id => $ref_id,
4866                                  old_repo_id => $repo_id,
4867                                  old_path => $path };
4868                }
4869        }
4870
4871        my @emptied;
4872        foreach my $url (keys %$new_urls) {
4873                # see if we can re-use an existing [svn-remote "repo_id"]
4874                # instead of creating a(n ugly) new section:
4875                my $repo_id = $root_repos->{$url} || $url;
4876
4877                my $fetch = $new_urls->{$url};
4878                foreach my $path (keys %$fetch) {
4879                        my $x = $fetch->{$path};
4880                        Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
4881                        my $pfx = "svn-remote.$x->{old_repo_id}";
4882
4883                        my $old_fetch = quotemeta("$x->{old_path}:".
4884                                                  "refs/remotes/$x->{ref_id}");
4885                        command_noisy(qw/config --unset/,
4886                                      "$pfx.fetch", '^'. $old_fetch . '$');
4887                        delete $r->{$x->{old_repo_id}}->
4888                               {fetch}->{$x->{old_path}};
4889                        if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
4890                                command_noisy(qw/config --unset/,
4891                                              "$pfx.url");
4892                                push @emptied, $x->{old_repo_id}
4893                        }
4894                }
4895        }
4896        if (@emptied) {
4897                my $file = $ENV{GIT_CONFIG} || $ENV{GIT_CONFIG_LOCAL} ||
4898                           "$ENV{GIT_DIR}/config";
4899                print STDERR <<EOF;
4900The following [svn-remote] sections in your config file ($file) are empty
4901and can be safely removed:
4902EOF
4903                print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
4904        }
4905}
4906
4907sub migration_check {
4908        migrate_from_v0();
4909        migrate_from_v1();
4910        migrate_from_v2();
4911        minimize_connections() if $_minimize;
4912}
4913
4914package Git::IndexInfo;
4915use strict;
4916use warnings;
4917use Git qw/command_input_pipe command_close_pipe/;
4918
4919sub new {
4920        my ($class) = @_;
4921        my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
4922        bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
4923}
4924
4925sub remove {
4926        my ($self, $path) = @_;
4927        if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
4928                return ++$self->{nr};
4929        }
4930        undef;
4931}
4932
4933sub update {
4934        my ($self, $mode, $hash, $path) = @_;
4935        if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
4936                return ++$self->{nr};
4937        }
4938        undef;
4939}
4940
4941sub DESTROY {
4942        my ($self) = @_;
4943        command_close_pipe($self->{gui}, $self->{ctx});
4944}
4945
4946package Git::SVN::GlobSpec;
4947use strict;
4948use warnings;
4949
4950sub new {
4951        my ($class, $glob) = @_;
4952        my $re = $glob;
4953        $re =~ s!/+$!!g; # no need for trailing slashes
4954        $re =~ m!^([^*]*)(\*(?:/\*)*)([^*]*)$!;
4955        my $temp = $re;
4956        my ($left, $right) = ($1, $3);
4957        $re = $2;
4958        my $depth = $re =~ tr/*/*/;
4959        if ($depth != $temp =~ tr/*/*/) {
4960                die "Only one set of wildcard directories " .
4961                        "(e.g. '*' or '*/*/*') is supported: '$glob'\n";
4962        }
4963        if ($depth == 0) {
4964                die "One '*' is needed for glob: '$glob'\n";
4965        }
4966        $re =~ s!\*!\[^/\]*!g;
4967        $re = quotemeta($left) . "($re)" . quotemeta($right);
4968        if (length $left && !($left =~ s!/+$!!g)) {
4969                die "Missing trailing '/' on left side of: '$glob' ($left)\n";
4970        }
4971        if (length $right && !($right =~ s!^/+!!g)) {
4972                die "Missing leading '/' on right side of: '$glob' ($right)\n";
4973        }
4974        my $left_re = qr/^\/\Q$left\E(\/|$)/;
4975        bless { left => $left, right => $right, left_regex => $left_re,
4976                regex => qr/$re/, glob => $glob, depth => $depth }, $class;
4977}
4978
4979sub full_path {
4980        my ($self, $path) = @_;
4981        return (length $self->{left} ? "$self->{left}/" : '') .
4982               $path . (length $self->{right} ? "/$self->{right}" : '');
4983}
4984
4985__END__
4986
4987Data structures:
4988
4989
4990$remotes = { # returned by read_all_remotes()
4991        'svn' => {
4992                # svn-remote.svn.url=https://svn.musicpd.org
4993                url => 'https://svn.musicpd.org',
4994                # svn-remote.svn.fetch=mpd/trunk:trunk
4995                fetch => {
4996                        'mpd/trunk' => 'trunk',
4997                },
4998                # svn-remote.svn.tags=mpd/tags/*:tags/*
4999                tags => {
5000                        path => {
5001                                left => 'mpd/tags',
5002                                right => '',
5003                                regex => qr!mpd/tags/([^/]+)$!,
5004                                glob => 'tags/*',
5005                        },
5006                        ref => {
5007                                left => 'tags',
5008                                right => '',
5009                                regex => qr!tags/([^/]+)$!,
5010                                glob => 'tags/*',
5011                        },
5012                }
5013        }
5014};
5015
5016$log_entry hashref as returned by libsvn_log_entry()
5017{
5018        log => 'whitespace-formatted log entry
5019',                                              # trailing newline is preserved
5020        revision => '8',                        # integer
5021        date => '2004-02-24T17:01:44.108345Z',  # commit date
5022        author => 'committer name'
5023};
5024
5025
5026# this is generated by generate_diff();
5027@mods = array of diff-index line hashes, each element represents one line
5028        of diff-index output
5029
5030diff-index line ($m hash)
5031{
5032        mode_a => first column of diff-index output, no leading ':',
5033        mode_b => second column of diff-index output,
5034        sha1_b => sha1sum of the final blob,
5035        chg => change type [MCRADT],
5036        file_a => original file name of a file (iff chg is 'C' or 'R')
5037        file_b => new/current file name of a file (any chg)
5038}
5039;
5040
5041# retval of read_url_paths{,_all}();
5042$l_map = {
5043        # repository root url
5044        'https://svn.musicpd.org' => {
5045                # repository path               # GIT_SVN_ID
5046                'mpd/trunk'             =>      'trunk',
5047                'mpd/tags/0.11.5'       =>      'tags/0.11.5',
5048        },
5049}
5050
5051Notes:
5052        I don't trust the each() function on unless I created %hash myself
5053        because the internal iterator may not have started at base.