git-svn.perlon commit Merge git://git.bogomips.org/git-svn (9716f21)
   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
   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 IO::File qw//;
  39use File::Basename qw/dirname basename/;
  40use File::Path qw/mkpath/;
  41use Getopt::Long qw/:config gnu_getopt no_ignore_case auto_abbrev/;
  42use IPC::Open3;
  43use Git;
  44
  45BEGIN {
  46        # import functions from Git into our packages, en masse
  47        no strict 'refs';
  48        foreach (qw/command command_oneline command_noisy command_output_pipe
  49                    command_input_pipe command_close_pipe/) {
  50                for my $package ( qw(SVN::Git::Editor SVN::Git::Fetcher
  51                        Git::SVN::Migration Git::SVN::Log Git::SVN),
  52                        __PACKAGE__) {
  53                        *{"${package}::$_"} = \&{"Git::$_"};
  54                }
  55        }
  56}
  57
  58my ($SVN);
  59
  60$sha1 = qr/[a-f\d]{40}/;
  61$sha1_short = qr/[a-f\d]{4,40}/;
  62my ($_stdin, $_help, $_edit,
  63        $_message, $_file,
  64        $_template, $_shared,
  65        $_version, $_fetch_all, $_no_rebase,
  66        $_merge, $_strategy, $_dry_run, $_local,
  67        $_prefix, $_no_checkout, $_verbose);
  68$Git::SVN::_follow_parent = 1;
  69my %remote_opts = ( 'username=s' => \$Git::SVN::Prompt::_username,
  70                    'config-dir=s' => \$Git::SVN::Ra::config_dir,
  71                    'no-auth-cache' => \$Git::SVN::Prompt::_no_auth_cache );
  72my %fc_opts = ( 'follow-parent|follow!' => \$Git::SVN::_follow_parent,
  73                'authors-file|A=s' => \$_authors,
  74                'repack:i' => \$Git::SVN::_repack,
  75                'noMetadata' => \$Git::SVN::_no_metadata,
  76                'useSvmProps' => \$Git::SVN::_use_svm_props,
  77                'useSvnsyncProps' => \$Git::SVN::_use_svnsync_props,
  78                'log-window-size=i' => \$Git::SVN::Ra::_log_window_size,
  79                'no-checkout' => \$_no_checkout,
  80                'quiet|q' => \$_q,
  81                'repack-flags|repack-args|repack-opts=s' =>
  82                   \$Git::SVN::_repack_flags,
  83                %remote_opts );
  84
  85my ($_trunk, $_tags, $_branches, $_stdlayout);
  86my %icv;
  87my %init_opts = ( 'template=s' => \$_template, 'shared:s' => \$_shared,
  88                  'trunk|T=s' => \$_trunk, 'tags|t=s' => \$_tags,
  89                  'branches|b=s' => \$_branches, 'prefix=s' => \$_prefix,
  90                  'stdlayout|s' => \$_stdlayout,
  91                  'minimize-url|m' => \$Git::SVN::_minimize_url,
  92                  'no-metadata' => sub { $icv{noMetadata} = 1 },
  93                  'use-svm-props' => sub { $icv{useSvmProps} = 1 },
  94                  'use-svnsync-props' => sub { $icv{useSvnsyncProps} = 1 },
  95                  'rewrite-root=s' => sub { $icv{rewriteRoot} = $_[1] },
  96                  %remote_opts );
  97my %cmt_opts = ( 'edit|e' => \$_edit,
  98                'rmdir' => \$SVN::Git::Editor::_rmdir,
  99                'find-copies-harder' => \$SVN::Git::Editor::_find_copies_harder,
 100                'l=i' => \$SVN::Git::Editor::_rename_limit,
 101                'copy-similarity|C=i'=> \$SVN::Git::Editor::_cp_similarity
 102);
 103
 104my %cmd = (
 105        fetch => [ \&cmd_fetch, "Download new revisions from SVN",
 106                        { 'revision|r=s' => \$_revision,
 107                          'fetch-all|all' => \$_fetch_all,
 108                           %fc_opts } ],
 109        clone => [ \&cmd_clone, "Initialize and fetch revisions",
 110                        { 'revision|r=s' => \$_revision,
 111                           %fc_opts, %init_opts } ],
 112        init => [ \&cmd_init, "Initialize a repo for tracking" .
 113                          " (requires URL argument)",
 114                          \%init_opts ],
 115        'multi-init' => [ \&cmd_multi_init,
 116                          "Deprecated alias for ".
 117                          "'$0 init -T<trunk> -b<branches> -t<tags>'",
 118                          \%init_opts ],
 119        dcommit => [ \&cmd_dcommit,
 120                     'Commit several diffs to merge with upstream',
 121                        { 'merge|m|M' => \$_merge,
 122                          'strategy|s=s' => \$_strategy,
 123                          'verbose|v' => \$_verbose,
 124                          'dry-run|n' => \$_dry_run,
 125                          'fetch-all|all' => \$_fetch_all,
 126                          'no-rebase' => \$_no_rebase,
 127                        %cmt_opts, %fc_opts } ],
 128        'set-tree' => [ \&cmd_set_tree,
 129                        "Set an SVN repository to a git tree-ish",
 130                        { 'stdin|' => \$_stdin, %cmt_opts, %fc_opts, } ],
 131        'create-ignore' => [ \&cmd_create_ignore,
 132                             'Create a .gitignore per svn:ignore',
 133                             { 'revision|r=i' => \$_revision
 134                             } ],
 135        'propget' => [ \&cmd_propget,
 136                       'Print the value of a property on a file or directory',
 137                       { 'revision|r=i' => \$_revision } ],
 138        'proplist' => [ \&cmd_proplist,
 139                       'List all properties of a file or directory',
 140                       { 'revision|r=i' => \$_revision } ],
 141        'show-ignore' => [ \&cmd_show_ignore, "Show svn:ignore listings",
 142                        { 'revision|r=i' => \$_revision
 143                        } ],
 144        'multi-fetch' => [ \&cmd_multi_fetch,
 145                           "Deprecated alias for $0 fetch --all",
 146                           { 'revision|r=s' => \$_revision, %fc_opts } ],
 147        'migrate' => [ sub { },
 148                       # no-op, we automatically run this anyways,
 149                       'Migrate configuration/metadata/layout from
 150                        previous versions of git-svn',
 151                       { 'minimize' => \$Git::SVN::Migration::_minimize,
 152                         %remote_opts } ],
 153        'log' => [ \&Git::SVN::Log::cmd_show_log, 'Show commit logs',
 154                        { 'limit=i' => \$Git::SVN::Log::limit,
 155                          'revision|r=s' => \$_revision,
 156                          'verbose|v' => \$Git::SVN::Log::verbose,
 157                          'incremental' => \$Git::SVN::Log::incremental,
 158                          'oneline' => \$Git::SVN::Log::oneline,
 159                          'show-commit' => \$Git::SVN::Log::show_commit,
 160                          'non-recursive' => \$Git::SVN::Log::non_recursive,
 161                          'authors-file|A=s' => \$_authors,
 162                          'color' => \$Git::SVN::Log::color,
 163                          'pager=s' => \$Git::SVN::Log::pager
 164                        } ],
 165        'find-rev' => [ \&cmd_find_rev, "Translate between SVN revision numbers and tree-ish",
 166                        {} ],
 167        'rebase' => [ \&cmd_rebase, "Fetch and rebase your working directory",
 168                        { 'merge|m|M' => \$_merge,
 169                          'verbose|v' => \$_verbose,
 170                          'strategy|s=s' => \$_strategy,
 171                          'local|l' => \$_local,
 172                          'fetch-all|all' => \$_fetch_all,
 173                          %fc_opts } ],
 174        'commit-diff' => [ \&cmd_commit_diff,
 175                           'Commit a diff between two trees',
 176                        { 'message|m=s' => \$_message,
 177                          'file|F=s' => \$_file,
 178                          'revision|r=s' => \$_revision,
 179                        %cmt_opts } ],
 180);
 181
 182my $cmd;
 183for (my $i = 0; $i < @ARGV; $i++) {
 184        if (defined $cmd{$ARGV[$i]}) {
 185                $cmd = $ARGV[$i];
 186                splice @ARGV, $i, 1;
 187                last;
 188        }
 189};
 190
 191my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
 192
 193read_repo_config(\%opts);
 194Getopt::Long::Configure('pass_through') if ($cmd && $cmd eq 'log');
 195my $rv = GetOptions(%opts, 'help|H|h' => \$_help, 'version|V' => \$_version,
 196                    'minimize-connections' => \$Git::SVN::Migration::_minimize,
 197                    'id|i=s' => \$Git::SVN::default_ref_id,
 198                    'svn-remote|remote|R=s' => sub {
 199                       $Git::SVN::no_reuse_existing = 1;
 200                       $Git::SVN::default_repo_id = $_[1] });
 201exit 1 if (!$rv && $cmd && $cmd ne 'log');
 202
 203usage(0) if $_help;
 204version() if $_version;
 205usage(1) unless defined $cmd;
 206load_authors() if $_authors;
 207
 208# make sure we're always running
 209unless ($cmd =~ /(?:clone|init|multi-init)$/) {
 210        unless (-d $ENV{GIT_DIR}) {
 211                if ($git_dir_user_set) {
 212                        die "GIT_DIR=$ENV{GIT_DIR} explicitly set, ",
 213                            "but it is not a directory\n";
 214                }
 215                my $git_dir = delete $ENV{GIT_DIR};
 216                chomp(my $cdup = command_oneline(qw/rev-parse --show-cdup/));
 217                unless (length $cdup) {
 218                        die "Already at toplevel, but $git_dir ",
 219                            "not found '$cdup'\n";
 220                }
 221                chdir $cdup or die "Unable to chdir up to '$cdup'\n";
 222                unless (-d $git_dir) {
 223                        die "$git_dir still not found after going to ",
 224                            "'$cdup'\n";
 225                }
 226                $ENV{GIT_DIR} = $git_dir;
 227        }
 228}
 229unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
 230        Git::SVN::Migration::migration_check();
 231}
 232Git::SVN::init_vars();
 233eval {
 234        Git::SVN::verify_remotes_sanity();
 235        $cmd{$cmd}->[0]->(@ARGV);
 236};
 237fatal $@ if $@;
 238post_fetch_checkout();
 239exit 0;
 240
 241####################### primary functions ######################
 242sub usage {
 243        my $exit = shift || 0;
 244        my $fd = $exit ? \*STDERR : \*STDOUT;
 245        print $fd <<"";
 246git-svn - bidirectional operations between a single Subversion tree and git
 247Usage: $0 <command> [options] [arguments]\n
 248
 249        print $fd "Available commands:\n" unless $cmd;
 250
 251        foreach (sort keys %cmd) {
 252                next if $cmd && $cmd ne $_;
 253                next if /^multi-/; # don't show deprecated commands
 254                print $fd '  ',pack('A17',$_),$cmd{$_}->[1],"\n";
 255                foreach (sort keys %{$cmd{$_}->[2]}) {
 256                        # mixed-case options are for .git/config only
 257                        next if /[A-Z]/ && /^[a-z]+$/i;
 258                        # prints out arguments as they should be passed:
 259                        my $x = s#[:=]s$## ? '<arg>' : s#[:=]i$## ? '<num>' : '';
 260                        print $fd ' ' x 21, join(', ', map { length $_ > 1 ?
 261                                                        "--$_" : "-$_" }
 262                                                split /\|/,$_)," $x\n";
 263                }
 264        }
 265        print $fd <<"";
 266\nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
 267arbitrary identifier if you're tracking multiple SVN branches/repositories in
 268one git repository and want to keep them separate.  See git-svn(1) for more
 269information.
 270
 271        exit $exit;
 272}
 273
 274sub version {
 275        print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
 276        exit 0;
 277}
 278
 279sub do_git_init_db {
 280        unless (-d $ENV{GIT_DIR}) {
 281                my @init_db = ('init');
 282                push @init_db, "--template=$_template" if defined $_template;
 283                if (defined $_shared) {
 284                        if ($_shared =~ /[a-z]/) {
 285                                push @init_db, "--shared=$_shared";
 286                        } else {
 287                                push @init_db, "--shared";
 288                        }
 289                }
 290                command_noisy(@init_db);
 291        }
 292        my $set;
 293        my $pfx = "svn-remote.$Git::SVN::default_repo_id";
 294        foreach my $i (keys %icv) {
 295                die "'$set' and '$i' cannot both be set\n" if $set;
 296                next unless defined $icv{$i};
 297                command_noisy('config', "$pfx.$i", $icv{$i});
 298                $set = $i;
 299        }
 300}
 301
 302sub init_subdir {
 303        my $repo_path = shift or return;
 304        mkpath([$repo_path]) unless -d $repo_path;
 305        chdir $repo_path or die "Couldn't chdir to $repo_path: $!\n";
 306        $ENV{GIT_DIR} = '.git';
 307}
 308
 309sub cmd_clone {
 310        my ($url, $path) = @_;
 311        if (!defined $path &&
 312            (defined $_trunk || defined $_branches || defined $_tags ||
 313             defined $_stdlayout) &&
 314            $url !~ m#^[a-z\+]+://#) {
 315                $path = $url;
 316        }
 317        $path = basename($url) if !defined $path || !length $path;
 318        cmd_init($url, $path);
 319        Git::SVN::fetch_all($Git::SVN::default_repo_id);
 320}
 321
 322sub cmd_init {
 323        if (defined $_stdlayout) {
 324                $_trunk = 'trunk' if (!defined $_trunk);
 325                $_tags = 'tags' if (!defined $_tags);
 326                $_branches = 'branches' if (!defined $_branches);
 327        }
 328        if (defined $_trunk || defined $_branches || defined $_tags) {
 329                return cmd_multi_init(@_);
 330        }
 331        my $url = shift or die "SVN repository location required ",
 332                               "as a command-line argument\n";
 333        init_subdir(@_);
 334        do_git_init_db();
 335
 336        Git::SVN->init($url);
 337}
 338
 339sub cmd_fetch {
 340        if (grep /^\d+=./, @_) {
 341                die "'<rev>=<commit>' fetch arguments are ",
 342                    "no longer supported.\n";
 343        }
 344        my ($remote) = @_;
 345        if (@_ > 1) {
 346                die "Usage: $0 fetch [--all] [svn-remote]\n";
 347        }
 348        $remote ||= $Git::SVN::default_repo_id;
 349        if ($_fetch_all) {
 350                cmd_multi_fetch();
 351        } else {
 352                Git::SVN::fetch_all($remote, Git::SVN::read_all_remotes());
 353        }
 354}
 355
 356sub cmd_set_tree {
 357        my (@commits) = @_;
 358        if ($_stdin || !@commits) {
 359                print "Reading from stdin...\n";
 360                @commits = ();
 361                while (<STDIN>) {
 362                        if (/\b($sha1_short)\b/o) {
 363                                unshift @commits, $1;
 364                        }
 365                }
 366        }
 367        my @revs;
 368        foreach my $c (@commits) {
 369                my @tmp = command('rev-parse',$c);
 370                if (scalar @tmp == 1) {
 371                        push @revs, $tmp[0];
 372                } elsif (scalar @tmp > 1) {
 373                        push @revs, reverse(command('rev-list',@tmp));
 374                } else {
 375                        fatal "Failed to rev-parse $c";
 376                }
 377        }
 378        my $gs = Git::SVN->new;
 379        my ($r_last, $cmt_last) = $gs->last_rev_commit;
 380        $gs->fetch;
 381        if (defined $gs->{last_rev} && $r_last != $gs->{last_rev}) {
 382                fatal "There are new revisions that were fetched ",
 383                      "and need to be merged (or acknowledged) ",
 384                      "before committing.\nlast rev: $r_last\n",
 385                      " current: $gs->{last_rev}";
 386        }
 387        $gs->set_tree($_) foreach @revs;
 388        print "Done committing ",scalar @revs," revisions to SVN\n";
 389}
 390
 391sub cmd_dcommit {
 392        my $head = shift;
 393        git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
 394                'Cannot dcommit with a dirty index.  Commit your changes first, '
 395                . "or stash them with `git stash'.\n";
 396        $head ||= 'HEAD';
 397        my @refs;
 398        my ($url, $rev, $uuid, $gs) = working_head_info($head, \@refs);
 399        print "Committing to $url ...\n";
 400        unless ($gs) {
 401                die "Unable to determine upstream SVN information from ",
 402                    "$head history\n";
 403        }
 404        my $last_rev;
 405        my ($linear_refs, $parents) = linearize_history($gs, \@refs);
 406        if ($_no_rebase && scalar(@$linear_refs) > 1) {
 407                warn "Attempting to commit more than one change while ",
 408                     "--no-rebase is enabled.\n",
 409                     "If these changes depend on each other, re-running ",
 410                     "without --no-rebase will be required."
 411        }
 412        while (1) {
 413                my $d = shift @$linear_refs or last;
 414                unless (defined $last_rev) {
 415                        (undef, $last_rev, undef) = cmt_metadata("$d~1");
 416                        unless (defined $last_rev) {
 417                                fatal "Unable to extract revision information ",
 418                                      "from commit $d~1";
 419                        }
 420                }
 421                if ($_dry_run) {
 422                        print "diff-tree $d~1 $d\n";
 423                } else {
 424                        my $cmt_rev;
 425                        my %ed_opts = ( r => $last_rev,
 426                                        log => get_commit_entry($d)->{log},
 427                                        ra => Git::SVN::Ra->new($gs->full_url),
 428                                        config => SVN::Core::config_get_config(
 429                                                $Git::SVN::Ra::config_dir
 430                                        ),
 431                                        tree_a => "$d~1",
 432                                        tree_b => $d,
 433                                        editor_cb => sub {
 434                                               print "Committed r$_[0]\n";
 435                                               $cmt_rev = $_[0];
 436                                        },
 437                                        svn_path => '');
 438                        if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
 439                                print "No changes\n$d~1 == $d\n";
 440                        } elsif ($parents->{$d} && @{$parents->{$d}}) {
 441                                $gs->{inject_parents_dcommit}->{$cmt_rev} =
 442                                                               $parents->{$d};
 443                        }
 444                        $_fetch_all ? $gs->fetch_all : $gs->fetch;
 445                        next if $_no_rebase;
 446
 447                        # we always want to rebase against the current HEAD,
 448                        # not any head that was passed to us
 449                        my @diff = command('diff-tree', $d,
 450                                           $gs->refname, '--');
 451                        my @finish;
 452                        if (@diff) {
 453                                @finish = rebase_cmd();
 454                                print STDERR "W: $d and ", $gs->refname,
 455                                             " differ, using @finish:\n",
 456                                             join("\n", @diff), "\n";
 457                        } else {
 458                                print "No changes between current HEAD and ",
 459                                      $gs->refname,
 460                                      "\nResetting to the latest ",
 461                                      $gs->refname, "\n";
 462                                @finish = qw/reset --mixed/;
 463                        }
 464                        command_noisy(@finish, $gs->refname);
 465                        if (@diff) {
 466                                @refs = ();
 467                                my ($url_, $rev_, $uuid_, $gs_) =
 468                                              working_head_info($head, \@refs);
 469                                my ($linear_refs_, $parents_) =
 470                                              linearize_history($gs_, \@refs);
 471                                if (scalar(@$linear_refs) !=
 472                                    scalar(@$linear_refs_)) {
 473                                        fatal "# of revisions changed ",
 474                                          "\nbefore:\n",
 475                                          join("\n", @$linear_refs),
 476                                          "\n\nafter:\n",
 477                                          join("\n", @$linear_refs_), "\n",
 478                                          'If you are attempting to commit ',
 479                                          "merges, try running:\n\t",
 480                                          'git rebase --interactive',
 481                                          '--preserve-merges ',
 482                                          $gs->refname,
 483                                          "\nBefore dcommitting";
 484                                }
 485                                if ($url_ ne $url) {
 486                                        fatal "URL mismatch after rebase: ",
 487                                              "$url_ != $url";
 488                                }
 489                                if ($uuid_ ne $uuid) {
 490                                        fatal "uuid mismatch after rebase: ",
 491                                              "$uuid_ != $uuid";
 492                                }
 493                                # remap parents
 494                                my (%p, @l, $i);
 495                                for ($i = 0; $i < scalar @$linear_refs; $i++) {
 496                                        my $new = $linear_refs_->[$i] or next;
 497                                        $p{$new} =
 498                                                $parents->{$linear_refs->[$i]};
 499                                        push @l, $new;
 500                                }
 501                                $parents = \%p;
 502                                $linear_refs = \@l;
 503                        }
 504                        $last_rev = $cmt_rev;
 505                }
 506        }
 507}
 508
 509sub cmd_find_rev {
 510        my $revision_or_hash = shift;
 511        my $result;
 512        if ($revision_or_hash =~ /^r\d+$/) {
 513                my $head = shift;
 514                $head ||= 'HEAD';
 515                my @refs;
 516                my (undef, undef, undef, $gs) = working_head_info($head, \@refs);
 517                unless ($gs) {
 518                        die "Unable to determine upstream SVN information from ",
 519                            "$head history\n";
 520                }
 521                my $desired_revision = substr($revision_or_hash, 1);
 522                $result = $gs->rev_db_get($desired_revision);
 523        } else {
 524                my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
 525                $result = $rev;
 526        }
 527        print "$result\n" if $result;
 528}
 529
 530sub cmd_rebase {
 531        command_noisy(qw/update-index --refresh/);
 532        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 533        unless ($gs) {
 534                die "Unable to determine upstream SVN information from ",
 535                    "working tree history\n";
 536        }
 537        if (command(qw/diff-index HEAD --/)) {
 538                print STDERR "Cannot rebase with uncommited changes:\n";
 539                command_noisy('status');
 540                exit 1;
 541        }
 542        unless ($_local) {
 543                $_fetch_all ? $gs->fetch_all : $gs->fetch;
 544        }
 545        command_noisy(rebase_cmd(), $gs->refname);
 546}
 547
 548sub cmd_show_ignore {
 549        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 550        $gs ||= Git::SVN->new;
 551        my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
 552        $gs->prop_walk($gs->{path}, $r, sub {
 553                my ($gs, $path, $props) = @_;
 554                print STDOUT "\n# $path\n";
 555                my $s = $props->{'svn:ignore'} or return;
 556                $s =~ s/[\r\n]+/\n/g;
 557                chomp $s;
 558                $s =~ s#^#$path#gm;
 559                print STDOUT "$s\n";
 560        });
 561}
 562
 563sub cmd_create_ignore {
 564        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 565        $gs ||= Git::SVN->new;
 566        my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
 567        $gs->prop_walk($gs->{path}, $r, sub {
 568                my ($gs, $path, $props) = @_;
 569                # $path is of the form /path/to/dir/
 570                my $ignore = '.' . $path . '.gitignore';
 571                my $s = $props->{'svn:ignore'} or return;
 572                open(GITIGNORE, '>', $ignore)
 573                  or fatal("Failed to open `$ignore' for writing: $!");
 574                $s =~ s/[\r\n]+/\n/g;
 575                chomp $s;
 576                # Prefix all patterns so that the ignore doesn't apply
 577                # to sub-directories.
 578                $s =~ s#^#/#gm;
 579                print GITIGNORE "$s\n";
 580                close(GITIGNORE)
 581                  or fatal("Failed to close `$ignore': $!");
 582                command_noisy('add', $ignore);
 583        });
 584}
 585
 586# get_svnprops(PATH)
 587# ------------------
 588# Helper for cmd_propget and cmd_proplist below.
 589sub get_svnprops {
 590        my $path = shift;
 591        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 592        $gs ||= Git::SVN->new;
 593
 594        # prefix THE PATH by the sub-directory from which the user
 595        # invoked us.
 596        $path = $cmd_dir_prefix . $path;
 597        fatal("No such file or directory: $path") unless -e $path;
 598        my $is_dir = -d $path ? 1 : 0;
 599        $path = $gs->{path} . '/' . $path;
 600
 601        # canonicalize the path (otherwise libsvn will abort or fail to
 602        # find the file)
 603        # File::Spec->canonpath doesn't collapse x/../y into y (for a
 604        # good reason), so let's do this manually.
 605        $path =~ s#/+#/#g;
 606        $path =~ s#/\.(?:/|$)#/#g;
 607        $path =~ s#/[^/]+/\.\.##g;
 608        $path =~ s#/$##g;
 609
 610        my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
 611        my $props;
 612        if ($is_dir) {
 613                (undef, undef, $props) = $gs->ra->get_dir($path, $r);
 614        }
 615        else {
 616                (undef, $props) = $gs->ra->get_file($path, $r, undef);
 617        }
 618        return $props;
 619}
 620
 621# cmd_propget (PROP, PATH)
 622# ------------------------
 623# Print the SVN property PROP for PATH.
 624sub cmd_propget {
 625        my ($prop, $path) = @_;
 626        $path = '.' if not defined $path;
 627        usage(1) if not defined $prop;
 628        my $props = get_svnprops($path);
 629        if (not defined $props->{$prop}) {
 630                fatal("`$path' does not have a `$prop' SVN property.");
 631        }
 632        print $props->{$prop} . "\n";
 633}
 634
 635# cmd_proplist (PATH)
 636# -------------------
 637# Print the list of SVN properties for PATH.
 638sub cmd_proplist {
 639        my $path = shift;
 640        $path = '.' if not defined $path;
 641        my $props = get_svnprops($path);
 642        print "Properties on '$path':\n";
 643        foreach (sort keys %{$props}) {
 644                print "  $_\n";
 645        }
 646}
 647
 648sub cmd_multi_init {
 649        my $url = shift;
 650        unless (defined $_trunk || defined $_branches || defined $_tags) {
 651                usage(1);
 652        }
 653
 654        # there are currently some bugs that prevent multi-init/multi-fetch
 655        # setups from working well without this.
 656        $Git::SVN::_minimize_url = 1;
 657
 658        $_prefix = '' unless defined $_prefix;
 659        if (defined $url) {
 660                $url =~ s#/+$##;
 661                init_subdir(@_);
 662        }
 663        do_git_init_db();
 664        if (defined $_trunk) {
 665                my $trunk_ref = $_prefix . 'trunk';
 666                # try both old-style and new-style lookups:
 667                my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
 668                unless ($gs_trunk) {
 669                        my ($trunk_url, $trunk_path) =
 670                                              complete_svn_url($url, $_trunk);
 671                        $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
 672                                                   undef, $trunk_ref);
 673                }
 674        }
 675        return unless defined $_branches || defined $_tags;
 676        my $ra = $url ? Git::SVN::Ra->new($url) : undef;
 677        complete_url_ls_init($ra, $_branches, '--branches/-b', $_prefix);
 678        complete_url_ls_init($ra, $_tags, '--tags/-t', $_prefix . 'tags/');
 679}
 680
 681sub cmd_multi_fetch {
 682        my $remotes = Git::SVN::read_all_remotes();
 683        foreach my $repo_id (sort keys %$remotes) {
 684                if ($remotes->{$repo_id}->{url}) {
 685                        Git::SVN::fetch_all($repo_id, $remotes);
 686                }
 687        }
 688}
 689
 690# this command is special because it requires no metadata
 691sub cmd_commit_diff {
 692        my ($ta, $tb, $url) = @_;
 693        my $usage = "Usage: $0 commit-diff -r<revision> ".
 694                    "<tree-ish> <tree-ish> [<URL>]";
 695        fatal($usage) if (!defined $ta || !defined $tb);
 696        my $svn_path;
 697        if (!defined $url) {
 698                my $gs = eval { Git::SVN->new };
 699                if (!$gs) {
 700                        fatal("Needed URL or usable git-svn --id in ",
 701                              "the command-line\n", $usage);
 702                }
 703                $url = $gs->{url};
 704                $svn_path = $gs->{path};
 705        }
 706        unless (defined $_revision) {
 707                fatal("-r|--revision is a required argument\n", $usage);
 708        }
 709        if (defined $_message && defined $_file) {
 710                fatal("Both --message/-m and --file/-F specified ",
 711                      "for the commit message.\n",
 712                      "I have no idea what you mean");
 713        }
 714        if (defined $_file) {
 715                $_message = file_to_s($_file);
 716        } else {
 717                $_message ||= get_commit_entry($tb)->{log};
 718        }
 719        my $ra ||= Git::SVN::Ra->new($url);
 720        $svn_path ||= $ra->{svn_path};
 721        my $r = $_revision;
 722        if ($r eq 'HEAD') {
 723                $r = $ra->get_latest_revnum;
 724        } elsif ($r !~ /^\d+$/) {
 725                die "revision argument: $r not understood by git-svn\n";
 726        }
 727        my %ed_opts = ( r => $r,
 728                        log => $_message,
 729                        ra => $ra,
 730                        tree_a => $ta,
 731                        tree_b => $tb,
 732                        editor_cb => sub { print "Committed r$_[0]\n" },
 733                        svn_path => $svn_path );
 734        if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
 735                print "No changes\n$ta == $tb\n";
 736        }
 737}
 738
 739########################### utility functions #########################
 740
 741sub rebase_cmd {
 742        my @cmd = qw/rebase/;
 743        push @cmd, '-v' if $_verbose;
 744        push @cmd, qw/--merge/ if $_merge;
 745        push @cmd, "--strategy=$_strategy" if $_strategy;
 746        @cmd;
 747}
 748
 749sub post_fetch_checkout {
 750        return if $_no_checkout;
 751        my $gs = $Git::SVN::_head or return;
 752        return if verify_ref('refs/heads/master^0');
 753
 754        my $valid_head = verify_ref('HEAD^0');
 755        command_noisy(qw(update-ref refs/heads/master), $gs->refname);
 756        return if ($valid_head || !verify_ref('HEAD^0'));
 757
 758        return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
 759        my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
 760        return if -f $index;
 761
 762        return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
 763        return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
 764        command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
 765        print STDERR "Checked out HEAD:\n  ",
 766                     $gs->full_url, " r", $gs->last_rev, "\n";
 767}
 768
 769sub complete_svn_url {
 770        my ($url, $path) = @_;
 771        $path =~ s#/+$##;
 772        if ($path !~ m#^[a-z\+]+://#) {
 773                if (!defined $url || $url !~ m#^[a-z\+]+://#) {
 774                        fatal("E: '$path' is not a complete URL ",
 775                              "and a separate URL is not specified");
 776                }
 777                return ($url, $path);
 778        }
 779        return ($path, '');
 780}
 781
 782sub complete_url_ls_init {
 783        my ($ra, $repo_path, $switch, $pfx) = @_;
 784        unless ($repo_path) {
 785                print STDERR "W: $switch not specified\n";
 786                return;
 787        }
 788        $repo_path =~ s#/+$##;
 789        if ($repo_path =~ m#^[a-z\+]+://#) {
 790                $ra = Git::SVN::Ra->new($repo_path);
 791                $repo_path = '';
 792        } else {
 793                $repo_path =~ s#^/+##;
 794                unless ($ra) {
 795                        fatal("E: '$repo_path' is not a complete URL ",
 796                              "and a separate URL is not specified");
 797                }
 798        }
 799        my $url = $ra->{url};
 800        my $gs = Git::SVN->init($url, undef, undef, undef, 1);
 801        my $k = "svn-remote.$gs->{repo_id}.url";
 802        my $orig_url = eval { command_oneline(qw/config --get/, $k) };
 803        if ($orig_url && ($orig_url ne $gs->{url})) {
 804                die "$k already set: $orig_url\n",
 805                    "wanted to set to: $gs->{url}\n";
 806        }
 807        command_oneline('config', $k, $gs->{url}) unless $orig_url;
 808        my $remote_path = "$ra->{svn_path}/$repo_path/*";
 809        $remote_path =~ s#/+#/#g;
 810        $remote_path =~ s#^/##g;
 811        my ($n) = ($switch =~ /^--(\w+)/);
 812        if (length $pfx && $pfx !~ m#/$#) {
 813                die "--prefix='$pfx' must have a trailing slash '/'\n";
 814        }
 815        command_noisy('config', "svn-remote.$gs->{repo_id}.$n",
 816                                "$remote_path:refs/remotes/$pfx*");
 817}
 818
 819sub verify_ref {
 820        my ($ref) = @_;
 821        eval { command_oneline([ 'rev-parse', '--verify', $ref ],
 822                               { STDERR => 0 }); };
 823}
 824
 825sub get_tree_from_treeish {
 826        my ($treeish) = @_;
 827        # $treeish can be a symbolic ref, too:
 828        my $type = command_oneline(qw/cat-file -t/, $treeish);
 829        my $expected;
 830        while ($type eq 'tag') {
 831                ($treeish, $type) = command(qw/cat-file tag/, $treeish);
 832        }
 833        if ($type eq 'commit') {
 834                $expected = (grep /^tree /, command(qw/cat-file commit/,
 835                                                    $treeish))[0];
 836                ($expected) = ($expected =~ /^tree ($sha1)$/o);
 837                die "Unable to get tree from $treeish\n" unless $expected;
 838        } elsif ($type eq 'tree') {
 839                $expected = $treeish;
 840        } else {
 841                die "$treeish is a $type, expected tree, tag or commit\n";
 842        }
 843        return $expected;
 844}
 845
 846sub get_commit_entry {
 847        my ($treeish) = shift;
 848        my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
 849        my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
 850        my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
 851        open my $log_fh, '>', $commit_editmsg or croak $!;
 852
 853        my $type = command_oneline(qw/cat-file -t/, $treeish);
 854        if ($type eq 'commit' || $type eq 'tag') {
 855                my ($msg_fh, $ctx) = command_output_pipe('cat-file',
 856                                                         $type, $treeish);
 857                my $in_msg = 0;
 858                while (<$msg_fh>) {
 859                        if (!$in_msg) {
 860                                $in_msg = 1 if (/^\s*$/);
 861                        } elsif (/^git-svn-id: /) {
 862                                # skip this for now, we regenerate the
 863                                # correct one on re-fetch anyways
 864                                # TODO: set *:merge properties or like...
 865                        } else {
 866                                print $log_fh $_ or croak $!;
 867                        }
 868                }
 869                command_close_pipe($msg_fh, $ctx);
 870        }
 871        close $log_fh or croak $!;
 872
 873        if ($_edit || ($type eq 'tree')) {
 874                my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
 875                # TODO: strip out spaces, comments, like git-commit.sh
 876                system($editor, $commit_editmsg);
 877        }
 878        rename $commit_editmsg, $commit_msg or croak $!;
 879        open $log_fh, '<', $commit_msg or croak $!;
 880        { local $/; chomp($log_entry{log} = <$log_fh>); }
 881        close $log_fh or croak $!;
 882        unlink $commit_msg;
 883        \%log_entry;
 884}
 885
 886sub s_to_file {
 887        my ($str, $file, $mode) = @_;
 888        open my $fd,'>',$file or croak $!;
 889        print $fd $str,"\n" or croak $!;
 890        close $fd or croak $!;
 891        chmod ($mode &~ umask, $file) if (defined $mode);
 892}
 893
 894sub file_to_s {
 895        my $file = shift;
 896        open my $fd,'<',$file or croak "$!: file: $file\n";
 897        local $/;
 898        my $ret = <$fd>;
 899        close $fd or croak $!;
 900        $ret =~ s/\s*$//s;
 901        return $ret;
 902}
 903
 904# '<svn username> = real-name <email address>' mapping based on git-svnimport:
 905sub load_authors {
 906        open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
 907        my $log = $cmd eq 'log';
 908        while (<$authors>) {
 909                chomp;
 910                next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
 911                my ($user, $name, $email) = ($1, $2, $3);
 912                if ($log) {
 913                        $Git::SVN::Log::rusers{"$name <$email>"} = $user;
 914                } else {
 915                        $users{$user} = [$name, $email];
 916                }
 917        }
 918        close $authors or croak $!;
 919}
 920
 921# convert GetOpt::Long specs for use by git-config
 922sub read_repo_config {
 923        return unless -d $ENV{GIT_DIR};
 924        my $opts = shift;
 925        my @config_only;
 926        foreach my $o (keys %$opts) {
 927                # if we have mixedCase and a long option-only, then
 928                # it's a config-only variable that we don't need for
 929                # the command-line.
 930                push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
 931                my $v = $opts->{$o};
 932                my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
 933                $key =~ s/-//g;
 934                my $arg = 'git-config';
 935                $arg .= ' --int' if ($o =~ /[:=]i$/);
 936                $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
 937                if (ref $v eq 'ARRAY') {
 938                        chomp(my @tmp = `$arg --get-all svn.$key`);
 939                        @$v = @tmp if @tmp;
 940                } else {
 941                        chomp(my $tmp = `$arg --get svn.$key`);
 942                        if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
 943                                $$v = $tmp;
 944                        }
 945                }
 946        }
 947        delete @$opts{@config_only} if @config_only;
 948}
 949
 950sub extract_metadata {
 951        my $id = shift or return (undef, undef, undef);
 952        my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
 953                                                        \s([a-f\d\-]+)$/x);
 954        if (!defined $rev || !$uuid || !$url) {
 955                # some of the original repositories I made had
 956                # identifiers like this:
 957                ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/);
 958        }
 959        return ($url, $rev, $uuid);
 960}
 961
 962sub cmt_metadata {
 963        return extract_metadata((grep(/^git-svn-id: /,
 964                command(qw/cat-file commit/, shift)))[-1]);
 965}
 966
 967sub working_head_info {
 968        my ($head, $refs) = @_;
 969        my @args = ('log', '--no-color', '--first-parent');
 970        my ($fh, $ctx) = command_output_pipe(@args, $head);
 971        my $hash;
 972        my %max;
 973        while (<$fh>) {
 974                if ( m{^commit ($::sha1)$} ) {
 975                        unshift @$refs, $hash if $hash and $refs;
 976                        $hash = $1;
 977                        next;
 978                }
 979                next unless s{^\s*(git-svn-id:)}{$1};
 980                my ($url, $rev, $uuid) = extract_metadata($_);
 981                if (defined $url && defined $rev) {
 982                        next if $max{$url} and $max{$url} < $rev;
 983                        if (my $gs = Git::SVN->find_by_url($url)) {
 984                                my $c = $gs->rev_db_get($rev);
 985                                if ($c && $c eq $hash) {
 986                                        close $fh; # break the pipe
 987                                        return ($url, $rev, $uuid, $gs);
 988                                } else {
 989                                        $max{$url} ||= $gs->rev_db_max;
 990                                }
 991                        }
 992                }
 993        }
 994        command_close_pipe($fh, $ctx);
 995        (undef, undef, undef, undef);
 996}
 997
 998sub read_commit_parents {
 999        my ($parents, $c) = @_;
1000        chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1001        $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1002        @{$parents->{$c}} = split(/ /, $p);
1003}
1004
1005sub linearize_history {
1006        my ($gs, $refs) = @_;
1007        my %parents;
1008        foreach my $c (@$refs) {
1009                read_commit_parents(\%parents, $c);
1010        }
1011
1012        my @linear_refs;
1013        my %skip = ();
1014        my $last_svn_commit = $gs->last_commit;
1015        foreach my $c (reverse @$refs) {
1016                next if $c eq $last_svn_commit;
1017                last if $skip{$c};
1018
1019                unshift @linear_refs, $c;
1020                $skip{$c} = 1;
1021
1022                # we only want the first parent to diff against for linear
1023                # history, we save the rest to inject when we finalize the
1024                # svn commit
1025                my $fp_a = verify_ref("$c~1");
1026                my $fp_b = shift @{$parents{$c}} if $parents{$c};
1027                if (!$fp_a || !$fp_b) {
1028                        die "Commit $c\n",
1029                            "has no parent commit, and therefore ",
1030                            "nothing to diff against.\n",
1031                            "You should be working from a repository ",
1032                            "originally created by git-svn\n";
1033                }
1034                if ($fp_a ne $fp_b) {
1035                        die "$c~1 = $fp_a, however parsing commit $c ",
1036                            "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1037                }
1038
1039                foreach my $p (@{$parents{$c}}) {
1040                        $skip{$p} = 1;
1041                }
1042        }
1043        (\@linear_refs, \%parents);
1044}
1045
1046package Git::SVN;
1047use strict;
1048use warnings;
1049use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1050            $_repack $_repack_flags $_use_svm_props $_head
1051            $_use_svnsync_props $no_reuse_existing $_minimize_url/;
1052use Carp qw/croak/;
1053use File::Path qw/mkpath/;
1054use File::Copy qw/copy/;
1055use IPC::Open3;
1056
1057my $_repack_nr;
1058# properties that we do not log:
1059my %SKIP_PROP;
1060BEGIN {
1061        %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1062                                        svn:special svn:executable
1063                                        svn:entry:committed-rev
1064                                        svn:entry:last-author
1065                                        svn:entry:uuid
1066                                        svn:entry:committed-date/;
1067
1068        # some options are read globally, but can be overridden locally
1069        # per [svn-remote "..."] section.  Command-line options will *NOT*
1070        # override options set in an [svn-remote "..."] section
1071        no strict 'refs';
1072        for my $option (qw/follow_parent no_metadata use_svm_props
1073                           use_svnsync_props/) {
1074                my $key = $option;
1075                $key =~ tr/_//d;
1076                my $prop = "-$option";
1077                *$option = sub {
1078                        my ($self) = @_;
1079                        return $self->{$prop} if exists $self->{$prop};
1080                        my $k = "svn-remote.$self->{repo_id}.$key";
1081                        eval { command_oneline(qw/config --get/, $k) };
1082                        if ($@) {
1083                                $self->{$prop} = ${"Git::SVN::_$option"};
1084                        } else {
1085                                my $v = command_oneline(qw/config --bool/,$k);
1086                                $self->{$prop} = $v eq 'false' ? 0 : 1;
1087                        }
1088                        return $self->{$prop};
1089                }
1090        }
1091}
1092
1093my %LOCKFILES;
1094END { unlink keys %LOCKFILES if %LOCKFILES }
1095
1096sub resolve_local_globs {
1097        my ($url, $fetch, $glob_spec) = @_;
1098        return unless defined $glob_spec;
1099        my $ref = $glob_spec->{ref};
1100        my $path = $glob_spec->{path};
1101        foreach (command(qw#for-each-ref --format=%(refname) refs/remotes#)) {
1102                next unless m#^refs/remotes/$ref->{regex}$#;
1103                my $p = $1;
1104                my $pathname = desanitize_refname($path->full_path($p));
1105                my $refname = desanitize_refname($ref->full_path($p));
1106                if (my $existing = $fetch->{$pathname}) {
1107                        if ($existing ne $refname) {
1108                                die "Refspec conflict:\n",
1109                                    "existing: refs/remotes/$existing\n",
1110                                    " globbed: refs/remotes/$refname\n";
1111                        }
1112                        my $u = (::cmt_metadata("refs/remotes/$refname"))[0];
1113                        $u =~ s!^\Q$url\E(/|$)!! or die
1114                          "refs/remotes/$refname: '$url' not found in '$u'\n";
1115                        if ($pathname ne $u) {
1116                                warn "W: Refspec glob conflict ",
1117                                     "(ref: refs/remotes/$refname):\n",
1118                                     "expected path: $pathname\n",
1119                                     "    real path: $u\n",
1120                                     "Continuing ahead with $u\n";
1121                                next;
1122                        }
1123                } else {
1124                        $fetch->{$pathname} = $refname;
1125                }
1126        }
1127}
1128
1129sub parse_revision_argument {
1130        my ($base, $head) = @_;
1131        if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1132                return ($base, $head);
1133        }
1134        return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1135        return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1136        return ($head, $head) if ($::_revision eq 'HEAD');
1137        return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1138        return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1139        die "revision argument: $::_revision not understood by git-svn\n";
1140}
1141
1142sub fetch_all {
1143        my ($repo_id, $remotes) = @_;
1144        if (ref $repo_id) {
1145                my $gs = $repo_id;
1146                $repo_id = undef;
1147                $repo_id = $gs->{repo_id};
1148        }
1149        $remotes ||= read_all_remotes();
1150        my $remote = $remotes->{$repo_id} or
1151                     die "[svn-remote \"$repo_id\"] unknown\n";
1152        my $fetch = $remote->{fetch};
1153        my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1154        my (@gs, @globs);
1155        my $ra = Git::SVN::Ra->new($url);
1156        my $uuid = $ra->get_uuid;
1157        my $head = $ra->get_latest_revnum;
1158        my $base = defined $fetch ? $head : 0;
1159
1160        # read the max revs for wildcard expansion (branches/*, tags/*)
1161        foreach my $t (qw/branches tags/) {
1162                defined $remote->{$t} or next;
1163                push @globs, $remote->{$t};
1164                my $max_rev = eval { tmp_config(qw/--int --get/,
1165                                         "svn-remote.$repo_id.${t}-maxRev") };
1166                if (defined $max_rev && ($max_rev < $base)) {
1167                        $base = $max_rev;
1168                } elsif (!defined $max_rev) {
1169                        $base = 0;
1170                }
1171        }
1172
1173        if ($fetch) {
1174                foreach my $p (sort keys %$fetch) {
1175                        my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1176                        my $lr = $gs->rev_db_max;
1177                        if (defined $lr) {
1178                                $base = $lr if ($lr < $base);
1179                        }
1180                        push @gs, $gs;
1181                }
1182        }
1183
1184        ($base, $head) = parse_revision_argument($base, $head);
1185        $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1186}
1187
1188sub read_all_remotes {
1189        my $r = {};
1190        foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1191                if (m!^(.+)\.fetch=\s*(.*)\s*:\s*refs/remotes/(.+)\s*$!) {
1192                        my ($remote, $local_ref, $remote_ref) = ($1, $2, $3);
1193                        $local_ref =~ s{^/}{};
1194                        $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1195                } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1196                        $r->{$1}->{url} = $2;
1197                } elsif (m!^(.+)\.(branches|tags)=
1198                           (.*):refs/remotes/(.+)\s*$/!x) {
1199                        my ($p, $g) = ($3, $4);
1200                        my $rs = $r->{$1}->{$2} = {
1201                                          t => $2,
1202                                          remote => $1,
1203                                          path => Git::SVN::GlobSpec->new($p),
1204                                          ref => Git::SVN::GlobSpec->new($g) };
1205                        if (length($rs->{ref}->{right}) != 0) {
1206                                die "The '*' glob character must be the last ",
1207                                    "character of '$g'\n";
1208                        }
1209                }
1210        }
1211        $r;
1212}
1213
1214sub init_vars {
1215        if (defined $_repack) {
1216                $_repack = 1000 if ($_repack <= 0);
1217                $_repack_nr = $_repack;
1218                $_repack_flags ||= '-d';
1219        }
1220}
1221
1222sub verify_remotes_sanity {
1223        return unless -d $ENV{GIT_DIR};
1224        my %seen;
1225        foreach (command(qw/config -l/)) {
1226                if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1227                        if ($seen{$1}) {
1228                                die "Remote ref refs/remote/$1 is tracked by",
1229                                    "\n  \"$_\"\nand\n  \"$seen{$1}\"\n",
1230                                    "Please resolve this ambiguity in ",
1231                                    "your git configuration file before ",
1232                                    "continuing\n";
1233                        }
1234                        $seen{$1} = $_;
1235                }
1236        }
1237}
1238
1239# we allow more chars than remotes2config.sh...
1240sub sanitize_remote_name {
1241        my ($name) = @_;
1242        $name =~ tr{A-Za-z0-9:,/+-}{.}c;
1243        $name;
1244}
1245
1246sub find_existing_remote {
1247        my ($url, $remotes) = @_;
1248        return undef if $no_reuse_existing;
1249        my $existing;
1250        foreach my $repo_id (keys %$remotes) {
1251                my $u = $remotes->{$repo_id}->{url} or next;
1252                next if $u ne $url;
1253                $existing = $repo_id;
1254                last;
1255        }
1256        $existing;
1257}
1258
1259sub init_remote_config {
1260        my ($self, $url, $no_write) = @_;
1261        $url =~ s!/+$!!; # strip trailing slash
1262        my $r = read_all_remotes();
1263        my $existing = find_existing_remote($url, $r);
1264        if ($existing) {
1265                unless ($no_write) {
1266                        print STDERR "Using existing ",
1267                                     "[svn-remote \"$existing\"]\n";
1268                }
1269                $self->{repo_id} = $existing;
1270        } elsif ($_minimize_url) {
1271                my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1272                $existing = find_existing_remote($min_url, $r);
1273                if ($existing) {
1274                        unless ($no_write) {
1275                                print STDERR "Using existing ",
1276                                             "[svn-remote \"$existing\"]\n";
1277                        }
1278                        $self->{repo_id} = $existing;
1279                }
1280                if ($min_url ne $url) {
1281                        unless ($no_write) {
1282                                print STDERR "Using higher level of URL: ",
1283                                             "$url => $min_url\n";
1284                        }
1285                        my $old_path = $self->{path};
1286                        $self->{path} = $url;
1287                        $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1288                        if (length $old_path) {
1289                                $self->{path} .= "/$old_path";
1290                        }
1291                        $url = $min_url;
1292                }
1293        }
1294        my $orig_url;
1295        if (!$existing) {
1296                # verify that we aren't overwriting anything:
1297                $orig_url = eval {
1298                        command_oneline('config', '--get',
1299                                        "svn-remote.$self->{repo_id}.url")
1300                };
1301                if ($orig_url && ($orig_url ne $url)) {
1302                        die "svn-remote.$self->{repo_id}.url already set: ",
1303                            "$orig_url\nwanted to set to: $url\n";
1304                }
1305        }
1306        my ($xrepo_id, $xpath) = find_ref($self->refname);
1307        if (defined $xpath) {
1308                die "svn-remote.$xrepo_id.fetch already set to track ",
1309                    "$xpath:refs/remotes/", $self->refname, "\n";
1310        }
1311        unless ($no_write) {
1312                command_noisy('config',
1313                              "svn-remote.$self->{repo_id}.url", $url);
1314                $self->{path} =~ s{^/}{};
1315                command_noisy('config', '--add',
1316                              "svn-remote.$self->{repo_id}.fetch",
1317                              "$self->{path}:".$self->refname);
1318        }
1319        $self->{url} = $url;
1320}
1321
1322sub find_by_url { # repos_root and, path are optional
1323        my ($class, $full_url, $repos_root, $path) = @_;
1324
1325        return undef unless defined $full_url;
1326        remove_username($full_url);
1327        remove_username($repos_root) if defined $repos_root;
1328        my $remotes = read_all_remotes();
1329        if (defined $full_url && defined $repos_root && !defined $path) {
1330                $path = $full_url;
1331                $path =~ s#^\Q$repos_root\E(?:/|$)##;
1332        }
1333        foreach my $repo_id (keys %$remotes) {
1334                my $u = $remotes->{$repo_id}->{url} or next;
1335                remove_username($u);
1336                next if defined $repos_root && $repos_root ne $u;
1337
1338                my $fetch = $remotes->{$repo_id}->{fetch} || {};
1339                foreach (qw/branches tags/) {
1340                        resolve_local_globs($u, $fetch,
1341                                            $remotes->{$repo_id}->{$_});
1342                }
1343                my $p = $path;
1344                unless (defined $p) {
1345                        $p = $full_url;
1346                        $p =~ s#^\Q$u\E(?:/|$)## or next;
1347                }
1348                foreach my $f (keys %$fetch) {
1349                        next if $f ne $p;
1350                        return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1351                }
1352        }
1353        undef;
1354}
1355
1356sub init {
1357        my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1358        my $self = _new($class, $repo_id, $ref_id, $path);
1359        if (defined $url) {
1360                $self->init_remote_config($url, $no_write);
1361        }
1362        $self;
1363}
1364
1365sub find_ref {
1366        my ($ref_id) = @_;
1367        foreach (command(qw/config -l/)) {
1368                next unless m!^svn-remote\.(.+)\.fetch=
1369                              \s*(.*)\s*:\s*refs/remotes/(.+)\s*$!x;
1370                my ($repo_id, $path, $ref) = ($1, $2, $3);
1371                if ($ref eq $ref_id) {
1372                        $path = '' if ($path =~ m#^\./?#);
1373                        return ($repo_id, $path);
1374                }
1375        }
1376        (undef, undef, undef);
1377}
1378
1379sub new {
1380        my ($class, $ref_id, $repo_id, $path) = @_;
1381        if (defined $ref_id && !defined $repo_id && !defined $path) {
1382                ($repo_id, $path) = find_ref($ref_id);
1383                if (!defined $repo_id) {
1384                        die "Could not find a \"svn-remote.*.fetch\" key ",
1385                            "in the repository configuration matching: ",
1386                            "refs/remotes/$ref_id\n";
1387                }
1388        }
1389        my $self = _new($class, $repo_id, $ref_id, $path);
1390        if (!defined $self->{path} || !length $self->{path}) {
1391                my $fetch = command_oneline('config', '--get',
1392                                            "svn-remote.$repo_id.fetch",
1393                                            ":refs/remotes/$ref_id\$") or
1394                     die "Failed to read \"svn-remote.$repo_id.fetch\" ",
1395                         "\":refs/remotes/$ref_id\$\" in config\n";
1396                ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
1397        }
1398        $self->{url} = command_oneline('config', '--get',
1399                                       "svn-remote.$repo_id.url") or
1400                  die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
1401        $self->rebuild;
1402        $self;
1403}
1404
1405sub refname {
1406        my ($refname) = "refs/remotes/$_[0]->{ref_id}" ;
1407
1408        # It cannot end with a slash /, we'll throw up on this because
1409        # SVN can't have directories with a slash in their name, either:
1410        if ($refname =~ m{/$}) {
1411                die "ref: '$refname' ends with a trailing slash, this is ",
1412                    "not permitted by git nor Subversion\n";
1413        }
1414
1415        # It cannot have ASCII control character space, tilde ~, caret ^,
1416        # colon :, question-mark ?, asterisk *, space, or open bracket [
1417        # anywhere.
1418        #
1419        # Additionally, % must be escaped because it is used for escaping
1420        # and we want our escaped refname to be reversible
1421        $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
1422
1423        # no slash-separated component can begin with a dot .
1424        # /.* becomes /%2E*
1425        $refname =~ s{/\.}{/%2E}g;
1426
1427        # It cannot have two consecutive dots .. anywhere
1428        # .. becomes %2E%2E
1429        $refname =~ s{\.\.}{%2E%2E}g;
1430
1431        return $refname;
1432}
1433
1434sub desanitize_refname {
1435        my ($refname) = @_;
1436        $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
1437        return $refname;
1438}
1439
1440sub svm_uuid {
1441        my ($self) = @_;
1442        return $self->{svm}->{uuid} if $self->svm;
1443        $self->ra;
1444        unless ($self->{svm}) {
1445                die "SVM UUID not cached, and reading remotely failed\n";
1446        }
1447        $self->{svm}->{uuid};
1448}
1449
1450sub svm {
1451        my ($self) = @_;
1452        return $self->{svm} if $self->{svm};
1453        my $svm;
1454        # see if we have it in our config, first:
1455        eval {
1456                my $section = "svn-remote.$self->{repo_id}";
1457                $svm = {
1458                  source => tmp_config('--get', "$section.svm-source"),
1459                  uuid => tmp_config('--get', "$section.svm-uuid"),
1460                  replace => tmp_config('--get', "$section.svm-replace"),
1461                }
1462        };
1463        if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
1464                $self->{svm} = $svm;
1465        }
1466        $self->{svm};
1467}
1468
1469sub _set_svm_vars {
1470        my ($self, $ra) = @_;
1471        return $ra if $self->svm;
1472
1473        my @err = ( "useSvmProps set, but failed to read SVM properties\n",
1474                    "(svm:source, svm:uuid) ",
1475                    "from the following URLs:\n" );
1476        sub read_svm_props {
1477                my ($self, $ra, $path, $r) = @_;
1478                my $props = ($ra->get_dir($path, $r))[2];
1479                my $src = $props->{'svm:source'};
1480                my $uuid = $props->{'svm:uuid'};
1481                return undef if (!$src || !$uuid);
1482
1483                chomp($src, $uuid);
1484
1485                $uuid =~ m{^[0-9a-f\-]{30,}$}
1486                    or die "doesn't look right - svm:uuid is '$uuid'\n";
1487
1488                # the '!' is used to mark the repos_root!/relative/path
1489                $src =~ s{/?!/?}{/};
1490                $src =~ s{/+$}{}; # no trailing slashes please
1491                # username is of no interest
1492                $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
1493
1494                my $replace = $ra->{url};
1495                $replace .= "/$path" if length $path;
1496
1497                my $section = "svn-remote.$self->{repo_id}";
1498                tmp_config("$section.svm-source", $src);
1499                tmp_config("$section.svm-replace", $replace);
1500                tmp_config("$section.svm-uuid", $uuid);
1501                $self->{svm} = {
1502                        source => $src,
1503                        uuid => $uuid,
1504                        replace => $replace
1505                };
1506        }
1507
1508        my $r = $ra->get_latest_revnum;
1509        my $path = $self->{path};
1510        my %tried;
1511        while (length $path) {
1512                unless ($tried{"$self->{url}/$path"}) {
1513                        return $ra if $self->read_svm_props($ra, $path, $r);
1514                        $tried{"$self->{url}/$path"} = 1;
1515                }
1516                $path =~ s#/?[^/]+$##;
1517        }
1518        die "Path: '$path' should be ''\n" if $path ne '';
1519        return $ra if $self->read_svm_props($ra, $path, $r);
1520        $tried{"$self->{url}/$path"} = 1;
1521
1522        if ($ra->{repos_root} eq $self->{url}) {
1523                die @err, (map { "  $_\n" } keys %tried), "\n";
1524        }
1525
1526        # nope, make sure we're connected to the repository root:
1527        my $ok;
1528        my @tried_b;
1529        $path = $ra->{svn_path};
1530        $ra = Git::SVN::Ra->new($ra->{repos_root});
1531        while (length $path) {
1532                unless ($tried{"$ra->{url}/$path"}) {
1533                        $ok = $self->read_svm_props($ra, $path, $r);
1534                        last if $ok;
1535                        $tried{"$ra->{url}/$path"} = 1;
1536                }
1537                $path =~ s#/?[^/]+$##;
1538        }
1539        die "Path: '$path' should be ''\n" if $path ne '';
1540        $ok ||= $self->read_svm_props($ra, $path, $r);
1541        $tried{"$ra->{url}/$path"} = 1;
1542        if (!$ok) {
1543                die @err, (map { "  $_\n" } keys %tried), "\n";
1544        }
1545        Git::SVN::Ra->new($self->{url});
1546}
1547
1548sub svnsync {
1549        my ($self) = @_;
1550        return $self->{svnsync} if $self->{svnsync};
1551
1552        if ($self->no_metadata) {
1553                die "Can't have both 'noMetadata' and ",
1554                    "'useSvnsyncProps' options set!\n";
1555        }
1556        if ($self->rewrite_root) {
1557                die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
1558                    "options set!\n";
1559        }
1560
1561        my $svnsync;
1562        # see if we have it in our config, first:
1563        eval {
1564                my $section = "svn-remote.$self->{repo_id}";
1565                $svnsync = {
1566                  url => tmp_config('--get', "$section.svnsync-url"),
1567                  uuid => tmp_config('--get', "$section.svnsync-uuid"),
1568                }
1569        };
1570        if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
1571                return $self->{svnsync} = $svnsync;
1572        }
1573
1574        my $err = "useSvnsyncProps set, but failed to read " .
1575                  "svnsync property: svn:sync-from-";
1576        my $rp = $self->ra->rev_proplist(0);
1577
1578        my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
1579        $url =~ m{^[a-z\+]+://} or
1580                   die "doesn't look right - svn:sync-from-url is '$url'\n";
1581
1582        my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
1583        $uuid =~ m{^[0-9a-f\-]{30,}$} or
1584                   die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1585
1586        my $section = "svn-remote.$self->{repo_id}";
1587        tmp_config('--add', "$section.svnsync-uuid", $uuid);
1588        tmp_config('--add', "$section.svnsync-url", $url);
1589        return $self->{svnsync} = { url => $url, uuid => $uuid };
1590}
1591
1592# this allows us to memoize our SVN::Ra UUID locally and avoid a
1593# remote lookup (useful for 'git svn log').
1594sub ra_uuid {
1595        my ($self) = @_;
1596        unless ($self->{ra_uuid}) {
1597                my $key = "svn-remote.$self->{repo_id}.uuid";
1598                my $uuid = eval { tmp_config('--get', $key) };
1599                if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/) {
1600                        $self->{ra_uuid} = $uuid;
1601                } else {
1602                        die "ra_uuid called without URL\n" unless $self->{url};
1603                        $self->{ra_uuid} = $self->ra->get_uuid;
1604                        tmp_config('--add', $key, $self->{ra_uuid});
1605                }
1606        }
1607        $self->{ra_uuid};
1608}
1609
1610sub ra {
1611        my ($self) = shift;
1612        my $ra = Git::SVN::Ra->new($self->{url});
1613        if ($self->use_svm_props && !$self->{svm}) {
1614                if ($self->no_metadata) {
1615                        die "Can't have both 'noMetadata' and ",
1616                            "'useSvmProps' options set!\n";
1617                } elsif ($self->use_svnsync_props) {
1618                        die "Can't have both 'useSvnsyncProps' and ",
1619                            "'useSvmProps' options set!\n";
1620                }
1621                $ra = $self->_set_svm_vars($ra);
1622                $self->{-want_revprops} = 1;
1623        }
1624        $ra;
1625}
1626
1627sub rel_path {
1628        my ($self) = @_;
1629        my $repos_root = $self->ra->{repos_root};
1630        return $self->{path} if ($self->{url} eq $repos_root);
1631        my $url = $self->{url} .
1632                  (length $self->{path} ? "/$self->{path}" : $self->{path});
1633        $url =~ s!^\Q$repos_root\E(?:/+|$)!!g;
1634        $url;
1635}
1636
1637# prop_walk(PATH, REV, SUB)
1638# -------------------------
1639# Recursively traverse PATH at revision REV and invoke SUB for each
1640# directory that contains a SVN property.  SUB will be invoked as
1641# follows:  &SUB(gs, path, props);  where `gs' is this instance of
1642# Git::SVN, `path' the path to the directory where the properties
1643# `props' were found.  The `path' will be relative to point of checkout,
1644# that is, if url://repo/trunk is the current Git branch, and that
1645# directory contains a sub-directory `d', SUB will be invoked with `/d/'
1646# as `path' (note the trailing `/').
1647sub prop_walk {
1648        my ($self, $path, $rev, $sub) = @_;
1649
1650        my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
1651        $path =~ s#^/*#/#g;
1652        my $p = $path;
1653        # Strip the irrelevant part of the path.
1654        $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
1655        # Ensure the path is terminated by a `/'.
1656        $p =~ s#/*$#/#;
1657
1658        # The properties contain all the internal SVN stuff nobody
1659        # (usually) cares about.
1660        my $interesting_props = 0;
1661        foreach (keys %{$props}) {
1662                # If it doesn't start with `svn:', it must be a
1663                # user-defined property.
1664                ++$interesting_props and next if $_ !~ /^svn:/;
1665                # FIXME: Fragile, if SVN adds new public properties,
1666                # this needs to be updated.
1667                ++$interesting_props if /^svn:(?:ignore|keywords|executable
1668                                                 |eol-style|mime-type
1669                                                 |externals|needs-lock)$/x;
1670        }
1671        &$sub($self, $p, $props) if $interesting_props;
1672
1673        foreach (sort keys %$dirent) {
1674                next if $dirent->{$_}->{kind} != $SVN::Node::dir;
1675                $self->prop_walk($path . '/' . $_, $rev, $sub);
1676        }
1677}
1678
1679sub last_rev { ($_[0]->last_rev_commit)[0] }
1680sub last_commit { ($_[0]->last_rev_commit)[1] }
1681
1682# returns the newest SVN revision number and newest commit SHA1
1683sub last_rev_commit {
1684        my ($self) = @_;
1685        if (defined $self->{last_rev} && defined $self->{last_commit}) {
1686                return ($self->{last_rev}, $self->{last_commit});
1687        }
1688        my $c = ::verify_ref($self->refname.'^0');
1689        if ($c && !$self->use_svm_props && !$self->no_metadata) {
1690                my $rev = (::cmt_metadata($c))[1];
1691                if (defined $rev) {
1692                        ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
1693                        return ($rev, $c);
1694                }
1695        }
1696        my $db_path = $self->db_path;
1697        unless (-e $db_path) {
1698                ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
1699                return (undef, undef);
1700        }
1701        my $offset = -41; # from tail
1702        my $rl;
1703        open my $fh, '<', $db_path or croak "$db_path not readable: $!\n";
1704        sysseek($fh, $offset, 2); # don't care for errors
1705        sysread($fh, $rl, 41) == 41 or return (undef, undef);
1706        chomp $rl;
1707        while (('0' x40) eq $rl && sysseek($fh, 0, 1) != 0) {
1708                $offset -= 41;
1709                sysseek($fh, $offset, 2); # don't care for errors
1710                sysread($fh, $rl, 41) == 41 or return (undef, undef);
1711                chomp $rl;
1712        }
1713        if ($c && $c ne $rl) {
1714                die "$db_path and ", $self->refname,
1715                    " inconsistent!:\n$c != $rl\n";
1716        }
1717        my $rev = sysseek($fh, 0, 1) or croak $!;
1718        $rev =  ($rev - 41) / 41;
1719        close $fh or croak $!;
1720        ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
1721        return ($rev, $c);
1722}
1723
1724sub get_fetch_range {
1725        my ($self, $min, $max) = @_;
1726        $max ||= $self->ra->get_latest_revnum;
1727        $min ||= $self->rev_db_max;
1728        (++$min, $max);
1729}
1730
1731sub tmp_config {
1732        my (@args) = @_;
1733        my $old_def_config = "$ENV{GIT_DIR}/svn/config";
1734        my $config = "$ENV{GIT_DIR}/svn/.metadata";
1735        if (! -f $config && -f $old_def_config) {
1736                rename $old_def_config, $config or
1737                       die "Failed rename $old_def_config => $config: $!\n";
1738        }
1739        my $old_config = $ENV{GIT_CONFIG};
1740        $ENV{GIT_CONFIG} = $config;
1741        $@ = undef;
1742        my @ret = eval {
1743                unless (-f $config) {
1744                        mkfile($config);
1745                        open my $fh, '>', $config or
1746                            die "Can't open $config: $!\n";
1747                        print $fh "; This file is used internally by ",
1748                                  "git-svn\n" or die
1749                                  "Couldn't write to $config: $!\n";
1750                        print $fh "; You should not have to edit it\n" or
1751                              die "Couldn't write to $config: $!\n";
1752                        close $fh or die "Couldn't close $config: $!\n";
1753                }
1754                command('config', @args);
1755        };
1756        my $err = $@;
1757        if (defined $old_config) {
1758                $ENV{GIT_CONFIG} = $old_config;
1759        } else {
1760                delete $ENV{GIT_CONFIG};
1761        }
1762        die $err if $err;
1763        wantarray ? @ret : $ret[0];
1764}
1765
1766sub tmp_index_do {
1767        my ($self, $sub) = @_;
1768        my $old_index = $ENV{GIT_INDEX_FILE};
1769        $ENV{GIT_INDEX_FILE} = $self->{index};
1770        $@ = undef;
1771        my @ret = eval {
1772                my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
1773                mkpath([$dir]) unless -d $dir;
1774                &$sub;
1775        };
1776        my $err = $@;
1777        if (defined $old_index) {
1778                $ENV{GIT_INDEX_FILE} = $old_index;
1779        } else {
1780                delete $ENV{GIT_INDEX_FILE};
1781        }
1782        die $err if $err;
1783        wantarray ? @ret : $ret[0];
1784}
1785
1786sub assert_index_clean {
1787        my ($self, $treeish) = @_;
1788
1789        $self->tmp_index_do(sub {
1790                command_noisy('read-tree', $treeish) unless -e $self->{index};
1791                my $x = command_oneline('write-tree');
1792                my ($y) = (command(qw/cat-file commit/, $treeish) =~
1793                           /^tree ($::sha1)/mo);
1794                return if $y eq $x;
1795
1796                warn "Index mismatch: $y != $x\nrereading $treeish\n";
1797                unlink $self->{index} or die "unlink $self->{index}: $!\n";
1798                command_noisy('read-tree', $treeish);
1799                $x = command_oneline('write-tree');
1800                if ($y ne $x) {
1801                        ::fatal "trees ($treeish) $y != $x\n",
1802                                "Something is seriously wrong...";
1803                }
1804        });
1805}
1806
1807sub get_commit_parents {
1808        my ($self, $log_entry) = @_;
1809        my (%seen, @ret, @tmp);
1810        # legacy support for 'set-tree'; this is only used by set_tree_cb:
1811        if (my $ip = $self->{inject_parents}) {
1812                if (my $commit = delete $ip->{$log_entry->{revision}}) {
1813                        push @tmp, $commit;
1814                }
1815        }
1816        if (my $cur = ::verify_ref($self->refname.'^0')) {
1817                push @tmp, $cur;
1818        }
1819        if (my $ipd = $self->{inject_parents_dcommit}) {
1820                if (my $commit = delete $ipd->{$log_entry->{revision}}) {
1821                        push @tmp, @$commit;
1822                }
1823        }
1824        push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
1825        while (my $p = shift @tmp) {
1826                next if $seen{$p};
1827                $seen{$p} = 1;
1828                push @ret, $p;
1829                # MAXPARENT is defined to 16 in commit-tree.c:
1830                last if @ret >= 16;
1831        }
1832        if (@tmp) {
1833                die "r$log_entry->{revision}: No room for parents:\n\t",
1834                    join("\n\t", @tmp), "\n";
1835        }
1836        @ret;
1837}
1838
1839sub rewrite_root {
1840        my ($self) = @_;
1841        return $self->{-rewrite_root} if exists $self->{-rewrite_root};
1842        my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
1843        my $rwr = eval { command_oneline(qw/config --get/, $k) };
1844        if ($rwr) {
1845                $rwr =~ s#/+$##;
1846                if ($rwr !~ m#^[a-z\+]+://#) {
1847                        die "$rwr is not a valid URL (key: $k)\n";
1848                }
1849        }
1850        $self->{-rewrite_root} = $rwr;
1851}
1852
1853sub metadata_url {
1854        my ($self) = @_;
1855        ($self->rewrite_root || $self->{url}) .
1856           (length $self->{path} ? '/' . $self->{path} : '');
1857}
1858
1859sub full_url {
1860        my ($self) = @_;
1861        $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
1862}
1863
1864sub do_git_commit {
1865        my ($self, $log_entry) = @_;
1866        my $lr = $self->last_rev;
1867        if (defined $lr && $lr >= $log_entry->{revision}) {
1868                die "Last fetched revision of ", $self->refname,
1869                    " was r$lr, but we are about to fetch: ",
1870                    "r$log_entry->{revision}!\n";
1871        }
1872        if (my $c = $self->rev_db_get($log_entry->{revision})) {
1873                croak "$log_entry->{revision} = $c already exists! ",
1874                      "Why are we refetching it?\n";
1875        }
1876        $ENV{GIT_AUTHOR_NAME} = $ENV{GIT_COMMITTER_NAME} = $log_entry->{name};
1877        $ENV{GIT_AUTHOR_EMAIL} = $ENV{GIT_COMMITTER_EMAIL} =
1878                                                          $log_entry->{email};
1879        $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
1880
1881        my $tree = $log_entry->{tree};
1882        if (!defined $tree) {
1883                $tree = $self->tmp_index_do(sub {
1884                                            command_oneline('write-tree') });
1885        }
1886        die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
1887
1888        my @exec = ('git-commit-tree', $tree);
1889        foreach ($self->get_commit_parents($log_entry)) {
1890                push @exec, '-p', $_;
1891        }
1892        defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
1893                                                                   or croak $!;
1894        print $msg_fh $log_entry->{log} or croak $!;
1895        unless ($self->no_metadata) {
1896                print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
1897                              or croak $!;
1898        }
1899        $msg_fh->flush == 0 or croak $!;
1900        close $msg_fh or croak $!;
1901        chomp(my $commit = do { local $/; <$out_fh> });
1902        close $out_fh or croak $!;
1903        waitpid $pid, 0;
1904        croak $? if $?;
1905        if ($commit !~ /^$::sha1$/o) {
1906                die "Failed to commit, invalid sha1: $commit\n";
1907        }
1908
1909        $self->rev_db_set($log_entry->{revision}, $commit, 1);
1910
1911        $self->{last_rev} = $log_entry->{revision};
1912        $self->{last_commit} = $commit;
1913        print "r$log_entry->{revision}";
1914        if (defined $log_entry->{svm_revision}) {
1915                 print " (\@$log_entry->{svm_revision})";
1916                 $self->rev_db_set($log_entry->{svm_revision}, $commit,
1917                                   0, $self->svm_uuid);
1918        }
1919        print " = $commit ($self->{ref_id})\n";
1920        if (defined $_repack && (--$_repack_nr == 0)) {
1921                $_repack_nr = $_repack;
1922                # repack doesn't use any arguments with spaces in them, does it?
1923                print "Running git repack $_repack_flags ...\n";
1924                command_noisy('repack', split(/\s+/, $_repack_flags));
1925                print "Done repacking\n";
1926        }
1927        return $commit;
1928}
1929
1930sub match_paths {
1931        my ($self, $paths, $r) = @_;
1932        return 1 if $self->{path} eq '';
1933        if (my $path = $paths->{"/$self->{path}"}) {
1934                return ($path->{action} eq 'D') ? 0 : 1;
1935        }
1936        $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
1937        if (grep /$self->{path_regex}/, keys %$paths) {
1938                return 1;
1939        }
1940        my $c = '';
1941        foreach (split m#/#, $self->{path}) {
1942                $c .= "/$_";
1943                next unless ($paths->{$c} &&
1944                             ($paths->{$c}->{action} =~ /^[AR]$/));
1945                if ($self->ra->check_path($self->{path}, $r) ==
1946                    $SVN::Node::dir) {
1947                        return 1;
1948                }
1949        }
1950        return 0;
1951}
1952
1953sub find_parent_branch {
1954        my ($self, $paths, $rev) = @_;
1955        return undef unless $self->follow_parent;
1956        unless (defined $paths) {
1957                my $err_handler = $SVN::Error::handler;
1958                $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
1959                $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1, sub {
1960                                   $paths =
1961                                      Git::SVN::Ra::dup_changed_paths($_[0]) });
1962                $SVN::Error::handler = $err_handler;
1963        }
1964        return undef unless defined $paths;
1965
1966        # look for a parent from another branch:
1967        my @b_path_components = split m#/#, $self->rel_path;
1968        my @a_path_components;
1969        my $i;
1970        while (@b_path_components) {
1971                $i = $paths->{'/'.join('/', @b_path_components)};
1972                last if $i && defined $i->{copyfrom_path};
1973                unshift(@a_path_components, pop(@b_path_components));
1974        }
1975        return undef unless defined $i && defined $i->{copyfrom_path};
1976        my $branch_from = $i->{copyfrom_path};
1977        if (@a_path_components) {
1978                print STDERR "branch_from: $branch_from => ";
1979                $branch_from .= '/'.join('/', @a_path_components);
1980                print STDERR $branch_from, "\n";
1981        }
1982        my $r = $i->{copyfrom_rev};
1983        my $repos_root = $self->ra->{repos_root};
1984        my $url = $self->ra->{url};
1985        my $new_url = $repos_root . $branch_from;
1986        print STDERR  "Found possible branch point: ",
1987                      "$new_url => ", $self->full_url, ", $r\n";
1988        $branch_from =~ s#^/##;
1989        my $gs = Git::SVN->find_by_url($new_url, $repos_root, $branch_from);
1990        unless ($gs) {
1991                my $ref_id = $self->{ref_id};
1992                $ref_id =~ s/\@\d+$//;
1993                $ref_id .= "\@$r";
1994                # just grow a tail if we're not unique enough :x
1995                $ref_id .= '-' while find_ref($ref_id);
1996                print STDERR "Initializing parent: $ref_id\n";
1997                $gs = Git::SVN->init($new_url, '', $ref_id, $ref_id, 1);
1998        }
1999        my ($r0, $parent) = $gs->find_rev_before($r, 1);
2000        if (!defined $r0 || !defined $parent) {
2001                my ($base, $head) = parse_revision_argument(0, $r);
2002                if ($base <= $r) {
2003                        $gs->fetch($base, $r);
2004                }
2005                ($r0, $parent) = $gs->last_rev_commit;
2006        }
2007        if (defined $r0 && defined $parent) {
2008                print STDERR "Found branch parent: ($self->{ref_id}) $parent\n";
2009                my $ed;
2010                if ($self->ra->can_do_switch) {
2011                        $self->assert_index_clean($parent);
2012                        print STDERR "Following parent with do_switch\n";
2013                        # do_switch works with svn/trunk >= r22312, but that
2014                        # is not included with SVN 1.4.3 (the latest version
2015                        # at the moment), so we can't rely on it
2016                        $self->{last_commit} = $parent;
2017                        $ed = SVN::Git::Fetcher->new($self);
2018                        $gs->ra->gs_do_switch($r0, $rev, $gs,
2019                                              $self->full_url, $ed)
2020                          or die "SVN connection failed somewhere...\n";
2021                } elsif ($self->ra->trees_match($new_url, $r0,
2022                                                $self->full_url, $rev)) {
2023                        print STDERR "Trees match:\n",
2024                                     "  $new_url\@$r0\n",
2025                                     "  ${\$self->full_url}\@$rev\n",
2026                                     "Following parent with no changes\n";
2027                        $self->tmp_index_do(sub {
2028                            command_noisy('read-tree', $parent);
2029                        });
2030                        $self->{last_commit} = $parent;
2031                } else {
2032                        print STDERR "Following parent with do_update\n";
2033                        $ed = SVN::Git::Fetcher->new($self);
2034                        $self->ra->gs_do_update($rev, $rev, $self, $ed)
2035                          or die "SVN connection failed somewhere...\n";
2036                }
2037                print STDERR "Successfully followed parent\n";
2038                return $self->make_log_entry($rev, [$parent], $ed);
2039        }
2040        return undef;
2041}
2042
2043sub do_fetch {
2044        my ($self, $paths, $rev) = @_;
2045        my $ed;
2046        my ($last_rev, @parents);
2047        if (my $lc = $self->last_commit) {
2048                # we can have a branch that was deleted, then re-added
2049                # under the same name but copied from another path, in
2050                # which case we'll have multiple parents (we don't
2051                # want to break the original ref, nor lose copypath info):
2052                if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2053                        push @{$log_entry->{parents}}, $lc;
2054                        return $log_entry;
2055                }
2056                $ed = SVN::Git::Fetcher->new($self);
2057                $last_rev = $self->{last_rev};
2058                $ed->{c} = $lc;
2059                @parents = ($lc);
2060        } else {
2061                $last_rev = $rev;
2062                if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2063                        return $log_entry;
2064                }
2065                $ed = SVN::Git::Fetcher->new($self);
2066        }
2067        unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2068                die "SVN connection failed somewhere...\n";
2069        }
2070        $self->make_log_entry($rev, \@parents, $ed);
2071}
2072
2073sub get_untracked {
2074        my ($self, $ed) = @_;
2075        my @out;
2076        my $h = $ed->{empty};
2077        foreach (sort keys %$h) {
2078                my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2079                push @out, "  $act: " . uri_encode($_);
2080                warn "W: $act: $_\n";
2081        }
2082        foreach my $t (qw/dir_prop file_prop/) {
2083                $h = $ed->{$t} or next;
2084                foreach my $path (sort keys %$h) {
2085                        my $ppath = $path eq '' ? '.' : $path;
2086                        foreach my $prop (sort keys %{$h->{$path}}) {
2087                                next if $SKIP_PROP{$prop};
2088                                my $v = $h->{$path}->{$prop};
2089                                my $t_ppath_prop = "$t: " .
2090                                                    uri_encode($ppath) . ' ' .
2091                                                    uri_encode($prop);
2092                                if (defined $v) {
2093                                        push @out, "  +$t_ppath_prop " .
2094                                                   uri_encode($v);
2095                                } else {
2096                                        push @out, "  -$t_ppath_prop";
2097                                }
2098                        }
2099                }
2100        }
2101        foreach my $t (qw/absent_file absent_directory/) {
2102                $h = $ed->{$t} or next;
2103                foreach my $parent (sort keys %$h) {
2104                        foreach my $path (sort @{$h->{$parent}}) {
2105                                push @out, "  $t: " .
2106                                           uri_encode("$parent/$path");
2107                                warn "W: $t: $parent/$path ",
2108                                     "Insufficient permissions?\n";
2109                        }
2110                }
2111        }
2112        \@out;
2113}
2114
2115sub parse_svn_date {
2116        my $date = shift || return '+0000 1970-01-01 00:00:00';
2117        my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2118                                            (\d\d)\:(\d\d)\:(\d\d).\d+Z$/x) or
2119                                         croak "Unable to parse date: $date\n";
2120        "+0000 $Y-$m-$d $H:$M:$S";
2121}
2122
2123sub check_author {
2124        my ($author) = @_;
2125        if (!defined $author || length $author == 0) {
2126                $author = '(no author)';
2127        }
2128        if (defined $::_authors && ! defined $::users{$author}) {
2129                die "Author: $author not defined in $::_authors file\n";
2130        }
2131        $author;
2132}
2133
2134sub make_log_entry {
2135        my ($self, $rev, $parents, $ed) = @_;
2136        my $untracked = $self->get_untracked($ed);
2137
2138        open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
2139        print $un "r$rev\n" or croak $!;
2140        print $un $_, "\n" foreach @$untracked;
2141        my %log_entry = ( parents => $parents || [], revision => $rev,
2142                          log => '');
2143
2144        my $headrev;
2145        my $logged = delete $self->{logged_rev_props};
2146        if (!$logged || $self->{-want_revprops}) {
2147                my $rp = $self->ra->rev_proplist($rev);
2148                foreach (sort keys %$rp) {
2149                        my $v = $rp->{$_};
2150                        if (/^svn:(author|date|log)$/) {
2151                                $log_entry{$1} = $v;
2152                        } elsif ($_ eq 'svm:headrev') {
2153                                $headrev = $v;
2154                        } else {
2155                                print $un "  rev_prop: ", uri_encode($_), ' ',
2156                                          uri_encode($v), "\n";
2157                        }
2158                }
2159        } else {
2160                map { $log_entry{$_} = $logged->{$_} } keys %$logged;
2161        }
2162        close $un or croak $!;
2163
2164        $log_entry{date} = parse_svn_date($log_entry{date});
2165        $log_entry{log} .= "\n";
2166        my $author = $log_entry{author} = check_author($log_entry{author});
2167        my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
2168                                                       : ($author, undef);
2169        if (defined $headrev && $self->use_svm_props) {
2170                if ($self->rewrite_root) {
2171                        die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
2172                            "options set!\n";
2173                }
2174                my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$};
2175                # we don't want "SVM: initializing mirror for junk" ...
2176                return undef if $r == 0;
2177                my $svm = $self->svm;
2178                if ($uuid ne $svm->{uuid}) {
2179                        die "UUID mismatch on SVM path:\n",
2180                            "expected: $svm->{uuid}\n",
2181                            "     got: $uuid\n";
2182                }
2183                my $full_url = $self->full_url;
2184                $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
2185                             die "Failed to replace '$svm->{replace}' with ",
2186                                 "'$svm->{source}' in $full_url\n";
2187                # throw away username for storing in records
2188                remove_username($full_url);
2189                $log_entry{metadata} = "$full_url\@$r $uuid";
2190                $log_entry{svm_revision} = $r;
2191                $email ||= "$author\@$uuid"
2192        } elsif ($self->use_svnsync_props) {
2193                my $full_url = $self->svnsync->{url};
2194                $full_url .= "/$self->{path}" if length $self->{path};
2195                remove_username($full_url);
2196                my $uuid = $self->svnsync->{uuid};
2197                $log_entry{metadata} = "$full_url\@$rev $uuid";
2198                $email ||= "$author\@$uuid"
2199        } else {
2200                my $url = $self->metadata_url;
2201                remove_username($url);
2202                $log_entry{metadata} = "$url\@$rev " .
2203                                       $self->ra->get_uuid;
2204                $email ||= "$author\@" . $self->ra->get_uuid;
2205        }
2206        $log_entry{name} = $name;
2207        $log_entry{email} = $email;
2208        \%log_entry;
2209}
2210
2211sub fetch {
2212        my ($self, $min_rev, $max_rev, @parents) = @_;
2213        my ($last_rev, $last_commit) = $self->last_rev_commit;
2214        my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
2215        $self->ra->gs_fetch_loop_common($base, $head, [$self]);
2216}
2217
2218sub set_tree_cb {
2219        my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
2220        $self->{inject_parents} = { $rev => $tree };
2221        $self->fetch(undef, undef);
2222}
2223
2224sub set_tree {
2225        my ($self, $tree) = (shift, shift);
2226        my $log_entry = ::get_commit_entry($tree);
2227        unless ($self->{last_rev}) {
2228                fatal("Must have an existing revision to commit");
2229        }
2230        my %ed_opts = ( r => $self->{last_rev},
2231                        log => $log_entry->{log},
2232                        ra => $self->ra,
2233                        tree_a => $self->{last_commit},
2234                        tree_b => $tree,
2235                        editor_cb => sub {
2236                               $self->set_tree_cb($log_entry, $tree, @_) },
2237                        svn_path => $self->{path} );
2238        if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
2239                print "No changes\nr$self->{last_rev} = $tree\n";
2240        }
2241}
2242
2243sub rebuild {
2244        my ($self) = @_;
2245        my $db_path = $self->db_path;
2246        return if (-e $db_path && ! -z $db_path);
2247        return unless ::verify_ref($self->refname.'^0');
2248        if (-f $self->{db_root}) {
2249                rename $self->{db_root}, $db_path or die
2250                     "rename $self->{db_root} => $db_path failed: $!\n";
2251                my ($dir, $base) = ($db_path =~ m#^(.*?)/?([^/]+)$#);
2252                symlink $base, $self->{db_root} or die
2253                     "symlink $base => $self->{db_root} failed: $!\n";
2254                return;
2255        }
2256        print "Rebuilding $db_path ...\n";
2257        my ($log, $ctx) = command_output_pipe("log", '--no-color', $self->refname);
2258        my $latest;
2259        my $full_url = $self->full_url;
2260        remove_username($full_url);
2261        my $svn_uuid;
2262        my $c;
2263        while (<$log>) {
2264                if ( m{^commit ($::sha1)$} ) {
2265                        $c = $1;
2266                        next;
2267                }
2268                next unless s{^\s*(git-svn-id:)}{$1};
2269                my ($url, $rev, $uuid) = ::extract_metadata($_);
2270                remove_username($url);
2271
2272                # ignore merges (from set-tree)
2273                next if (!defined $rev || !$uuid);
2274
2275                # if we merged or otherwise started elsewhere, this is
2276                # how we break out of it
2277                if ((defined $svn_uuid && ($uuid ne $svn_uuid)) ||
2278                    ($full_url && $url && ($url ne $full_url))) {
2279                        next;
2280                }
2281                $latest ||= $rev;
2282                $svn_uuid ||= $uuid;
2283
2284                $self->rev_db_set($rev, $c);
2285                print "r$rev = $c\n";
2286        }
2287        command_close_pipe($log, $ctx);
2288        print "Done rebuilding $db_path\n";
2289}
2290
2291# rev_db:
2292# Tie::File seems to be prone to offset errors if revisions get sparse,
2293# it's not that fast, either.  Tie::File is also not in Perl 5.6.  So
2294# one of my favorite modules is out :<  Next up would be one of the DBM
2295# modules, but I'm not sure which is most portable...  So I'll just
2296# go with something that's plain-text, but still capable of
2297# being randomly accessed.  So here's my ultra-simple fixed-width
2298# database.  All records are 40 characters + "\n", so it's easy to seek
2299# to a revision: (41 * rev) is the byte offset.
2300# A record of 40 0s denotes an empty revision.
2301# And yes, it's still pretty fast (faster than Tie::File).
2302# These files are disposable unless noMetadata or useSvmProps is set
2303
2304sub _rev_db_set {
2305        my ($fh, $rev, $commit) = @_;
2306        my $offset = $rev * 41;
2307        # assume that append is the common case:
2308        seek $fh, 0, 2 or croak $!;
2309        my $pos = tell $fh;
2310        if ($pos < $offset) {
2311                for (1 .. (($offset - $pos) / 41)) {
2312                        print $fh (('0' x 40),"\n") or croak $!;
2313                }
2314        }
2315        seek $fh, $offset, 0 or croak $!;
2316        print $fh $commit,"\n" or croak $!;
2317}
2318
2319sub mkfile {
2320        my ($path) = @_;
2321        unless (-e $path) {
2322                my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
2323                mkpath([$dir]) unless -d $dir;
2324                open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
2325                close $fh or die "Couldn't close (create) $path: $!\n";
2326        }
2327}
2328
2329sub rev_db_set {
2330        my ($self, $rev, $commit, $update_ref, $uuid) = @_;
2331        length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
2332        my $db = $self->db_path($uuid);
2333        my $db_lock = "$db.lock";
2334        my $sig;
2335        if ($update_ref) {
2336                $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2337                            $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
2338        }
2339        mkfile($db);
2340
2341        $LOCKFILES{$db_lock} = 1;
2342        my $sync;
2343        # both of these options make our .rev_db file very, very important
2344        # and we can't afford to lose it because rebuild() won't work
2345        if ($self->use_svm_props || $self->no_metadata) {
2346                $sync = 1;
2347                copy($db, $db_lock) or die "rev_db_set(@_): ",
2348                                           "Failed to copy: ",
2349                                           "$db => $db_lock ($!)\n";
2350        } else {
2351                rename $db, $db_lock or die "rev_db_set(@_): ",
2352                                            "Failed to rename: ",
2353                                            "$db => $db_lock ($!)\n";
2354        }
2355        open my $fh, '+<', $db_lock or die "Couldn't open $db_lock: $!\n";
2356        _rev_db_set($fh, $rev, $commit);
2357        if ($sync) {
2358                $fh->flush or die "Couldn't flush $db_lock: $!\n";
2359                $fh->sync or die "Couldn't sync $db_lock: $!\n";
2360        }
2361        close $fh or croak $!;
2362        if ($update_ref) {
2363                $_head = $self;
2364                command_noisy('update-ref', '-m', "r$rev",
2365                              $self->refname, $commit);
2366        }
2367        rename $db_lock, $db or die "rev_db_set(@_): ", "Failed to rename: ",
2368                                    "$db_lock => $db ($!)\n";
2369        delete $LOCKFILES{$db_lock};
2370        if ($update_ref) {
2371                $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2372                            $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
2373                kill $sig, $$ if defined $sig;
2374        }
2375}
2376
2377sub rev_db_max {
2378        my ($self) = @_;
2379        $self->rebuild;
2380        my $db_path = $self->db_path;
2381        my @stat = stat $db_path or return 0;
2382        ($stat[7] % 41) == 0 or die "$db_path inconsistent size: $stat[7]\n";
2383        my $max = $stat[7] / 41;
2384        (($max > 0) ? $max - 1 : 0);
2385}
2386
2387sub rev_db_get {
2388        my ($self, $rev, $uuid) = @_;
2389        my $ret;
2390        my $offset = $rev * 41;
2391        my $db_path = $self->db_path($uuid);
2392        return undef unless -e $db_path;
2393        open my $fh, '<', $db_path or croak $!;
2394        if (sysseek($fh, $offset, 0) == $offset) {
2395                my $read = sysread($fh, $ret, 40);
2396                $ret = undef if ($read != 40 || $ret eq ('0'x40));
2397        }
2398        close $fh or croak $!;
2399        $ret;
2400}
2401
2402# Finds the first svn revision that exists on (if $eq_ok is true) or
2403# before $rev for the current branch.  It will not search any lower
2404# than $min_rev.  Returns the git commit hash and svn revision number
2405# if found, else (undef, undef).
2406sub find_rev_before {
2407        my ($self, $rev, $eq_ok, $min_rev) = @_;
2408        --$rev unless $eq_ok;
2409        $min_rev ||= 1;
2410        while ($rev >= $min_rev) {
2411                if (my $c = $self->rev_db_get($rev)) {
2412                        return ($rev, $c);
2413                }
2414                --$rev;
2415        }
2416        return (undef, undef);
2417}
2418
2419# Finds the first svn revision that exists on (if $eq_ok is true) or
2420# after $rev for the current branch.  It will not search any higher
2421# than $max_rev.  Returns the git commit hash and svn revision number
2422# if found, else (undef, undef).
2423sub find_rev_after {
2424        my ($self, $rev, $eq_ok, $max_rev) = @_;
2425        ++$rev unless $eq_ok;
2426        $max_rev ||= $self->rev_db_max();
2427        while ($rev <= $max_rev) {
2428                if (my $c = $self->rev_db_get($rev)) {
2429                        return ($rev, $c);
2430                }
2431                ++$rev;
2432        }
2433        return (undef, undef);
2434}
2435
2436sub _new {
2437        my ($class, $repo_id, $ref_id, $path) = @_;
2438        unless (defined $repo_id && length $repo_id) {
2439                $repo_id = $Git::SVN::default_repo_id;
2440        }
2441        unless (defined $ref_id && length $ref_id) {
2442                $_[2] = $ref_id = $Git::SVN::default_ref_id;
2443        }
2444        $_[1] = $repo_id = sanitize_remote_name($repo_id);
2445        my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
2446        $_[3] = $path = '' unless (defined $path);
2447        mkpath(["$ENV{GIT_DIR}/svn"]);
2448        bless {
2449                ref_id => $ref_id, dir => $dir, index => "$dir/index",
2450                path => $path, config => "$ENV{GIT_DIR}/svn/config",
2451                db_root => "$dir/.rev_db", repo_id => $repo_id }, $class;
2452}
2453
2454sub db_path {
2455        my ($self, $uuid) = @_;
2456        $uuid ||= $self->ra_uuid;
2457        "$self->{db_root}.$uuid";
2458}
2459
2460sub uri_encode {
2461        my ($f) = @_;
2462        $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
2463        $f
2464}
2465
2466sub remove_username {
2467        $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
2468}
2469
2470package Git::SVN::Prompt;
2471use strict;
2472use warnings;
2473require SVN::Core;
2474use vars qw/$_no_auth_cache $_username/;
2475
2476sub simple {
2477        my ($cred, $realm, $default_username, $may_save, $pool) = @_;
2478        $may_save = undef if $_no_auth_cache;
2479        $default_username = $_username if defined $_username;
2480        if (defined $default_username && length $default_username) {
2481                if (defined $realm && length $realm) {
2482                        print STDERR "Authentication realm: $realm\n";
2483                        STDERR->flush;
2484                }
2485                $cred->username($default_username);
2486        } else {
2487                username($cred, $realm, $may_save, $pool);
2488        }
2489        $cred->password(_read_password("Password for '" .
2490                                       $cred->username . "': ", $realm));
2491        $cred->may_save($may_save);
2492        $SVN::_Core::SVN_NO_ERROR;
2493}
2494
2495sub ssl_server_trust {
2496        my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
2497        $may_save = undef if $_no_auth_cache;
2498        print STDERR "Error validating server certificate for '$realm':\n";
2499        {
2500                no warnings 'once';
2501                # All variables SVN::Auth::SSL::* are used only once,
2502                # so we're shutting up Perl warnings about this.
2503                if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
2504                        print STDERR " - The certificate is not issued ",
2505                            "by a trusted authority. Use the\n",
2506                            "   fingerprint to validate ",
2507                            "the certificate manually!\n";
2508                }
2509                if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
2510                        print STDERR " - The certificate hostname ",
2511                            "does not match.\n";
2512                }
2513                if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
2514                        print STDERR " - The certificate is not yet valid.\n";
2515                }
2516                if ($failures & $SVN::Auth::SSL::EXPIRED) {
2517                        print STDERR " - The certificate has expired.\n";
2518                }
2519                if ($failures & $SVN::Auth::SSL::OTHER) {
2520                        print STDERR " - The certificate has ",
2521                            "an unknown error.\n";
2522                }
2523        } # no warnings 'once'
2524        printf STDERR
2525                "Certificate information:\n".
2526                " - Hostname: %s\n".
2527                " - Valid: from %s until %s\n".
2528                " - Issuer: %s\n".
2529                " - Fingerprint: %s\n",
2530                map $cert_info->$_, qw(hostname valid_from valid_until
2531                                       issuer_dname fingerprint);
2532        my $choice;
2533prompt:
2534        print STDERR $may_save ?
2535              "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
2536              "(R)eject or accept (t)emporarily? ";
2537        STDERR->flush;
2538        $choice = lc(substr(<STDIN> || 'R', 0, 1));
2539        if ($choice =~ /^t$/i) {
2540                $cred->may_save(undef);
2541        } elsif ($choice =~ /^r$/i) {
2542                return -1;
2543        } elsif ($may_save && $choice =~ /^p$/i) {
2544                $cred->may_save($may_save);
2545        } else {
2546                goto prompt;
2547        }
2548        $cred->accepted_failures($failures);
2549        $SVN::_Core::SVN_NO_ERROR;
2550}
2551
2552sub ssl_client_cert {
2553        my ($cred, $realm, $may_save, $pool) = @_;
2554        $may_save = undef if $_no_auth_cache;
2555        print STDERR "Client certificate filename: ";
2556        STDERR->flush;
2557        chomp(my $filename = <STDIN>);
2558        $cred->cert_file($filename);
2559        $cred->may_save($may_save);
2560        $SVN::_Core::SVN_NO_ERROR;
2561}
2562
2563sub ssl_client_cert_pw {
2564        my ($cred, $realm, $may_save, $pool) = @_;
2565        $may_save = undef if $_no_auth_cache;
2566        $cred->password(_read_password("Password: ", $realm));
2567        $cred->may_save($may_save);
2568        $SVN::_Core::SVN_NO_ERROR;
2569}
2570
2571sub username {
2572        my ($cred, $realm, $may_save, $pool) = @_;
2573        $may_save = undef if $_no_auth_cache;
2574        if (defined $realm && length $realm) {
2575                print STDERR "Authentication realm: $realm\n";
2576        }
2577        my $username;
2578        if (defined $_username) {
2579                $username = $_username;
2580        } else {
2581                print STDERR "Username: ";
2582                STDERR->flush;
2583                chomp($username = <STDIN>);
2584        }
2585        $cred->username($username);
2586        $cred->may_save($may_save);
2587        $SVN::_Core::SVN_NO_ERROR;
2588}
2589
2590sub _read_password {
2591        my ($prompt, $realm) = @_;
2592        print STDERR $prompt;
2593        STDERR->flush;
2594        require Term::ReadKey;
2595        Term::ReadKey::ReadMode('noecho');
2596        my $password = '';
2597        while (defined(my $key = Term::ReadKey::ReadKey(0))) {
2598                last if $key =~ /[\012\015]/; # \n\r
2599                $password .= $key;
2600        }
2601        Term::ReadKey::ReadMode('restore');
2602        print STDERR "\n";
2603        STDERR->flush;
2604        $password;
2605}
2606
2607package SVN::Git::Fetcher;
2608use vars qw/@ISA/;
2609use strict;
2610use warnings;
2611use Carp qw/croak/;
2612use IO::File qw//;
2613use Digest::MD5;
2614
2615# file baton members: path, mode_a, mode_b, pool, fh, blob, base
2616sub new {
2617        my ($class, $git_svn) = @_;
2618        my $self = SVN::Delta::Editor->new;
2619        bless $self, $class;
2620        $self->{c} = $git_svn->{last_commit} if exists $git_svn->{last_commit};
2621        $self->{empty} = {};
2622        $self->{dir_prop} = {};
2623        $self->{file_prop} = {};
2624        $self->{absent_dir} = {};
2625        $self->{absent_file} = {};
2626        $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
2627        $self;
2628}
2629
2630sub set_path_strip {
2631        my ($self, $path) = @_;
2632        $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
2633}
2634
2635sub open_root {
2636        { path => '' };
2637}
2638
2639sub open_directory {
2640        my ($self, $path, $pb, $rev) = @_;
2641        { path => $path };
2642}
2643
2644sub git_path {
2645        my ($self, $path) = @_;
2646        if ($self->{path_strip}) {
2647                $path =~ s!$self->{path_strip}!! or
2648                  die "Failed to strip path '$path' ($self->{path_strip})\n";
2649        }
2650        $path;
2651}
2652
2653sub delete_entry {
2654        my ($self, $path, $rev, $pb) = @_;
2655
2656        my $gpath = $self->git_path($path);
2657        return undef if ($gpath eq '');
2658
2659        # remove entire directories.
2660        if (command('ls-tree', $self->{c}, '--', $gpath) =~ /^040000 tree/) {
2661                my ($ls, $ctx) = command_output_pipe(qw/ls-tree
2662                                                     -r --name-only -z/,
2663                                                     $self->{c}, '--', $gpath);
2664                local $/ = "\0";
2665                while (<$ls>) {
2666                        chomp;
2667                        $self->{gii}->remove($_);
2668                        print "\tD\t$_\n" unless $::_q;
2669                }
2670                print "\tD\t$gpath/\n" unless $::_q;
2671                command_close_pipe($ls, $ctx);
2672                $self->{empty}->{$path} = 0
2673        } else {
2674                $self->{gii}->remove($gpath);
2675                print "\tD\t$gpath\n" unless $::_q;
2676        }
2677        undef;
2678}
2679
2680sub open_file {
2681        my ($self, $path, $pb, $rev) = @_;
2682        my $gpath = $self->git_path($path);
2683        my ($mode, $blob) = (command('ls-tree', $self->{c}, '--', $gpath)
2684                             =~ /^(\d{6}) blob ([a-f\d]{40})\t/);
2685        unless (defined $mode && defined $blob) {
2686                die "$path was not found in commit $self->{c} (r$rev)\n";
2687        }
2688        { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
2689          pool => SVN::Pool->new, action => 'M' };
2690}
2691
2692sub add_file {
2693        my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
2694        my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
2695        delete $self->{empty}->{$dir};
2696        { path => $path, mode_a => 100644, mode_b => 100644,
2697          pool => SVN::Pool->new, action => 'A' };
2698}
2699
2700sub add_directory {
2701        my ($self, $path, $cp_path, $cp_rev) = @_;
2702        my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
2703        delete $self->{empty}->{$dir};
2704        $self->{empty}->{$path} = 1;
2705        { path => $path };
2706}
2707
2708sub change_dir_prop {
2709        my ($self, $db, $prop, $value) = @_;
2710        $self->{dir_prop}->{$db->{path}} ||= {};
2711        $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
2712        undef;
2713}
2714
2715sub absent_directory {
2716        my ($self, $path, $pb) = @_;
2717        $self->{absent_dir}->{$pb->{path}} ||= [];
2718        push @{$self->{absent_dir}->{$pb->{path}}}, $path;
2719        undef;
2720}
2721
2722sub absent_file {
2723        my ($self, $path, $pb) = @_;
2724        $self->{absent_file}->{$pb->{path}} ||= [];
2725        push @{$self->{absent_file}->{$pb->{path}}}, $path;
2726        undef;
2727}
2728
2729sub change_file_prop {
2730        my ($self, $fb, $prop, $value) = @_;
2731        if ($prop eq 'svn:executable') {
2732                if ($fb->{mode_b} != 120000) {
2733                        $fb->{mode_b} = defined $value ? 100755 : 100644;
2734                }
2735        } elsif ($prop eq 'svn:special') {
2736                $fb->{mode_b} = defined $value ? 120000 : 100644;
2737        } else {
2738                $self->{file_prop}->{$fb->{path}} ||= {};
2739                $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
2740        }
2741        undef;
2742}
2743
2744sub apply_textdelta {
2745        my ($self, $fb, $exp) = @_;
2746        my $fh = IO::File->new_tmpfile;
2747        $fh->autoflush(1);
2748        # $fh gets auto-closed() by SVN::TxDelta::apply(),
2749        # (but $base does not,) so dup() it for reading in close_file
2750        open my $dup, '<&', $fh or croak $!;
2751        my $base = IO::File->new_tmpfile;
2752        $base->autoflush(1);
2753        if ($fb->{blob}) {
2754                defined (my $pid = fork) or croak $!;
2755                if (!$pid) {
2756                        open STDOUT, '>&', $base or croak $!;
2757                        print STDOUT 'link ' if ($fb->{mode_a} == 120000);
2758                        exec qw/git-cat-file blob/, $fb->{blob} or croak $!;
2759                }
2760                waitpid $pid, 0;
2761                croak $? if $?;
2762
2763                if (defined $exp) {
2764                        seek $base, 0, 0 or croak $!;
2765                        my $md5 = Digest::MD5->new;
2766                        $md5->addfile($base);
2767                        my $got = $md5->hexdigest;
2768                        die "Checksum mismatch: $fb->{path} $fb->{blob}\n",
2769                            "expected: $exp\n",
2770                            "     got: $got\n" if ($got ne $exp);
2771                }
2772        }
2773        seek $base, 0, 0 or croak $!;
2774        $fb->{fh} = $dup;
2775        $fb->{base} = $base;
2776        [ SVN::TxDelta::apply($base, $fh, undef, $fb->{path}, $fb->{pool}) ];
2777}
2778
2779sub close_file {
2780        my ($self, $fb, $exp) = @_;
2781        my $hash;
2782        my $path = $self->git_path($fb->{path});
2783        if (my $fh = $fb->{fh}) {
2784                if (defined $exp) {
2785                        seek($fh, 0, 0) or croak $!;
2786                        my $md5 = Digest::MD5->new;
2787                        $md5->addfile($fh);
2788                        my $got = $md5->hexdigest;
2789                        if ($got ne $exp) {
2790                                die "Checksum mismatch: $path\n",
2791                                    "expected: $exp\n    got: $got\n";
2792                        }
2793                }
2794                sysseek($fh, 0, 0) or croak $!;
2795                if ($fb->{mode_b} == 120000) {
2796                        sysread($fh, my $buf, 5) == 5 or croak $!;
2797                        $buf eq 'link ' or die "$path has mode 120000",
2798                                               "but is not a link\n";
2799                }
2800                defined(my $pid = open my $out,'-|') or die "Can't fork: $!\n";
2801                if (!$pid) {
2802                        open STDIN, '<&', $fh or croak $!;
2803                        exec qw/git-hash-object -w --stdin/ or croak $!;
2804                }
2805                chomp($hash = do { local $/; <$out> });
2806                close $out or croak $!;
2807                close $fh or croak $!;
2808                $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
2809                close $fb->{base} or croak $!;
2810        } else {
2811                $hash = $fb->{blob} or die "no blob information\n";
2812        }
2813        $fb->{pool}->clear;
2814        $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
2815        print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
2816        undef;
2817}
2818
2819sub abort_edit {
2820        my $self = shift;
2821        $self->{nr} = $self->{gii}->{nr};
2822        delete $self->{gii};
2823        $self->SUPER::abort_edit(@_);
2824}
2825
2826sub close_edit {
2827        my $self = shift;
2828        $self->{git_commit_ok} = 1;
2829        $self->{nr} = $self->{gii}->{nr};
2830        delete $self->{gii};
2831        $self->SUPER::close_edit(@_);
2832}
2833
2834package SVN::Git::Editor;
2835use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
2836use strict;
2837use warnings;
2838use Carp qw/croak/;
2839use IO::File;
2840use Digest::MD5;
2841
2842sub new {
2843        my ($class, $opts) = @_;
2844        foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
2845                die "$_ required!\n" unless (defined $opts->{$_});
2846        }
2847
2848        my $pool = SVN::Pool->new;
2849        my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
2850        my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
2851                                     $opts->{r}, $mods);
2852
2853        # $opts->{ra} functions should not be used after this:
2854        my @ce  = $opts->{ra}->get_commit_editor($opts->{log},
2855                                                $opts->{editor_cb}, $pool);
2856        my $self = SVN::Delta::Editor->new(@ce, $pool);
2857        bless $self, $class;
2858        foreach (qw/svn_path r tree_a tree_b/) {
2859                $self->{$_} = $opts->{$_};
2860        }
2861        $self->{url} = $opts->{ra}->{url};
2862        $self->{mods} = $mods;
2863        $self->{types} = $types;
2864        $self->{pool} = $pool;
2865        $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
2866        $self->{rm} = { };
2867        $self->{path_prefix} = length $self->{svn_path} ?
2868                               "$self->{svn_path}/" : '';
2869        return $self;
2870}
2871
2872sub generate_diff {
2873        my ($tree_a, $tree_b) = @_;
2874        my @diff_tree = qw(diff-tree -z -r);
2875        if ($_cp_similarity) {
2876                push @diff_tree, "-C$_cp_similarity";
2877        } else {
2878                push @diff_tree, '-C';
2879        }
2880        push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
2881        push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
2882        push @diff_tree, $tree_a, $tree_b;
2883        my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
2884        local $/ = "\0";
2885        my $state = 'meta';
2886        my @mods;
2887        while (<$diff_fh>) {
2888                chomp $_; # this gets rid of the trailing "\0"
2889                if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
2890                                        $::sha1\s($::sha1)\s
2891                                        ([MTCRAD])\d*$/xo) {
2892                        push @mods, {   mode_a => $1, mode_b => $2,
2893                                        sha1_b => $3, chg => $4 };
2894                        if ($4 =~ /^(?:C|R)$/) {
2895                                $state = 'file_a';
2896                        } else {
2897                                $state = 'file_b';
2898                        }
2899                } elsif ($state eq 'file_a') {
2900                        my $x = $mods[$#mods] or croak "Empty array\n";
2901                        if ($x->{chg} !~ /^(?:C|R)$/) {
2902                                croak "Error parsing $_, $x->{chg}\n";
2903                        }
2904                        $x->{file_a} = $_;
2905                        $state = 'file_b';
2906                } elsif ($state eq 'file_b') {
2907                        my $x = $mods[$#mods] or croak "Empty array\n";
2908                        if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
2909                                croak "Error parsing $_, $x->{chg}\n";
2910                        }
2911                        if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
2912                                croak "Error parsing $_, $x->{chg}\n";
2913                        }
2914                        $x->{file_b} = $_;
2915                        $state = 'meta';
2916                } else {
2917                        croak "Error parsing $_\n";
2918                }
2919        }
2920        command_close_pipe($diff_fh, $ctx);
2921        \@mods;
2922}
2923
2924sub check_diff_paths {
2925        my ($ra, $pfx, $rev, $mods) = @_;
2926        my %types;
2927        $pfx .= '/' if length $pfx;
2928
2929        sub type_diff_paths {
2930                my ($ra, $types, $path, $rev) = @_;
2931                my @p = split m#/+#, $path;
2932                my $c = shift @p;
2933                unless (defined $types->{$c}) {
2934                        $types->{$c} = $ra->check_path($c, $rev);
2935                }
2936                while (@p) {
2937                        $c .= '/' . shift @p;
2938                        next if defined $types->{$c};
2939                        $types->{$c} = $ra->check_path($c, $rev);
2940                }
2941        }
2942
2943        foreach my $m (@$mods) {
2944                foreach my $f (qw/file_a file_b/) {
2945                        next unless defined $m->{$f};
2946                        my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
2947                        if (length $pfx.$dir && ! defined $types{$dir}) {
2948                                type_diff_paths($ra, \%types, $pfx.$dir, $rev);
2949                        }
2950                }
2951        }
2952        \%types;
2953}
2954
2955sub split_path {
2956        return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
2957}
2958
2959sub repo_path {
2960        my ($self, $path) = @_;
2961        $self->{path_prefix}.(defined $path ? $path : '');
2962}
2963
2964sub url_path {
2965        my ($self, $path) = @_;
2966        if ($self->{url} =~ m#^https?://#) {
2967                $path =~ s/([^a-zA-Z0-9_.-])/uc sprintf("%%%02x",ord($1))/eg;
2968        }
2969        $self->{url} . '/' . $self->repo_path($path);
2970}
2971
2972sub rmdirs {
2973        my ($self) = @_;
2974        my $rm = $self->{rm};
2975        delete $rm->{''}; # we never delete the url we're tracking
2976        return unless %$rm;
2977
2978        foreach (keys %$rm) {
2979                my @d = split m#/#, $_;
2980                my $c = shift @d;
2981                $rm->{$c} = 1;
2982                while (@d) {
2983                        $c .= '/' . shift @d;
2984                        $rm->{$c} = 1;
2985                }
2986        }
2987        delete $rm->{$self->{svn_path}};
2988        delete $rm->{''}; # we never delete the url we're tracking
2989        return unless %$rm;
2990
2991        my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
2992                                             $self->{tree_b});
2993        local $/ = "\0";
2994        while (<$fh>) {
2995                chomp;
2996                my @dn = split m#/#, $_;
2997                while (pop @dn) {
2998                        delete $rm->{join '/', @dn};
2999                }
3000                unless (%$rm) {
3001                        close $fh;
3002                        return;
3003                }
3004        }
3005        command_close_pipe($fh, $ctx);
3006
3007        my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
3008        foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
3009                $self->close_directory($bat->{$d}, $p);
3010                my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
3011                print "\tD+\t$d/\n" unless $::_q;
3012                $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
3013                delete $bat->{$d};
3014        }
3015}
3016
3017sub open_or_add_dir {
3018        my ($self, $full_path, $baton) = @_;
3019        my $t = $self->{types}->{$full_path};
3020        if (!defined $t) {
3021                die "$full_path not known in r$self->{r} or we have a bug!\n";
3022        }
3023        {
3024                no warnings 'once';
3025                # SVN::Node::none and SVN::Node::file are used only once,
3026                # so we're shutting up Perl's warnings about them.
3027                if ($t == $SVN::Node::none) {
3028                        return $self->add_directory($full_path, $baton,
3029                            undef, -1, $self->{pool});
3030                } elsif ($t == $SVN::Node::dir) {
3031                        return $self->open_directory($full_path, $baton,
3032                            $self->{r}, $self->{pool});
3033                } # no warnings 'once'
3034                print STDERR "$full_path already exists in repository at ",
3035                    "r$self->{r} and it is not a directory (",
3036                    ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
3037        } # no warnings 'once'
3038        exit 1;
3039}
3040
3041sub ensure_path {
3042        my ($self, $path) = @_;
3043        my $bat = $self->{bat};
3044        my $repo_path = $self->repo_path($path);
3045        return $bat->{''} unless (length $repo_path);
3046        my @p = split m#/+#, $repo_path;
3047        my $c = shift @p;
3048        $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
3049        while (@p) {
3050                my $c0 = $c;
3051                $c .= '/' . shift @p;
3052                $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
3053        }
3054        return $bat->{$c};
3055}
3056
3057sub A {
3058        my ($self, $m) = @_;
3059        my ($dir, $file) = split_path($m->{file_b});
3060        my $pbat = $self->ensure_path($dir);
3061        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3062                                        undef, -1);
3063        print "\tA\t$m->{file_b}\n" unless $::_q;
3064        $self->chg_file($fbat, $m);
3065        $self->close_file($fbat,undef,$self->{pool});
3066}
3067
3068sub C {
3069        my ($self, $m) = @_;
3070        my ($dir, $file) = split_path($m->{file_b});
3071        my $pbat = $self->ensure_path($dir);
3072        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3073                                $self->url_path($m->{file_a}), $self->{r});
3074        print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3075        $self->chg_file($fbat, $m);
3076        $self->close_file($fbat,undef,$self->{pool});
3077}
3078
3079sub delete_entry {
3080        my ($self, $path, $pbat) = @_;
3081        my $rpath = $self->repo_path($path);
3082        my ($dir, $file) = split_path($rpath);
3083        $self->{rm}->{$dir} = 1;
3084        $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
3085}
3086
3087sub R {
3088        my ($self, $m) = @_;
3089        my ($dir, $file) = split_path($m->{file_b});
3090        my $pbat = $self->ensure_path($dir);
3091        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3092                                $self->url_path($m->{file_a}), $self->{r});
3093        print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3094        $self->chg_file($fbat, $m);
3095        $self->close_file($fbat,undef,$self->{pool});
3096
3097        ($dir, $file) = split_path($m->{file_a});
3098        $pbat = $self->ensure_path($dir);
3099        $self->delete_entry($m->{file_a}, $pbat);
3100}
3101
3102sub M {
3103        my ($self, $m) = @_;
3104        my ($dir, $file) = split_path($m->{file_b});
3105        my $pbat = $self->ensure_path($dir);
3106        my $fbat = $self->open_file($self->repo_path($m->{file_b}),
3107                                $pbat,$self->{r},$self->{pool});
3108        print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
3109        $self->chg_file($fbat, $m);
3110        $self->close_file($fbat,undef,$self->{pool});
3111}
3112
3113sub T { shift->M(@_) }
3114
3115sub change_file_prop {
3116        my ($self, $fbat, $pname, $pval) = @_;
3117        $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
3118}
3119
3120sub chg_file {
3121        my ($self, $fbat, $m) = @_;
3122        if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
3123                $self->change_file_prop($fbat,'svn:executable','*');
3124        } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
3125                $self->change_file_prop($fbat,'svn:executable',undef);
3126        }
3127        my $fh = IO::File->new_tmpfile or croak $!;
3128        if ($m->{mode_b} =~ /^120/) {
3129                print $fh 'link ' or croak $!;
3130                $self->change_file_prop($fbat,'svn:special','*');
3131        } elsif ($m->{mode_a} =~ /^120/ && $m->{mode_b} !~ /^120/) {
3132                $self->change_file_prop($fbat,'svn:special',undef);
3133        }
3134        defined(my $pid = fork) or croak $!;
3135        if (!$pid) {
3136                open STDOUT, '>&', $fh or croak $!;
3137                exec qw/git-cat-file blob/, $m->{sha1_b} or croak $!;
3138        }
3139        waitpid $pid, 0;
3140        croak $? if $?;
3141        $fh->flush == 0 or croak $!;
3142        seek $fh, 0, 0 or croak $!;
3143
3144        my $md5 = Digest::MD5->new;
3145        $md5->addfile($fh) or croak $!;
3146        seek $fh, 0, 0 or croak $!;
3147
3148        my $exp = $md5->hexdigest;
3149        my $pool = SVN::Pool->new;
3150        my $atd = $self->apply_textdelta($fbat, undef, $pool);
3151        my $got = SVN::TxDelta::send_stream($fh, @$atd, $pool);
3152        die "Checksum mismatch\nexpected: $exp\ngot: $got\n" if ($got ne $exp);
3153        $pool->clear;
3154
3155        close $fh or croak $!;
3156}
3157
3158sub D {
3159        my ($self, $m) = @_;
3160        my ($dir, $file) = split_path($m->{file_b});
3161        my $pbat = $self->ensure_path($dir);
3162        print "\tD\t$m->{file_b}\n" unless $::_q;
3163        $self->delete_entry($m->{file_b}, $pbat);
3164}
3165
3166sub close_edit {
3167        my ($self) = @_;
3168        my ($p,$bat) = ($self->{pool}, $self->{bat});
3169        foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
3170                next if $_ eq '';
3171                $self->close_directory($bat->{$_}, $p);
3172        }
3173        $self->close_directory($bat->{''}, $p);
3174        $self->SUPER::close_edit($p);
3175        $p->clear;
3176}
3177
3178sub abort_edit {
3179        my ($self) = @_;
3180        $self->SUPER::abort_edit($self->{pool});
3181}
3182
3183sub DESTROY {
3184        my $self = shift;
3185        $self->SUPER::DESTROY(@_);
3186        $self->{pool}->clear;
3187}
3188
3189# this drives the editor
3190sub apply_diff {
3191        my ($self) = @_;
3192        my $mods = $self->{mods};
3193        my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
3194        foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
3195                my $f = $m->{chg};
3196                if (defined $o{$f}) {
3197                        $self->$f($m);
3198                } else {
3199                        fatal("Invalid change type: $f");
3200                }
3201        }
3202        $self->rmdirs if $_rmdir;
3203        if (@$mods == 0) {
3204                $self->abort_edit;
3205        } else {
3206                $self->close_edit;
3207        }
3208        return scalar @$mods;
3209}
3210
3211package Git::SVN::Ra;
3212use vars qw/@ISA $config_dir $_log_window_size/;
3213use strict;
3214use warnings;
3215my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
3216
3217BEGIN {
3218        # enforce temporary pool usage for some simple functions
3219        no strict 'refs';
3220        for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root/) {
3221                my $SUPER = "SUPER::$f";
3222                *$f = sub {
3223                        my $self = shift;
3224                        my $pool = SVN::Pool->new;
3225                        my @ret = $self->$SUPER(@_,$pool);
3226                        $pool->clear;
3227                        wantarray ? @ret : $ret[0];
3228                };
3229        }
3230}
3231
3232sub _auth_providers () {
3233        [
3234          SVN::Client::get_simple_provider(),
3235          SVN::Client::get_ssl_server_trust_file_provider(),
3236          SVN::Client::get_simple_prompt_provider(
3237            \&Git::SVN::Prompt::simple, 2),
3238          SVN::Client::get_ssl_client_cert_file_provider(),
3239          SVN::Client::get_ssl_client_cert_prompt_provider(
3240            \&Git::SVN::Prompt::ssl_client_cert, 2),
3241          SVN::Client::get_ssl_client_cert_pw_prompt_provider(
3242            \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
3243          SVN::Client::get_username_provider(),
3244          SVN::Client::get_ssl_server_trust_prompt_provider(
3245            \&Git::SVN::Prompt::ssl_server_trust),
3246          SVN::Client::get_username_prompt_provider(
3247            \&Git::SVN::Prompt::username, 2)
3248        ]
3249}
3250
3251sub escape_uri_only {
3252        my ($uri) = @_;
3253        my @tmp;
3254        foreach (split m{/}, $uri) {
3255                s/([^\w.-])/sprintf("%%%02X",ord($1))/eg;
3256                push @tmp, $_;
3257        }
3258        join('/', @tmp);
3259}
3260
3261sub escape_url {
3262        my ($url) = @_;
3263        if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
3264                my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
3265                $url = "$scheme://$domain$uri";
3266        }
3267        $url;
3268}
3269
3270sub new {
3271        my ($class, $url) = @_;
3272        $url =~ s!/+$!!;
3273        return $RA if ($RA && $RA->{url} eq $url);
3274
3275        SVN::_Core::svn_config_ensure($config_dir, undef);
3276        my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
3277        my $config = SVN::Core::config_get_config($config_dir);
3278        $RA = undef;
3279        my $dont_store_passwords = 1;
3280        my $conf_t = ${$config}{'config'};
3281        {
3282                no warnings 'once';
3283                # The usage of $SVN::_Core::SVN_CONFIG_* variables
3284                # produces warnings that variables are used only once.
3285                # I had not found the better way to shut them up, so
3286                # the warnings of type 'once' are disabled in this block.
3287                if (SVN::_Core::svn_config_get_bool($conf_t,
3288                    $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3289                    $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
3290                    1) == 0) {
3291                        SVN::_Core::svn_auth_set_parameter($baton,
3292                            $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
3293                            bless (\$dont_store_passwords, "_p_void"));
3294                }
3295                if (SVN::_Core::svn_config_get_bool($conf_t,
3296                    $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
3297                    $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
3298                    1) == 0) {
3299                        $Git::SVN::Prompt::_no_auth_cache = 1;
3300                }
3301        } # no warnings 'once'
3302        my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
3303                              config => $config,
3304                              pool => SVN::Pool->new,
3305                              auth_provider_callbacks => $callbacks);
3306        $self->{url} = $url;
3307        $self->{svn_path} = $url;
3308        $self->{repos_root} = $self->get_repos_root;
3309        $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
3310        $self->{cache} = { check_path => { r => 0, data => {} },
3311                           get_dir => { r => 0, data => {} } };
3312        $RA = bless $self, $class;
3313}
3314
3315sub check_path {
3316        my ($self, $path, $r) = @_;
3317        my $cache = $self->{cache}->{check_path};
3318        if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
3319                return $cache->{data}->{$path};
3320        }
3321        my $pool = SVN::Pool->new;
3322        my $t = $self->SUPER::check_path($path, $r, $pool);
3323        $pool->clear;
3324        if ($r != $cache->{r}) {
3325                %{$cache->{data}} = ();
3326                $cache->{r} = $r;
3327        }
3328        $cache->{data}->{$path} = $t;
3329}
3330
3331sub get_dir {
3332        my ($self, $dir, $r) = @_;
3333        my $cache = $self->{cache}->{get_dir};
3334        if ($r == $cache->{r}) {
3335                if (my $x = $cache->{data}->{$dir}) {
3336                        return wantarray ? @$x : $x->[0];
3337                }
3338        }
3339        my $pool = SVN::Pool->new;
3340        my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
3341        my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
3342        $pool->clear;
3343        if ($r != $cache->{r}) {
3344                %{$cache->{data}} = ();
3345                $cache->{r} = $r;
3346        }
3347        $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
3348        wantarray ? (\%dirents, $r, $props) : \%dirents;
3349}
3350
3351sub DESTROY {
3352        # do not call the real DESTROY since we store ourselves in $RA
3353}
3354
3355sub get_log {
3356        my ($self, @args) = @_;
3357        my $pool = SVN::Pool->new;
3358        splice(@args, 3, 1) if ($SVN::Core::VERSION le '1.2.0');
3359        my $ret = $self->SUPER::get_log(@args, $pool);
3360        $pool->clear;
3361        $ret;
3362}
3363
3364sub trees_match {
3365        my ($self, $url1, $rev1, $url2, $rev2) = @_;
3366        my $ctx = SVN::Client->new(auth => _auth_providers);
3367        my $out = IO::File->new_tmpfile;
3368
3369        # older SVN (1.1.x) doesn't take $pool as the last parameter for
3370        # $ctx->diff(), so we'll create a default one
3371        my $pool = SVN::Pool->new_default_sub;
3372
3373        $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
3374        $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
3375        $out->flush;
3376        my $ret = (($out->stat)[7] == 0);
3377        close $out or croak $!;
3378
3379        $ret;
3380}
3381
3382sub get_commit_editor {
3383        my ($self, $log, $cb, $pool) = @_;
3384        my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
3385        $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
3386}
3387
3388sub gs_do_update {
3389        my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
3390        my $new = ($rev_a == $rev_b);
3391        my $path = $gs->{path};
3392
3393        if ($new && -e $gs->{index}) {
3394                unlink $gs->{index} or die
3395                  "Couldn't unlink index: $gs->{index}: $!\n";
3396        }
3397        my $pool = SVN::Pool->new;
3398        $editor->set_path_strip($path);
3399        my (@pc) = split m#/#, $path;
3400        my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
3401                                        1, $editor, $pool);
3402        my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3403
3404        # Since we can't rely on svn_ra_reparent being available, we'll
3405        # just have to do some magic with set_path to make it so
3406        # we only want a partial path.
3407        my $sp = '';
3408        my $final = join('/', @pc);
3409        while (@pc) {
3410                $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
3411                $sp .= '/' if length $sp;
3412                $sp .= shift @pc;
3413        }
3414        die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
3415
3416        $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
3417
3418        $reporter->finish_report($pool);
3419        $pool->clear;
3420        $editor->{git_commit_ok};
3421}
3422
3423# this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
3424# svn_ra_reparent didn't work before 1.4)
3425sub gs_do_switch {
3426        my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
3427        my $path = $gs->{path};
3428        my $pool = SVN::Pool->new;
3429
3430        my $full_url = $self->{url};
3431        my $old_url = $full_url;
3432        $full_url .= '/' . escape_uri_only($path) if length $path;
3433        my ($ra, $reparented);
3434        if ($old_url ne $full_url) {
3435                if ($old_url !~ m#^svn(\+ssh)?://#) {
3436                        SVN::_Ra::svn_ra_reparent($self->{session}, $full_url,
3437                                                  $pool);
3438                        $self->{url} = $full_url;
3439                        $reparented = 1;
3440                } else {
3441                        $_[0] = undef;
3442                        $self = undef;
3443                        $RA = undef;
3444                        $ra = Git::SVN::Ra->new($full_url);
3445                        $ra_invalid = 1;
3446                }
3447        }
3448        $ra ||= $self;
3449        my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
3450        my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
3451        $reporter->set_path('', $rev_a, 0, @lock, $pool);
3452        $reporter->finish_report($pool);
3453
3454        if ($reparented) {
3455                SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
3456                $self->{url} = $old_url;
3457        }
3458
3459        $pool->clear;
3460        $editor->{git_commit_ok};
3461}
3462
3463sub longest_common_path {
3464        my ($gsv, $globs) = @_;
3465        my %common;
3466        my $common_max = scalar @$gsv;
3467
3468        foreach my $gs (@$gsv) {
3469                my @tmp = split m#/#, $gs->{path};
3470                my $p = '';
3471                foreach (@tmp) {
3472                        $p .= length($p) ? "/$_" : $_;
3473                        $common{$p} ||= 0;
3474                        $common{$p}++;
3475                }
3476        }
3477        $globs ||= [];
3478        $common_max += scalar @$globs;
3479        foreach my $glob (@$globs) {
3480                my @tmp = split m#/#, $glob->{path}->{left};
3481                my $p = '';
3482                foreach (@tmp) {
3483                        $p .= length($p) ? "/$_" : $_;
3484                        $common{$p} ||= 0;
3485                        $common{$p}++;
3486                }
3487        }
3488
3489        my $longest_path = '';
3490        foreach (sort {length $b <=> length $a} keys %common) {
3491                if ($common{$_} == $common_max) {
3492                        $longest_path = $_;
3493                        last;
3494                }
3495        }
3496        $longest_path;
3497}
3498
3499sub gs_fetch_loop_common {
3500        my ($self, $base, $head, $gsv, $globs) = @_;
3501        return if ($base > $head);
3502        my $inc = $_log_window_size;
3503        my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
3504        my $longest_path = longest_common_path($gsv, $globs);
3505        my $ra_url = $self->{url};
3506        while (1) {
3507                my %revs;
3508                my $err;
3509                my $err_handler = $SVN::Error::handler;
3510                $SVN::Error::handler = sub {
3511                        ($err) = @_;
3512                        skip_unknown_revs($err);
3513                };
3514                sub _cb {
3515                        my ($paths, $r, $author, $date, $log) = @_;
3516                        [ dup_changed_paths($paths),
3517                          { author => $author, date => $date, log => $log } ];
3518                }
3519                $self->get_log([$longest_path], $min, $max, 0, 1, 1,
3520                               sub { $revs{$_[1]} = _cb(@_) });
3521                if ($err && $max >= $head) {
3522                        print STDERR "Path '$longest_path' ",
3523                                     "was probably deleted:\n",
3524                                     $err->expanded_message,
3525                                     "\nWill attempt to follow ",
3526                                     "revisions r$min .. r$max ",
3527                                     "committed before the deletion\n";
3528                        my $hi = $max;
3529                        while (--$hi >= $min) {
3530                                my $ok;
3531                                $self->get_log([$longest_path], $min, $hi,
3532                                               0, 1, 1, sub {
3533                                               $ok ||= $_[1];
3534                                               $revs{$_[1]} = _cb(@_) });
3535                                if ($ok) {
3536                                        print STDERR "r$min .. r$ok OK\n";
3537                                        last;
3538                                }
3539                        }
3540                }
3541                $SVN::Error::handler = $err_handler;
3542
3543                my %exists = map { $_->{path} => $_ } @$gsv;
3544                foreach my $r (sort {$a <=> $b} keys %revs) {
3545                        my ($paths, $logged) = @{$revs{$r}};
3546
3547                        foreach my $gs ($self->match_globs(\%exists, $paths,
3548                                                           $globs, $r)) {
3549                                if ($gs->rev_db_max >= $r) {
3550                                        next;
3551                                }
3552                                next unless $gs->match_paths($paths, $r);
3553                                $gs->{logged_rev_props} = $logged;
3554                                if (my $last_commit = $gs->last_commit) {
3555                                        $gs->assert_index_clean($last_commit);
3556                                }
3557                                my $log_entry = $gs->do_fetch($paths, $r);
3558                                if ($log_entry) {
3559                                        $gs->do_git_commit($log_entry);
3560                                }
3561                        }
3562                        foreach my $g (@$globs) {
3563                                my $k = "svn-remote.$g->{remote}." .
3564                                        "$g->{t}-maxRev";
3565                                Git::SVN::tmp_config($k, $r);
3566                        }
3567                        if ($ra_invalid) {
3568                                $_[0] = undef;
3569                                $self = undef;
3570                                $RA = undef;
3571                                $self = Git::SVN::Ra->new($ra_url);
3572                                $ra_invalid = undef;
3573                        }
3574                }
3575                # pre-fill the .rev_db since it'll eventually get filled in
3576                # with '0' x40 if something new gets committed
3577                foreach my $gs (@$gsv) {
3578                        next if defined $gs->rev_db_get($max);
3579                        $gs->rev_db_set($max, 0 x40);
3580                }
3581                foreach my $g (@$globs) {
3582                        my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
3583                        Git::SVN::tmp_config($k, $max);
3584                }
3585                last if $max >= $head;
3586                $min = $max + 1;
3587                $max += $inc;
3588                $max = $head if ($max > $head);
3589        }
3590}
3591
3592sub match_globs {
3593        my ($self, $exists, $paths, $globs, $r) = @_;
3594
3595        sub get_dir_check {
3596                my ($self, $exists, $g, $r) = @_;
3597                my @x = eval { $self->get_dir($g->{path}->{left}, $r) };
3598                return unless scalar @x == 3;
3599                my $dirents = $x[0];
3600                foreach my $de (keys %$dirents) {
3601                        next if $dirents->{$de}->{kind} != $SVN::Node::dir;
3602                        my $p = $g->{path}->full_path($de);
3603                        next if $exists->{$p};
3604                        next if (length $g->{path}->{right} &&
3605                                 ($self->check_path($p, $r) !=
3606                                  $SVN::Node::dir));
3607                        $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
3608                                         $g->{ref}->full_path($de), 1);
3609                }
3610        }
3611        foreach my $g (@$globs) {
3612                if (my $path = $paths->{"/$g->{path}->{left}"}) {
3613                        if ($path->{action} =~ /^[AR]$/) {
3614                                get_dir_check($self, $exists, $g, $r);
3615                        }
3616                }
3617                foreach (keys %$paths) {
3618                        if (/$g->{path}->{left_regex}/ &&
3619                            !/$g->{path}->{regex}/) {
3620                                next if $paths->{$_}->{action} !~ /^[AR]$/;
3621                                get_dir_check($self, $exists, $g, $r);
3622                        }
3623                        next unless /$g->{path}->{regex}/;
3624                        my $p = $1;
3625                        my $pathname = $g->{path}->full_path($p);
3626                        next if $exists->{$pathname};
3627                        next if ($self->check_path($pathname, $r) !=
3628                                 $SVN::Node::dir);
3629                        $exists->{$pathname} = Git::SVN->init(
3630                                              $self->{url}, $pathname, undef,
3631                                              $g->{ref}->full_path($p), 1);
3632                }
3633                my $c = '';
3634                foreach (split m#/#, $g->{path}->{left}) {
3635                        $c .= "/$_";
3636                        next unless ($paths->{$c} &&
3637                                     ($paths->{$c}->{action} =~ /^[AR]$/));
3638                        get_dir_check($self, $exists, $g, $r);
3639                }
3640        }
3641        values %$exists;
3642}
3643
3644sub minimize_url {
3645        my ($self) = @_;
3646        return $self->{url} if ($self->{url} eq $self->{repos_root});
3647        my $url = $self->{repos_root};
3648        my @components = split(m!/!, $self->{svn_path});
3649        my $c = '';
3650        do {
3651                $url .= "/$c" if length $c;
3652                eval { (ref $self)->new($url)->get_latest_revnum };
3653        } while ($@ && ($c = shift @components));
3654        $url;
3655}
3656
3657sub can_do_switch {
3658        my $self = shift;
3659        unless (defined $can_do_switch) {
3660                my $pool = SVN::Pool->new;
3661                my $rep = eval {
3662                        $self->do_switch(1, '', 0, $self->{url},
3663                                         SVN::Delta::Editor->new, $pool);
3664                };
3665                if ($@) {
3666                        $can_do_switch = 0;
3667                } else {
3668                        $rep->abort_report($pool);
3669                        $can_do_switch = 1;
3670                }
3671                $pool->clear;
3672        }
3673        $can_do_switch;
3674}
3675
3676sub skip_unknown_revs {
3677        my ($err) = @_;
3678        my $errno = $err->apr_err();
3679        # Maybe the branch we're tracking didn't
3680        # exist when the repo started, so it's
3681        # not an error if it doesn't, just continue
3682        #
3683        # Wonderfully consistent library, eh?
3684        # 160013 - svn:// and file://
3685        # 175002 - http(s)://
3686        # 175007 - http(s):// (this repo required authorization, too...)
3687        #   More codes may be discovered later...
3688        if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
3689                my $err_key = $err->expanded_message;
3690                # revision numbers change every time, filter them out
3691                $err_key =~ s/\d+/\0/g;
3692                $err_key = "$errno\0$err_key";
3693                unless ($ignored_err{$err_key}) {
3694                        warn "W: Ignoring error from SVN, path probably ",
3695                             "does not exist: ($errno): ",
3696                             $err->expanded_message,"\n";
3697                        $ignored_err{$err_key} = 1;
3698                }
3699                return;
3700        }
3701        die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
3702}
3703
3704# svn_log_changed_path_t objects passed to get_log are likely to be
3705# overwritten even if only the refs are copied to an external variable,
3706# so we should dup the structures in their entirety.  Using an externally
3707# passed pool (instead of our temporary and quickly cleared pool in
3708# Git::SVN::Ra) does not help matters at all...
3709sub dup_changed_paths {
3710        my ($paths) = @_;
3711        return undef unless $paths;
3712        my %ret;
3713        foreach my $p (keys %$paths) {
3714                my $i = $paths->{$p};
3715                my %s = map { $_ => $i->$_ }
3716                              qw/copyfrom_path copyfrom_rev action/;
3717                $ret{$p} = \%s;
3718        }
3719        \%ret;
3720}
3721
3722package Git::SVN::Log;
3723use strict;
3724use warnings;
3725use POSIX qw/strftime/;
3726use constant commit_log_separator => ('-' x 72) . "\n";
3727use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
3728            %rusers $show_commit $incremental/;
3729my $l_fmt;
3730
3731sub cmt_showable {
3732        my ($c) = @_;
3733        return 1 if defined $c->{r};
3734
3735        # big commit message got truncated by the 16k pretty buffer in rev-list
3736        if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
3737                                $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
3738                @{$c->{l}} = ();
3739                my @log = command(qw/cat-file commit/, $c->{c});
3740
3741                # shift off the headers
3742                shift @log while ($log[0] ne '');
3743                shift @log;
3744
3745                # TODO: make $c->{l} not have a trailing newline in the future
3746                @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
3747
3748                (undef, $c->{r}, undef) = ::extract_metadata(
3749                                (grep(/^git-svn-id: /, @log))[-1]);
3750        }
3751        return defined $c->{r};
3752}
3753
3754sub log_use_color {
3755        return 1 if $color;
3756        my ($dc, $dcvar);
3757        $dcvar = 'color.diff';
3758        $dc = `git-config --get $dcvar`;
3759        if ($dc eq '') {
3760                # nothing at all; fallback to "diff.color"
3761                $dcvar = 'diff.color';
3762                $dc = `git-config --get $dcvar`;
3763        }
3764        chomp($dc);
3765        if ($dc eq 'auto') {
3766                my $pc;
3767                $pc = `git-config --get color.pager`;
3768                if ($pc eq '') {
3769                        # does not have it -- fallback to pager.color
3770                        $pc = `git-config --bool --get pager.color`;
3771                }
3772                else {
3773                        $pc = `git-config --bool --get color.pager`;
3774                        if ($?) {
3775                                $pc = 'false';
3776                        }
3777                }
3778                chomp($pc);
3779                if (-t *STDOUT || (defined $pager && $pc eq 'true')) {
3780                        return ($ENV{TERM} && $ENV{TERM} ne 'dumb');
3781                }
3782                return 0;
3783        }
3784        return 0 if $dc eq 'never';
3785        return 1 if $dc eq 'always';
3786        chomp($dc = `git-config --bool --get $dcvar`);
3787        return ($dc eq 'true');
3788}
3789
3790sub git_svn_log_cmd {
3791        my ($r_min, $r_max, @args) = @_;
3792        my $head = 'HEAD';
3793        my (@files, @log_opts);
3794        foreach my $x (@args) {
3795                if ($x eq '--' || @files) {
3796                        push @files, $x;
3797                } else {
3798                        if (::verify_ref("$x^0")) {
3799                                $head = $x;
3800                        } else {
3801                                push @log_opts, $x;
3802                        }
3803                }
3804        }
3805
3806        my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
3807        $gs ||= Git::SVN->_new;
3808        my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
3809                   $gs->refname);
3810        push @cmd, '-r' unless $non_recursive;
3811        push @cmd, qw/--raw --name-status/ if $verbose;
3812        push @cmd, '--color' if log_use_color();
3813        push @cmd, @log_opts;
3814        if (defined $r_max && $r_max == $r_min) {
3815                push @cmd, '--max-count=1';
3816                if (my $c = $gs->rev_db_get($r_max)) {
3817                        push @cmd, $c;
3818                }
3819        } elsif (defined $r_max) {
3820                if ($r_max < $r_min) {
3821                        ($r_min, $r_max) = ($r_max, $r_min);
3822                }
3823                my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
3824                my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
3825                # If there are no commits in the range, both $c_max and $c_min
3826                # will be undefined.  If there is at least 1 commit in the
3827                # range, both will be defined.
3828                return () if !defined $c_min || !defined $c_max;
3829                if ($c_min eq $c_max) {
3830                        push @cmd, '--max-count=1', $c_min;
3831                } else {
3832                        push @cmd, '--boundary', "$c_min..$c_max";
3833                }
3834        }
3835        return (@cmd, @files);
3836}
3837
3838# adapted from pager.c
3839sub config_pager {
3840        $pager ||= $ENV{GIT_PAGER} || $ENV{PAGER};
3841        if (!defined $pager) {
3842                $pager = 'less';
3843        } elsif (length $pager == 0 || $pager eq 'cat') {
3844                $pager = undef;
3845        }
3846}
3847
3848sub run_pager {
3849        return unless -t *STDOUT && defined $pager;
3850        pipe my $rfd, my $wfd or return;
3851        defined(my $pid = fork) or ::fatal "Can't fork: $!";
3852        if (!$pid) {
3853                open STDOUT, '>&', $wfd or
3854                                     ::fatal "Can't redirect to stdout: $!";
3855                return;
3856        }
3857        open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
3858        $ENV{LESS} ||= 'FRSX';
3859        exec $pager or ::fatal "Can't run pager: $! ($pager)";
3860}
3861
3862sub tz_to_s_offset {
3863        my ($tz) = @_;
3864        $tz =~ s/(\d\d)$//;
3865        return ($1 * 60) + ($tz * 3600);
3866}
3867
3868sub get_author_info {
3869        my ($dest, $author, $t, $tz) = @_;
3870        $author =~ s/(?:^\s*|\s*$)//g;
3871        $dest->{a_raw} = $author;
3872        my $au;
3873        if ($::_authors) {
3874                $au = $rusers{$author} || undef;
3875        }
3876        if (!$au) {
3877                ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
3878        }
3879        $dest->{t} = $t;
3880        $dest->{tz} = $tz;
3881        $dest->{a} = $au;
3882        # Date::Parse isn't in the standard Perl distro :(
3883        if ($tz =~ s/^\+//) {
3884                $t += tz_to_s_offset($tz);
3885        } elsif ($tz =~ s/^\-//) {
3886                $t -= tz_to_s_offset($tz);
3887        }
3888        $dest->{t_utc} = $t;
3889}
3890
3891sub process_commit {
3892        my ($c, $r_min, $r_max, $defer) = @_;
3893        if (defined $r_min && defined $r_max) {
3894                if ($r_min == $c->{r} && $r_min == $r_max) {
3895                        show_commit($c);
3896                        return 0;
3897                }
3898                return 1 if $r_min == $r_max;
3899                if ($r_min < $r_max) {
3900                        # we need to reverse the print order
3901                        return 0 if (defined $limit && --$limit < 0);
3902                        push @$defer, $c;
3903                        return 1;
3904                }
3905                if ($r_min != $r_max) {
3906                        return 1 if ($r_min < $c->{r});
3907                        return 1 if ($r_max > $c->{r});
3908                }
3909        }
3910        return 0 if (defined $limit && --$limit < 0);
3911        show_commit($c);
3912        return 1;
3913}
3914
3915sub show_commit {
3916        my $c = shift;
3917        if ($oneline) {
3918                my $x = "\n";
3919                if (my $l = $c->{l}) {
3920                        while ($l->[0] =~ /^\s*$/) { shift @$l }
3921                        $x = $l->[0];
3922                }
3923                $l_fmt ||= 'A' . length($c->{r});
3924                print 'r',pack($l_fmt, $c->{r}),' | ';
3925                print "$c->{c} | " if $show_commit;
3926                print $x;
3927        } else {
3928                show_commit_normal($c);
3929        }
3930}
3931
3932sub show_commit_changed_paths {
3933        my ($c) = @_;
3934        return unless $c->{changed};
3935        print "Changed paths:\n", @{$c->{changed}};
3936}
3937
3938sub show_commit_normal {
3939        my ($c) = @_;
3940        print commit_log_separator, "r$c->{r} | ";
3941        print "$c->{c} | " if $show_commit;
3942        print "$c->{a} | ", strftime("%Y-%m-%d %H:%M:%S %z (%a, %d %b %Y)",
3943                                 localtime($c->{t_utc})), ' | ';
3944        my $nr_line = 0;
3945
3946        if (my $l = $c->{l}) {
3947                while ($l->[$#$l] eq "\n" && $#$l > 0
3948                                          && $l->[($#$l - 1)] eq "\n") {
3949                        pop @$l;
3950                }
3951                $nr_line = scalar @$l;
3952                if (!$nr_line) {
3953                        print "1 line\n\n\n";
3954                } else {
3955                        if ($nr_line == 1) {
3956                                $nr_line = '1 line';
3957                        } else {
3958                                $nr_line .= ' lines';
3959                        }
3960                        print $nr_line, "\n";
3961                        show_commit_changed_paths($c);
3962                        print "\n";
3963                        print $_ foreach @$l;
3964                }
3965        } else {
3966                print "1 line\n";
3967                show_commit_changed_paths($c);
3968                print "\n";
3969
3970        }
3971        foreach my $x (qw/raw stat diff/) {
3972                if ($c->{$x}) {
3973                        print "\n";
3974                        print $_ foreach @{$c->{$x}}
3975                }
3976        }
3977}
3978
3979sub cmd_show_log {
3980        my (@args) = @_;
3981        my ($r_min, $r_max);
3982        my $r_last = -1; # prevent dupes
3983        if (defined $TZ) {
3984                $ENV{TZ} = $TZ;
3985        } else {
3986                delete $ENV{TZ};
3987        }
3988        if (defined $::_revision) {
3989                if ($::_revision =~ /^(\d+):(\d+)$/) {
3990                        ($r_min, $r_max) = ($1, $2);
3991                } elsif ($::_revision =~ /^\d+$/) {
3992                        $r_min = $r_max = $::_revision;
3993                } else {
3994                        ::fatal "-r$::_revision is not supported, use ",
3995                                "standard 'git log' arguments instead";
3996                }
3997        }
3998
3999        config_pager();
4000        @args = git_svn_log_cmd($r_min, $r_max, @args);
4001        if (!@args) {
4002                print commit_log_separator unless $incremental || $oneline;
4003                return;
4004        }
4005        my $log = command_output_pipe(@args);
4006        run_pager();
4007        my (@k, $c, $d, $stat);
4008        my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
4009        while (<$log>) {
4010                if (/^${esc_color}commit -?($::sha1_short)/o) {
4011                        my $cmt = $1;
4012                        if ($c && cmt_showable($c) && $c->{r} != $r_last) {
4013                                $r_last = $c->{r};
4014                                process_commit($c, $r_min, $r_max, \@k) or
4015                                                                goto out;
4016                        }
4017                        $d = undef;
4018                        $c = { c => $cmt };
4019                } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
4020                        get_author_info($c, $1, $2, $3);
4021                } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
4022                        # ignore
4023                } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
4024                        push @{$c->{raw}}, $_;
4025                } elsif (/^${esc_color}[ACRMDT]\t/) {
4026                        # we could add $SVN->{svn_path} here, but that requires
4027                        # remote access at the moment (repo_path_split)...
4028                        s#^(${esc_color})([ACRMDT])\t#$1   $2 #o;
4029                        push @{$c->{changed}}, $_;
4030                } elsif (/^${esc_color}diff /o) {
4031                        $d = 1;
4032                        push @{$c->{diff}}, $_;
4033                } elsif ($d) {
4034                        push @{$c->{diff}}, $_;
4035                } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
4036                          $esc_color*[\+\-]*$esc_color$/x) {
4037                        $stat = 1;
4038                        push @{$c->{stat}}, $_;
4039                } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
4040                        push @{$c->{stat}}, $_;
4041                        $stat = undef;
4042                } elsif (/^${esc_color}    (git-svn-id:.+)$/o) {
4043                        ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
4044                } elsif (s/^${esc_color}    //o) {
4045                        push @{$c->{l}}, $_;
4046                }
4047        }
4048        if ($c && defined $c->{r} && $c->{r} != $r_last) {
4049                $r_last = $c->{r};
4050                process_commit($c, $r_min, $r_max, \@k);
4051        }
4052        if (@k) {
4053                ($r_min, $r_max) = ($r_max, $r_min);
4054                process_commit($_, $r_min, $r_max) foreach reverse @k;
4055        }
4056out:
4057        close $log;
4058        print commit_log_separator unless $incremental || $oneline;
4059}
4060
4061package Git::SVN::Migration;
4062# these version numbers do NOT correspond to actual version numbers
4063# of git nor git-svn.  They are just relative.
4064#
4065# v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
4066#
4067# v1 layout: .git/$id/info/url, refs/remotes/$id
4068#
4069# v2 layout: .git/svn/$id/info/url, refs/remotes/$id
4070#
4071# v3 layout: .git/svn/$id, refs/remotes/$id
4072#            - info/url may remain for backwards compatibility
4073#            - this is what we migrate up to this layout automatically,
4074#            - this will be used by git svn init on single branches
4075# v3.1 layout (auto migrated):
4076#            - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
4077#              for backwards compatibility
4078#
4079# v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
4080#            - this is only created for newly multi-init-ed
4081#              repositories.  Similar in spirit to the
4082#              --use-separate-remotes option in git-clone (now default)
4083#            - we do not automatically migrate to this (following
4084#              the example set by core git)
4085use strict;
4086use warnings;
4087use Carp qw/croak/;
4088use File::Path qw/mkpath/;
4089use File::Basename qw/dirname basename/;
4090use vars qw/$_minimize/;
4091
4092sub migrate_from_v0 {
4093        my $git_dir = $ENV{GIT_DIR};
4094        return undef unless -d $git_dir;
4095        my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4096        my $migrated = 0;
4097        while (<$fh>) {
4098                chomp;
4099                my ($id, $orig_ref) = ($_, $_);
4100                next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
4101                next unless -f "$git_dir/$id/info/url";
4102                my $new_ref = "refs/remotes/$id";
4103                if (::verify_ref("$new_ref^0")) {
4104                        print STDERR "W: $orig_ref is probably an old ",
4105                                     "branch used by an ancient version of ",
4106                                     "git-svn.\n",
4107                                     "However, $new_ref also exists.\n",
4108                                     "We will not be able ",
4109                                     "to use this branch until this ",
4110                                     "ambiguity is resolved.\n";
4111                        next;
4112                }
4113                print STDERR "Migrating from v0 layout...\n" if !$migrated;
4114                print STDERR "Renaming ref: $orig_ref => $new_ref\n";
4115                command_noisy('update-ref', $new_ref, $orig_ref);
4116                command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
4117                $migrated++;
4118        }
4119        command_close_pipe($fh, $ctx);
4120        print STDERR "Done migrating from v0 layout...\n" if $migrated;
4121        $migrated;
4122}
4123
4124sub migrate_from_v1 {
4125        my $git_dir = $ENV{GIT_DIR};
4126        my $migrated = 0;
4127        return $migrated unless -d $git_dir;
4128        my $svn_dir = "$git_dir/svn";
4129
4130        # just in case somebody used 'svn' as their $id at some point...
4131        return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
4132
4133        print STDERR "Migrating from a git-svn v1 layout...\n";
4134        mkpath([$svn_dir]);
4135        print STDERR "Data from a previous version of git-svn exists, but\n\t",
4136                     "$svn_dir\n\t(required for this version ",
4137                     "($::VERSION) of git-svn) does not. exist\n";
4138        my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
4139        while (<$fh>) {
4140                my $x = $_;
4141                next unless $x =~ s#^refs/remotes/##;
4142                chomp $x;
4143                next unless -f "$git_dir/$x/info/url";
4144                my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
4145                next unless $u;
4146                my $dn = dirname("$git_dir/svn/$x");
4147                mkpath([$dn]) unless -d $dn;
4148                if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
4149                        mkpath(["$git_dir/svn/svn"]);
4150                        print STDERR " - $git_dir/$x/info => ",
4151                                        "$git_dir/svn/$x/info\n";
4152                        rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
4153                               croak "$!: $x";
4154                        # don't worry too much about these, they probably
4155                        # don't exist with repos this old (save for index,
4156                        # and we can easily regenerate that)
4157                        foreach my $f (qw/unhandled.log index .rev_db/) {
4158                                rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
4159                        }
4160                } else {
4161                        print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
4162                        rename "$git_dir/$x", "$git_dir/svn/$x" or
4163                               croak "$!: $x";
4164                }
4165                $migrated++;
4166        }
4167        command_close_pipe($fh, $ctx);
4168        print STDERR "Done migrating from a git-svn v1 layout\n";
4169        $migrated;
4170}
4171
4172sub read_old_urls {
4173        my ($l_map, $pfx, $path) = @_;
4174        my @dir;
4175        foreach (<$path/*>) {
4176                if (-r "$_/info/url") {
4177                        $pfx .= '/' if $pfx && $pfx !~ m!/$!;
4178                        my $ref_id = $pfx . basename $_;
4179                        my $url = ::file_to_s("$_/info/url");
4180                        $l_map->{$ref_id} = $url;
4181                } elsif (-d $_) {
4182                        push @dir, $_;
4183                }
4184        }
4185        foreach (@dir) {
4186                my $x = $_;
4187                $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
4188                read_old_urls($l_map, $x, $_);
4189        }
4190}
4191
4192sub migrate_from_v2 {
4193        my @cfg = command(qw/config -l/);
4194        return if grep /^svn-remote\..+\.url=/, @cfg;
4195        my %l_map;
4196        read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
4197        my $migrated = 0;
4198
4199        foreach my $ref_id (sort keys %l_map) {
4200                eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
4201                if ($@) {
4202                        Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
4203                }
4204                $migrated++;
4205        }
4206        $migrated;
4207}
4208
4209sub minimize_connections {
4210        my $r = Git::SVN::read_all_remotes();
4211        my $new_urls = {};
4212        my $root_repos = {};
4213        foreach my $repo_id (keys %$r) {
4214                my $url = $r->{$repo_id}->{url} or next;
4215                my $fetch = $r->{$repo_id}->{fetch} or next;
4216                my $ra = Git::SVN::Ra->new($url);
4217
4218                # skip existing cases where we already connect to the root
4219                if (($ra->{url} eq $ra->{repos_root}) ||
4220                    (Git::SVN::sanitize_remote_name($ra->{repos_root}) eq
4221                     $repo_id)) {
4222                        $root_repos->{$ra->{url}} = $repo_id;
4223                        next;
4224                }
4225
4226                my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
4227                my $root_path = $ra->{url};
4228                $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
4229                foreach my $path (keys %$fetch) {
4230                        my $ref_id = $fetch->{$path};
4231                        my $gs = Git::SVN->new($ref_id, $repo_id, $path);
4232
4233                        # make sure we can read when connecting to
4234                        # a higher level of a repository
4235                        my ($last_rev, undef) = $gs->last_rev_commit;
4236                        if (!defined $last_rev) {
4237                                $last_rev = eval {
4238                                        $root_ra->get_latest_revnum;
4239                                };
4240                                next if $@;
4241                        }
4242                        my $new = $root_path;
4243                        $new .= length $path ? "/$path" : '';
4244                        eval {
4245                                $root_ra->get_log([$new], $last_rev, $last_rev,
4246                                                  0, 0, 1, sub { });
4247                        };
4248                        next if $@;
4249                        $new_urls->{$ra->{repos_root}}->{$new} =
4250                                { ref_id => $ref_id,
4251                                  old_repo_id => $repo_id,
4252                                  old_path => $path };
4253                }
4254        }
4255
4256        my @emptied;
4257        foreach my $url (keys %$new_urls) {
4258                # see if we can re-use an existing [svn-remote "repo_id"]
4259                # instead of creating a(n ugly) new section:
4260                my $repo_id = $root_repos->{$url} ||
4261                              Git::SVN::sanitize_remote_name($url);
4262
4263                my $fetch = $new_urls->{$url};
4264                foreach my $path (keys %$fetch) {
4265                        my $x = $fetch->{$path};
4266                        Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
4267                        my $pfx = "svn-remote.$x->{old_repo_id}";
4268
4269                        my $old_fetch = quotemeta("$x->{old_path}:".
4270                                                  "refs/remotes/$x->{ref_id}");
4271                        command_noisy(qw/config --unset/,
4272                                      "$pfx.fetch", '^'. $old_fetch . '$');
4273                        delete $r->{$x->{old_repo_id}}->
4274                               {fetch}->{$x->{old_path}};
4275                        if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
4276                                command_noisy(qw/config --unset/,
4277                                              "$pfx.url");
4278                                push @emptied, $x->{old_repo_id}
4279                        }
4280                }
4281        }
4282        if (@emptied) {
4283                my $file = $ENV{GIT_CONFIG} || $ENV{GIT_CONFIG_LOCAL} ||
4284                           "$ENV{GIT_DIR}/config";
4285                print STDERR <<EOF;
4286The following [svn-remote] sections in your config file ($file) are empty
4287and can be safely removed:
4288EOF
4289                print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
4290        }
4291}
4292
4293sub migration_check {
4294        migrate_from_v0();
4295        migrate_from_v1();
4296        migrate_from_v2();
4297        minimize_connections() if $_minimize;
4298}
4299
4300package Git::IndexInfo;
4301use strict;
4302use warnings;
4303use Git qw/command_input_pipe command_close_pipe/;
4304
4305sub new {
4306        my ($class) = @_;
4307        my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
4308        bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
4309}
4310
4311sub remove {
4312        my ($self, $path) = @_;
4313        if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
4314                return ++$self->{nr};
4315        }
4316        undef;
4317}
4318
4319sub update {
4320        my ($self, $mode, $hash, $path) = @_;
4321        if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
4322                return ++$self->{nr};
4323        }
4324        undef;
4325}
4326
4327sub DESTROY {
4328        my ($self) = @_;
4329        command_close_pipe($self->{gui}, $self->{ctx});
4330}
4331
4332package Git::SVN::GlobSpec;
4333use strict;
4334use warnings;
4335
4336sub new {
4337        my ($class, $glob) = @_;
4338        my $re = $glob;
4339        $re =~ s!/+$!!g; # no need for trailing slashes
4340        my $nr = ($re =~ s!^(.*)\*(.*)$!\(\[^/\]+\)!g);
4341        my ($left, $right) = ($1, $2);
4342        if ($nr > 1) {
4343                die "Only one '*' wildcard expansion ",
4344                    "is supported (got $nr): '$glob'\n";
4345        } elsif ($nr == 0) {
4346                die "One '*' is needed for glob: '$glob'\n";
4347        }
4348        $re = quotemeta($left) . $re . quotemeta($right);
4349        if (length $left && !($left =~ s!/+$!!g)) {
4350                die "Missing trailing '/' on left side of: '$glob' ($left)\n";
4351        }
4352        if (length $right && !($right =~ s!^/+!!g)) {
4353                die "Missing leading '/' on right side of: '$glob' ($right)\n";
4354        }
4355        my $left_re = qr/^\/\Q$left\E(\/|$)/;
4356        bless { left => $left, right => $right, left_regex => $left_re,
4357                regex => qr/$re/, glob => $glob }, $class;
4358}
4359
4360sub full_path {
4361        my ($self, $path) = @_;
4362        return (length $self->{left} ? "$self->{left}/" : '') .
4363               $path . (length $self->{right} ? "/$self->{right}" : '');
4364}
4365
4366__END__
4367
4368Data structures:
4369
4370
4371$remotes = { # returned by read_all_remotes()
4372        'svn' => {
4373                # svn-remote.svn.url=https://svn.musicpd.org
4374                url => 'https://svn.musicpd.org',
4375                # svn-remote.svn.fetch=mpd/trunk:trunk
4376                fetch => {
4377                        'mpd/trunk' => 'trunk',
4378                },
4379                # svn-remote.svn.tags=mpd/tags/*:tags/*
4380                tags => {
4381                        path => {
4382                                left => 'mpd/tags',
4383                                right => '',
4384                                regex => qr!mpd/tags/([^/]+)$!,
4385                                glob => 'tags/*',
4386                        },
4387                        ref => {
4388                                left => 'tags',
4389                                right => '',
4390                                regex => qr!tags/([^/]+)$!,
4391                                glob => 'tags/*',
4392                        },
4393                }
4394        }
4395};
4396
4397$log_entry hashref as returned by libsvn_log_entry()
4398{
4399        log => 'whitespace-formatted log entry
4400',                                              # trailing newline is preserved
4401        revision => '8',                        # integer
4402        date => '2004-02-24T17:01:44.108345Z',  # commit date
4403        author => 'committer name'
4404};
4405
4406
4407# this is generated by generate_diff();
4408@mods = array of diff-index line hashes, each element represents one line
4409        of diff-index output
4410
4411diff-index line ($m hash)
4412{
4413        mode_a => first column of diff-index output, no leading ':',
4414        mode_b => second column of diff-index output,
4415        sha1_b => sha1sum of the final blob,
4416        chg => change type [MCRADT],
4417        file_a => original file name of a file (iff chg is 'C' or 'R')
4418        file_b => new/current file name of a file (any chg)
4419}
4420;
4421
4422# retval of read_url_paths{,_all}();
4423$l_map = {
4424        # repository root url
4425        'https://svn.musicpd.org' => {
4426                # repository path               # GIT_SVN_ID
4427                'mpd/trunk'             =>      'trunk',
4428                'mpd/tags/0.11.5'       =>      'tags/0.11.5',
4429        },
4430}
4431
4432Notes:
4433        I don't trust the each() function on unless I created %hash myself
4434        because the internal iterator may not have started at base.