git-svn.perlon commit git-rebase.txt: --interactive does not work with --whitespace (7fe5438)
   1#!/usr/bin/env perl
   2# Copyright (C) 2006, Eric Wong <normalperson@yhbt.net>
   3# License: GPL v2 or later
   4use warnings;
   5use strict;
   6use vars qw/    $AUTHOR $VERSION
   7                $sha1 $sha1_short $_revision $_repository
   8                $_q $_authors %users/;
   9$AUTHOR = 'Eric Wong <normalperson@yhbt.net>';
  10$VERSION = '@@GIT_VERSION@@';
  11
  12# From which subdir have we been invoked?
  13my $cmd_dir_prefix = eval {
  14        command_oneline([qw/rev-parse --show-prefix/], STDERR => 0)
  15} || '';
  16
  17my $git_dir_user_set = 1 if defined $ENV{GIT_DIR};
  18$ENV{GIT_DIR} ||= '.git';
  19$Git::SVN::default_repo_id = 'svn';
  20$Git::SVN::default_ref_id = $ENV{GIT_SVN_ID} || 'git-svn';
  21$Git::SVN::Ra::_log_window_size = 100;
  22
  23$Git::SVN::Log::TZ = $ENV{TZ};
  24$ENV{TZ} = 'UTC';
  25$| = 1; # unbuffer STDOUT
  26
  27sub fatal (@) { print STDERR "@_\n"; exit 1 }
  28require SVN::Core; # use()-ing this causes segfaults for me... *shrug*
  29require SVN::Ra;
  30require SVN::Delta;
  31if ($SVN::Core::VERSION lt '1.1.0') {
  32        fatal "Need SVN::Core 1.1.0 or better (got $SVN::Core::VERSION)";
  33}
  34push @Git::SVN::Ra::ISA, 'SVN::Ra';
  35push @SVN::Git::Editor::ISA, 'SVN::Delta::Editor';
  36push @SVN::Git::Fetcher::ISA, 'SVN::Delta::Editor';
  37use Carp qw/croak/;
  38use Digest::MD5;
  39use IO::File qw//;
  40use File::Basename qw/dirname basename/;
  41use File::Path qw/mkpath/;
  42use Getopt::Long qw/:config gnu_getopt no_ignore_case auto_abbrev/;
  43use IPC::Open3;
  44use Git;
  45
  46BEGIN {
  47        # import functions from Git into our packages, en masse
  48        no strict 'refs';
  49        foreach (qw/command command_oneline command_noisy command_output_pipe
  50                    command_input_pipe command_close_pipe/) {
  51                for my $package ( qw(SVN::Git::Editor SVN::Git::Fetcher
  52                        Git::SVN::Migration Git::SVN::Log Git::SVN),
  53                        __PACKAGE__) {
  54                        *{"${package}::$_"} = \&{"Git::$_"};
  55                }
  56        }
  57}
  58
  59my ($SVN);
  60
  61$sha1 = qr/[a-f\d]{40}/;
  62$sha1_short = qr/[a-f\d]{4,40}/;
  63my ($_stdin, $_help, $_edit,
  64        $_message, $_file,
  65        $_template, $_shared,
  66        $_version, $_fetch_all, $_no_rebase,
  67        $_merge, $_strategy, $_dry_run, $_local,
  68        $_prefix, $_no_checkout, $_url, $_verbose,
  69        $_git_format, $_commit_url, $_tag);
  70$Git::SVN::_follow_parent = 1;
  71my %remote_opts = ( 'username=s' => \$Git::SVN::Prompt::_username,
  72                    'config-dir=s' => \$Git::SVN::Ra::config_dir,
  73                    'no-auth-cache' => \$Git::SVN::Prompt::_no_auth_cache,
  74                    'ignore-paths=s' => \$SVN::Git::Fetcher::_ignore_regex );
  75my %fc_opts = ( 'follow-parent|follow!' => \$Git::SVN::_follow_parent,
  76                'authors-file|A=s' => \$_authors,
  77                'repack:i' => \$Git::SVN::_repack,
  78                'noMetadata' => \$Git::SVN::_no_metadata,
  79                'useSvmProps' => \$Git::SVN::_use_svm_props,
  80                'useSvnsyncProps' => \$Git::SVN::_use_svnsync_props,
  81                'log-window-size=i' => \$Git::SVN::Ra::_log_window_size,
  82                'no-checkout' => \$_no_checkout,
  83                'quiet|q' => \$_q,
  84                'repack-flags|repack-args|repack-opts=s' =>
  85                   \$Git::SVN::_repack_flags,
  86                'use-log-author' => \$Git::SVN::_use_log_author,
  87                'add-author-from' => \$Git::SVN::_add_author_from,
  88                'localtime' => \$Git::SVN::_localtime,
  89                %remote_opts );
  90
  91my ($_trunk, $_tags, $_branches, $_stdlayout);
  92my %icv;
  93my %init_opts = ( 'template=s' => \$_template, 'shared:s' => \$_shared,
  94                  'trunk|T=s' => \$_trunk, 'tags|t=s' => \$_tags,
  95                  'branches|b=s' => \$_branches, 'prefix=s' => \$_prefix,
  96                  'stdlayout|s' => \$_stdlayout,
  97                  'minimize-url|m' => \$Git::SVN::_minimize_url,
  98                  'no-metadata' => sub { $icv{noMetadata} = 1 },
  99                  'use-svm-props' => sub { $icv{useSvmProps} = 1 },
 100                  'use-svnsync-props' => sub { $icv{useSvnsyncProps} = 1 },
 101                  'rewrite-root=s' => sub { $icv{rewriteRoot} = $_[1] },
 102                  %remote_opts );
 103my %cmt_opts = ( 'edit|e' => \$_edit,
 104                'rmdir' => \$SVN::Git::Editor::_rmdir,
 105                'find-copies-harder' => \$SVN::Git::Editor::_find_copies_harder,
 106                'l=i' => \$SVN::Git::Editor::_rename_limit,
 107                'copy-similarity|C=i'=> \$SVN::Git::Editor::_cp_similarity
 108);
 109
 110my %cmd = (
 111        fetch => [ \&cmd_fetch, "Download new revisions from SVN",
 112                        { 'revision|r=s' => \$_revision,
 113                          'fetch-all|all' => \$_fetch_all,
 114                           %fc_opts } ],
 115        clone => [ \&cmd_clone, "Initialize and fetch revisions",
 116                        { 'revision|r=s' => \$_revision,
 117                           %fc_opts, %init_opts } ],
 118        init => [ \&cmd_init, "Initialize a repo for tracking" .
 119                          " (requires URL argument)",
 120                          \%init_opts ],
 121        'multi-init' => [ \&cmd_multi_init,
 122                          "Deprecated alias for ".
 123                          "'$0 init -T<trunk> -b<branches> -t<tags>'",
 124                          \%init_opts ],
 125        dcommit => [ \&cmd_dcommit,
 126                     'Commit several diffs to merge with upstream',
 127                        { 'merge|m|M' => \$_merge,
 128                          'strategy|s=s' => \$_strategy,
 129                          'verbose|v' => \$_verbose,
 130                          'dry-run|n' => \$_dry_run,
 131                          'fetch-all|all' => \$_fetch_all,
 132                          'commit-url=s' => \$_commit_url,
 133                          'revision|r=i' => \$_revision,
 134                          'no-rebase' => \$_no_rebase,
 135                        %cmt_opts, %fc_opts } ],
 136        branch => [ \&cmd_branch,
 137                    'Create a branch in the SVN repository',
 138                    { 'message|m=s' => \$_message,
 139                      'dry-run|n' => \$_dry_run,
 140                      'tag|t' => \$_tag } ],
 141        tag => [ sub { $_tag = 1; cmd_branch(@_) },
 142                 'Create a tag in the SVN repository',
 143                 { 'message|m=s' => \$_message,
 144                   'dry-run|n' => \$_dry_run } ],
 145        'set-tree' => [ \&cmd_set_tree,
 146                        "Set an SVN repository to a git tree-ish",
 147                        { 'stdin|' => \$_stdin, %cmt_opts, %fc_opts, } ],
 148        'create-ignore' => [ \&cmd_create_ignore,
 149                             'Create a .gitignore per svn:ignore',
 150                             { 'revision|r=i' => \$_revision
 151                             } ],
 152        'propget' => [ \&cmd_propget,
 153                       'Print the value of a property on a file or directory',
 154                       { 'revision|r=i' => \$_revision } ],
 155        'proplist' => [ \&cmd_proplist,
 156                       'List all properties of a file or directory',
 157                       { 'revision|r=i' => \$_revision } ],
 158        'show-ignore' => [ \&cmd_show_ignore, "Show svn:ignore listings",
 159                        { 'revision|r=i' => \$_revision
 160                        } ],
 161        'show-externals' => [ \&cmd_show_externals, "Show svn:externals listings",
 162                        { 'revision|r=i' => \$_revision
 163                        } ],
 164        'multi-fetch' => [ \&cmd_multi_fetch,
 165                           "Deprecated alias for $0 fetch --all",
 166                           { 'revision|r=s' => \$_revision, %fc_opts } ],
 167        'migrate' => [ sub { },
 168                       # no-op, we automatically run this anyways,
 169                       'Migrate configuration/metadata/layout from
 170                        previous versions of git-svn',
 171                       { 'minimize' => \$Git::SVN::Migration::_minimize,
 172                         %remote_opts } ],
 173        'log' => [ \&Git::SVN::Log::cmd_show_log, 'Show commit logs',
 174                        { 'limit=i' => \$Git::SVN::Log::limit,
 175                          'revision|r=s' => \$_revision,
 176                          'verbose|v' => \$Git::SVN::Log::verbose,
 177                          'incremental' => \$Git::SVN::Log::incremental,
 178                          'oneline' => \$Git::SVN::Log::oneline,
 179                          'show-commit' => \$Git::SVN::Log::show_commit,
 180                          'non-recursive' => \$Git::SVN::Log::non_recursive,
 181                          'authors-file|A=s' => \$_authors,
 182                          'color' => \$Git::SVN::Log::color,
 183                          'pager=s' => \$Git::SVN::Log::pager
 184                        } ],
 185        'find-rev' => [ \&cmd_find_rev,
 186                        "Translate between SVN revision numbers and tree-ish",
 187                        {} ],
 188        'rebase' => [ \&cmd_rebase, "Fetch and rebase your working directory",
 189                        { 'merge|m|M' => \$_merge,
 190                          'verbose|v' => \$_verbose,
 191                          'strategy|s=s' => \$_strategy,
 192                          'local|l' => \$_local,
 193                          'fetch-all|all' => \$_fetch_all,
 194                          'dry-run|n' => \$_dry_run,
 195                          %fc_opts } ],
 196        'commit-diff' => [ \&cmd_commit_diff,
 197                           'Commit a diff between two trees',
 198                        { 'message|m=s' => \$_message,
 199                          'file|F=s' => \$_file,
 200                          'revision|r=s' => \$_revision,
 201                        %cmt_opts } ],
 202        'info' => [ \&cmd_info,
 203                    "Show info about the latest SVN revision
 204                     on the current branch",
 205                    { 'url' => \$_url, } ],
 206        'blame' => [ \&Git::SVN::Log::cmd_blame,
 207                    "Show what revision and author last modified each line of a file",
 208                    { 'git-format' => \$_git_format } ],
 209);
 210
 211my $cmd;
 212for (my $i = 0; $i < @ARGV; $i++) {
 213        if (defined $cmd{$ARGV[$i]}) {
 214                $cmd = $ARGV[$i];
 215                splice @ARGV, $i, 1;
 216                last;
 217        }
 218};
 219
 220# make sure we're always running at the top-level working directory
 221unless ($cmd && $cmd =~ /(?:clone|init|multi-init)$/) {
 222        unless (-d $ENV{GIT_DIR}) {
 223                if ($git_dir_user_set) {
 224                        die "GIT_DIR=$ENV{GIT_DIR} explicitly set, ",
 225                            "but it is not a directory\n";
 226                }
 227                my $git_dir = delete $ENV{GIT_DIR};
 228                my $cdup = undef;
 229                git_cmd_try {
 230                        $cdup = command_oneline(qw/rev-parse --show-cdup/);
 231                        $git_dir = '.' unless ($cdup);
 232                        chomp $cdup if ($cdup);
 233                        $cdup = "." unless ($cdup && length $cdup);
 234                } "Already at toplevel, but $git_dir not found\n";
 235                chdir $cdup or die "Unable to chdir up to '$cdup'\n";
 236                unless (-d $git_dir) {
 237                        die "$git_dir still not found after going to ",
 238                            "'$cdup'\n";
 239                }
 240                $ENV{GIT_DIR} = $git_dir;
 241        }
 242        $_repository = Git->repository(Repository => $ENV{GIT_DIR});
 243}
 244
 245my %opts = %{$cmd{$cmd}->[2]} if (defined $cmd);
 246
 247read_repo_config(\%opts);
 248if ($cmd && ($cmd eq 'log' || $cmd eq 'blame')) {
 249        Getopt::Long::Configure('pass_through');
 250}
 251my $rv = GetOptions(%opts, 'help|H|h' => \$_help, 'version|V' => \$_version,
 252                    'minimize-connections' => \$Git::SVN::Migration::_minimize,
 253                    'id|i=s' => \$Git::SVN::default_ref_id,
 254                    'svn-remote|remote|R=s' => sub {
 255                       $Git::SVN::no_reuse_existing = 1;
 256                       $Git::SVN::default_repo_id = $_[1] });
 257exit 1 if (!$rv && $cmd && $cmd ne 'log');
 258
 259usage(0) if $_help;
 260version() if $_version;
 261usage(1) unless defined $cmd;
 262load_authors() if $_authors;
 263
 264unless ($cmd =~ /^(?:clone|init|multi-init|commit-diff)$/) {
 265        Git::SVN::Migration::migration_check();
 266}
 267Git::SVN::init_vars();
 268eval {
 269        Git::SVN::verify_remotes_sanity();
 270        $cmd{$cmd}->[0]->(@ARGV);
 271};
 272fatal $@ if $@;
 273post_fetch_checkout();
 274exit 0;
 275
 276####################### primary functions ######################
 277sub usage {
 278        my $exit = shift || 0;
 279        my $fd = $exit ? \*STDERR : \*STDOUT;
 280        print $fd <<"";
 281git-svn - bidirectional operations between a single Subversion tree and git
 282Usage: git svn <command> [options] [arguments]\n
 283
 284        print $fd "Available commands:\n" unless $cmd;
 285
 286        foreach (sort keys %cmd) {
 287                next if $cmd && $cmd ne $_;
 288                next if /^multi-/; # don't show deprecated commands
 289                print $fd '  ',pack('A17',$_),$cmd{$_}->[1],"\n";
 290                foreach (sort keys %{$cmd{$_}->[2]}) {
 291                        # mixed-case options are for .git/config only
 292                        next if /[A-Z]/ && /^[a-z]+$/i;
 293                        # prints out arguments as they should be passed:
 294                        my $x = s#[:=]s$## ? '<arg>' : s#[:=]i$## ? '<num>' : '';
 295                        print $fd ' ' x 21, join(', ', map { length $_ > 1 ?
 296                                                        "--$_" : "-$_" }
 297                                                split /\|/,$_)," $x\n";
 298                }
 299        }
 300        print $fd <<"";
 301\nGIT_SVN_ID may be set in the environment or via the --id/-i switch to an
 302arbitrary identifier if you're tracking multiple SVN branches/repositories in
 303one git repository and want to keep them separate.  See git-svn(1) for more
 304information.
 305
 306        exit $exit;
 307}
 308
 309sub version {
 310        print "git-svn version $VERSION (svn $SVN::Core::VERSION)\n";
 311        exit 0;
 312}
 313
 314sub do_git_init_db {
 315        unless (-d $ENV{GIT_DIR}) {
 316                my @init_db = ('init');
 317                push @init_db, "--template=$_template" if defined $_template;
 318                if (defined $_shared) {
 319                        if ($_shared =~ /[a-z]/) {
 320                                push @init_db, "--shared=$_shared";
 321                        } else {
 322                                push @init_db, "--shared";
 323                        }
 324                }
 325                command_noisy(@init_db);
 326                $_repository = Git->repository(Repository => ".git");
 327        }
 328        my $set;
 329        my $pfx = "svn-remote.$Git::SVN::default_repo_id";
 330        foreach my $i (keys %icv) {
 331                die "'$set' and '$i' cannot both be set\n" if $set;
 332                next unless defined $icv{$i};
 333                command_noisy('config', "$pfx.$i", $icv{$i});
 334                $set = $i;
 335        }
 336}
 337
 338sub init_subdir {
 339        my $repo_path = shift or return;
 340        mkpath([$repo_path]) unless -d $repo_path;
 341        chdir $repo_path or die "Couldn't chdir to $repo_path: $!\n";
 342        $ENV{GIT_DIR} = '.git';
 343        $_repository = Git->repository(Repository => $ENV{GIT_DIR});
 344}
 345
 346sub cmd_clone {
 347        my ($url, $path) = @_;
 348        if (!defined $path &&
 349            (defined $_trunk || defined $_branches || defined $_tags ||
 350             defined $_stdlayout) &&
 351            $url !~ m#^[a-z\+]+://#) {
 352                $path = $url;
 353        }
 354        $path = basename($url) if !defined $path || !length $path;
 355        cmd_init($url, $path);
 356        Git::SVN::fetch_all($Git::SVN::default_repo_id);
 357}
 358
 359sub cmd_init {
 360        if (defined $_stdlayout) {
 361                $_trunk = 'trunk' if (!defined $_trunk);
 362                $_tags = 'tags' if (!defined $_tags);
 363                $_branches = 'branches' if (!defined $_branches);
 364        }
 365        if (defined $_trunk || defined $_branches || defined $_tags) {
 366                return cmd_multi_init(@_);
 367        }
 368        my $url = shift or die "SVN repository location required ",
 369                               "as a command-line argument\n";
 370        init_subdir(@_);
 371        do_git_init_db();
 372
 373        Git::SVN->init($url);
 374}
 375
 376sub cmd_fetch {
 377        if (grep /^\d+=./, @_) {
 378                die "'<rev>=<commit>' fetch arguments are ",
 379                    "no longer supported.\n";
 380        }
 381        my ($remote) = @_;
 382        if (@_ > 1) {
 383                die "Usage: $0 fetch [--all] [svn-remote]\n";
 384        }
 385        $remote ||= $Git::SVN::default_repo_id;
 386        if ($_fetch_all) {
 387                cmd_multi_fetch();
 388        } else {
 389                Git::SVN::fetch_all($remote, Git::SVN::read_all_remotes());
 390        }
 391}
 392
 393sub cmd_set_tree {
 394        my (@commits) = @_;
 395        if ($_stdin || !@commits) {
 396                print "Reading from stdin...\n";
 397                @commits = ();
 398                while (<STDIN>) {
 399                        if (/\b($sha1_short)\b/o) {
 400                                unshift @commits, $1;
 401                        }
 402                }
 403        }
 404        my @revs;
 405        foreach my $c (@commits) {
 406                my @tmp = command('rev-parse',$c);
 407                if (scalar @tmp == 1) {
 408                        push @revs, $tmp[0];
 409                } elsif (scalar @tmp > 1) {
 410                        push @revs, reverse(command('rev-list',@tmp));
 411                } else {
 412                        fatal "Failed to rev-parse $c";
 413                }
 414        }
 415        my $gs = Git::SVN->new;
 416        my ($r_last, $cmt_last) = $gs->last_rev_commit;
 417        $gs->fetch;
 418        if (defined $gs->{last_rev} && $r_last != $gs->{last_rev}) {
 419                fatal "There are new revisions that were fetched ",
 420                      "and need to be merged (or acknowledged) ",
 421                      "before committing.\nlast rev: $r_last\n",
 422                      " current: $gs->{last_rev}";
 423        }
 424        $gs->set_tree($_) foreach @revs;
 425        print "Done committing ",scalar @revs," revisions to SVN\n";
 426        unlink $gs->{index};
 427}
 428
 429sub cmd_dcommit {
 430        my $head = shift;
 431        git_cmd_try { command_oneline(qw/diff-index --quiet HEAD/) }
 432                'Cannot dcommit with a dirty index.  Commit your changes first, '
 433                . "or stash them with `git stash'.\n";
 434        $head ||= 'HEAD';
 435        my @refs;
 436        my ($url, $rev, $uuid, $gs) = working_head_info($head, \@refs);
 437        unless ($gs) {
 438                die "Unable to determine upstream SVN information from ",
 439                    "$head history.\nPerhaps the repository is empty.";
 440        }
 441        $url = defined $_commit_url ? $_commit_url : $gs->full_url;
 442        my $last_rev = $_revision if defined $_revision;
 443        if ($url) {
 444                print "Committing to $url ...\n";
 445        }
 446        my ($linear_refs, $parents) = linearize_history($gs, \@refs);
 447        if ($_no_rebase && scalar(@$linear_refs) > 1) {
 448                warn "Attempting to commit more than one change while ",
 449                     "--no-rebase is enabled.\n",
 450                     "If these changes depend on each other, re-running ",
 451                     "without --no-rebase may be required."
 452        }
 453        my $expect_url = $url;
 454        Git::SVN::remove_username($expect_url);
 455        while (1) {
 456                my $d = shift @$linear_refs or last;
 457                unless (defined $last_rev) {
 458                        (undef, $last_rev, undef) = cmt_metadata("$d~1");
 459                        unless (defined $last_rev) {
 460                                fatal "Unable to extract revision information ",
 461                                      "from commit $d~1";
 462                        }
 463                }
 464                if ($_dry_run) {
 465                        print "diff-tree $d~1 $d\n";
 466                } else {
 467                        my $cmt_rev;
 468                        my %ed_opts = ( r => $last_rev,
 469                                        log => get_commit_entry($d)->{log},
 470                                        ra => Git::SVN::Ra->new($url),
 471                                        config => SVN::Core::config_get_config(
 472                                                $Git::SVN::Ra::config_dir
 473                                        ),
 474                                        tree_a => "$d~1",
 475                                        tree_b => $d,
 476                                        editor_cb => sub {
 477                                               print "Committed r$_[0]\n";
 478                                               $cmt_rev = $_[0];
 479                                        },
 480                                        svn_path => '');
 481                        if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
 482                                print "No changes\n$d~1 == $d\n";
 483                        } elsif ($parents->{$d} && @{$parents->{$d}}) {
 484                                $gs->{inject_parents_dcommit}->{$cmt_rev} =
 485                                                               $parents->{$d};
 486                        }
 487                        $_fetch_all ? $gs->fetch_all : $gs->fetch;
 488                        $last_rev = $cmt_rev;
 489                        next if $_no_rebase;
 490
 491                        # we always want to rebase against the current HEAD,
 492                        # not any head that was passed to us
 493                        my @diff = command('diff-tree', $d,
 494                                           $gs->refname, '--');
 495                        my @finish;
 496                        if (@diff) {
 497                                @finish = rebase_cmd();
 498                                print STDERR "W: $d and ", $gs->refname,
 499                                             " differ, using @finish:\n",
 500                                             join("\n", @diff), "\n";
 501                        } else {
 502                                print "No changes between current HEAD and ",
 503                                      $gs->refname,
 504                                      "\nResetting to the latest ",
 505                                      $gs->refname, "\n";
 506                                @finish = qw/reset --mixed/;
 507                        }
 508                        command_noisy(@finish, $gs->refname);
 509                        if (@diff) {
 510                                @refs = ();
 511                                my ($url_, $rev_, $uuid_, $gs_) =
 512                                              working_head_info($head, \@refs);
 513                                my ($linear_refs_, $parents_) =
 514                                              linearize_history($gs_, \@refs);
 515                                if (scalar(@$linear_refs) !=
 516                                    scalar(@$linear_refs_)) {
 517                                        fatal "# of revisions changed ",
 518                                          "\nbefore:\n",
 519                                          join("\n", @$linear_refs),
 520                                          "\n\nafter:\n",
 521                                          join("\n", @$linear_refs_), "\n",
 522                                          'If you are attempting to commit ',
 523                                          "merges, try running:\n\t",
 524                                          'git rebase --interactive',
 525                                          '--preserve-merges ',
 526                                          $gs->refname,
 527                                          "\nBefore dcommitting";
 528                                }
 529                                if ($url_ ne $expect_url) {
 530                                        fatal "URL mismatch after rebase: ",
 531                                              "$url_ != $expect_url";
 532                                }
 533                                if ($uuid_ ne $uuid) {
 534                                        fatal "uuid mismatch after rebase: ",
 535                                              "$uuid_ != $uuid";
 536                                }
 537                                # remap parents
 538                                my (%p, @l, $i);
 539                                for ($i = 0; $i < scalar @$linear_refs; $i++) {
 540                                        my $new = $linear_refs_->[$i] or next;
 541                                        $p{$new} =
 542                                                $parents->{$linear_refs->[$i]};
 543                                        push @l, $new;
 544                                }
 545                                $parents = \%p;
 546                                $linear_refs = \@l;
 547                        }
 548                }
 549        }
 550        unlink $gs->{index};
 551}
 552
 553sub cmd_branch {
 554        my ($branch_name, $head) = @_;
 555
 556        unless (defined $branch_name && length $branch_name) {
 557                die(($_tag ? "tag" : "branch") . " name required\n");
 558        }
 559        $head ||= 'HEAD';
 560
 561        my ($src, $rev, undef, $gs) = working_head_info($head);
 562
 563        my $remote = Git::SVN::read_all_remotes()->{$gs->{repo_id}};
 564        my $glob = $remote->{ $_tag ? 'tags' : 'branches' };
 565        my ($lft, $rgt) = @{ $glob->{path} }{qw/left right/};
 566        my $dst = join '/', $remote->{url}, $lft, $branch_name, ($rgt || ());
 567
 568        my $ctx = SVN::Client->new(
 569                auth    => Git::SVN::Ra::_auth_providers(),
 570                log_msg => sub {
 571                        ${ $_[0] } = defined $_message
 572                                ? $_message
 573                                : 'Create ' . ($_tag ? 'tag ' : 'branch ' )
 574                                . $branch_name;
 575                },
 576        );
 577
 578        eval {
 579                $ctx->ls($dst, 'HEAD', 0);
 580        } and die "branch ${branch_name} already exists\n";
 581
 582        print "Copying ${src} at r${rev} to ${dst}...\n";
 583        $ctx->copy($src, $rev, $dst)
 584                unless $_dry_run;
 585
 586        $gs->fetch_all;
 587}
 588
 589sub cmd_find_rev {
 590        my $revision_or_hash = shift or die "SVN or git revision required ",
 591                                            "as a command-line argument\n";
 592        my $result;
 593        if ($revision_or_hash =~ /^r\d+$/) {
 594                my $head = shift;
 595                $head ||= 'HEAD';
 596                my @refs;
 597                my (undef, undef, $uuid, $gs) = working_head_info($head, \@refs);
 598                unless ($gs) {
 599                        die "Unable to determine upstream SVN information from ",
 600                            "$head history\n";
 601                }
 602                my $desired_revision = substr($revision_or_hash, 1);
 603                $result = $gs->rev_map_get($desired_revision, $uuid);
 604        } else {
 605                my (undef, $rev, undef) = cmt_metadata($revision_or_hash);
 606                $result = $rev;
 607        }
 608        print "$result\n" if $result;
 609}
 610
 611sub cmd_rebase {
 612        command_noisy(qw/update-index --refresh/);
 613        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 614        unless ($gs) {
 615                die "Unable to determine upstream SVN information from ",
 616                    "working tree history\n";
 617        }
 618        if ($_dry_run) {
 619                print "Remote Branch: " . $gs->refname . "\n";
 620                print "SVN URL: " . $url . "\n";
 621                return;
 622        }
 623        if (command(qw/diff-index HEAD --/)) {
 624                print STDERR "Cannot rebase with uncommited changes:\n";
 625                command_noisy('status');
 626                exit 1;
 627        }
 628        unless ($_local) {
 629                # rebase will checkout for us, so no need to do it explicitly
 630                $_no_checkout = 'true';
 631                $_fetch_all ? $gs->fetch_all : $gs->fetch;
 632        }
 633        command_noisy(rebase_cmd(), $gs->refname);
 634}
 635
 636sub cmd_show_ignore {
 637        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 638        $gs ||= Git::SVN->new;
 639        my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
 640        $gs->prop_walk($gs->{path}, $r, sub {
 641                my ($gs, $path, $props) = @_;
 642                print STDOUT "\n# $path\n";
 643                my $s = $props->{'svn:ignore'} or return;
 644                $s =~ s/[\r\n]+/\n/g;
 645                chomp $s;
 646                $s =~ s#^#$path#gm;
 647                print STDOUT "$s\n";
 648        });
 649}
 650
 651sub cmd_show_externals {
 652        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 653        $gs ||= Git::SVN->new;
 654        my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
 655        $gs->prop_walk($gs->{path}, $r, sub {
 656                my ($gs, $path, $props) = @_;
 657                print STDOUT "\n# $path\n";
 658                my $s = $props->{'svn:externals'} or return;
 659                $s =~ s/[\r\n]+/\n/g;
 660                chomp $s;
 661                $s =~ s#^#$path#gm;
 662                print STDOUT "$s\n";
 663        });
 664}
 665
 666sub cmd_create_ignore {
 667        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 668        $gs ||= Git::SVN->new;
 669        my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
 670        $gs->prop_walk($gs->{path}, $r, sub {
 671                my ($gs, $path, $props) = @_;
 672                # $path is of the form /path/to/dir/
 673                my $ignore = '.' . $path . '.gitignore';
 674                my $s = $props->{'svn:ignore'} or return;
 675                open(GITIGNORE, '>', $ignore)
 676                  or fatal("Failed to open `$ignore' for writing: $!");
 677                $s =~ s/[\r\n]+/\n/g;
 678                chomp $s;
 679                # Prefix all patterns so that the ignore doesn't apply
 680                # to sub-directories.
 681                $s =~ s#^#/#gm;
 682                print GITIGNORE "$s\n";
 683                close(GITIGNORE)
 684                  or fatal("Failed to close `$ignore': $!");
 685                command_noisy('add', '-f', $ignore);
 686        });
 687}
 688
 689sub canonicalize_path {
 690        my ($path) = @_;
 691        my $dot_slash_added = 0;
 692        if (substr($path, 0, 1) ne "/") {
 693                $path = "./" . $path;
 694                $dot_slash_added = 1;
 695        }
 696        # File::Spec->canonpath doesn't collapse x/../y into y (for a
 697        # good reason), so let's do this manually.
 698        $path =~ s#/+#/#g;
 699        $path =~ s#/\.(?:/|$)#/#g;
 700        $path =~ s#/[^/]+/\.\.##g;
 701        $path =~ s#/$##g;
 702        $path =~ s#^\./## if $dot_slash_added;
 703        $path =~ s#^/##;
 704        $path =~ s#^\.$##;
 705        return $path;
 706}
 707
 708# get_svnprops(PATH)
 709# ------------------
 710# Helper for cmd_propget and cmd_proplist below.
 711sub get_svnprops {
 712        my $path = shift;
 713        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 714        $gs ||= Git::SVN->new;
 715
 716        # prefix THE PATH by the sub-directory from which the user
 717        # invoked us.
 718        $path = $cmd_dir_prefix . $path;
 719        fatal("No such file or directory: $path") unless -e $path;
 720        my $is_dir = -d $path ? 1 : 0;
 721        $path = $gs->{path} . '/' . $path;
 722
 723        # canonicalize the path (otherwise libsvn will abort or fail to
 724        # find the file)
 725        $path = canonicalize_path($path);
 726
 727        my $r = (defined $_revision ? $_revision : $gs->ra->get_latest_revnum);
 728        my $props;
 729        if ($is_dir) {
 730                (undef, undef, $props) = $gs->ra->get_dir($path, $r);
 731        }
 732        else {
 733                (undef, $props) = $gs->ra->get_file($path, $r, undef);
 734        }
 735        return $props;
 736}
 737
 738# cmd_propget (PROP, PATH)
 739# ------------------------
 740# Print the SVN property PROP for PATH.
 741sub cmd_propget {
 742        my ($prop, $path) = @_;
 743        $path = '.' if not defined $path;
 744        usage(1) if not defined $prop;
 745        my $props = get_svnprops($path);
 746        if (not defined $props->{$prop}) {
 747                fatal("`$path' does not have a `$prop' SVN property.");
 748        }
 749        print $props->{$prop} . "\n";
 750}
 751
 752# cmd_proplist (PATH)
 753# -------------------
 754# Print the list of SVN properties for PATH.
 755sub cmd_proplist {
 756        my $path = shift;
 757        $path = '.' if not defined $path;
 758        my $props = get_svnprops($path);
 759        print "Properties on '$path':\n";
 760        foreach (sort keys %{$props}) {
 761                print "  $_\n";
 762        }
 763}
 764
 765sub cmd_multi_init {
 766        my $url = shift;
 767        unless (defined $_trunk || defined $_branches || defined $_tags) {
 768                usage(1);
 769        }
 770
 771        # there are currently some bugs that prevent multi-init/multi-fetch
 772        # setups from working well without this.
 773        $Git::SVN::_minimize_url = 1;
 774
 775        $_prefix = '' unless defined $_prefix;
 776        if (defined $url) {
 777                $url =~ s#/+$##;
 778                init_subdir(@_);
 779        }
 780        do_git_init_db();
 781        if (defined $_trunk) {
 782                my $trunk_ref = $_prefix . 'trunk';
 783                # try both old-style and new-style lookups:
 784                my $gs_trunk = eval { Git::SVN->new($trunk_ref) };
 785                unless ($gs_trunk) {
 786                        my ($trunk_url, $trunk_path) =
 787                                              complete_svn_url($url, $_trunk);
 788                        $gs_trunk = Git::SVN->init($trunk_url, $trunk_path,
 789                                                   undef, $trunk_ref);
 790                }
 791        }
 792        return unless defined $_branches || defined $_tags;
 793        my $ra = $url ? Git::SVN::Ra->new($url) : undef;
 794        complete_url_ls_init($ra, $_branches, '--branches/-b', $_prefix);
 795        complete_url_ls_init($ra, $_tags, '--tags/-t', $_prefix . 'tags/');
 796}
 797
 798sub cmd_multi_fetch {
 799        my $remotes = Git::SVN::read_all_remotes();
 800        foreach my $repo_id (sort keys %$remotes) {
 801                if ($remotes->{$repo_id}->{url}) {
 802                        Git::SVN::fetch_all($repo_id, $remotes);
 803                }
 804        }
 805}
 806
 807# this command is special because it requires no metadata
 808sub cmd_commit_diff {
 809        my ($ta, $tb, $url) = @_;
 810        my $usage = "Usage: $0 commit-diff -r<revision> ".
 811                    "<tree-ish> <tree-ish> [<URL>]";
 812        fatal($usage) if (!defined $ta || !defined $tb);
 813        my $svn_path = '';
 814        if (!defined $url) {
 815                my $gs = eval { Git::SVN->new };
 816                if (!$gs) {
 817                        fatal("Needed URL or usable git-svn --id in ",
 818                              "the command-line\n", $usage);
 819                }
 820                $url = $gs->{url};
 821                $svn_path = $gs->{path};
 822        }
 823        unless (defined $_revision) {
 824                fatal("-r|--revision is a required argument\n", $usage);
 825        }
 826        if (defined $_message && defined $_file) {
 827                fatal("Both --message/-m and --file/-F specified ",
 828                      "for the commit message.\n",
 829                      "I have no idea what you mean");
 830        }
 831        if (defined $_file) {
 832                $_message = file_to_s($_file);
 833        } else {
 834                $_message ||= get_commit_entry($tb)->{log};
 835        }
 836        my $ra ||= Git::SVN::Ra->new($url);
 837        my $r = $_revision;
 838        if ($r eq 'HEAD') {
 839                $r = $ra->get_latest_revnum;
 840        } elsif ($r !~ /^\d+$/) {
 841                die "revision argument: $r not understood by git-svn\n";
 842        }
 843        my %ed_opts = ( r => $r,
 844                        log => $_message,
 845                        ra => $ra,
 846                        tree_a => $ta,
 847                        tree_b => $tb,
 848                        editor_cb => sub { print "Committed r$_[0]\n" },
 849                        svn_path => $svn_path );
 850        if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
 851                print "No changes\n$ta == $tb\n";
 852        }
 853}
 854
 855sub escape_uri_only {
 856        my ($uri) = @_;
 857        my @tmp;
 858        foreach (split m{/}, $uri) {
 859                s/([^~\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
 860                push @tmp, $_;
 861        }
 862        join('/', @tmp);
 863}
 864
 865sub escape_url {
 866        my ($url) = @_;
 867        if ($url =~ m#^([^:]+)://([^/]*)(.*)$#) {
 868                my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
 869                $url = "$scheme://$domain$uri";
 870        }
 871        $url;
 872}
 873
 874sub cmd_info {
 875        my $path = canonicalize_path(defined($_[0]) ? $_[0] : ".");
 876        my $fullpath = canonicalize_path($cmd_dir_prefix . $path);
 877        if (exists $_[1]) {
 878                die "Too many arguments specified\n";
 879        }
 880
 881        my ($file_type, $diff_status) = find_file_type_and_diff_status($path);
 882
 883        if (!$file_type && !$diff_status) {
 884                print STDERR "svn: '$path' is not under version control\n";
 885                exit 1;
 886        }
 887
 888        my ($url, $rev, $uuid, $gs) = working_head_info('HEAD');
 889        unless ($gs) {
 890                die "Unable to determine upstream SVN information from ",
 891                    "working tree history\n";
 892        }
 893
 894        # canonicalize_path() will return "" to make libsvn 1.5.x happy,
 895        $path = "." if $path eq "";
 896
 897        my $full_url = $url . ($fullpath eq "" ? "" : "/$fullpath");
 898
 899        if ($_url) {
 900                print escape_url($full_url), "\n";
 901                return;
 902        }
 903
 904        my $result = "Path: $path\n";
 905        $result .= "Name: " . basename($path) . "\n" if $file_type ne "dir";
 906        $result .= "URL: " . escape_url($full_url) . "\n";
 907
 908        eval {
 909                my $repos_root = $gs->repos_root;
 910                Git::SVN::remove_username($repos_root);
 911                $result .= "Repository Root: " . escape_url($repos_root) . "\n";
 912        };
 913        if ($@) {
 914                $result .= "Repository Root: (offline)\n";
 915        }
 916        $result .= "Repository UUID: $uuid\n" unless $diff_status eq "A" &&
 917                ($SVN::Core::VERSION le '1.5.4' || $file_type ne "dir");
 918        $result .= "Revision: " . ($diff_status eq "A" ? 0 : $rev) . "\n";
 919
 920        $result .= "Node Kind: " .
 921                   ($file_type eq "dir" ? "directory" : "file") . "\n";
 922
 923        my $schedule = $diff_status eq "A"
 924                       ? "add"
 925                       : ($diff_status eq "D" ? "delete" : "normal");
 926        $result .= "Schedule: $schedule\n";
 927
 928        if ($diff_status eq "A") {
 929                print $result, "\n";
 930                return;
 931        }
 932
 933        my ($lc_author, $lc_rev, $lc_date_utc);
 934        my @args = Git::SVN::Log::git_svn_log_cmd($rev, $rev, "--", $fullpath);
 935        my $log = command_output_pipe(@args);
 936        my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
 937        while (<$log>) {
 938                if (/^${esc_color}author (.+) <[^>]+> (\d+) ([\-\+]?\d+)$/o) {
 939                        $lc_author = $1;
 940                        $lc_date_utc = Git::SVN::Log::parse_git_date($2, $3);
 941                } elsif (/^${esc_color}    (git-svn-id:.+)$/o) {
 942                        (undef, $lc_rev, undef) = ::extract_metadata($1);
 943                }
 944        }
 945        close $log;
 946
 947        Git::SVN::Log::set_local_timezone();
 948
 949        $result .= "Last Changed Author: $lc_author\n";
 950        $result .= "Last Changed Rev: $lc_rev\n";
 951        $result .= "Last Changed Date: " .
 952                   Git::SVN::Log::format_svn_date($lc_date_utc) . "\n";
 953
 954        if ($file_type ne "dir") {
 955                my $text_last_updated_date =
 956                    ($diff_status eq "D" ? $lc_date_utc : (stat $path)[9]);
 957                $result .=
 958                    "Text Last Updated: " .
 959                    Git::SVN::Log::format_svn_date($text_last_updated_date) .
 960                    "\n";
 961                my $checksum;
 962                if ($diff_status eq "D") {
 963                        my ($fh, $ctx) =
 964                            command_output_pipe(qw(cat-file blob), "HEAD:$path");
 965                        if ($file_type eq "link") {
 966                                my $file_name = <$fh>;
 967                                $checksum = md5sum("link $file_name");
 968                        } else {
 969                                $checksum = md5sum($fh);
 970                        }
 971                        command_close_pipe($fh, $ctx);
 972                } elsif ($file_type eq "link") {
 973                        my $file_name =
 974                            command(qw(cat-file blob), "HEAD:$path");
 975                        $checksum =
 976                            md5sum("link " . $file_name);
 977                } else {
 978                        open FILE, "<", $path or die $!;
 979                        $checksum = md5sum(\*FILE);
 980                        close FILE or die $!;
 981                }
 982                $result .= "Checksum: " . $checksum . "\n";
 983        }
 984
 985        print $result, "\n";
 986}
 987
 988########################### utility functions #########################
 989
 990sub rebase_cmd {
 991        my @cmd = qw/rebase/;
 992        push @cmd, '-v' if $_verbose;
 993        push @cmd, qw/--merge/ if $_merge;
 994        push @cmd, "--strategy=$_strategy" if $_strategy;
 995        @cmd;
 996}
 997
 998sub post_fetch_checkout {
 999        return if $_no_checkout;
1000        my $gs = $Git::SVN::_head or return;
1001        return if verify_ref('refs/heads/master^0');
1002
1003        my $valid_head = verify_ref('HEAD^0');
1004        command_noisy(qw(update-ref refs/heads/master), $gs->refname);
1005        return if ($valid_head || !verify_ref('HEAD^0'));
1006
1007        return if $ENV{GIT_DIR} !~ m#^(?:.*/)?\.git$#;
1008        my $index = $ENV{GIT_INDEX_FILE} || "$ENV{GIT_DIR}/index";
1009        return if -f $index;
1010
1011        return if command_oneline(qw/rev-parse --is-inside-work-tree/) eq 'false';
1012        return if command_oneline(qw/rev-parse --is-inside-git-dir/) eq 'true';
1013        command_noisy(qw/read-tree -m -u -v HEAD HEAD/);
1014        print STDERR "Checked out HEAD:\n  ",
1015                     $gs->full_url, " r", $gs->last_rev, "\n";
1016}
1017
1018sub complete_svn_url {
1019        my ($url, $path) = @_;
1020        $path =~ s#/+$##;
1021        if ($path !~ m#^[a-z\+]+://#) {
1022                if (!defined $url || $url !~ m#^[a-z\+]+://#) {
1023                        fatal("E: '$path' is not a complete URL ",
1024                              "and a separate URL is not specified");
1025                }
1026                return ($url, $path);
1027        }
1028        return ($path, '');
1029}
1030
1031sub complete_url_ls_init {
1032        my ($ra, $repo_path, $switch, $pfx) = @_;
1033        unless ($repo_path) {
1034                print STDERR "W: $switch not specified\n";
1035                return;
1036        }
1037        $repo_path =~ s#/+$##;
1038        if ($repo_path =~ m#^[a-z\+]+://#) {
1039                $ra = Git::SVN::Ra->new($repo_path);
1040                $repo_path = '';
1041        } else {
1042                $repo_path =~ s#^/+##;
1043                unless ($ra) {
1044                        fatal("E: '$repo_path' is not a complete URL ",
1045                              "and a separate URL is not specified");
1046                }
1047        }
1048        my $url = $ra->{url};
1049        my $gs = Git::SVN->init($url, undef, undef, undef, 1);
1050        my $k = "svn-remote.$gs->{repo_id}.url";
1051        my $orig_url = eval { command_oneline(qw/config --get/, $k) };
1052        if ($orig_url && ($orig_url ne $gs->{url})) {
1053                die "$k already set: $orig_url\n",
1054                    "wanted to set to: $gs->{url}\n";
1055        }
1056        command_oneline('config', $k, $gs->{url}) unless $orig_url;
1057        my $remote_path = "$ra->{svn_path}/$repo_path";
1058        $remote_path =~ s#/+#/#g;
1059        $remote_path =~ s#^/##g;
1060        $remote_path .= "/*" if $remote_path !~ /\*/;
1061        my ($n) = ($switch =~ /^--(\w+)/);
1062        if (length $pfx && $pfx !~ m#/$#) {
1063                die "--prefix='$pfx' must have a trailing slash '/'\n";
1064        }
1065        command_noisy('config',
1066                      "svn-remote.$gs->{repo_id}.$n",
1067                      "$remote_path:refs/remotes/$pfx*" .
1068                        ('/*' x (($remote_path =~ tr/*/*/) - 1)) );
1069}
1070
1071sub verify_ref {
1072        my ($ref) = @_;
1073        eval { command_oneline([ 'rev-parse', '--verify', $ref ],
1074                               { STDERR => 0 }); };
1075}
1076
1077sub get_tree_from_treeish {
1078        my ($treeish) = @_;
1079        # $treeish can be a symbolic ref, too:
1080        my $type = command_oneline(qw/cat-file -t/, $treeish);
1081        my $expected;
1082        while ($type eq 'tag') {
1083                ($treeish, $type) = command(qw/cat-file tag/, $treeish);
1084        }
1085        if ($type eq 'commit') {
1086                $expected = (grep /^tree /, command(qw/cat-file commit/,
1087                                                    $treeish))[0];
1088                ($expected) = ($expected =~ /^tree ($sha1)$/o);
1089                die "Unable to get tree from $treeish\n" unless $expected;
1090        } elsif ($type eq 'tree') {
1091                $expected = $treeish;
1092        } else {
1093                die "$treeish is a $type, expected tree, tag or commit\n";
1094        }
1095        return $expected;
1096}
1097
1098sub get_commit_entry {
1099        my ($treeish) = shift;
1100        my %log_entry = ( log => '', tree => get_tree_from_treeish($treeish) );
1101        my $commit_editmsg = "$ENV{GIT_DIR}/COMMIT_EDITMSG";
1102        my $commit_msg = "$ENV{GIT_DIR}/COMMIT_MSG";
1103        open my $log_fh, '>', $commit_editmsg or croak $!;
1104
1105        my $type = command_oneline(qw/cat-file -t/, $treeish);
1106        if ($type eq 'commit' || $type eq 'tag') {
1107                my ($msg_fh, $ctx) = command_output_pipe('cat-file',
1108                                                         $type, $treeish);
1109                my $in_msg = 0;
1110                my $author;
1111                my $saw_from = 0;
1112                my $msgbuf = "";
1113                while (<$msg_fh>) {
1114                        if (!$in_msg) {
1115                                $in_msg = 1 if (/^\s*$/);
1116                                $author = $1 if (/^author (.*>)/);
1117                        } elsif (/^git-svn-id: /) {
1118                                # skip this for now, we regenerate the
1119                                # correct one on re-fetch anyways
1120                                # TODO: set *:merge properties or like...
1121                        } else {
1122                                if (/^From:/ || /^Signed-off-by:/) {
1123                                        $saw_from = 1;
1124                                }
1125                                $msgbuf .= $_;
1126                        }
1127                }
1128                $msgbuf =~ s/\s+$//s;
1129                if ($Git::SVN::_add_author_from && defined($author)
1130                    && !$saw_from) {
1131                        $msgbuf .= "\n\nFrom: $author";
1132                }
1133                print $log_fh $msgbuf or croak $!;
1134                command_close_pipe($msg_fh, $ctx);
1135        }
1136        close $log_fh or croak $!;
1137
1138        if ($_edit || ($type eq 'tree')) {
1139                my $editor = $ENV{VISUAL} || $ENV{EDITOR} || 'vi';
1140                # TODO: strip out spaces, comments, like git-commit.sh
1141                system($editor, $commit_editmsg);
1142        }
1143        rename $commit_editmsg, $commit_msg or croak $!;
1144        {
1145                # SVN requires messages to be UTF-8 when entering the repo
1146                local $/;
1147                open $log_fh, '<', $commit_msg or croak $!;
1148                binmode $log_fh;
1149                chomp($log_entry{log} = <$log_fh>);
1150
1151                if (my $enc = Git::config('i18n.commitencoding')) {
1152                        require Encode;
1153                        Encode::from_to($log_entry{log}, $enc, 'UTF-8');
1154                }
1155                close $log_fh or croak $!;
1156        }
1157        unlink $commit_msg;
1158        \%log_entry;
1159}
1160
1161sub s_to_file {
1162        my ($str, $file, $mode) = @_;
1163        open my $fd,'>',$file or croak $!;
1164        print $fd $str,"\n" or croak $!;
1165        close $fd or croak $!;
1166        chmod ($mode &~ umask, $file) if (defined $mode);
1167}
1168
1169sub file_to_s {
1170        my $file = shift;
1171        open my $fd,'<',$file or croak "$!: file: $file\n";
1172        local $/;
1173        my $ret = <$fd>;
1174        close $fd or croak $!;
1175        $ret =~ s/\s*$//s;
1176        return $ret;
1177}
1178
1179# '<svn username> = real-name <email address>' mapping based on git-svnimport:
1180sub load_authors {
1181        open my $authors, '<', $_authors or die "Can't open $_authors $!\n";
1182        my $log = $cmd eq 'log';
1183        while (<$authors>) {
1184                chomp;
1185                next unless /^(.+?|\(no author\))\s*=\s*(.+?)\s*<(.+)>\s*$/;
1186                my ($user, $name, $email) = ($1, $2, $3);
1187                if ($log) {
1188                        $Git::SVN::Log::rusers{"$name <$email>"} = $user;
1189                } else {
1190                        $users{$user} = [$name, $email];
1191                }
1192        }
1193        close $authors or croak $!;
1194}
1195
1196# convert GetOpt::Long specs for use by git-config
1197sub read_repo_config {
1198        return unless -d $ENV{GIT_DIR};
1199        my $opts = shift;
1200        my @config_only;
1201        foreach my $o (keys %$opts) {
1202                # if we have mixedCase and a long option-only, then
1203                # it's a config-only variable that we don't need for
1204                # the command-line.
1205                push @config_only, $o if ($o =~ /[A-Z]/ && $o =~ /^[a-z]+$/i);
1206                my $v = $opts->{$o};
1207                my ($key) = ($o =~ /^([a-zA-Z\-]+)/);
1208                $key =~ s/-//g;
1209                my $arg = 'git config';
1210                $arg .= ' --int' if ($o =~ /[:=]i$/);
1211                $arg .= ' --bool' if ($o !~ /[:=][sfi]$/);
1212                if (ref $v eq 'ARRAY') {
1213                        chomp(my @tmp = `$arg --get-all svn.$key`);
1214                        @$v = @tmp if @tmp;
1215                } else {
1216                        chomp(my $tmp = `$arg --get svn.$key`);
1217                        if ($tmp && !($arg =~ / --bool/ && $tmp eq 'false')) {
1218                                $$v = $tmp;
1219                        }
1220                }
1221        }
1222        delete @$opts{@config_only} if @config_only;
1223}
1224
1225sub extract_metadata {
1226        my $id = shift or return (undef, undef, undef);
1227        my ($url, $rev, $uuid) = ($id =~ /^\s*git-svn-id:\s+(.*)\@(\d+)
1228                                                        \s([a-f\d\-]+)$/x);
1229        if (!defined $rev || !$uuid || !$url) {
1230                # some of the original repositories I made had
1231                # identifiers like this:
1232                ($rev, $uuid) = ($id =~/^\s*git-svn-id:\s(\d+)\@([a-f\d\-]+)/);
1233        }
1234        return ($url, $rev, $uuid);
1235}
1236
1237sub cmt_metadata {
1238        return extract_metadata((grep(/^git-svn-id: /,
1239                command(qw/cat-file commit/, shift)))[-1]);
1240}
1241
1242sub working_head_info {
1243        my ($head, $refs) = @_;
1244        my @args = ('log', '--no-color', '--first-parent', '--pretty=medium');
1245        my ($fh, $ctx) = command_output_pipe(@args, $head);
1246        my $hash;
1247        my %max;
1248        while (<$fh>) {
1249                if ( m{^commit ($::sha1)$} ) {
1250                        unshift @$refs, $hash if $hash and $refs;
1251                        $hash = $1;
1252                        next;
1253                }
1254                next unless s{^\s*(git-svn-id:)}{$1};
1255                my ($url, $rev, $uuid) = extract_metadata($_);
1256                if (defined $url && defined $rev) {
1257                        next if $max{$url} and $max{$url} < $rev;
1258                        if (my $gs = Git::SVN->find_by_url($url)) {
1259                                my $c = $gs->rev_map_get($rev, $uuid);
1260                                if ($c && $c eq $hash) {
1261                                        close $fh; # break the pipe
1262                                        return ($url, $rev, $uuid, $gs);
1263                                } else {
1264                                        $max{$url} ||= $gs->rev_map_max;
1265                                }
1266                        }
1267                }
1268        }
1269        command_close_pipe($fh, $ctx);
1270        (undef, undef, undef, undef);
1271}
1272
1273sub read_commit_parents {
1274        my ($parents, $c) = @_;
1275        chomp(my $p = command_oneline(qw/rev-list --parents -1/, $c));
1276        $p =~ s/^($c)\s*// or die "rev-list --parents -1 $c failed!\n";
1277        @{$parents->{$c}} = split(/ /, $p);
1278}
1279
1280sub linearize_history {
1281        my ($gs, $refs) = @_;
1282        my %parents;
1283        foreach my $c (@$refs) {
1284                read_commit_parents(\%parents, $c);
1285        }
1286
1287        my @linear_refs;
1288        my %skip = ();
1289        my $last_svn_commit = $gs->last_commit;
1290        foreach my $c (reverse @$refs) {
1291                next if $c eq $last_svn_commit;
1292                last if $skip{$c};
1293
1294                unshift @linear_refs, $c;
1295                $skip{$c} = 1;
1296
1297                # we only want the first parent to diff against for linear
1298                # history, we save the rest to inject when we finalize the
1299                # svn commit
1300                my $fp_a = verify_ref("$c~1");
1301                my $fp_b = shift @{$parents{$c}} if $parents{$c};
1302                if (!$fp_a || !$fp_b) {
1303                        die "Commit $c\n",
1304                            "has no parent commit, and therefore ",
1305                            "nothing to diff against.\n",
1306                            "You should be working from a repository ",
1307                            "originally created by git-svn\n";
1308                }
1309                if ($fp_a ne $fp_b) {
1310                        die "$c~1 = $fp_a, however parsing commit $c ",
1311                            "revealed that:\n$c~1 = $fp_b\nBUG!\n";
1312                }
1313
1314                foreach my $p (@{$parents{$c}}) {
1315                        $skip{$p} = 1;
1316                }
1317        }
1318        (\@linear_refs, \%parents);
1319}
1320
1321sub find_file_type_and_diff_status {
1322        my ($path) = @_;
1323        return ('dir', '') if $path eq '';
1324
1325        my $diff_output =
1326            command_oneline(qw(diff --cached --name-status --), $path) || "";
1327        my $diff_status = (split(' ', $diff_output))[0] || "";
1328
1329        my $ls_tree = command_oneline(qw(ls-tree HEAD), $path) || "";
1330
1331        return (undef, undef) if !$diff_status && !$ls_tree;
1332
1333        if ($diff_status eq "A") {
1334                return ("link", $diff_status) if -l $path;
1335                return ("dir", $diff_status) if -d $path;
1336                return ("file", $diff_status);
1337        }
1338
1339        my $mode = (split(' ', $ls_tree))[0] || "";
1340
1341        return ("link", $diff_status) if $mode eq "120000";
1342        return ("dir", $diff_status) if $mode eq "040000";
1343        return ("file", $diff_status);
1344}
1345
1346sub md5sum {
1347        my $arg = shift;
1348        my $ref = ref $arg;
1349        my $md5 = Digest::MD5->new();
1350        if ($ref eq 'GLOB' || $ref eq 'IO::File' || $ref eq 'File::Temp') {
1351                $md5->addfile($arg) or croak $!;
1352        } elsif ($ref eq 'SCALAR') {
1353                $md5->add($$arg) or croak $!;
1354        } elsif (!$ref) {
1355                $md5->add($arg) or croak $!;
1356        } else {
1357                ::fatal "Can't provide MD5 hash for unknown ref type: '", $ref, "'";
1358        }
1359        return $md5->hexdigest();
1360}
1361
1362package Git::SVN;
1363use strict;
1364use warnings;
1365use Fcntl qw/:DEFAULT :seek/;
1366use constant rev_map_fmt => 'NH40';
1367use vars qw/$default_repo_id $default_ref_id $_no_metadata $_follow_parent
1368            $_repack $_repack_flags $_use_svm_props $_head
1369            $_use_svnsync_props $no_reuse_existing $_minimize_url
1370            $_use_log_author $_add_author_from $_localtime/;
1371use Carp qw/croak/;
1372use File::Path qw/mkpath/;
1373use File::Copy qw/copy/;
1374use IPC::Open3;
1375
1376my ($_gc_nr, $_gc_period);
1377
1378# properties that we do not log:
1379my %SKIP_PROP;
1380BEGIN {
1381        %SKIP_PROP = map { $_ => 1 } qw/svn:wc:ra_dav:version-url
1382                                        svn:special svn:executable
1383                                        svn:entry:committed-rev
1384                                        svn:entry:last-author
1385                                        svn:entry:uuid
1386                                        svn:entry:committed-date/;
1387
1388        # some options are read globally, but can be overridden locally
1389        # per [svn-remote "..."] section.  Command-line options will *NOT*
1390        # override options set in an [svn-remote "..."] section
1391        no strict 'refs';
1392        for my $option (qw/follow_parent no_metadata use_svm_props
1393                           use_svnsync_props/) {
1394                my $key = $option;
1395                $key =~ tr/_//d;
1396                my $prop = "-$option";
1397                *$option = sub {
1398                        my ($self) = @_;
1399                        return $self->{$prop} if exists $self->{$prop};
1400                        my $k = "svn-remote.$self->{repo_id}.$key";
1401                        eval { command_oneline(qw/config --get/, $k) };
1402                        if ($@) {
1403                                $self->{$prop} = ${"Git::SVN::_$option"};
1404                        } else {
1405                                my $v = command_oneline(qw/config --bool/,$k);
1406                                $self->{$prop} = $v eq 'false' ? 0 : 1;
1407                        }
1408                        return $self->{$prop};
1409                }
1410        }
1411}
1412
1413
1414my (%LOCKFILES, %INDEX_FILES);
1415END {
1416        unlink keys %LOCKFILES if %LOCKFILES;
1417        unlink keys %INDEX_FILES if %INDEX_FILES;
1418}
1419
1420sub resolve_local_globs {
1421        my ($url, $fetch, $glob_spec) = @_;
1422        return unless defined $glob_spec;
1423        my $ref = $glob_spec->{ref};
1424        my $path = $glob_spec->{path};
1425        foreach (command(qw#for-each-ref --format=%(refname) refs/remotes#)) {
1426                next unless m#^refs/remotes/$ref->{regex}$#;
1427                my $p = $1;
1428                my $pathname = desanitize_refname($path->full_path($p));
1429                my $refname = desanitize_refname($ref->full_path($p));
1430                if (my $existing = $fetch->{$pathname}) {
1431                        if ($existing ne $refname) {
1432                                die "Refspec conflict:\n",
1433                                    "existing: refs/remotes/$existing\n",
1434                                    " globbed: refs/remotes/$refname\n";
1435                        }
1436                        my $u = (::cmt_metadata("refs/remotes/$refname"))[0];
1437                        $u =~ s!^\Q$url\E(/|$)!! or die
1438                          "refs/remotes/$refname: '$url' not found in '$u'\n";
1439                        if ($pathname ne $u) {
1440                                warn "W: Refspec glob conflict ",
1441                                     "(ref: refs/remotes/$refname):\n",
1442                                     "expected path: $pathname\n",
1443                                     "    real path: $u\n",
1444                                     "Continuing ahead with $u\n";
1445                                next;
1446                        }
1447                } else {
1448                        $fetch->{$pathname} = $refname;
1449                }
1450        }
1451}
1452
1453sub parse_revision_argument {
1454        my ($base, $head) = @_;
1455        if (!defined $::_revision || $::_revision eq 'BASE:HEAD') {
1456                return ($base, $head);
1457        }
1458        return ($1, $2) if ($::_revision =~ /^(\d+):(\d+)$/);
1459        return ($::_revision, $::_revision) if ($::_revision =~ /^\d+$/);
1460        return ($head, $head) if ($::_revision eq 'HEAD');
1461        return ($base, $1) if ($::_revision =~ /^BASE:(\d+)$/);
1462        return ($1, $head) if ($::_revision =~ /^(\d+):HEAD$/);
1463        die "revision argument: $::_revision not understood by git-svn\n";
1464}
1465
1466sub fetch_all {
1467        my ($repo_id, $remotes) = @_;
1468        if (ref $repo_id) {
1469                my $gs = $repo_id;
1470                $repo_id = undef;
1471                $repo_id = $gs->{repo_id};
1472        }
1473        $remotes ||= read_all_remotes();
1474        my $remote = $remotes->{$repo_id} or
1475                     die "[svn-remote \"$repo_id\"] unknown\n";
1476        my $fetch = $remote->{fetch};
1477        my $url = $remote->{url} or die "svn-remote.$repo_id.url not defined\n";
1478        my (@gs, @globs);
1479        my $ra = Git::SVN::Ra->new($url);
1480        my $uuid = $ra->get_uuid;
1481        my $head = $ra->get_latest_revnum;
1482        my $base = defined $fetch ? $head : 0;
1483
1484        # read the max revs for wildcard expansion (branches/*, tags/*)
1485        foreach my $t (qw/branches tags/) {
1486                defined $remote->{$t} or next;
1487                push @globs, $remote->{$t};
1488                my $max_rev = eval { tmp_config(qw/--int --get/,
1489                                         "svn-remote.$repo_id.${t}-maxRev") };
1490                if (defined $max_rev && ($max_rev < $base)) {
1491                        $base = $max_rev;
1492                } elsif (!defined $max_rev) {
1493                        $base = 0;
1494                }
1495        }
1496
1497        if ($fetch) {
1498                foreach my $p (sort keys %$fetch) {
1499                        my $gs = Git::SVN->new($fetch->{$p}, $repo_id, $p);
1500                        my $lr = $gs->rev_map_max;
1501                        if (defined $lr) {
1502                                $base = $lr if ($lr < $base);
1503                        }
1504                        push @gs, $gs;
1505                }
1506        }
1507
1508        ($base, $head) = parse_revision_argument($base, $head);
1509        $ra->gs_fetch_loop_common($base, $head, \@gs, \@globs);
1510}
1511
1512sub read_all_remotes {
1513        my $r = {};
1514        my $use_svm_props = eval { command_oneline(qw/config --bool
1515            svn.useSvmProps/) };
1516        $use_svm_props = $use_svm_props eq 'true' if $use_svm_props;
1517        foreach (grep { s/^svn-remote\.// } command(qw/config -l/)) {
1518                if (m!^(.+)\.fetch=\s*(.*)\s*:\s*(.+)\s*$!) {
1519                        my ($remote, $local_ref, $_remote_ref) = ($1, $2, $3);
1520                        die("svn-remote.$remote: remote ref '$_remote_ref' "
1521                            . "must start with 'refs/remotes/'\n")
1522                                unless $_remote_ref =~ m{^refs/remotes/(.+)};
1523                        my $remote_ref = $1;
1524                        $local_ref =~ s{^/}{};
1525                        $r->{$remote}->{fetch}->{$local_ref} = $remote_ref;
1526                        $r->{$remote}->{svm} = {} if $use_svm_props;
1527                } elsif (m!^(.+)\.usesvmprops=\s*(.*)\s*$!) {
1528                        $r->{$1}->{svm} = {};
1529                } elsif (m!^(.+)\.url=\s*(.*)\s*$!) {
1530                        $r->{$1}->{url} = $2;
1531                } elsif (m!^(.+)\.(branches|tags)=
1532                           (.*):refs/remotes/(.+)\s*$/!x) {
1533                        my ($p, $g) = ($3, $4);
1534                        my $rs = $r->{$1}->{$2} = {
1535                                          t => $2,
1536                                          remote => $1,
1537                                          path => Git::SVN::GlobSpec->new($p),
1538                                          ref => Git::SVN::GlobSpec->new($g) };
1539                        if (length($rs->{ref}->{right}) != 0) {
1540                                die "The '*' glob character must be the last ",
1541                                    "character of '$g'\n";
1542                        }
1543                }
1544        }
1545
1546        map {
1547                if (defined $r->{$_}->{svm}) {
1548                        my $svm;
1549                        eval {
1550                                my $section = "svn-remote.$_";
1551                                $svm = {
1552                                        source => tmp_config('--get',
1553                                            "$section.svm-source"),
1554                                        replace => tmp_config('--get',
1555                                            "$section.svm-replace"),
1556                                }
1557                        };
1558                        $r->{$_}->{svm} = $svm;
1559                }
1560        } keys %$r;
1561
1562        $r;
1563}
1564
1565sub init_vars {
1566        $_gc_nr = $_gc_period = 1000;
1567        if (defined $_repack || defined $_repack_flags) {
1568               warn "Repack options are obsolete; they have no effect.\n";
1569        }
1570}
1571
1572sub verify_remotes_sanity {
1573        return unless -d $ENV{GIT_DIR};
1574        my %seen;
1575        foreach (command(qw/config -l/)) {
1576                if (m!^svn-remote\.(?:.+)\.fetch=.*:refs/remotes/(\S+)\s*$!) {
1577                        if ($seen{$1}) {
1578                                die "Remote ref refs/remote/$1 is tracked by",
1579                                    "\n  \"$_\"\nand\n  \"$seen{$1}\"\n",
1580                                    "Please resolve this ambiguity in ",
1581                                    "your git configuration file before ",
1582                                    "continuing\n";
1583                        }
1584                        $seen{$1} = $_;
1585                }
1586        }
1587}
1588
1589sub find_existing_remote {
1590        my ($url, $remotes) = @_;
1591        return undef if $no_reuse_existing;
1592        my $existing;
1593        foreach my $repo_id (keys %$remotes) {
1594                my $u = $remotes->{$repo_id}->{url} or next;
1595                next if $u ne $url;
1596                $existing = $repo_id;
1597                last;
1598        }
1599        $existing;
1600}
1601
1602sub init_remote_config {
1603        my ($self, $url, $no_write) = @_;
1604        $url =~ s!/+$!!; # strip trailing slash
1605        my $r = read_all_remotes();
1606        my $existing = find_existing_remote($url, $r);
1607        if ($existing) {
1608                unless ($no_write) {
1609                        print STDERR "Using existing ",
1610                                     "[svn-remote \"$existing\"]\n";
1611                }
1612                $self->{repo_id} = $existing;
1613        } elsif ($_minimize_url) {
1614                my $min_url = Git::SVN::Ra->new($url)->minimize_url;
1615                $existing = find_existing_remote($min_url, $r);
1616                if ($existing) {
1617                        unless ($no_write) {
1618                                print STDERR "Using existing ",
1619                                             "[svn-remote \"$existing\"]\n";
1620                        }
1621                        $self->{repo_id} = $existing;
1622                }
1623                if ($min_url ne $url) {
1624                        unless ($no_write) {
1625                                print STDERR "Using higher level of URL: ",
1626                                             "$url => $min_url\n";
1627                        }
1628                        my $old_path = $self->{path};
1629                        $self->{path} = $url;
1630                        $self->{path} =~ s!^\Q$min_url\E(/|$)!!;
1631                        if (length $old_path) {
1632                                $self->{path} .= "/$old_path";
1633                        }
1634                        $url = $min_url;
1635                }
1636        }
1637        my $orig_url;
1638        if (!$existing) {
1639                # verify that we aren't overwriting anything:
1640                $orig_url = eval {
1641                        command_oneline('config', '--get',
1642                                        "svn-remote.$self->{repo_id}.url")
1643                };
1644                if ($orig_url && ($orig_url ne $url)) {
1645                        die "svn-remote.$self->{repo_id}.url already set: ",
1646                            "$orig_url\nwanted to set to: $url\n";
1647                }
1648        }
1649        my ($xrepo_id, $xpath) = find_ref($self->refname);
1650        if (defined $xpath) {
1651                die "svn-remote.$xrepo_id.fetch already set to track ",
1652                    "$xpath:refs/remotes/", $self->refname, "\n";
1653        }
1654        unless ($no_write) {
1655                command_noisy('config',
1656                              "svn-remote.$self->{repo_id}.url", $url);
1657                $self->{path} =~ s{^/}{};
1658                command_noisy('config', '--add',
1659                              "svn-remote.$self->{repo_id}.fetch",
1660                              "$self->{path}:".$self->refname);
1661        }
1662        $self->{url} = $url;
1663}
1664
1665sub find_by_url { # repos_root and, path are optional
1666        my ($class, $full_url, $repos_root, $path) = @_;
1667
1668        return undef unless defined $full_url;
1669        remove_username($full_url);
1670        remove_username($repos_root) if defined $repos_root;
1671        my $remotes = read_all_remotes();
1672        if (defined $full_url && defined $repos_root && !defined $path) {
1673                $path = $full_url;
1674                $path =~ s#^\Q$repos_root\E(?:/|$)##;
1675        }
1676        foreach my $repo_id (keys %$remotes) {
1677                my $u = $remotes->{$repo_id}->{url} or next;
1678                remove_username($u);
1679                next if defined $repos_root && $repos_root ne $u;
1680
1681                my $fetch = $remotes->{$repo_id}->{fetch} || {};
1682                foreach (qw/branches tags/) {
1683                        resolve_local_globs($u, $fetch,
1684                                            $remotes->{$repo_id}->{$_});
1685                }
1686                my $p = $path;
1687                my $rwr = rewrite_root({repo_id => $repo_id});
1688                my $svm = $remotes->{$repo_id}->{svm}
1689                        if defined $remotes->{$repo_id}->{svm};
1690                unless (defined $p) {
1691                        $p = $full_url;
1692                        my $z = $u;
1693                        my $prefix = '';
1694                        if ($rwr) {
1695                                $z = $rwr;
1696                        } elsif (defined $svm) {
1697                                $z = $svm->{source};
1698                                $prefix = $svm->{replace};
1699                                $prefix =~ s#^\Q$u\E(?:/|$)##;
1700                                $prefix =~ s#/$##;
1701                        }
1702                        $p =~ s#^\Q$z\E(?:/|$)#$prefix# or next;
1703                }
1704                foreach my $f (keys %$fetch) {
1705                        next if $f ne $p;
1706                        return Git::SVN->new($fetch->{$f}, $repo_id, $f);
1707                }
1708        }
1709        undef;
1710}
1711
1712sub init {
1713        my ($class, $url, $path, $repo_id, $ref_id, $no_write) = @_;
1714        my $self = _new($class, $repo_id, $ref_id, $path);
1715        if (defined $url) {
1716                $self->init_remote_config($url, $no_write);
1717        }
1718        $self;
1719}
1720
1721sub find_ref {
1722        my ($ref_id) = @_;
1723        foreach (command(qw/config -l/)) {
1724                next unless m!^svn-remote\.(.+)\.fetch=
1725                              \s*(.*)\s*:\s*refs/remotes/(.+)\s*$!x;
1726                my ($repo_id, $path, $ref) = ($1, $2, $3);
1727                if ($ref eq $ref_id) {
1728                        $path = '' if ($path =~ m#^\./?#);
1729                        return ($repo_id, $path);
1730                }
1731        }
1732        (undef, undef, undef);
1733}
1734
1735sub new {
1736        my ($class, $ref_id, $repo_id, $path) = @_;
1737        if (defined $ref_id && !defined $repo_id && !defined $path) {
1738                ($repo_id, $path) = find_ref($ref_id);
1739                if (!defined $repo_id) {
1740                        die "Could not find a \"svn-remote.*.fetch\" key ",
1741                            "in the repository configuration matching: ",
1742                            "refs/remotes/$ref_id\n";
1743                }
1744        }
1745        my $self = _new($class, $repo_id, $ref_id, $path);
1746        if (!defined $self->{path} || !length $self->{path}) {
1747                my $fetch = command_oneline('config', '--get',
1748                                            "svn-remote.$repo_id.fetch",
1749                                            ":refs/remotes/$ref_id\$") or
1750                     die "Failed to read \"svn-remote.$repo_id.fetch\" ",
1751                         "\":refs/remotes/$ref_id\$\" in config\n";
1752                ($self->{path}, undef) = split(/\s*:\s*/, $fetch);
1753        }
1754        $self->{url} = command_oneline('config', '--get',
1755                                       "svn-remote.$repo_id.url") or
1756                  die "Failed to read \"svn-remote.$repo_id.url\" in config\n";
1757        $self->rebuild;
1758        $self;
1759}
1760
1761sub refname {
1762        my ($refname) = "refs/remotes/$_[0]->{ref_id}" ;
1763
1764        # It cannot end with a slash /, we'll throw up on this because
1765        # SVN can't have directories with a slash in their name, either:
1766        if ($refname =~ m{/$}) {
1767                die "ref: '$refname' ends with a trailing slash, this is ",
1768                    "not permitted by git nor Subversion\n";
1769        }
1770
1771        # It cannot have ASCII control character space, tilde ~, caret ^,
1772        # colon :, question-mark ?, asterisk *, space, or open bracket [
1773        # anywhere.
1774        #
1775        # Additionally, % must be escaped because it is used for escaping
1776        # and we want our escaped refname to be reversible
1777        $refname =~ s{([ \%~\^:\?\*\[\t])}{uc sprintf('%%%02x',ord($1))}eg;
1778
1779        # no slash-separated component can begin with a dot .
1780        # /.* becomes /%2E*
1781        $refname =~ s{/\.}{/%2E}g;
1782
1783        # It cannot have two consecutive dots .. anywhere
1784        # .. becomes %2E%2E
1785        $refname =~ s{\.\.}{%2E%2E}g;
1786
1787        return $refname;
1788}
1789
1790sub desanitize_refname {
1791        my ($refname) = @_;
1792        $refname =~ s{%(?:([0-9A-F]{2}))}{chr hex($1)}eg;
1793        return $refname;
1794}
1795
1796sub svm_uuid {
1797        my ($self) = @_;
1798        return $self->{svm}->{uuid} if $self->svm;
1799        $self->ra;
1800        unless ($self->{svm}) {
1801                die "SVM UUID not cached, and reading remotely failed\n";
1802        }
1803        $self->{svm}->{uuid};
1804}
1805
1806sub svm {
1807        my ($self) = @_;
1808        return $self->{svm} if $self->{svm};
1809        my $svm;
1810        # see if we have it in our config, first:
1811        eval {
1812                my $section = "svn-remote.$self->{repo_id}";
1813                $svm = {
1814                  source => tmp_config('--get', "$section.svm-source"),
1815                  uuid => tmp_config('--get', "$section.svm-uuid"),
1816                  replace => tmp_config('--get', "$section.svm-replace"),
1817                }
1818        };
1819        if ($svm && $svm->{source} && $svm->{uuid} && $svm->{replace}) {
1820                $self->{svm} = $svm;
1821        }
1822        $self->{svm};
1823}
1824
1825sub _set_svm_vars {
1826        my ($self, $ra) = @_;
1827        return $ra if $self->svm;
1828
1829        my @err = ( "useSvmProps set, but failed to read SVM properties\n",
1830                    "(svm:source, svm:uuid) ",
1831                    "from the following URLs:\n" );
1832        sub read_svm_props {
1833                my ($self, $ra, $path, $r) = @_;
1834                my $props = ($ra->get_dir($path, $r))[2];
1835                my $src = $props->{'svm:source'};
1836                my $uuid = $props->{'svm:uuid'};
1837                return undef if (!$src || !$uuid);
1838
1839                chomp($src, $uuid);
1840
1841                $uuid =~ m{^[0-9a-f\-]{30,}$}
1842                    or die "doesn't look right - svm:uuid is '$uuid'\n";
1843
1844                # the '!' is used to mark the repos_root!/relative/path
1845                $src =~ s{/?!/?}{/};
1846                $src =~ s{/+$}{}; # no trailing slashes please
1847                # username is of no interest
1848                $src =~ s{(^[a-z\+]*://)[^/@]*@}{$1};
1849
1850                my $replace = $ra->{url};
1851                $replace .= "/$path" if length $path;
1852
1853                my $section = "svn-remote.$self->{repo_id}";
1854                tmp_config("$section.svm-source", $src);
1855                tmp_config("$section.svm-replace", $replace);
1856                tmp_config("$section.svm-uuid", $uuid);
1857                $self->{svm} = {
1858                        source => $src,
1859                        uuid => $uuid,
1860                        replace => $replace
1861                };
1862        }
1863
1864        my $r = $ra->get_latest_revnum;
1865        my $path = $self->{path};
1866        my %tried;
1867        while (length $path) {
1868                unless ($tried{"$self->{url}/$path"}) {
1869                        return $ra if $self->read_svm_props($ra, $path, $r);
1870                        $tried{"$self->{url}/$path"} = 1;
1871                }
1872                $path =~ s#/?[^/]+$##;
1873        }
1874        die "Path: '$path' should be ''\n" if $path ne '';
1875        return $ra if $self->read_svm_props($ra, $path, $r);
1876        $tried{"$self->{url}/$path"} = 1;
1877
1878        if ($ra->{repos_root} eq $self->{url}) {
1879                die @err, (map { "  $_\n" } keys %tried), "\n";
1880        }
1881
1882        # nope, make sure we're connected to the repository root:
1883        my $ok;
1884        my @tried_b;
1885        $path = $ra->{svn_path};
1886        $ra = Git::SVN::Ra->new($ra->{repos_root});
1887        while (length $path) {
1888                unless ($tried{"$ra->{url}/$path"}) {
1889                        $ok = $self->read_svm_props($ra, $path, $r);
1890                        last if $ok;
1891                        $tried{"$ra->{url}/$path"} = 1;
1892                }
1893                $path =~ s#/?[^/]+$##;
1894        }
1895        die "Path: '$path' should be ''\n" if $path ne '';
1896        $ok ||= $self->read_svm_props($ra, $path, $r);
1897        $tried{"$ra->{url}/$path"} = 1;
1898        if (!$ok) {
1899                die @err, (map { "  $_\n" } keys %tried), "\n";
1900        }
1901        Git::SVN::Ra->new($self->{url});
1902}
1903
1904sub svnsync {
1905        my ($self) = @_;
1906        return $self->{svnsync} if $self->{svnsync};
1907
1908        if ($self->no_metadata) {
1909                die "Can't have both 'noMetadata' and ",
1910                    "'useSvnsyncProps' options set!\n";
1911        }
1912        if ($self->rewrite_root) {
1913                die "Can't have both 'useSvnsyncProps' and 'rewriteRoot' ",
1914                    "options set!\n";
1915        }
1916
1917        my $svnsync;
1918        # see if we have it in our config, first:
1919        eval {
1920                my $section = "svn-remote.$self->{repo_id}";
1921
1922                my $url = tmp_config('--get', "$section.svnsync-url");
1923                ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1924                   die "doesn't look right - svn:sync-from-url is '$url'\n";
1925
1926                my $uuid = tmp_config('--get', "$section.svnsync-uuid");
1927                ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1928                   die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1929
1930                $svnsync = { url => $url, uuid => $uuid }
1931        };
1932        if ($svnsync && $svnsync->{url} && $svnsync->{uuid}) {
1933                return $self->{svnsync} = $svnsync;
1934        }
1935
1936        my $err = "useSvnsyncProps set, but failed to read " .
1937                  "svnsync property: svn:sync-from-";
1938        my $rp = $self->ra->rev_proplist(0);
1939
1940        my $url = $rp->{'svn:sync-from-url'} or die $err . "url\n";
1941        ($url) = ($url =~ m{^([a-z\+]+://\S+)$}) or
1942                   die "doesn't look right - svn:sync-from-url is '$url'\n";
1943
1944        my $uuid = $rp->{'svn:sync-from-uuid'} or die $err . "uuid\n";
1945        ($uuid) = ($uuid =~ m{^([0-9a-f\-]{30,})$}) or
1946                   die "doesn't look right - svn:sync-from-uuid is '$uuid'\n";
1947
1948        my $section = "svn-remote.$self->{repo_id}";
1949        tmp_config('--add', "$section.svnsync-uuid", $uuid);
1950        tmp_config('--add', "$section.svnsync-url", $url);
1951        return $self->{svnsync} = { url => $url, uuid => $uuid };
1952}
1953
1954# this allows us to memoize our SVN::Ra UUID locally and avoid a
1955# remote lookup (useful for 'git svn log').
1956sub ra_uuid {
1957        my ($self) = @_;
1958        unless ($self->{ra_uuid}) {
1959                my $key = "svn-remote.$self->{repo_id}.uuid";
1960                my $uuid = eval { tmp_config('--get', $key) };
1961                if (!$@ && $uuid && $uuid =~ /^([a-f\d\-]{30,})$/) {
1962                        $self->{ra_uuid} = $uuid;
1963                } else {
1964                        die "ra_uuid called without URL\n" unless $self->{url};
1965                        $self->{ra_uuid} = $self->ra->get_uuid;
1966                        tmp_config('--add', $key, $self->{ra_uuid});
1967                }
1968        }
1969        $self->{ra_uuid};
1970}
1971
1972sub _set_repos_root {
1973        my ($self, $repos_root) = @_;
1974        my $k = "svn-remote.$self->{repo_id}.reposRoot";
1975        $repos_root ||= $self->ra->{repos_root};
1976        tmp_config($k, $repos_root);
1977        $repos_root;
1978}
1979
1980sub repos_root {
1981        my ($self) = @_;
1982        my $k = "svn-remote.$self->{repo_id}.reposRoot";
1983        eval { tmp_config('--get', $k) } || $self->_set_repos_root;
1984}
1985
1986sub ra {
1987        my ($self) = shift;
1988        my $ra = Git::SVN::Ra->new($self->{url});
1989        $self->_set_repos_root($ra->{repos_root});
1990        if ($self->use_svm_props && !$self->{svm}) {
1991                if ($self->no_metadata) {
1992                        die "Can't have both 'noMetadata' and ",
1993                            "'useSvmProps' options set!\n";
1994                } elsif ($self->use_svnsync_props) {
1995                        die "Can't have both 'useSvnsyncProps' and ",
1996                            "'useSvmProps' options set!\n";
1997                }
1998                $ra = $self->_set_svm_vars($ra);
1999                $self->{-want_revprops} = 1;
2000        }
2001        $ra;
2002}
2003
2004sub rel_path {
2005        my ($self) = @_;
2006        my $repos_root = $self->ra->{repos_root};
2007        return $self->{path} if ($self->{url} eq $repos_root);
2008        my $url = $self->{url} .
2009                  (length $self->{path} ? "/$self->{path}" : $self->{path});
2010        $url =~ s!^\Q$repos_root\E(?:/+|$)!!g;
2011        $url;
2012}
2013
2014# prop_walk(PATH, REV, SUB)
2015# -------------------------
2016# Recursively traverse PATH at revision REV and invoke SUB for each
2017# directory that contains a SVN property.  SUB will be invoked as
2018# follows:  &SUB(gs, path, props);  where `gs' is this instance of
2019# Git::SVN, `path' the path to the directory where the properties
2020# `props' were found.  The `path' will be relative to point of checkout,
2021# that is, if url://repo/trunk is the current Git branch, and that
2022# directory contains a sub-directory `d', SUB will be invoked with `/d/'
2023# as `path' (note the trailing `/').
2024sub prop_walk {
2025        my ($self, $path, $rev, $sub) = @_;
2026
2027        $path =~ s#^/##;
2028        my ($dirent, undef, $props) = $self->ra->get_dir($path, $rev);
2029        $path =~ s#^/*#/#g;
2030        my $p = $path;
2031        # Strip the irrelevant part of the path.
2032        $p =~ s#^/+\Q$self->{path}\E(/|$)#/#;
2033        # Ensure the path is terminated by a `/'.
2034        $p =~ s#/*$#/#;
2035
2036        # The properties contain all the internal SVN stuff nobody
2037        # (usually) cares about.
2038        my $interesting_props = 0;
2039        foreach (keys %{$props}) {
2040                # If it doesn't start with `svn:', it must be a
2041                # user-defined property.
2042                ++$interesting_props and next if $_ !~ /^svn:/;
2043                # FIXME: Fragile, if SVN adds new public properties,
2044                # this needs to be updated.
2045                ++$interesting_props if /^svn:(?:ignore|keywords|executable
2046                                                 |eol-style|mime-type
2047                                                 |externals|needs-lock)$/x;
2048        }
2049        &$sub($self, $p, $props) if $interesting_props;
2050
2051        foreach (sort keys %$dirent) {
2052                next if $dirent->{$_}->{kind} != $SVN::Node::dir;
2053                $self->prop_walk($self->{path} . $p . $_, $rev, $sub);
2054        }
2055}
2056
2057sub last_rev { ($_[0]->last_rev_commit)[0] }
2058sub last_commit { ($_[0]->last_rev_commit)[1] }
2059
2060# returns the newest SVN revision number and newest commit SHA1
2061sub last_rev_commit {
2062        my ($self) = @_;
2063        if (defined $self->{last_rev} && defined $self->{last_commit}) {
2064                return ($self->{last_rev}, $self->{last_commit});
2065        }
2066        my $c = ::verify_ref($self->refname.'^0');
2067        if ($c && !$self->use_svm_props && !$self->no_metadata) {
2068                my $rev = (::cmt_metadata($c))[1];
2069                if (defined $rev) {
2070                        ($self->{last_rev}, $self->{last_commit}) = ($rev, $c);
2071                        return ($rev, $c);
2072                }
2073        }
2074        my $map_path = $self->map_path;
2075        unless (-e $map_path) {
2076                ($self->{last_rev}, $self->{last_commit}) = (undef, undef);
2077                return (undef, undef);
2078        }
2079        my ($rev, $commit) = $self->rev_map_max(1);
2080        ($self->{last_rev}, $self->{last_commit}) = ($rev, $commit);
2081        return ($rev, $commit);
2082}
2083
2084sub get_fetch_range {
2085        my ($self, $min, $max) = @_;
2086        $max ||= $self->ra->get_latest_revnum;
2087        $min ||= $self->rev_map_max;
2088        (++$min, $max);
2089}
2090
2091sub tmp_config {
2092        my (@args) = @_;
2093        my $old_def_config = "$ENV{GIT_DIR}/svn/config";
2094        my $config = "$ENV{GIT_DIR}/svn/.metadata";
2095        if (! -f $config && -f $old_def_config) {
2096                rename $old_def_config, $config or
2097                       die "Failed rename $old_def_config => $config: $!\n";
2098        }
2099        my $old_config = $ENV{GIT_CONFIG};
2100        $ENV{GIT_CONFIG} = $config;
2101        $@ = undef;
2102        my @ret = eval {
2103                unless (-f $config) {
2104                        mkfile($config);
2105                        open my $fh, '>', $config or
2106                            die "Can't open $config: $!\n";
2107                        print $fh "; This file is used internally by ",
2108                                  "git-svn\n" or die
2109                                  "Couldn't write to $config: $!\n";
2110                        print $fh "; You should not have to edit it\n" or
2111                              die "Couldn't write to $config: $!\n";
2112                        close $fh or die "Couldn't close $config: $!\n";
2113                }
2114                command('config', @args);
2115        };
2116        my $err = $@;
2117        if (defined $old_config) {
2118                $ENV{GIT_CONFIG} = $old_config;
2119        } else {
2120                delete $ENV{GIT_CONFIG};
2121        }
2122        die $err if $err;
2123        wantarray ? @ret : $ret[0];
2124}
2125
2126sub tmp_index_do {
2127        my ($self, $sub) = @_;
2128        my $old_index = $ENV{GIT_INDEX_FILE};
2129        $ENV{GIT_INDEX_FILE} = $self->{index};
2130        $@ = undef;
2131        my @ret = eval {
2132                my ($dir, $base) = ($self->{index} =~ m#^(.*?)/?([^/]+)$#);
2133                mkpath([$dir]) unless -d $dir;
2134                &$sub;
2135        };
2136        my $err = $@;
2137        if (defined $old_index) {
2138                $ENV{GIT_INDEX_FILE} = $old_index;
2139        } else {
2140                delete $ENV{GIT_INDEX_FILE};
2141        }
2142        die $err if $err;
2143        wantarray ? @ret : $ret[0];
2144}
2145
2146sub assert_index_clean {
2147        my ($self, $treeish) = @_;
2148
2149        $self->tmp_index_do(sub {
2150                command_noisy('read-tree', $treeish) unless -e $self->{index};
2151                my $x = command_oneline('write-tree');
2152                my ($y) = (command(qw/cat-file commit/, $treeish) =~
2153                           /^tree ($::sha1)/mo);
2154                return if $y eq $x;
2155
2156                warn "Index mismatch: $y != $x\nrereading $treeish\n";
2157                unlink $self->{index} or die "unlink $self->{index}: $!\n";
2158                command_noisy('read-tree', $treeish);
2159                $x = command_oneline('write-tree');
2160                if ($y ne $x) {
2161                        ::fatal "trees ($treeish) $y != $x\n",
2162                                "Something is seriously wrong...";
2163                }
2164        });
2165}
2166
2167sub get_commit_parents {
2168        my ($self, $log_entry) = @_;
2169        my (%seen, @ret, @tmp);
2170        # legacy support for 'set-tree'; this is only used by set_tree_cb:
2171        if (my $ip = $self->{inject_parents}) {
2172                if (my $commit = delete $ip->{$log_entry->{revision}}) {
2173                        push @tmp, $commit;
2174                }
2175        }
2176        if (my $cur = ::verify_ref($self->refname.'^0')) {
2177                push @tmp, $cur;
2178        }
2179        if (my $ipd = $self->{inject_parents_dcommit}) {
2180                if (my $commit = delete $ipd->{$log_entry->{revision}}) {
2181                        push @tmp, @$commit;
2182                }
2183        }
2184        push @tmp, $_ foreach (@{$log_entry->{parents}}, @tmp);
2185        while (my $p = shift @tmp) {
2186                next if $seen{$p};
2187                $seen{$p} = 1;
2188                push @ret, $p;
2189                # MAXPARENT is defined to 16 in commit-tree.c:
2190                last if @ret >= 16;
2191        }
2192        if (@tmp) {
2193                die "r$log_entry->{revision}: No room for parents:\n\t",
2194                    join("\n\t", @tmp), "\n";
2195        }
2196        @ret;
2197}
2198
2199sub rewrite_root {
2200        my ($self) = @_;
2201        return $self->{-rewrite_root} if exists $self->{-rewrite_root};
2202        my $k = "svn-remote.$self->{repo_id}.rewriteRoot";
2203        my $rwr = eval { command_oneline(qw/config --get/, $k) };
2204        if ($rwr) {
2205                $rwr =~ s#/+$##;
2206                if ($rwr !~ m#^[a-z\+]+://#) {
2207                        die "$rwr is not a valid URL (key: $k)\n";
2208                }
2209        }
2210        $self->{-rewrite_root} = $rwr;
2211}
2212
2213sub metadata_url {
2214        my ($self) = @_;
2215        ($self->rewrite_root || $self->{url}) .
2216           (length $self->{path} ? '/' . $self->{path} : '');
2217}
2218
2219sub full_url {
2220        my ($self) = @_;
2221        $self->{url} . (length $self->{path} ? '/' . $self->{path} : '');
2222}
2223
2224
2225sub set_commit_header_env {
2226        my ($log_entry) = @_;
2227        my %env;
2228        foreach my $ned (qw/NAME EMAIL DATE/) {
2229                foreach my $ac (qw/AUTHOR COMMITTER/) {
2230                        $env{"GIT_${ac}_${ned}"} = $ENV{"GIT_${ac}_${ned}"};
2231                }
2232        }
2233
2234        $ENV{GIT_AUTHOR_NAME} = $log_entry->{name};
2235        $ENV{GIT_AUTHOR_EMAIL} = $log_entry->{email};
2236        $ENV{GIT_AUTHOR_DATE} = $ENV{GIT_COMMITTER_DATE} = $log_entry->{date};
2237
2238        $ENV{GIT_COMMITTER_NAME} = (defined $log_entry->{commit_name})
2239                                                ? $log_entry->{commit_name}
2240                                                : $log_entry->{name};
2241        $ENV{GIT_COMMITTER_EMAIL} = (defined $log_entry->{commit_email})
2242                                                ? $log_entry->{commit_email}
2243                                                : $log_entry->{email};
2244        \%env;
2245}
2246
2247sub restore_commit_header_env {
2248        my ($env) = @_;
2249        foreach my $ned (qw/NAME EMAIL DATE/) {
2250                foreach my $ac (qw/AUTHOR COMMITTER/) {
2251                        my $k = "GIT_${ac}_${ned}";
2252                        if (defined $env->{$k}) {
2253                                $ENV{$k} = $env->{$k};
2254                        } else {
2255                                delete $ENV{$k};
2256                        }
2257                }
2258        }
2259}
2260
2261sub gc {
2262        command_noisy('gc', '--auto');
2263};
2264
2265sub do_git_commit {
2266        my ($self, $log_entry) = @_;
2267        my $lr = $self->last_rev;
2268        if (defined $lr && $lr >= $log_entry->{revision}) {
2269                die "Last fetched revision of ", $self->refname,
2270                    " was r$lr, but we are about to fetch: ",
2271                    "r$log_entry->{revision}!\n";
2272        }
2273        if (my $c = $self->rev_map_get($log_entry->{revision})) {
2274                croak "$log_entry->{revision} = $c already exists! ",
2275                      "Why are we refetching it?\n";
2276        }
2277        my $old_env = set_commit_header_env($log_entry);
2278        my $tree = $log_entry->{tree};
2279        if (!defined $tree) {
2280                $tree = $self->tmp_index_do(sub {
2281                                            command_oneline('write-tree') });
2282        }
2283        die "Tree is not a valid sha1: $tree\n" if $tree !~ /^$::sha1$/o;
2284
2285        my @exec = ('git', 'commit-tree', $tree);
2286        foreach ($self->get_commit_parents($log_entry)) {
2287                push @exec, '-p', $_;
2288        }
2289        defined(my $pid = open3(my $msg_fh, my $out_fh, '>&STDERR', @exec))
2290                                                                   or croak $!;
2291        binmode $msg_fh;
2292
2293        # we always get UTF-8 from SVN, but we may want our commits in
2294        # a different encoding.
2295        if (my $enc = Git::config('i18n.commitencoding')) {
2296                require Encode;
2297                Encode::from_to($log_entry->{log}, 'UTF-8', $enc);
2298        }
2299        print $msg_fh $log_entry->{log} or croak $!;
2300        restore_commit_header_env($old_env);
2301        unless ($self->no_metadata) {
2302                print $msg_fh "\ngit-svn-id: $log_entry->{metadata}\n"
2303                              or croak $!;
2304        }
2305        $msg_fh->flush == 0 or croak $!;
2306        close $msg_fh or croak $!;
2307        chomp(my $commit = do { local $/; <$out_fh> });
2308        close $out_fh or croak $!;
2309        waitpid $pid, 0;
2310        croak $? if $?;
2311        if ($commit !~ /^$::sha1$/o) {
2312                die "Failed to commit, invalid sha1: $commit\n";
2313        }
2314
2315        $self->rev_map_set($log_entry->{revision}, $commit, 1);
2316
2317        $self->{last_rev} = $log_entry->{revision};
2318        $self->{last_commit} = $commit;
2319        print "r$log_entry->{revision}";
2320        if (defined $log_entry->{svm_revision}) {
2321                 print " (\@$log_entry->{svm_revision})";
2322                 $self->rev_map_set($log_entry->{svm_revision}, $commit,
2323                                   0, $self->svm_uuid);
2324        }
2325        print " = $commit ($self->{ref_id})\n";
2326        if (--$_gc_nr == 0) {
2327                $_gc_nr = $_gc_period;
2328                gc();
2329        }
2330        return $commit;
2331}
2332
2333sub match_paths {
2334        my ($self, $paths, $r) = @_;
2335        return 1 if $self->{path} eq '';
2336        if (my $path = $paths->{"/$self->{path}"}) {
2337                return ($path->{action} eq 'D') ? 0 : 1;
2338        }
2339        $self->{path_regex} ||= qr/^\/\Q$self->{path}\E\//;
2340        if (grep /$self->{path_regex}/, keys %$paths) {
2341                return 1;
2342        }
2343        my $c = '';
2344        foreach (split m#/#, $self->{path}) {
2345                $c .= "/$_";
2346                next unless ($paths->{$c} &&
2347                             ($paths->{$c}->{action} =~ /^[AR]$/));
2348                if ($self->ra->check_path($self->{path}, $r) ==
2349                    $SVN::Node::dir) {
2350                        return 1;
2351                }
2352        }
2353        return 0;
2354}
2355
2356sub find_parent_branch {
2357        my ($self, $paths, $rev) = @_;
2358        return undef unless $self->follow_parent;
2359        unless (defined $paths) {
2360                my $err_handler = $SVN::Error::handler;
2361                $SVN::Error::handler = \&Git::SVN::Ra::skip_unknown_revs;
2362                $self->ra->get_log([$self->{path}], $rev, $rev, 0, 1, 1, sub {
2363                                   $paths =
2364                                      Git::SVN::Ra::dup_changed_paths($_[0]) });
2365                $SVN::Error::handler = $err_handler;
2366        }
2367        return undef unless defined $paths;
2368
2369        # look for a parent from another branch:
2370        my @b_path_components = split m#/#, $self->rel_path;
2371        my @a_path_components;
2372        my $i;
2373        while (@b_path_components) {
2374                $i = $paths->{'/'.join('/', @b_path_components)};
2375                last if $i && defined $i->{copyfrom_path};
2376                unshift(@a_path_components, pop(@b_path_components));
2377        }
2378        return undef unless defined $i && defined $i->{copyfrom_path};
2379        my $branch_from = $i->{copyfrom_path};
2380        if (@a_path_components) {
2381                print STDERR "branch_from: $branch_from => ";
2382                $branch_from .= '/'.join('/', @a_path_components);
2383                print STDERR $branch_from, "\n";
2384        }
2385        my $r = $i->{copyfrom_rev};
2386        my $repos_root = $self->ra->{repos_root};
2387        my $url = $self->ra->{url};
2388        my $new_url = $repos_root . $branch_from;
2389        print STDERR  "Found possible branch point: ",
2390                      "$new_url => ", $self->full_url, ", $r\n";
2391        $branch_from =~ s#^/##;
2392        my $gs = $self->other_gs($new_url, $url, $repos_root,
2393                                 $branch_from, $r, $self->{ref_id});
2394        my ($r0, $parent) = $gs->find_rev_before($r, 1);
2395        {
2396                my ($base, $head);
2397                if (!defined $r0 || !defined $parent) {
2398                        ($base, $head) = parse_revision_argument(0, $r);
2399                } else {
2400                        if ($r0 < $r) {
2401                                $gs->ra->get_log([$gs->{path}], $r0 + 1, $r, 1,
2402                                        0, 1, sub { $base = $_[1] - 1 });
2403                        }
2404                }
2405                if (defined $base && $base <= $r) {
2406                        $gs->fetch($base, $r);
2407                }
2408                ($r0, $parent) = $gs->find_rev_before($r, 1);
2409        }
2410        if (defined $r0 && defined $parent) {
2411                print STDERR "Found branch parent: ($self->{ref_id}) $parent\n";
2412                my $ed;
2413                if ($self->ra->can_do_switch) {
2414                        $self->assert_index_clean($parent);
2415                        print STDERR "Following parent with do_switch\n";
2416                        # do_switch works with svn/trunk >= r22312, but that
2417                        # is not included with SVN 1.4.3 (the latest version
2418                        # at the moment), so we can't rely on it
2419                        $self->{last_commit} = $parent;
2420                        $ed = SVN::Git::Fetcher->new($self, $gs->{path});
2421                        $gs->ra->gs_do_switch($r0, $rev, $gs,
2422                                              $self->full_url, $ed)
2423                          or die "SVN connection failed somewhere...\n";
2424                } elsif ($self->ra->trees_match($new_url, $r0,
2425                                                $self->full_url, $rev)) {
2426                        print STDERR "Trees match:\n",
2427                                     "  $new_url\@$r0\n",
2428                                     "  ${\$self->full_url}\@$rev\n",
2429                                     "Following parent with no changes\n";
2430                        $self->tmp_index_do(sub {
2431                            command_noisy('read-tree', $parent);
2432                        });
2433                        $self->{last_commit} = $parent;
2434                } else {
2435                        print STDERR "Following parent with do_update\n";
2436                        $ed = SVN::Git::Fetcher->new($self);
2437                        $self->ra->gs_do_update($rev, $rev, $self, $ed)
2438                          or die "SVN connection failed somewhere...\n";
2439                }
2440                print STDERR "Successfully followed parent\n";
2441                return $self->make_log_entry($rev, [$parent], $ed);
2442        }
2443        return undef;
2444}
2445
2446sub do_fetch {
2447        my ($self, $paths, $rev) = @_;
2448        my $ed;
2449        my ($last_rev, @parents);
2450        if (my $lc = $self->last_commit) {
2451                # we can have a branch that was deleted, then re-added
2452                # under the same name but copied from another path, in
2453                # which case we'll have multiple parents (we don't
2454                # want to break the original ref, nor lose copypath info):
2455                if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2456                        push @{$log_entry->{parents}}, $lc;
2457                        return $log_entry;
2458                }
2459                $ed = SVN::Git::Fetcher->new($self);
2460                $last_rev = $self->{last_rev};
2461                $ed->{c} = $lc;
2462                @parents = ($lc);
2463        } else {
2464                $last_rev = $rev;
2465                if (my $log_entry = $self->find_parent_branch($paths, $rev)) {
2466                        return $log_entry;
2467                }
2468                $ed = SVN::Git::Fetcher->new($self);
2469        }
2470        unless ($self->ra->gs_do_update($last_rev, $rev, $self, $ed)) {
2471                die "SVN connection failed somewhere...\n";
2472        }
2473        $self->make_log_entry($rev, \@parents, $ed);
2474}
2475
2476sub get_untracked {
2477        my ($self, $ed) = @_;
2478        my @out;
2479        my $h = $ed->{empty};
2480        foreach (sort keys %$h) {
2481                my $act = $h->{$_} ? '+empty_dir' : '-empty_dir';
2482                push @out, "  $act: " . uri_encode($_);
2483                warn "W: $act: $_\n";
2484        }
2485        foreach my $t (qw/dir_prop file_prop/) {
2486                $h = $ed->{$t} or next;
2487                foreach my $path (sort keys %$h) {
2488                        my $ppath = $path eq '' ? '.' : $path;
2489                        foreach my $prop (sort keys %{$h->{$path}}) {
2490                                next if $SKIP_PROP{$prop};
2491                                my $v = $h->{$path}->{$prop};
2492                                my $t_ppath_prop = "$t: " .
2493                                                    uri_encode($ppath) . ' ' .
2494                                                    uri_encode($prop);
2495                                if (defined $v) {
2496                                        push @out, "  +$t_ppath_prop " .
2497                                                   uri_encode($v);
2498                                } else {
2499                                        push @out, "  -$t_ppath_prop";
2500                                }
2501                        }
2502                }
2503        }
2504        foreach my $t (qw/absent_file absent_directory/) {
2505                $h = $ed->{$t} or next;
2506                foreach my $parent (sort keys %$h) {
2507                        foreach my $path (sort @{$h->{$parent}}) {
2508                                push @out, "  $t: " .
2509                                           uri_encode("$parent/$path");
2510                                warn "W: $t: $parent/$path ",
2511                                     "Insufficient permissions?\n";
2512                        }
2513                }
2514        }
2515        \@out;
2516}
2517
2518# parse_svn_date(DATE)
2519# --------------------
2520# Given a date (in UTC) from Subversion, return a string in the format
2521# "<TZ Offset> <local date/time>" that Git will use.
2522#
2523# By default the parsed date will be in UTC; if $Git::SVN::_localtime
2524# is true we'll convert it to the local timezone instead.
2525sub parse_svn_date {
2526        my $date = shift || return '+0000 1970-01-01 00:00:00';
2527        my ($Y,$m,$d,$H,$M,$S) = ($date =~ /^(\d{4})\-(\d\d)\-(\d\d)T
2528                                            (\d\d)\:(\d\d)\:(\d\d).\d+Z$/x) or
2529                                         croak "Unable to parse date: $date\n";
2530        my $parsed_date;    # Set next.
2531
2532        if ($Git::SVN::_localtime) {
2533                # Translate the Subversion datetime to an epoch time.
2534                # Begin by switching ourselves to $date's timezone, UTC.
2535                my $old_env_TZ = $ENV{TZ};
2536                $ENV{TZ} = 'UTC';
2537
2538                my $epoch_in_UTC =
2539                    POSIX::strftime('%s', $S, $M, $H, $d, $m - 1, $Y - 1900);
2540
2541                # Determine our local timezone (including DST) at the
2542                # time of $epoch_in_UTC.  $Git::SVN::Log::TZ stored the
2543                # value of TZ, if any, at the time we were run.
2544                if (defined $Git::SVN::Log::TZ) {
2545                        $ENV{TZ} = $Git::SVN::Log::TZ;
2546                } else {
2547                        delete $ENV{TZ};
2548                }
2549
2550                my $our_TZ =
2551                    POSIX::strftime('%Z', $S, $M, $H, $d, $m - 1, $Y - 1900);
2552
2553                # This converts $epoch_in_UTC into our local timezone.
2554                my ($sec, $min, $hour, $mday, $mon, $year,
2555                    $wday, $yday, $isdst) = localtime($epoch_in_UTC);
2556
2557                $parsed_date = sprintf('%s %04d-%02d-%02d %02d:%02d:%02d',
2558                                       $our_TZ, $year + 1900, $mon + 1,
2559                                       $mday, $hour, $min, $sec);
2560
2561                # Reset us to the timezone in effect when we entered
2562                # this routine.
2563                if (defined $old_env_TZ) {
2564                        $ENV{TZ} = $old_env_TZ;
2565                } else {
2566                        delete $ENV{TZ};
2567                }
2568        } else {
2569                $parsed_date = "+0000 $Y-$m-$d $H:$M:$S";
2570        }
2571
2572        return $parsed_date;
2573}
2574
2575sub other_gs {
2576        my ($self, $new_url, $url, $repos_root,
2577            $branch_from, $r, $old_ref_id) = @_;
2578        my $gs = Git::SVN->find_by_url($new_url, $repos_root, $branch_from);
2579        unless ($gs) {
2580                my $ref_id = $old_ref_id;
2581                $ref_id =~ s/\@\d+$//;
2582                $ref_id .= "\@$r";
2583                # just grow a tail if we're not unique enough :x
2584                $ref_id .= '-' while find_ref($ref_id);
2585                print STDERR "Initializing parent: $ref_id\n";
2586                my ($u, $p, $repo_id) = ($new_url, '', $ref_id);
2587                if ($u =~ s#^\Q$url\E(/|$)##) {
2588                        $p = $u;
2589                        $u = $url;
2590                        $repo_id = $self->{repo_id};
2591                }
2592                $gs = Git::SVN->init($u, $p, $repo_id, $ref_id, 1);
2593        }
2594        $gs
2595}
2596
2597sub check_author {
2598        my ($author) = @_;
2599        if (!defined $author || length $author == 0) {
2600                $author = '(no author)';
2601        } elsif (defined $::_authors && ! defined $::users{$author}) {
2602                die "Author: $author not defined in $::_authors file\n";
2603        }
2604        $author;
2605}
2606
2607sub make_log_entry {
2608        my ($self, $rev, $parents, $ed) = @_;
2609        my $untracked = $self->get_untracked($ed);
2610
2611        open my $un, '>>', "$self->{dir}/unhandled.log" or croak $!;
2612        print $un "r$rev\n" or croak $!;
2613        print $un $_, "\n" foreach @$untracked;
2614        my %log_entry = ( parents => $parents || [], revision => $rev,
2615                          log => '');
2616
2617        my $headrev;
2618        my $logged = delete $self->{logged_rev_props};
2619        if (!$logged || $self->{-want_revprops}) {
2620                my $rp = $self->ra->rev_proplist($rev);
2621                foreach (sort keys %$rp) {
2622                        my $v = $rp->{$_};
2623                        if (/^svn:(author|date|log)$/) {
2624                                $log_entry{$1} = $v;
2625                        } elsif ($_ eq 'svm:headrev') {
2626                                $headrev = $v;
2627                        } else {
2628                                print $un "  rev_prop: ", uri_encode($_), ' ',
2629                                          uri_encode($v), "\n";
2630                        }
2631                }
2632        } else {
2633                map { $log_entry{$_} = $logged->{$_} } keys %$logged;
2634        }
2635        close $un or croak $!;
2636
2637        $log_entry{date} = parse_svn_date($log_entry{date});
2638        $log_entry{log} .= "\n";
2639        my $author = $log_entry{author} = check_author($log_entry{author});
2640        my ($name, $email) = defined $::users{$author} ? @{$::users{$author}}
2641                                                       : ($author, undef);
2642
2643        my ($commit_name, $commit_email) = ($name, $email);
2644        if ($_use_log_author) {
2645                my $name_field;
2646                if ($log_entry{log} =~ /From:\s+(.*\S)\s*\n/i) {
2647                        $name_field = $1;
2648                } elsif ($log_entry{log} =~ /Signed-off-by:\s+(.*\S)\s*\n/i) {
2649                        $name_field = $1;
2650                }
2651                if (!defined $name_field) {
2652                        if (!defined $email) {
2653                                $email = $name;
2654                        }
2655                } elsif ($name_field =~ /(.*?)\s+<(.*)>/) {
2656                        ($name, $email) = ($1, $2);
2657                } elsif ($name_field =~ /(.*)@/) {
2658                        ($name, $email) = ($1, $name_field);
2659                } else {
2660                        ($name, $email) = ($name_field, $name_field);
2661                }
2662        }
2663        if (defined $headrev && $self->use_svm_props) {
2664                if ($self->rewrite_root) {
2665                        die "Can't have both 'useSvmProps' and 'rewriteRoot' ",
2666                            "options set!\n";
2667                }
2668                my ($uuid, $r) = $headrev =~ m{^([a-f\d\-]{30,}):(\d+)$};
2669                # we don't want "SVM: initializing mirror for junk" ...
2670                return undef if $r == 0;
2671                my $svm = $self->svm;
2672                if ($uuid ne $svm->{uuid}) {
2673                        die "UUID mismatch on SVM path:\n",
2674                            "expected: $svm->{uuid}\n",
2675                            "     got: $uuid\n";
2676                }
2677                my $full_url = $self->full_url;
2678                $full_url =~ s#^\Q$svm->{replace}\E(/|$)#$svm->{source}$1# or
2679                             die "Failed to replace '$svm->{replace}' with ",
2680                                 "'$svm->{source}' in $full_url\n";
2681                # throw away username for storing in records
2682                remove_username($full_url);
2683                $log_entry{metadata} = "$full_url\@$r $uuid";
2684                $log_entry{svm_revision} = $r;
2685                $email ||= "$author\@$uuid";
2686                $commit_email ||= "$author\@$uuid";
2687        } elsif ($self->use_svnsync_props) {
2688                my $full_url = $self->svnsync->{url};
2689                $full_url .= "/$self->{path}" if length $self->{path};
2690                remove_username($full_url);
2691                my $uuid = $self->svnsync->{uuid};
2692                $log_entry{metadata} = "$full_url\@$rev $uuid";
2693                $email ||= "$author\@$uuid";
2694                $commit_email ||= "$author\@$uuid";
2695        } else {
2696                my $url = $self->metadata_url;
2697                remove_username($url);
2698                $log_entry{metadata} = "$url\@$rev " .
2699                                       $self->ra->get_uuid;
2700                $email ||= "$author\@" . $self->ra->get_uuid;
2701                $commit_email ||= "$author\@" . $self->ra->get_uuid;
2702        }
2703        $log_entry{name} = $name;
2704        $log_entry{email} = $email;
2705        $log_entry{commit_name} = $commit_name;
2706        $log_entry{commit_email} = $commit_email;
2707        \%log_entry;
2708}
2709
2710sub fetch {
2711        my ($self, $min_rev, $max_rev, @parents) = @_;
2712        my ($last_rev, $last_commit) = $self->last_rev_commit;
2713        my ($base, $head) = $self->get_fetch_range($min_rev, $max_rev);
2714        $self->ra->gs_fetch_loop_common($base, $head, [$self]);
2715}
2716
2717sub set_tree_cb {
2718        my ($self, $log_entry, $tree, $rev, $date, $author) = @_;
2719        $self->{inject_parents} = { $rev => $tree };
2720        $self->fetch(undef, undef);
2721}
2722
2723sub set_tree {
2724        my ($self, $tree) = (shift, shift);
2725        my $log_entry = ::get_commit_entry($tree);
2726        unless ($self->{last_rev}) {
2727                ::fatal("Must have an existing revision to commit");
2728        }
2729        my %ed_opts = ( r => $self->{last_rev},
2730                        log => $log_entry->{log},
2731                        ra => $self->ra,
2732                        tree_a => $self->{last_commit},
2733                        tree_b => $tree,
2734                        editor_cb => sub {
2735                               $self->set_tree_cb($log_entry, $tree, @_) },
2736                        svn_path => $self->{path} );
2737        if (!SVN::Git::Editor->new(\%ed_opts)->apply_diff) {
2738                print "No changes\nr$self->{last_rev} = $tree\n";
2739        }
2740}
2741
2742sub rebuild_from_rev_db {
2743        my ($self, $path) = @_;
2744        my $r = -1;
2745        open my $fh, '<', $path or croak "open: $!";
2746        binmode $fh or croak "binmode: $!";
2747        while (<$fh>) {
2748                length($_) == 41 or croak "inconsistent size in ($_) != 41";
2749                chomp($_);
2750                ++$r;
2751                next if $_ eq ('0' x 40);
2752                $self->rev_map_set($r, $_);
2753                print "r$r = $_\n";
2754        }
2755        close $fh or croak "close: $!";
2756        unlink $path or croak "unlink: $!";
2757}
2758
2759sub rebuild {
2760        my ($self) = @_;
2761        my $map_path = $self->map_path;
2762        my $partial = (-e $map_path && ! -z $map_path);
2763        return unless ::verify_ref($self->refname.'^0');
2764        if (!$partial && ($self->use_svm_props || $self->no_metadata)) {
2765                my $rev_db = $self->rev_db_path;
2766                $self->rebuild_from_rev_db($rev_db);
2767                if ($self->use_svm_props) {
2768                        my $svm_rev_db = $self->rev_db_path($self->svm_uuid);
2769                        $self->rebuild_from_rev_db($svm_rev_db);
2770                }
2771                $self->unlink_rev_db_symlink;
2772                return;
2773        }
2774        print "Rebuilding $map_path ...\n" if (!$partial);
2775        my ($base_rev, $head) = ($partial ? $self->rev_map_max_norebuild(1) :
2776                (undef, undef));
2777        my ($log, $ctx) =
2778            command_output_pipe(qw/rev-list --pretty=raw --no-color --reverse/,
2779                                ($head ? "$head.." : "") . $self->refname,
2780                                '--');
2781        my $metadata_url = $self->metadata_url;
2782        remove_username($metadata_url);
2783        my $svn_uuid = $self->ra_uuid;
2784        my $c;
2785        while (<$log>) {
2786                if ( m{^commit ($::sha1)$} ) {
2787                        $c = $1;
2788                        next;
2789                }
2790                next unless s{^\s*(git-svn-id:)}{$1};
2791                my ($url, $rev, $uuid) = ::extract_metadata($_);
2792                remove_username($url);
2793
2794                # ignore merges (from set-tree)
2795                next if (!defined $rev || !$uuid);
2796
2797                # if we merged or otherwise started elsewhere, this is
2798                # how we break out of it
2799                if (($uuid ne $svn_uuid) ||
2800                    ($metadata_url && $url && ($url ne $metadata_url))) {
2801                        next;
2802                }
2803                if ($partial && $head) {
2804                        print "Partial-rebuilding $map_path ...\n";
2805                        print "Currently at $base_rev = $head\n";
2806                        $head = undef;
2807                }
2808
2809                $self->rev_map_set($rev, $c);
2810                print "r$rev = $c\n";
2811        }
2812        command_close_pipe($log, $ctx);
2813        print "Done rebuilding $map_path\n" if (!$partial || !$head);
2814        my $rev_db_path = $self->rev_db_path;
2815        if (-f $self->rev_db_path) {
2816                unlink $self->rev_db_path or croak "unlink: $!";
2817        }
2818        $self->unlink_rev_db_symlink;
2819}
2820
2821# rev_map:
2822# Tie::File seems to be prone to offset errors if revisions get sparse,
2823# it's not that fast, either.  Tie::File is also not in Perl 5.6.  So
2824# one of my favorite modules is out :<  Next up would be one of the DBM
2825# modules, but I'm not sure which is most portable...
2826#
2827# This is the replacement for the rev_db format, which was too big
2828# and inefficient for large repositories with a lot of sparse history
2829# (mainly tags)
2830#
2831# The format is this:
2832#   - 24 bytes for every record,
2833#     * 4 bytes for the integer representing an SVN revision number
2834#     * 20 bytes representing the sha1 of a git commit
2835#   - No empty padding records like the old format
2836#     (except the last record, which can be overwritten)
2837#   - new records are written append-only since SVN revision numbers
2838#     increase monotonically
2839#   - lookups on SVN revision number are done via a binary search
2840#   - Piping the file to xxd -c24 is a good way of dumping it for
2841#     viewing or editing (piped back through xxd -r), should the need
2842#     ever arise.
2843#   - The last record can be padding revision with an all-zero sha1
2844#     This is used to optimize fetch performance when using multiple
2845#     "fetch" directives in .git/config
2846#
2847# These files are disposable unless noMetadata or useSvmProps is set
2848
2849sub _rev_map_set {
2850        my ($fh, $rev, $commit) = @_;
2851
2852        binmode $fh or croak "binmode: $!";
2853        my $size = (stat($fh))[7];
2854        ($size % 24) == 0 or croak "inconsistent size: $size";
2855
2856        my $wr_offset = 0;
2857        if ($size > 0) {
2858                sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2859                my $read = sysread($fh, my $buf, 24) or croak "read: $!";
2860                $read == 24 or croak "read only $read bytes (!= 24)";
2861                my ($last_rev, $last_commit) = unpack(rev_map_fmt, $buf);
2862                if ($last_commit eq ('0' x40)) {
2863                        if ($size >= 48) {
2864                                sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2865                                $read = sysread($fh, $buf, 24) or
2866                                    croak "read: $!";
2867                                $read == 24 or
2868                                    croak "read only $read bytes (!= 24)";
2869                                ($last_rev, $last_commit) =
2870                                    unpack(rev_map_fmt, $buf);
2871                                if ($last_commit eq ('0' x40)) {
2872                                        croak "inconsistent .rev_map\n";
2873                                }
2874                        }
2875                        if ($last_rev >= $rev) {
2876                                croak "last_rev is higher!: $last_rev >= $rev";
2877                        }
2878                        $wr_offset = -24;
2879                }
2880        }
2881        sysseek($fh, $wr_offset, SEEK_END) or croak "seek: $!";
2882        syswrite($fh, pack(rev_map_fmt, $rev, $commit), 24) == 24 or
2883          croak "write: $!";
2884}
2885
2886sub mkfile {
2887        my ($path) = @_;
2888        unless (-e $path) {
2889                my ($dir, $base) = ($path =~ m#^(.*?)/?([^/]+)$#);
2890                mkpath([$dir]) unless -d $dir;
2891                open my $fh, '>>', $path or die "Couldn't create $path: $!\n";
2892                close $fh or die "Couldn't close (create) $path: $!\n";
2893        }
2894}
2895
2896sub rev_map_set {
2897        my ($self, $rev, $commit, $update_ref, $uuid) = @_;
2898        length $commit == 40 or die "arg3 must be a full SHA1 hexsum\n";
2899        my $db = $self->map_path($uuid);
2900        my $db_lock = "$db.lock";
2901        my $sig;
2902        if ($update_ref) {
2903                $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2904                            $SIG{USR1} = $SIG{USR2} = sub { $sig = $_[0] };
2905        }
2906        mkfile($db);
2907
2908        $LOCKFILES{$db_lock} = 1;
2909        my $sync;
2910        # both of these options make our .rev_db file very, very important
2911        # and we can't afford to lose it because rebuild() won't work
2912        if ($self->use_svm_props || $self->no_metadata) {
2913                $sync = 1;
2914                copy($db, $db_lock) or die "rev_map_set(@_): ",
2915                                           "Failed to copy: ",
2916                                           "$db => $db_lock ($!)\n";
2917        } else {
2918                rename $db, $db_lock or die "rev_map_set(@_): ",
2919                                            "Failed to rename: ",
2920                                            "$db => $db_lock ($!)\n";
2921        }
2922
2923        sysopen(my $fh, $db_lock, O_RDWR | O_CREAT)
2924             or croak "Couldn't open $db_lock: $!\n";
2925        _rev_map_set($fh, $rev, $commit);
2926        if ($sync) {
2927                $fh->flush or die "Couldn't flush $db_lock: $!\n";
2928                $fh->sync or die "Couldn't sync $db_lock: $!\n";
2929        }
2930        close $fh or croak $!;
2931        if ($update_ref) {
2932                $_head = $self;
2933                command_noisy('update-ref', '-m', "r$rev",
2934                              $self->refname, $commit);
2935        }
2936        rename $db_lock, $db or die "rev_map_set(@_): ", "Failed to rename: ",
2937                                    "$db_lock => $db ($!)\n";
2938        delete $LOCKFILES{$db_lock};
2939        if ($update_ref) {
2940                $SIG{INT} = $SIG{HUP} = $SIG{TERM} = $SIG{ALRM} = $SIG{PIPE} =
2941                            $SIG{USR1} = $SIG{USR2} = 'DEFAULT';
2942                kill $sig, $$ if defined $sig;
2943        }
2944}
2945
2946# If want_commit, this will return an array of (rev, commit) where
2947# commit _must_ be a valid commit in the archive.
2948# Otherwise, it'll return the max revision (whether or not the
2949# commit is valid or just a 0x40 placeholder).
2950sub rev_map_max {
2951        my ($self, $want_commit) = @_;
2952        $self->rebuild;
2953        my ($r, $c) = $self->rev_map_max_norebuild($want_commit);
2954        $want_commit ? ($r, $c) : $r;
2955}
2956
2957sub rev_map_max_norebuild {
2958        my ($self, $want_commit) = @_;
2959        my $map_path = $self->map_path;
2960        stat $map_path or return $want_commit ? (0, undef) : 0;
2961        sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2962        binmode $fh or croak "binmode: $!";
2963        my $size = (stat($fh))[7];
2964        ($size % 24) == 0 or croak "inconsistent size: $size";
2965
2966        if ($size == 0) {
2967                close $fh or croak "close: $!";
2968                return $want_commit ? (0, undef) : 0;
2969        }
2970
2971        sysseek($fh, -24, SEEK_END) or croak "seek: $!";
2972        sysread($fh, my $buf, 24) == 24 or croak "read: $!";
2973        my ($r, $c) = unpack(rev_map_fmt, $buf);
2974        if ($want_commit && $c eq ('0' x40)) {
2975                if ($size < 48) {
2976                        return $want_commit ? (0, undef) : 0;
2977                }
2978                sysseek($fh, -48, SEEK_END) or croak "seek: $!";
2979                sysread($fh, $buf, 24) == 24 or croak "read: $!";
2980                ($r, $c) = unpack(rev_map_fmt, $buf);
2981                if ($c eq ('0'x40)) {
2982                        croak "Penultimate record is all-zeroes in $map_path";
2983                }
2984        }
2985        close $fh or croak "close: $!";
2986        $want_commit ? ($r, $c) : $r;
2987}
2988
2989sub rev_map_get {
2990        my ($self, $rev, $uuid) = @_;
2991        my $map_path = $self->map_path($uuid);
2992        return undef unless -e $map_path;
2993
2994        sysopen(my $fh, $map_path, O_RDONLY) or croak "open: $!";
2995        binmode $fh or croak "binmode: $!";
2996        my $size = (stat($fh))[7];
2997        ($size % 24) == 0 or croak "inconsistent size: $size";
2998
2999        if ($size == 0) {
3000                close $fh or croak "close: $fh";
3001                return undef;
3002        }
3003
3004        my ($l, $u) = (0, $size - 24);
3005        my ($r, $c, $buf);
3006
3007        while ($l <= $u) {
3008                my $i = int(($l/24 + $u/24) / 2) * 24;
3009                sysseek($fh, $i, SEEK_SET) or croak "seek: $!";
3010                sysread($fh, my $buf, 24) == 24 or croak "read: $!";
3011                my ($r, $c) = unpack('NH40', $buf);
3012
3013                if ($r < $rev) {
3014                        $l = $i + 24;
3015                } elsif ($r > $rev) {
3016                        $u = $i - 24;
3017                } else { # $r == $rev
3018                        close($fh) or croak "close: $!";
3019                        return $c eq ('0' x 40) ? undef : $c;
3020                }
3021        }
3022        close($fh) or croak "close: $!";
3023        undef;
3024}
3025
3026# Finds the first svn revision that exists on (if $eq_ok is true) or
3027# before $rev for the current branch.  It will not search any lower
3028# than $min_rev.  Returns the git commit hash and svn revision number
3029# if found, else (undef, undef).
3030sub find_rev_before {
3031        my ($self, $rev, $eq_ok, $min_rev) = @_;
3032        --$rev unless $eq_ok;
3033        $min_rev ||= 1;
3034        while ($rev >= $min_rev) {
3035                if (my $c = $self->rev_map_get($rev)) {
3036                        return ($rev, $c);
3037                }
3038                --$rev;
3039        }
3040        return (undef, undef);
3041}
3042
3043# Finds the first svn revision that exists on (if $eq_ok is true) or
3044# after $rev for the current branch.  It will not search any higher
3045# than $max_rev.  Returns the git commit hash and svn revision number
3046# if found, else (undef, undef).
3047sub find_rev_after {
3048        my ($self, $rev, $eq_ok, $max_rev) = @_;
3049        ++$rev unless $eq_ok;
3050        $max_rev ||= $self->rev_map_max;
3051        while ($rev <= $max_rev) {
3052                if (my $c = $self->rev_map_get($rev)) {
3053                        return ($rev, $c);
3054                }
3055                ++$rev;
3056        }
3057        return (undef, undef);
3058}
3059
3060sub _new {
3061        my ($class, $repo_id, $ref_id, $path) = @_;
3062        unless (defined $repo_id && length $repo_id) {
3063                $repo_id = $Git::SVN::default_repo_id;
3064        }
3065        unless (defined $ref_id && length $ref_id) {
3066                $_[2] = $ref_id = $Git::SVN::default_ref_id;
3067        }
3068        $_[1] = $repo_id;
3069        my $dir = "$ENV{GIT_DIR}/svn/$ref_id";
3070        $_[3] = $path = '' unless (defined $path);
3071        mkpath(["$ENV{GIT_DIR}/svn"]);
3072        bless {
3073                ref_id => $ref_id, dir => $dir, index => "$dir/index",
3074                path => $path, config => "$ENV{GIT_DIR}/svn/config",
3075                map_root => "$dir/.rev_map", repo_id => $repo_id }, $class;
3076}
3077
3078# for read-only access of old .rev_db formats
3079sub unlink_rev_db_symlink {
3080        my ($self) = @_;
3081        my $link = $self->rev_db_path;
3082        $link =~ s/\.[\w-]+$// or croak "missing UUID at the end of $link";
3083        if (-l $link) {
3084                unlink $link or croak "unlink: $link failed!";
3085        }
3086}
3087
3088sub rev_db_path {
3089        my ($self, $uuid) = @_;
3090        my $db_path = $self->map_path($uuid);
3091        $db_path =~ s{/\.rev_map\.}{/\.rev_db\.}
3092            or croak "map_path: $db_path does not contain '/.rev_map.' !";
3093        $db_path;
3094}
3095
3096# the new replacement for .rev_db
3097sub map_path {
3098        my ($self, $uuid) = @_;
3099        $uuid ||= $self->ra_uuid;
3100        "$self->{map_root}.$uuid";
3101}
3102
3103sub uri_encode {
3104        my ($f) = @_;
3105        $f =~ s#([^a-zA-Z0-9\*!\:_\./\-])#uc sprintf("%%%02x",ord($1))#eg;
3106        $f
3107}
3108
3109sub remove_username {
3110        $_[0] =~ s{^([^:]*://)[^@]+@}{$1};
3111}
3112
3113package Git::SVN::Prompt;
3114use strict;
3115use warnings;
3116require SVN::Core;
3117use vars qw/$_no_auth_cache $_username/;
3118
3119sub simple {
3120        my ($cred, $realm, $default_username, $may_save, $pool) = @_;
3121        $may_save = undef if $_no_auth_cache;
3122        $default_username = $_username if defined $_username;
3123        if (defined $default_username && length $default_username) {
3124                if (defined $realm && length $realm) {
3125                        print STDERR "Authentication realm: $realm\n";
3126                        STDERR->flush;
3127                }
3128                $cred->username($default_username);
3129        } else {
3130                username($cred, $realm, $may_save, $pool);
3131        }
3132        $cred->password(_read_password("Password for '" .
3133                                       $cred->username . "': ", $realm));
3134        $cred->may_save($may_save);
3135        $SVN::_Core::SVN_NO_ERROR;
3136}
3137
3138sub ssl_server_trust {
3139        my ($cred, $realm, $failures, $cert_info, $may_save, $pool) = @_;
3140        $may_save = undef if $_no_auth_cache;
3141        print STDERR "Error validating server certificate for '$realm':\n";
3142        {
3143                no warnings 'once';
3144                # All variables SVN::Auth::SSL::* are used only once,
3145                # so we're shutting up Perl warnings about this.
3146                if ($failures & $SVN::Auth::SSL::UNKNOWNCA) {
3147                        print STDERR " - The certificate is not issued ",
3148                            "by a trusted authority. Use the\n",
3149                            "   fingerprint to validate ",
3150                            "the certificate manually!\n";
3151                }
3152                if ($failures & $SVN::Auth::SSL::CNMISMATCH) {
3153                        print STDERR " - The certificate hostname ",
3154                            "does not match.\n";
3155                }
3156                if ($failures & $SVN::Auth::SSL::NOTYETVALID) {
3157                        print STDERR " - The certificate is not yet valid.\n";
3158                }
3159                if ($failures & $SVN::Auth::SSL::EXPIRED) {
3160                        print STDERR " - The certificate has expired.\n";
3161                }
3162                if ($failures & $SVN::Auth::SSL::OTHER) {
3163                        print STDERR " - The certificate has ",
3164                            "an unknown error.\n";
3165                }
3166        } # no warnings 'once'
3167        printf STDERR
3168                "Certificate information:\n".
3169                " - Hostname: %s\n".
3170                " - Valid: from %s until %s\n".
3171                " - Issuer: %s\n".
3172                " - Fingerprint: %s\n",
3173                map $cert_info->$_, qw(hostname valid_from valid_until
3174                                       issuer_dname fingerprint);
3175        my $choice;
3176prompt:
3177        print STDERR $may_save ?
3178              "(R)eject, accept (t)emporarily or accept (p)ermanently? " :
3179              "(R)eject or accept (t)emporarily? ";
3180        STDERR->flush;
3181        $choice = lc(substr(<STDIN> || 'R', 0, 1));
3182        if ($choice =~ /^t$/i) {
3183                $cred->may_save(undef);
3184        } elsif ($choice =~ /^r$/i) {
3185                return -1;
3186        } elsif ($may_save && $choice =~ /^p$/i) {
3187                $cred->may_save($may_save);
3188        } else {
3189                goto prompt;
3190        }
3191        $cred->accepted_failures($failures);
3192        $SVN::_Core::SVN_NO_ERROR;
3193}
3194
3195sub ssl_client_cert {
3196        my ($cred, $realm, $may_save, $pool) = @_;
3197        $may_save = undef if $_no_auth_cache;
3198        print STDERR "Client certificate filename: ";
3199        STDERR->flush;
3200        chomp(my $filename = <STDIN>);
3201        $cred->cert_file($filename);
3202        $cred->may_save($may_save);
3203        $SVN::_Core::SVN_NO_ERROR;
3204}
3205
3206sub ssl_client_cert_pw {
3207        my ($cred, $realm, $may_save, $pool) = @_;
3208        $may_save = undef if $_no_auth_cache;
3209        $cred->password(_read_password("Password: ", $realm));
3210        $cred->may_save($may_save);
3211        $SVN::_Core::SVN_NO_ERROR;
3212}
3213
3214sub username {
3215        my ($cred, $realm, $may_save, $pool) = @_;
3216        $may_save = undef if $_no_auth_cache;
3217        if (defined $realm && length $realm) {
3218                print STDERR "Authentication realm: $realm\n";
3219        }
3220        my $username;
3221        if (defined $_username) {
3222                $username = $_username;
3223        } else {
3224                print STDERR "Username: ";
3225                STDERR->flush;
3226                chomp($username = <STDIN>);
3227        }
3228        $cred->username($username);
3229        $cred->may_save($may_save);
3230        $SVN::_Core::SVN_NO_ERROR;
3231}
3232
3233sub _read_password {
3234        my ($prompt, $realm) = @_;
3235        print STDERR $prompt;
3236        STDERR->flush;
3237        require Term::ReadKey;
3238        Term::ReadKey::ReadMode('noecho');
3239        my $password = '';
3240        while (defined(my $key = Term::ReadKey::ReadKey(0))) {
3241                last if $key =~ /[\012\015]/; # \n\r
3242                $password .= $key;
3243        }
3244        Term::ReadKey::ReadMode('restore');
3245        print STDERR "\n";
3246        STDERR->flush;
3247        $password;
3248}
3249
3250package SVN::Git::Fetcher;
3251use vars qw/@ISA/;
3252use strict;
3253use warnings;
3254use Carp qw/croak/;
3255use File::Temp qw/tempfile/;
3256use IO::File qw//;
3257use vars qw/$_ignore_regex/;
3258
3259# file baton members: path, mode_a, mode_b, pool, fh, blob, base
3260sub new {
3261        my ($class, $git_svn, $switch_path) = @_;
3262        my $self = SVN::Delta::Editor->new;
3263        bless $self, $class;
3264        if (exists $git_svn->{last_commit}) {
3265                $self->{c} = $git_svn->{last_commit};
3266                $self->{empty_symlinks} =
3267                                  _mark_empty_symlinks($git_svn, $switch_path);
3268        }
3269        $self->{empty} = {};
3270        $self->{dir_prop} = {};
3271        $self->{file_prop} = {};
3272        $self->{absent_dir} = {};
3273        $self->{absent_file} = {};
3274        $self->{gii} = $git_svn->tmp_index_do(sub { Git::IndexInfo->new });
3275        $self;
3276}
3277
3278# this uses the Ra object, so it must be called before do_{switch,update},
3279# not inside them (when the Git::SVN::Fetcher object is passed) to
3280# do_{switch,update}
3281sub _mark_empty_symlinks {
3282        my ($git_svn, $switch_path) = @_;
3283        my $bool = Git::config_bool('svn.brokenSymlinkWorkaround');
3284        return {} if (defined($bool) && ! $bool);
3285
3286        my %ret;
3287        my ($rev, $cmt) = $git_svn->last_rev_commit;
3288        return {} unless ($rev && $cmt);
3289
3290        # allow the warning to be printed for each revision we fetch to
3291        # ensure the user sees it.  The user can also disable the workaround
3292        # on the repository even while git svn is running and the next
3293        # revision fetched will skip this expensive function.
3294        my $printed_warning;
3295        chomp(my $empty_blob = `git hash-object -t blob --stdin < /dev/null`);
3296        my ($ls, $ctx) = command_output_pipe(qw/ls-tree -r -z/, $cmt);
3297        local $/ = "\0";
3298        my $pfx = defined($switch_path) ? $switch_path : $git_svn->{path};
3299        $pfx .= '/' if length($pfx);
3300        while (<$ls>) {
3301                chomp;
3302                s/\A100644 blob $empty_blob\t//o or next;
3303                unless ($printed_warning) {
3304                        print STDERR "Scanning for empty symlinks, ",
3305                                     "this may take a while if you have ",
3306                                     "many empty files\n",
3307                                     "You may disable this with `",
3308                                     "git config svn.brokenSymlinkWorkaround ",
3309                                     "false'.\n",
3310                                     "This may be done in a different ",
3311                                     "terminal without restarting ",
3312                                     "git svn\n";
3313                        $printed_warning = 1;
3314                }
3315                my $path = $_;
3316                my (undef, $props) =
3317                               $git_svn->ra->get_file($pfx.$path, $rev, undef);
3318                if ($props->{'svn:special'}) {
3319                        $ret{$path} = 1;
3320                }
3321        }
3322        command_close_pipe($ls, $ctx);
3323        \%ret;
3324}
3325
3326# returns true if a given path is inside a ".git" directory
3327sub in_dot_git {
3328        $_[0] =~ m{(?:^|/)\.git(?:/|$)};
3329}
3330
3331# return value: 0 -- don't ignore, 1 -- ignore
3332sub is_path_ignored {
3333        my ($path) = @_;
3334        return 1 if in_dot_git($path);
3335        return 0 unless defined($_ignore_regex);
3336        return 1 if $path =~ m!$_ignore_regex!o;
3337        return 0;
3338}
3339
3340sub set_path_strip {
3341        my ($self, $path) = @_;
3342        $self->{path_strip} = qr/^\Q$path\E(\/|$)/ if length $path;
3343}
3344
3345sub open_root {
3346        { path => '' };
3347}
3348
3349sub open_directory {
3350        my ($self, $path, $pb, $rev) = @_;
3351        { path => $path };
3352}
3353
3354sub git_path {
3355        my ($self, $path) = @_;
3356        if ($self->{path_strip}) {
3357                $path =~ s!$self->{path_strip}!! or
3358                  die "Failed to strip path '$path' ($self->{path_strip})\n";
3359        }
3360        $path;
3361}
3362
3363sub delete_entry {
3364        my ($self, $path, $rev, $pb) = @_;
3365        return undef if is_path_ignored($path);
3366
3367        my $gpath = $self->git_path($path);
3368        return undef if ($gpath eq '');
3369
3370        # remove entire directories.
3371        if (command('ls-tree', $self->{c}, '--', $gpath) =~ /^040000 tree/) {
3372                my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3373                                                     -r --name-only -z/,
3374                                                     $self->{c}, '--', $gpath);
3375                local $/ = "\0";
3376                while (<$ls>) {
3377                        chomp;
3378                        $self->{gii}->remove($_);
3379                        print "\tD\t$_\n" unless $::_q;
3380                }
3381                print "\tD\t$gpath/\n" unless $::_q;
3382                command_close_pipe($ls, $ctx);
3383                $self->{empty}->{$path} = 0
3384        } else {
3385                $self->{gii}->remove($gpath);
3386                print "\tD\t$gpath\n" unless $::_q;
3387        }
3388        undef;
3389}
3390
3391sub open_file {
3392        my ($self, $path, $pb, $rev) = @_;
3393        my ($mode, $blob);
3394
3395        goto out if is_path_ignored($path);
3396
3397        my $gpath = $self->git_path($path);
3398        ($mode, $blob) = (command('ls-tree', $self->{c}, '--', $gpath)
3399                             =~ /^(\d{6}) blob ([a-f\d]{40})\t/);
3400        unless (defined $mode && defined $blob) {
3401                die "$path was not found in commit $self->{c} (r$rev)\n";
3402        }
3403        if ($mode eq '100644' && $self->{empty_symlinks}->{$path}) {
3404                $mode = '120000';
3405        }
3406out:
3407        { path => $path, mode_a => $mode, mode_b => $mode, blob => $blob,
3408          pool => SVN::Pool->new, action => 'M' };
3409}
3410
3411sub add_file {
3412        my ($self, $path, $pb, $cp_path, $cp_rev) = @_;
3413        my $mode;
3414
3415        if (!is_path_ignored($path)) {
3416                my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3417                delete $self->{empty}->{$dir};
3418                $mode = '100644';
3419        }
3420        { path => $path, mode_a => $mode, mode_b => $mode,
3421          pool => SVN::Pool->new, action => 'A' };
3422}
3423
3424sub add_directory {
3425        my ($self, $path, $cp_path, $cp_rev) = @_;
3426        goto out if is_path_ignored($path);
3427        my $gpath = $self->git_path($path);
3428        if ($gpath eq '') {
3429                my ($ls, $ctx) = command_output_pipe(qw/ls-tree
3430                                                     -r --name-only -z/,
3431                                                     $self->{c});
3432                local $/ = "\0";
3433                while (<$ls>) {
3434                        chomp;
3435                        $self->{gii}->remove($_);
3436                        print "\tD\t$_\n" unless $::_q;
3437                }
3438                command_close_pipe($ls, $ctx);
3439                $self->{empty}->{$path} = 0;
3440        }
3441        my ($dir, $file) = ($path =~ m#^(.*?)/?([^/]+)$#);
3442        delete $self->{empty}->{$dir};
3443        $self->{empty}->{$path} = 1;
3444out:
3445        { path => $path };
3446}
3447
3448sub change_dir_prop {
3449        my ($self, $db, $prop, $value) = @_;
3450        return undef if is_path_ignored($db->{path});
3451        $self->{dir_prop}->{$db->{path}} ||= {};
3452        $self->{dir_prop}->{$db->{path}}->{$prop} = $value;
3453        undef;
3454}
3455
3456sub absent_directory {
3457        my ($self, $path, $pb) = @_;
3458        return undef if is_path_ignored($path);
3459        $self->{absent_dir}->{$pb->{path}} ||= [];
3460        push @{$self->{absent_dir}->{$pb->{path}}}, $path;
3461        undef;
3462}
3463
3464sub absent_file {
3465        my ($self, $path, $pb) = @_;
3466        return undef if is_path_ignored($path);
3467        $self->{absent_file}->{$pb->{path}} ||= [];
3468        push @{$self->{absent_file}->{$pb->{path}}}, $path;
3469        undef;
3470}
3471
3472sub change_file_prop {
3473        my ($self, $fb, $prop, $value) = @_;
3474        return undef if is_path_ignored($fb->{path});
3475        if ($prop eq 'svn:executable') {
3476                if ($fb->{mode_b} != 120000) {
3477                        $fb->{mode_b} = defined $value ? 100755 : 100644;
3478                }
3479        } elsif ($prop eq 'svn:special') {
3480                $fb->{mode_b} = defined $value ? 120000 : 100644;
3481        } else {
3482                $self->{file_prop}->{$fb->{path}} ||= {};
3483                $self->{file_prop}->{$fb->{path}}->{$prop} = $value;
3484        }
3485        undef;
3486}
3487
3488sub apply_textdelta {
3489        my ($self, $fb, $exp) = @_;
3490        return undef if is_path_ignored($fb->{path});
3491        my $fh = $::_repository->temp_acquire('svn_delta');
3492        # $fh gets auto-closed() by SVN::TxDelta::apply(),
3493        # (but $base does not,) so dup() it for reading in close_file
3494        open my $dup, '<&', $fh or croak $!;
3495        my $base = $::_repository->temp_acquire('git_blob');
3496
3497        if ($fb->{blob}) {
3498                my ($base_is_link, $size);
3499
3500                if ($fb->{mode_a} eq '120000' &&
3501                    ! $self->{empty_symlinks}->{$fb->{path}}) {
3502                        print $base 'link ' or die "print $!\n";
3503                        $base_is_link = 1;
3504                }
3505        retry:
3506                $size = $::_repository->cat_blob($fb->{blob}, $base);
3507                die "Failed to read object $fb->{blob}" if ($size < 0);
3508
3509                if (defined $exp) {
3510                        seek $base, 0, 0 or croak $!;
3511                        my $got = ::md5sum($base);
3512                        if ($got ne $exp) {
3513                                my $err = "Checksum mismatch: ".
3514                                       "$fb->{path} $fb->{blob}\n" .
3515                                       "expected: $exp\n" .
3516                                       "     got: $got\n";
3517                                if ($base_is_link) {
3518                                        warn $err,
3519                                             "Retrying... (possibly ",
3520                                             "a bad symlink from SVN)\n";
3521                                        $::_repository->temp_reset($base);
3522                                        $base_is_link = 0;
3523                                        goto retry;
3524                                }
3525                                die $err;
3526                        }
3527                }
3528        }
3529        seek $base, 0, 0 or croak $!;
3530        $fb->{fh} = $fh;
3531        $fb->{base} = $base;
3532        [ SVN::TxDelta::apply($base, $dup, undef, $fb->{path}, $fb->{pool}) ];
3533}
3534
3535sub close_file {
3536        my ($self, $fb, $exp) = @_;
3537        return undef if is_path_ignored($fb->{path});
3538
3539        my $hash;
3540        my $path = $self->git_path($fb->{path});
3541        if (my $fh = $fb->{fh}) {
3542                if (defined $exp) {
3543                        seek($fh, 0, 0) or croak $!;
3544                        my $got = ::md5sum($fh);
3545                        if ($got ne $exp) {
3546                                die "Checksum mismatch: $path\n",
3547                                    "expected: $exp\n    got: $got\n";
3548                        }
3549                }
3550                if ($fb->{mode_b} == 120000) {
3551                        sysseek($fh, 0, 0) or croak $!;
3552                        my $rd = sysread($fh, my $buf, 5);
3553
3554                        if (!defined $rd) {
3555                                croak "sysread: $!\n";
3556                        } elsif ($rd == 0) {
3557                                warn "$path has mode 120000",
3558                                     " but it points to nothing\n",
3559                                     "converting to an empty file with mode",
3560                                     " 100644\n";
3561                                $fb->{mode_b} = '100644';
3562                        } elsif ($buf ne 'link ') {
3563                                warn "$path has mode 120000",
3564                                     " but is not a link\n";
3565                        } else {
3566                                my $tmp_fh = $::_repository->temp_acquire(
3567                                        'svn_hash');
3568                                my $res;
3569                                while ($res = sysread($fh, my $str, 1024)) {
3570                                        my $out = syswrite($tmp_fh, $str, $res);
3571                                        defined($out) && $out == $res
3572                                                or croak("write ",
3573                                                        Git::temp_path($tmp_fh),
3574                                                        ": $!\n");
3575                                }
3576                                defined $res or croak $!;
3577
3578                                ($fh, $tmp_fh) = ($tmp_fh, $fh);
3579                                Git::temp_release($tmp_fh, 1);
3580                        }
3581                }
3582
3583                $hash = $::_repository->hash_and_insert_object(
3584                                Git::temp_path($fh));
3585                $hash =~ /^[a-f\d]{40}$/ or die "not a sha1: $hash\n";
3586
3587                Git::temp_release($fb->{base}, 1);
3588                Git::temp_release($fh, 1);
3589        } else {
3590                $hash = $fb->{blob} or die "no blob information\n";
3591        }
3592        $fb->{pool}->clear;
3593        $self->{gii}->update($fb->{mode_b}, $hash, $path) or croak $!;
3594        print "\t$fb->{action}\t$path\n" if $fb->{action} && ! $::_q;
3595        undef;
3596}
3597
3598sub abort_edit {
3599        my $self = shift;
3600        $self->{nr} = $self->{gii}->{nr};
3601        delete $self->{gii};
3602        $self->SUPER::abort_edit(@_);
3603}
3604
3605sub close_edit {
3606        my $self = shift;
3607        $self->{git_commit_ok} = 1;
3608        $self->{nr} = $self->{gii}->{nr};
3609        delete $self->{gii};
3610        $self->SUPER::close_edit(@_);
3611}
3612
3613package SVN::Git::Editor;
3614use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
3615use strict;
3616use warnings;
3617use Carp qw/croak/;
3618use IO::File;
3619
3620sub new {
3621        my ($class, $opts) = @_;
3622        foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
3623                die "$_ required!\n" unless (defined $opts->{$_});
3624        }
3625
3626        my $pool = SVN::Pool->new;
3627        my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
3628        my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
3629                                     $opts->{r}, $mods);
3630
3631        # $opts->{ra} functions should not be used after this:
3632        my @ce  = $opts->{ra}->get_commit_editor($opts->{log},
3633                                                $opts->{editor_cb}, $pool);
3634        my $self = SVN::Delta::Editor->new(@ce, $pool);
3635        bless $self, $class;
3636        foreach (qw/svn_path r tree_a tree_b/) {
3637                $self->{$_} = $opts->{$_};
3638        }
3639        $self->{url} = $opts->{ra}->{url};
3640        $self->{mods} = $mods;
3641        $self->{types} = $types;
3642        $self->{pool} = $pool;
3643        $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
3644        $self->{rm} = { };
3645        $self->{path_prefix} = length $self->{svn_path} ?
3646                               "$self->{svn_path}/" : '';
3647        $self->{config} = $opts->{config};
3648        return $self;
3649}
3650
3651sub generate_diff {
3652        my ($tree_a, $tree_b) = @_;
3653        my @diff_tree = qw(diff-tree -z -r);
3654        if ($_cp_similarity) {
3655                push @diff_tree, "-C$_cp_similarity";
3656        } else {
3657                push @diff_tree, '-C';
3658        }
3659        push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
3660        push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
3661        push @diff_tree, $tree_a, $tree_b;
3662        my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
3663        local $/ = "\0";
3664        my $state = 'meta';
3665        my @mods;
3666        while (<$diff_fh>) {
3667                chomp $_; # this gets rid of the trailing "\0"
3668                if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
3669                                        ($::sha1)\s($::sha1)\s
3670                                        ([MTCRAD])\d*$/xo) {
3671                        push @mods, {   mode_a => $1, mode_b => $2,
3672                                        sha1_a => $3, sha1_b => $4,
3673                                        chg => $5 };
3674                        if ($5 =~ /^(?:C|R)$/) {
3675                                $state = 'file_a';
3676                        } else {
3677                                $state = 'file_b';
3678                        }
3679                } elsif ($state eq 'file_a') {
3680                        my $x = $mods[$#mods] or croak "Empty array\n";
3681                        if ($x->{chg} !~ /^(?:C|R)$/) {
3682                                croak "Error parsing $_, $x->{chg}\n";
3683                        }
3684                        $x->{file_a} = $_;
3685                        $state = 'file_b';
3686                } elsif ($state eq 'file_b') {
3687                        my $x = $mods[$#mods] or croak "Empty array\n";
3688                        if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
3689                                croak "Error parsing $_, $x->{chg}\n";
3690                        }
3691                        if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
3692                                croak "Error parsing $_, $x->{chg}\n";
3693                        }
3694                        $x->{file_b} = $_;
3695                        $state = 'meta';
3696                } else {
3697                        croak "Error parsing $_\n";
3698                }
3699        }
3700        command_close_pipe($diff_fh, $ctx);
3701        \@mods;
3702}
3703
3704sub check_diff_paths {
3705        my ($ra, $pfx, $rev, $mods) = @_;
3706        my %types;
3707        $pfx .= '/' if length $pfx;
3708
3709        sub type_diff_paths {
3710                my ($ra, $types, $path, $rev) = @_;
3711                my @p = split m#/+#, $path;
3712                my $c = shift @p;
3713                unless (defined $types->{$c}) {
3714                        $types->{$c} = $ra->check_path($c, $rev);
3715                }
3716                while (@p) {
3717                        $c .= '/' . shift @p;
3718                        next if defined $types->{$c};
3719                        $types->{$c} = $ra->check_path($c, $rev);
3720                }
3721        }
3722
3723        foreach my $m (@$mods) {
3724                foreach my $f (qw/file_a file_b/) {
3725                        next unless defined $m->{$f};
3726                        my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
3727                        if (length $pfx.$dir && ! defined $types{$dir}) {
3728                                type_diff_paths($ra, \%types, $pfx.$dir, $rev);
3729                        }
3730                }
3731        }
3732        \%types;
3733}
3734
3735sub split_path {
3736        return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
3737}
3738
3739sub repo_path {
3740        my ($self, $path) = @_;
3741        $self->{path_prefix}.(defined $path ? $path : '');
3742}
3743
3744sub url_path {
3745        my ($self, $path) = @_;
3746        if ($self->{url} =~ m#^https?://#) {
3747                $path =~ s/([^~a-zA-Z0-9_.-])/uc sprintf("%%%02x",ord($1))/eg;
3748        }
3749        $self->{url} . '/' . $self->repo_path($path);
3750}
3751
3752sub rmdirs {
3753        my ($self) = @_;
3754        my $rm = $self->{rm};
3755        delete $rm->{''}; # we never delete the url we're tracking
3756        return unless %$rm;
3757
3758        foreach (keys %$rm) {
3759                my @d = split m#/#, $_;
3760                my $c = shift @d;
3761                $rm->{$c} = 1;
3762                while (@d) {
3763                        $c .= '/' . shift @d;
3764                        $rm->{$c} = 1;
3765                }
3766        }
3767        delete $rm->{$self->{svn_path}};
3768        delete $rm->{''}; # we never delete the url we're tracking
3769        return unless %$rm;
3770
3771        my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
3772                                             $self->{tree_b});
3773        local $/ = "\0";
3774        while (<$fh>) {
3775                chomp;
3776                my @dn = split m#/#, $_;
3777                while (pop @dn) {
3778                        delete $rm->{join '/', @dn};
3779                }
3780                unless (%$rm) {
3781                        close $fh;
3782                        return;
3783                }
3784        }
3785        command_close_pipe($fh, $ctx);
3786
3787        my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
3788        foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
3789                $self->close_directory($bat->{$d}, $p);
3790                my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
3791                print "\tD+\t$d/\n" unless $::_q;
3792                $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
3793                delete $bat->{$d};
3794        }
3795}
3796
3797sub open_or_add_dir {
3798        my ($self, $full_path, $baton) = @_;
3799        my $t = $self->{types}->{$full_path};
3800        if (!defined $t) {
3801                die "$full_path not known in r$self->{r} or we have a bug!\n";
3802        }
3803        {
3804                no warnings 'once';
3805                # SVN::Node::none and SVN::Node::file are used only once,
3806                # so we're shutting up Perl's warnings about them.
3807                if ($t == $SVN::Node::none) {
3808                        return $self->add_directory($full_path, $baton,
3809                            undef, -1, $self->{pool});
3810                } elsif ($t == $SVN::Node::dir) {
3811                        return $self->open_directory($full_path, $baton,
3812                            $self->{r}, $self->{pool});
3813                } # no warnings 'once'
3814                print STDERR "$full_path already exists in repository at ",
3815                    "r$self->{r} and it is not a directory (",
3816                    ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
3817        } # no warnings 'once'
3818        exit 1;
3819}
3820
3821sub ensure_path {
3822        my ($self, $path) = @_;
3823        my $bat = $self->{bat};
3824        my $repo_path = $self->repo_path($path);
3825        return $bat->{''} unless (length $repo_path);
3826        my @p = split m#/+#, $repo_path;
3827        my $c = shift @p;
3828        $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''});
3829        while (@p) {
3830                my $c0 = $c;
3831                $c .= '/' . shift @p;
3832                $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0});
3833        }
3834        return $bat->{$c};
3835}
3836
3837# Subroutine to convert a globbing pattern to a regular expression.
3838# From perl cookbook.
3839sub glob2pat {
3840        my $globstr = shift;
3841        my %patmap = ('*' => '.*', '?' => '.', '[' => '[', ']' => ']');
3842        $globstr =~ s{(.)} { $patmap{$1} || "\Q$1" }ge;
3843        return '^' . $globstr . '$';
3844}
3845
3846sub check_autoprop {
3847        my ($self, $pattern, $properties, $file, $fbat) = @_;
3848        # Convert the globbing pattern to a regular expression.
3849        my $regex = glob2pat($pattern);
3850        # Check if the pattern matches the file name.
3851        if($file =~ m/($regex)/) {
3852                # Parse the list of properties to set.
3853                my @props = split(/;/, $properties);
3854                foreach my $prop (@props) {
3855                        # Parse 'name=value' syntax and set the property.
3856                        if ($prop =~ /([^=]+)=(.*)/) {
3857                                my ($n,$v) = ($1,$2);
3858                                for ($n, $v) {
3859                                        s/^\s+//; s/\s+$//;
3860                                }
3861                                $self->change_file_prop($fbat, $n, $v);
3862                        }
3863                }
3864        }
3865}
3866
3867sub apply_autoprops {
3868        my ($self, $file, $fbat) = @_;
3869        my $conf_t = ${$self->{config}}{'config'};
3870        no warnings 'once';
3871        # Check [miscellany]/enable-auto-props in svn configuration.
3872        if (SVN::_Core::svn_config_get_bool(
3873                $conf_t,
3874                $SVN::_Core::SVN_CONFIG_SECTION_MISCELLANY,
3875                $SVN::_Core::SVN_CONFIG_OPTION_ENABLE_AUTO_PROPS,
3876                0)) {
3877                # Auto-props are enabled.  Enumerate them to look for matches.
3878                my $callback = sub {
3879                        $self->check_autoprop($_[0], $_[1], $file, $fbat);
3880                };
3881                SVN::_Core::svn_config_enumerate(
3882                        $conf_t,
3883                        $SVN::_Core::SVN_CONFIG_SECTION_AUTO_PROPS,
3884                        $callback);
3885        }
3886}
3887
3888sub A {
3889        my ($self, $m) = @_;
3890        my ($dir, $file) = split_path($m->{file_b});
3891        my $pbat = $self->ensure_path($dir);
3892        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3893                                        undef, -1);
3894        print "\tA\t$m->{file_b}\n" unless $::_q;
3895        $self->apply_autoprops($file, $fbat);
3896        $self->chg_file($fbat, $m);
3897        $self->close_file($fbat,undef,$self->{pool});
3898}
3899
3900sub C {
3901        my ($self, $m) = @_;
3902        my ($dir, $file) = split_path($m->{file_b});
3903        my $pbat = $self->ensure_path($dir);
3904        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3905                                $self->url_path($m->{file_a}), $self->{r});
3906        print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3907        $self->chg_file($fbat, $m);
3908        $self->close_file($fbat,undef,$self->{pool});
3909}
3910
3911sub delete_entry {
3912        my ($self, $path, $pbat) = @_;
3913        my $rpath = $self->repo_path($path);
3914        my ($dir, $file) = split_path($rpath);
3915        $self->{rm}->{$dir} = 1;
3916        $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
3917}
3918
3919sub R {
3920        my ($self, $m) = @_;
3921        my ($dir, $file) = split_path($m->{file_b});
3922        my $pbat = $self->ensure_path($dir);
3923        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
3924                                $self->url_path($m->{file_a}), $self->{r});
3925        print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
3926        $self->apply_autoprops($file, $fbat);
3927        $self->chg_file($fbat, $m);
3928        $self->close_file($fbat,undef,$self->{pool});
3929
3930        ($dir, $file) = split_path($m->{file_a});
3931        $pbat = $self->ensure_path($dir);
3932        $self->delete_entry($m->{file_a}, $pbat);
3933}
3934
3935sub M {
3936        my ($self, $m) = @_;
3937        my ($dir, $file) = split_path($m->{file_b});
3938        my $pbat = $self->ensure_path($dir);
3939        my $fbat = $self->open_file($self->repo_path($m->{file_b}),
3940                                $pbat,$self->{r},$self->{pool});
3941        print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
3942        $self->chg_file($fbat, $m);
3943        $self->close_file($fbat,undef,$self->{pool});
3944}
3945
3946sub T { shift->M(@_) }
3947
3948sub change_file_prop {
3949        my ($self, $fbat, $pname, $pval) = @_;
3950        $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
3951}
3952
3953sub _chg_file_get_blob ($$$$) {
3954        my ($self, $fbat, $m, $which) = @_;
3955        my $fh = $::_repository->temp_acquire("git_blob_$which");
3956        if ($m->{"mode_$which"} =~ /^120/) {
3957                print $fh 'link ' or croak $!;
3958                $self->change_file_prop($fbat,'svn:special','*');
3959        } elsif ($m->{mode_a} =~ /^120/ && $m->{"mode_$which"} !~ /^120/) {
3960                $self->change_file_prop($fbat,'svn:special',undef);
3961        }
3962        my $blob = $m->{"sha1_$which"};
3963        return ($fh,) if ($blob =~ /^0{40}$/);
3964        my $size = $::_repository->cat_blob($blob, $fh);
3965        croak "Failed to read object $blob" if ($size < 0);
3966        $fh->flush == 0 or croak $!;
3967        seek $fh, 0, 0 or croak $!;
3968
3969        my $exp = ::md5sum($fh);
3970        seek $fh, 0, 0 or croak $!;
3971        return ($fh, $exp);
3972}
3973
3974sub chg_file {
3975        my ($self, $fbat, $m) = @_;
3976        if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
3977                $self->change_file_prop($fbat,'svn:executable','*');
3978        } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
3979                $self->change_file_prop($fbat,'svn:executable',undef);
3980        }
3981        my ($fh_a, $exp_a) = _chg_file_get_blob $self, $fbat, $m, 'a';
3982        my ($fh_b, $exp_b) = _chg_file_get_blob $self, $fbat, $m, 'b';
3983        my $pool = SVN::Pool->new;
3984        my $atd = $self->apply_textdelta($fbat, $exp_a, $pool);
3985        if (-s $fh_a) {
3986                my $txstream = SVN::TxDelta::new ($fh_a, $fh_b, $pool);
3987                my $res = SVN::TxDelta::send_txstream($txstream, @$atd, $pool);
3988                if (defined $res) {
3989                        die "Unexpected result from send_txstream: $res\n",
3990                            "(SVN::Core::VERSION: $SVN::Core::VERSION)\n";
3991                }
3992        } else {
3993                my $got = SVN::TxDelta::send_stream($fh_b, @$atd, $pool);
3994                die "Checksum mismatch\nexpected: $exp_b\ngot: $got\n"
3995                    if ($got ne $exp_b);
3996        }
3997        Git::temp_release($fh_b, 1);
3998        Git::temp_release($fh_a, 1);
3999        $pool->clear;
4000}
4001
4002sub D {
4003        my ($self, $m) = @_;
4004        my ($dir, $file) = split_path($m->{file_b});
4005        my $pbat = $self->ensure_path($dir);
4006        print "\tD\t$m->{file_b}\n" unless $::_q;
4007        $self->delete_entry($m->{file_b}, $pbat);
4008}
4009
4010sub close_edit {
4011        my ($self) = @_;
4012        my ($p,$bat) = ($self->{pool}, $self->{bat});
4013        foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
4014                next if $_ eq '';
4015                $self->close_directory($bat->{$_}, $p);
4016        }
4017        $self->close_directory($bat->{''}, $p);
4018        $self->SUPER::close_edit($p);
4019        $p->clear;
4020}
4021
4022sub abort_edit {
4023        my ($self) = @_;
4024        $self->SUPER::abort_edit($self->{pool});
4025}
4026
4027sub DESTROY {
4028        my $self = shift;
4029        $self->SUPER::DESTROY(@_);
4030        $self->{pool}->clear;
4031}
4032
4033# this drives the editor
4034sub apply_diff {
4035        my ($self) = @_;
4036        my $mods = $self->{mods};
4037        my %o = ( D => 1, R => 0, C => -1, A => 3, M => 3, T => 3 );
4038        foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
4039                my $f = $m->{chg};
4040                if (defined $o{$f}) {
4041                        $self->$f($m);
4042                } else {
4043                        fatal("Invalid change type: $f");
4044                }
4045        }
4046        $self->rmdirs if $_rmdir;
4047        if (@$mods == 0) {
4048                $self->abort_edit;
4049        } else {
4050                $self->close_edit;
4051        }
4052        return scalar @$mods;
4053}
4054
4055package Git::SVN::Ra;
4056use vars qw/@ISA $config_dir $_log_window_size/;
4057use strict;
4058use warnings;
4059my ($ra_invalid, $can_do_switch, %ignored_err, $RA);
4060
4061BEGIN {
4062        # enforce temporary pool usage for some simple functions
4063        no strict 'refs';
4064        for my $f (qw/rev_proplist get_latest_revnum get_uuid get_repos_root
4065                      get_file/) {
4066                my $SUPER = "SUPER::$f";
4067                *$f = sub {
4068                        my $self = shift;
4069                        my $pool = SVN::Pool->new;
4070                        my @ret = $self->$SUPER(@_,$pool);
4071                        $pool->clear;
4072                        wantarray ? @ret : $ret[0];
4073                };
4074        }
4075}
4076
4077sub _auth_providers () {
4078        [
4079          SVN::Client::get_simple_provider(),
4080          SVN::Client::get_ssl_server_trust_file_provider(),
4081          SVN::Client::get_simple_prompt_provider(
4082            \&Git::SVN::Prompt::simple, 2),
4083          SVN::Client::get_ssl_client_cert_file_provider(),
4084          SVN::Client::get_ssl_client_cert_prompt_provider(
4085            \&Git::SVN::Prompt::ssl_client_cert, 2),
4086          SVN::Client::get_ssl_client_cert_pw_file_provider(),
4087          SVN::Client::get_ssl_client_cert_pw_prompt_provider(
4088            \&Git::SVN::Prompt::ssl_client_cert_pw, 2),
4089          SVN::Client::get_username_provider(),
4090          SVN::Client::get_ssl_server_trust_prompt_provider(
4091            \&Git::SVN::Prompt::ssl_server_trust),
4092          SVN::Client::get_username_prompt_provider(
4093            \&Git::SVN::Prompt::username, 2)
4094        ]
4095}
4096
4097sub escape_uri_only {
4098        my ($uri) = @_;
4099        my @tmp;
4100        foreach (split m{/}, $uri) {
4101                s/([^~\w.%+-]|%(?![a-fA-F0-9]{2}))/sprintf("%%%02X",ord($1))/eg;
4102                push @tmp, $_;
4103        }
4104        join('/', @tmp);
4105}
4106
4107sub escape_url {
4108        my ($url) = @_;
4109        if ($url =~ m#^(https?)://([^/]+)(.*)$#) {
4110                my ($scheme, $domain, $uri) = ($1, $2, escape_uri_only($3));
4111                $url = "$scheme://$domain$uri";
4112        }
4113        $url;
4114}
4115
4116sub new {
4117        my ($class, $url) = @_;
4118        $url =~ s!/+$!!;
4119        return $RA if ($RA && $RA->{url} eq $url);
4120
4121        SVN::_Core::svn_config_ensure($config_dir, undef);
4122        my ($baton, $callbacks) = SVN::Core::auth_open_helper(_auth_providers);
4123        my $config = SVN::Core::config_get_config($config_dir);
4124        $RA = undef;
4125        my $dont_store_passwords = 1;
4126        my $conf_t = ${$config}{'config'};
4127        {
4128                no warnings 'once';
4129                # The usage of $SVN::_Core::SVN_CONFIG_* variables
4130                # produces warnings that variables are used only once.
4131                # I had not found the better way to shut them up, so
4132                # the warnings of type 'once' are disabled in this block.
4133                if (SVN::_Core::svn_config_get_bool($conf_t,
4134                    $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
4135                    $SVN::_Core::SVN_CONFIG_OPTION_STORE_PASSWORDS,
4136                    1) == 0) {
4137                        SVN::_Core::svn_auth_set_parameter($baton,
4138                            $SVN::_Core::SVN_AUTH_PARAM_DONT_STORE_PASSWORDS,
4139                            bless (\$dont_store_passwords, "_p_void"));
4140                }
4141                if (SVN::_Core::svn_config_get_bool($conf_t,
4142                    $SVN::_Core::SVN_CONFIG_SECTION_AUTH,
4143                    $SVN::_Core::SVN_CONFIG_OPTION_STORE_AUTH_CREDS,
4144                    1) == 0) {
4145                        $Git::SVN::Prompt::_no_auth_cache = 1;
4146                }
4147        } # no warnings 'once'
4148        my $self = SVN::Ra->new(url => escape_url($url), auth => $baton,
4149                              config => $config,
4150                              pool => SVN::Pool->new,
4151                              auth_provider_callbacks => $callbacks);
4152        $self->{url} = $url;
4153        $self->{svn_path} = $url;
4154        $self->{repos_root} = $self->get_repos_root;
4155        $self->{svn_path} =~ s#^\Q$self->{repos_root}\E(/|$)##;
4156        $self->{cache} = { check_path => { r => 0, data => {} },
4157                           get_dir => { r => 0, data => {} } };
4158        $RA = bless $self, $class;
4159}
4160
4161sub check_path {
4162        my ($self, $path, $r) = @_;
4163        my $cache = $self->{cache}->{check_path};
4164        if ($r == $cache->{r} && exists $cache->{data}->{$path}) {
4165                return $cache->{data}->{$path};
4166        }
4167        my $pool = SVN::Pool->new;
4168        my $t = $self->SUPER::check_path($path, $r, $pool);
4169        $pool->clear;
4170        if ($r != $cache->{r}) {
4171                %{$cache->{data}} = ();
4172                $cache->{r} = $r;
4173        }
4174        $cache->{data}->{$path} = $t;
4175}
4176
4177sub get_dir {
4178        my ($self, $dir, $r) = @_;
4179        my $cache = $self->{cache}->{get_dir};
4180        if ($r == $cache->{r}) {
4181                if (my $x = $cache->{data}->{$dir}) {
4182                        return wantarray ? @$x : $x->[0];
4183                }
4184        }
4185        my $pool = SVN::Pool->new;
4186        my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);
4187        my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;
4188        $pool->clear;
4189        if ($r != $cache->{r}) {
4190                %{$cache->{data}} = ();
4191                $cache->{r} = $r;
4192        }
4193        $cache->{data}->{$dir} = [ \%dirents, $r, $props ];
4194        wantarray ? (\%dirents, $r, $props) : \%dirents;
4195}
4196
4197sub DESTROY {
4198        # do not call the real DESTROY since we store ourselves in $RA
4199}
4200
4201# get_log(paths, start, end, limit,
4202#         discover_changed_paths, strict_node_history, receiver)
4203sub get_log {
4204        my ($self, @args) = @_;
4205        my $pool = SVN::Pool->new;
4206
4207        # the limit parameter was not supported in SVN 1.1.x, so we
4208        # drop it.  Therefore, the receiver callback passed to it
4209        # is made aware of this limitation by being wrapped if
4210        # the limit passed to is being wrapped.
4211        if ($SVN::Core::VERSION le '1.2.0') {
4212                my $limit = splice(@args, 3, 1);
4213                if ($limit > 0) {
4214                        my $receiver = pop @args;
4215                        push(@args, sub { &$receiver(@_) if (--$limit >= 0) });
4216                }
4217        }
4218        my $ret = $self->SUPER::get_log(@args, $pool);
4219        $pool->clear;
4220        $ret;
4221}
4222
4223sub trees_match {
4224        my ($self, $url1, $rev1, $url2, $rev2) = @_;
4225        my $ctx = SVN::Client->new(auth => _auth_providers);
4226        my $out = IO::File->new_tmpfile;
4227
4228        # older SVN (1.1.x) doesn't take $pool as the last parameter for
4229        # $ctx->diff(), so we'll create a default one
4230        my $pool = SVN::Pool->new_default_sub;
4231
4232        $ra_invalid = 1; # this will open a new SVN::Ra connection to $url1
4233        $ctx->diff([], $url1, $rev1, $url2, $rev2, 1, 1, 0, $out, $out);
4234        $out->flush;
4235        my $ret = (($out->stat)[7] == 0);
4236        close $out or croak $!;
4237
4238        $ret;
4239}
4240
4241sub get_commit_editor {
4242        my ($self, $log, $cb, $pool) = @_;
4243        my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef, 0) : ();
4244        $self->SUPER::get_commit_editor($log, $cb, @lock, $pool);
4245}
4246
4247sub gs_do_update {
4248        my ($self, $rev_a, $rev_b, $gs, $editor) = @_;
4249        my $new = ($rev_a == $rev_b);
4250        my $path = $gs->{path};
4251
4252        if ($new && -e $gs->{index}) {
4253                unlink $gs->{index} or die
4254                  "Couldn't unlink index: $gs->{index}: $!\n";
4255        }
4256        my $pool = SVN::Pool->new;
4257        $editor->set_path_strip($path);
4258        my (@pc) = split m#/#, $path;
4259        my $reporter = $self->do_update($rev_b, (@pc ? shift @pc : ''),
4260                                        1, $editor, $pool);
4261        my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
4262
4263        # Since we can't rely on svn_ra_reparent being available, we'll
4264        # just have to do some magic with set_path to make it so
4265        # we only want a partial path.
4266        my $sp = '';
4267        my $final = join('/', @pc);
4268        while (@pc) {
4269                $reporter->set_path($sp, $rev_b, 0, @lock, $pool);
4270                $sp .= '/' if length $sp;
4271                $sp .= shift @pc;
4272        }
4273        die "BUG: '$sp' != '$final'\n" if ($sp ne $final);
4274
4275        $reporter->set_path($sp, $rev_a, $new, @lock, $pool);
4276
4277        $reporter->finish_report($pool);
4278        $pool->clear;
4279        $editor->{git_commit_ok};
4280}
4281
4282# this requires SVN 1.4.3 or later (do_switch didn't work before 1.4.3, and
4283# svn_ra_reparent didn't work before 1.4)
4284sub gs_do_switch {
4285        my ($self, $rev_a, $rev_b, $gs, $url_b, $editor) = @_;
4286        my $path = $gs->{path};
4287        my $pool = SVN::Pool->new;
4288
4289        my $full_url = $self->{url};
4290        my $old_url = $full_url;
4291        $full_url .= '/' . escape_uri_only($path) if length $path;
4292        my ($ra, $reparented);
4293
4294        if ($old_url =~ m#^svn(\+ssh)?://#) {
4295                $_[0] = undef;
4296                $self = undef;
4297                $RA = undef;
4298                $ra = Git::SVN::Ra->new($full_url);
4299                $ra_invalid = 1;
4300        } elsif ($old_url ne $full_url) {
4301                SVN::_Ra::svn_ra_reparent($self->{session}, $full_url, $pool);
4302                $self->{url} = $full_url;
4303                $reparented = 1;
4304        }
4305
4306        $ra ||= $self;
4307        $url_b = escape_url($url_b);
4308        my $reporter = $ra->do_switch($rev_b, '', 1, $url_b, $editor, $pool);
4309        my @lock = $SVN::Core::VERSION ge '1.2.0' ? (undef) : ();
4310        $reporter->set_path('', $rev_a, 0, @lock, $pool);
4311        $reporter->finish_report($pool);
4312
4313        if ($reparented) {
4314                SVN::_Ra::svn_ra_reparent($self->{session}, $old_url, $pool);
4315                $self->{url} = $old_url;
4316        }
4317
4318        $pool->clear;
4319        $editor->{git_commit_ok};
4320}
4321
4322sub longest_common_path {
4323        my ($gsv, $globs) = @_;
4324        my %common;
4325        my $common_max = scalar @$gsv;
4326
4327        foreach my $gs (@$gsv) {
4328                my @tmp = split m#/#, $gs->{path};
4329                my $p = '';
4330                foreach (@tmp) {
4331                        $p .= length($p) ? "/$_" : $_;
4332                        $common{$p} ||= 0;
4333                        $common{$p}++;
4334                }
4335        }
4336        $globs ||= [];
4337        $common_max += scalar @$globs;
4338        foreach my $glob (@$globs) {
4339                my @tmp = split m#/#, $glob->{path}->{left};
4340                my $p = '';
4341                foreach (@tmp) {
4342                        $p .= length($p) ? "/$_" : $_;
4343                        $common{$p} ||= 0;
4344                        $common{$p}++;
4345                }
4346        }
4347
4348        my $longest_path = '';
4349        foreach (sort {length $b <=> length $a} keys %common) {
4350                if ($common{$_} == $common_max) {
4351                        $longest_path = $_;
4352                        last;
4353                }
4354        }
4355        $longest_path;
4356}
4357
4358sub gs_fetch_loop_common {
4359        my ($self, $base, $head, $gsv, $globs) = @_;
4360        return if ($base > $head);
4361        my $inc = $_log_window_size;
4362        my ($min, $max) = ($base, $head < $base + $inc ? $head : $base + $inc);
4363        my $longest_path = longest_common_path($gsv, $globs);
4364        my $ra_url = $self->{url};
4365        while (1) {
4366                my %revs;
4367                my $err;
4368                my $err_handler = $SVN::Error::handler;
4369                $SVN::Error::handler = sub {
4370                        ($err) = @_;
4371                        skip_unknown_revs($err);
4372                };
4373                sub _cb {
4374                        my ($paths, $r, $author, $date, $log) = @_;
4375                        [ dup_changed_paths($paths),
4376                          { author => $author, date => $date, log => $log } ];
4377                }
4378                $self->get_log([$longest_path], $min, $max, 0, 1, 1,
4379                               sub { $revs{$_[1]} = _cb(@_) });
4380                if ($err) {
4381                        print "Checked through r$max\r";
4382                }
4383                if ($err && $max >= $head) {
4384                        print STDERR "Path '$longest_path' ",
4385                                     "was probably deleted:\n",
4386                                     $err->expanded_message,
4387                                     "\nWill attempt to follow ",
4388                                     "revisions r$min .. r$max ",
4389                                     "committed before the deletion\n";
4390                        my $hi = $max;
4391                        while (--$hi >= $min) {
4392                                my $ok;
4393                                $self->get_log([$longest_path], $min, $hi,
4394                                               0, 1, 1, sub {
4395                                               $ok ||= $_[1];
4396                                               $revs{$_[1]} = _cb(@_) });
4397                                if ($ok) {
4398                                        print STDERR "r$min .. r$ok OK\n";
4399                                        last;
4400                                }
4401                        }
4402                }
4403                $SVN::Error::handler = $err_handler;
4404
4405                my %exists = map { $_->{path} => $_ } @$gsv;
4406                foreach my $r (sort {$a <=> $b} keys %revs) {
4407                        my ($paths, $logged) = @{$revs{$r}};
4408
4409                        foreach my $gs ($self->match_globs(\%exists, $paths,
4410                                                           $globs, $r)) {
4411                                if ($gs->rev_map_max >= $r) {
4412                                        next;
4413                                }
4414                                next unless $gs->match_paths($paths, $r);
4415                                $gs->{logged_rev_props} = $logged;
4416                                if (my $last_commit = $gs->last_commit) {
4417                                        $gs->assert_index_clean($last_commit);
4418                                }
4419                                my $log_entry = $gs->do_fetch($paths, $r);
4420                                if ($log_entry) {
4421                                        $gs->do_git_commit($log_entry);
4422                                }
4423                                $INDEX_FILES{$gs->{index}} = 1;
4424                        }
4425                        foreach my $g (@$globs) {
4426                                my $k = "svn-remote.$g->{remote}." .
4427                                        "$g->{t}-maxRev";
4428                                Git::SVN::tmp_config($k, $r);
4429                        }
4430                        if ($ra_invalid) {
4431                                $_[0] = undef;
4432                                $self = undef;
4433                                $RA = undef;
4434                                $self = Git::SVN::Ra->new($ra_url);
4435                                $ra_invalid = undef;
4436                        }
4437                }
4438                # pre-fill the .rev_db since it'll eventually get filled in
4439                # with '0' x40 if something new gets committed
4440                foreach my $gs (@$gsv) {
4441                        next if $gs->rev_map_max >= $max;
4442                        next if defined $gs->rev_map_get($max);
4443                        $gs->rev_map_set($max, 0 x40);
4444                }
4445                foreach my $g (@$globs) {
4446                        my $k = "svn-remote.$g->{remote}.$g->{t}-maxRev";
4447                        Git::SVN::tmp_config($k, $max);
4448                }
4449                last if $max >= $head;
4450                $min = $max + 1;
4451                $max += $inc;
4452                $max = $head if ($max > $head);
4453        }
4454        Git::SVN::gc();
4455}
4456
4457sub get_dir_globbed {
4458        my ($self, $left, $depth, $r) = @_;
4459
4460        my @x = eval { $self->get_dir($left, $r) };
4461        return unless scalar @x == 3;
4462        my $dirents = $x[0];
4463        my @finalents;
4464        foreach my $de (keys %$dirents) {
4465                next if $dirents->{$de}->{kind} != $SVN::Node::dir;
4466                if ($depth > 1) {
4467                        my @args = ("$left/$de", $depth - 1, $r);
4468                        foreach my $dir ($self->get_dir_globbed(@args)) {
4469                                push @finalents, "$de/$dir";
4470                        }
4471                } else {
4472                        push @finalents, $de;
4473                }
4474        }
4475        @finalents;
4476}
4477
4478sub match_globs {
4479        my ($self, $exists, $paths, $globs, $r) = @_;
4480
4481        sub get_dir_check {
4482                my ($self, $exists, $g, $r) = @_;
4483
4484                my @dirs = $self->get_dir_globbed($g->{path}->{left},
4485                                                  $g->{path}->{depth},
4486                                                  $r);
4487
4488                foreach my $de (@dirs) {
4489                        my $p = $g->{path}->full_path($de);
4490                        next if $exists->{$p};
4491                        next if (length $g->{path}->{right} &&
4492                                 ($self->check_path($p, $r) !=
4493                                  $SVN::Node::dir));
4494                        $exists->{$p} = Git::SVN->init($self->{url}, $p, undef,
4495                                         $g->{ref}->full_path($de), 1);
4496                }
4497        }
4498        foreach my $g (@$globs) {
4499                if (my $path = $paths->{"/$g->{path}->{left}"}) {
4500                        if ($path->{action} =~ /^[AR]$/) {
4501                                get_dir_check($self, $exists, $g, $r);
4502                        }
4503                }
4504                foreach (keys %$paths) {
4505                        if (/$g->{path}->{left_regex}/ &&
4506                            !/$g->{path}->{regex}/) {
4507                                next if $paths->{$_}->{action} !~ /^[AR]$/;
4508                                get_dir_check($self, $exists, $g, $r);
4509                        }
4510                        next unless /$g->{path}->{regex}/;
4511                        my $p = $1;
4512                        my $pathname = $g->{path}->full_path($p);
4513                        next if $exists->{$pathname};
4514                        next if ($self->check_path($pathname, $r) !=
4515                                 $SVN::Node::dir);
4516                        $exists->{$pathname} = Git::SVN->init(
4517                                              $self->{url}, $pathname, undef,
4518                                              $g->{ref}->full_path($p), 1);
4519                }
4520                my $c = '';
4521                foreach (split m#/#, $g->{path}->{left}) {
4522                        $c .= "/$_";
4523                        next unless ($paths->{$c} &&
4524                                     ($paths->{$c}->{action} =~ /^[AR]$/));
4525                        get_dir_check($self, $exists, $g, $r);
4526                }
4527        }
4528        values %$exists;
4529}
4530
4531sub minimize_url {
4532        my ($self) = @_;
4533        return $self->{url} if ($self->{url} eq $self->{repos_root});
4534        my $url = $self->{repos_root};
4535        my @components = split(m!/!, $self->{svn_path});
4536        my $c = '';
4537        do {
4538                $url .= "/$c" if length $c;
4539                eval { (ref $self)->new($url)->get_latest_revnum };
4540        } while ($@ && ($c = shift @components));
4541        $url;
4542}
4543
4544sub can_do_switch {
4545        my $self = shift;
4546        unless (defined $can_do_switch) {
4547                my $pool = SVN::Pool->new;
4548                my $rep = eval {
4549                        $self->do_switch(1, '', 0, $self->{url},
4550                                         SVN::Delta::Editor->new, $pool);
4551                };
4552                if ($@) {
4553                        $can_do_switch = 0;
4554                } else {
4555                        $rep->abort_report($pool);
4556                        $can_do_switch = 1;
4557                }
4558                $pool->clear;
4559        }
4560        $can_do_switch;
4561}
4562
4563sub skip_unknown_revs {
4564        my ($err) = @_;
4565        my $errno = $err->apr_err();
4566        # Maybe the branch we're tracking didn't
4567        # exist when the repo started, so it's
4568        # not an error if it doesn't, just continue
4569        #
4570        # Wonderfully consistent library, eh?
4571        # 160013 - svn:// and file://
4572        # 175002 - http(s)://
4573        # 175007 - http(s):// (this repo required authorization, too...)
4574        #   More codes may be discovered later...
4575        if ($errno == 175007 || $errno == 175002 || $errno == 160013) {
4576                my $err_key = $err->expanded_message;
4577                # revision numbers change every time, filter them out
4578                $err_key =~ s/\d+/\0/g;
4579                $err_key = "$errno\0$err_key";
4580                unless ($ignored_err{$err_key}) {
4581                        warn "W: Ignoring error from SVN, path probably ",
4582                             "does not exist: ($errno): ",
4583                             $err->expanded_message,"\n";
4584                        warn "W: Do not be alarmed at the above message ",
4585                             "git-svn is just searching aggressively for ",
4586                             "old history.\n",
4587                             "This may take a while on large repositories\n";
4588                        $ignored_err{$err_key} = 1;
4589                }
4590                return;
4591        }
4592        die "Error from SVN, ($errno): ", $err->expanded_message,"\n";
4593}
4594
4595# svn_log_changed_path_t objects passed to get_log are likely to be
4596# overwritten even if only the refs are copied to an external variable,
4597# so we should dup the structures in their entirety.  Using an externally
4598# passed pool (instead of our temporary and quickly cleared pool in
4599# Git::SVN::Ra) does not help matters at all...
4600sub dup_changed_paths {
4601        my ($paths) = @_;
4602        return undef unless $paths;
4603        my %ret;
4604        foreach my $p (keys %$paths) {
4605                my $i = $paths->{$p};
4606                my %s = map { $_ => $i->$_ }
4607                              qw/copyfrom_path copyfrom_rev action/;
4608                $ret{$p} = \%s;
4609        }
4610        \%ret;
4611}
4612
4613package Git::SVN::Log;
4614use strict;
4615use warnings;
4616use POSIX qw/strftime/;
4617use constant commit_log_separator => ('-' x 72) . "\n";
4618use vars qw/$TZ $limit $color $pager $non_recursive $verbose $oneline
4619            %rusers $show_commit $incremental/;
4620my $l_fmt;
4621
4622sub cmt_showable {
4623        my ($c) = @_;
4624        return 1 if defined $c->{r};
4625
4626        # big commit message got truncated by the 16k pretty buffer in rev-list
4627        if ($c->{l} && $c->{l}->[-1] eq "...\n" &&
4628                                $c->{a_raw} =~ /\@([a-f\d\-]+)>$/) {
4629                @{$c->{l}} = ();
4630                my @log = command(qw/cat-file commit/, $c->{c});
4631
4632                # shift off the headers
4633                shift @log while ($log[0] ne '');
4634                shift @log;
4635
4636                # TODO: make $c->{l} not have a trailing newline in the future
4637                @{$c->{l}} = map { "$_\n" } grep !/^git-svn-id: /, @log;
4638
4639                (undef, $c->{r}, undef) = ::extract_metadata(
4640                                (grep(/^git-svn-id: /, @log))[-1]);
4641        }
4642        return defined $c->{r};
4643}
4644
4645sub log_use_color {
4646        return $color || Git->repository->get_colorbool('color.diff');
4647}
4648
4649sub git_svn_log_cmd {
4650        my ($r_min, $r_max, @args) = @_;
4651        my $head = 'HEAD';
4652        my (@files, @log_opts);
4653        foreach my $x (@args) {
4654                if ($x eq '--' || @files) {
4655                        push @files, $x;
4656                } else {
4657                        if (::verify_ref("$x^0")) {
4658                                $head = $x;
4659                        } else {
4660                                push @log_opts, $x;
4661                        }
4662                }
4663        }
4664
4665        my ($url, $rev, $uuid, $gs) = ::working_head_info($head);
4666        $gs ||= Git::SVN->_new;
4667        my @cmd = (qw/log --abbrev-commit --pretty=raw --default/,
4668                   $gs->refname);
4669        push @cmd, '-r' unless $non_recursive;
4670        push @cmd, qw/--raw --name-status/ if $verbose;
4671        push @cmd, '--color' if log_use_color();
4672        push @cmd, @log_opts;
4673        if (defined $r_max && $r_max == $r_min) {
4674                push @cmd, '--max-count=1';
4675                if (my $c = $gs->rev_map_get($r_max)) {
4676                        push @cmd, $c;
4677                }
4678        } elsif (defined $r_max) {
4679                if ($r_max < $r_min) {
4680                        ($r_min, $r_max) = ($r_max, $r_min);
4681                }
4682                my (undef, $c_max) = $gs->find_rev_before($r_max, 1, $r_min);
4683                my (undef, $c_min) = $gs->find_rev_after($r_min, 1, $r_max);
4684                # If there are no commits in the range, both $c_max and $c_min
4685                # will be undefined.  If there is at least 1 commit in the
4686                # range, both will be defined.
4687                return () if !defined $c_min || !defined $c_max;
4688                if ($c_min eq $c_max) {
4689                        push @cmd, '--max-count=1', $c_min;
4690                } else {
4691                        push @cmd, '--boundary', "$c_min..$c_max";
4692                }
4693        }
4694        return (@cmd, @files);
4695}
4696
4697# adapted from pager.c
4698sub config_pager {
4699        $pager ||= $ENV{GIT_PAGER} || $ENV{PAGER};
4700        if (!defined $pager) {
4701                $pager = 'less';
4702        } elsif (length $pager == 0 || $pager eq 'cat') {
4703                $pager = undef;
4704        }
4705        $ENV{GIT_PAGER_IN_USE} = defined($pager);
4706}
4707
4708sub run_pager {
4709        return unless -t *STDOUT && defined $pager;
4710        pipe my ($rfd, $wfd) or return;
4711        defined(my $pid = fork) or ::fatal "Can't fork: $!";
4712        if (!$pid) {
4713                open STDOUT, '>&', $wfd or
4714                                     ::fatal "Can't redirect to stdout: $!";
4715                return;
4716        }
4717        open STDIN, '<&', $rfd or ::fatal "Can't redirect stdin: $!";
4718        $ENV{LESS} ||= 'FRSX';
4719        exec $pager or ::fatal "Can't run pager: $! ($pager)";
4720}
4721
4722sub format_svn_date {
4723        return strftime("%Y-%m-%d %H:%M:%S %z (%a, %d %b %Y)", localtime(shift));
4724}
4725
4726sub parse_git_date {
4727        my ($t, $tz) = @_;
4728        # Date::Parse isn't in the standard Perl distro :(
4729        if ($tz =~ s/^\+//) {
4730                $t += tz_to_s_offset($tz);
4731        } elsif ($tz =~ s/^\-//) {
4732                $t -= tz_to_s_offset($tz);
4733        }
4734        return $t;
4735}
4736
4737sub set_local_timezone {
4738        if (defined $TZ) {
4739                $ENV{TZ} = $TZ;
4740        } else {
4741                delete $ENV{TZ};
4742        }
4743}
4744
4745sub tz_to_s_offset {
4746        my ($tz) = @_;
4747        $tz =~ s/(\d\d)$//;
4748        return ($1 * 60) + ($tz * 3600);
4749}
4750
4751sub get_author_info {
4752        my ($dest, $author, $t, $tz) = @_;
4753        $author =~ s/(?:^\s*|\s*$)//g;
4754        $dest->{a_raw} = $author;
4755        my $au;
4756        if ($::_authors) {
4757                $au = $rusers{$author} || undef;
4758        }
4759        if (!$au) {
4760                ($au) = ($author =~ /<([^>]+)\@[^>]+>$/);
4761        }
4762        $dest->{t} = $t;
4763        $dest->{tz} = $tz;
4764        $dest->{a} = $au;
4765        $dest->{t_utc} = parse_git_date($t, $tz);
4766}
4767
4768sub process_commit {
4769        my ($c, $r_min, $r_max, $defer) = @_;
4770        if (defined $r_min && defined $r_max) {
4771                if ($r_min == $c->{r} && $r_min == $r_max) {
4772                        show_commit($c);
4773                        return 0;
4774                }
4775                return 1 if $r_min == $r_max;
4776                if ($r_min < $r_max) {
4777                        # we need to reverse the print order
4778                        return 0 if (defined $limit && --$limit < 0);
4779                        push @$defer, $c;
4780                        return 1;
4781                }
4782                if ($r_min != $r_max) {
4783                        return 1 if ($r_min < $c->{r});
4784                        return 1 if ($r_max > $c->{r});
4785                }
4786        }
4787        return 0 if (defined $limit && --$limit < 0);
4788        show_commit($c);
4789        return 1;
4790}
4791
4792sub show_commit {
4793        my $c = shift;
4794        if ($oneline) {
4795                my $x = "\n";
4796                if (my $l = $c->{l}) {
4797                        while ($l->[0] =~ /^\s*$/) { shift @$l }
4798                        $x = $l->[0];
4799                }
4800                $l_fmt ||= 'A' . length($c->{r});
4801                print 'r',pack($l_fmt, $c->{r}),' | ';
4802                print "$c->{c} | " if $show_commit;
4803                print $x;
4804        } else {
4805                show_commit_normal($c);
4806        }
4807}
4808
4809sub show_commit_changed_paths {
4810        my ($c) = @_;
4811        return unless $c->{changed};
4812        print "Changed paths:\n", @{$c->{changed}};
4813}
4814
4815sub show_commit_normal {
4816        my ($c) = @_;
4817        print commit_log_separator, "r$c->{r} | ";
4818        print "$c->{c} | " if $show_commit;
4819        print "$c->{a} | ", format_svn_date($c->{t_utc}), ' | ';
4820        my $nr_line = 0;
4821
4822        if (my $l = $c->{l}) {
4823                while ($l->[$#$l] eq "\n" && $#$l > 0
4824                                          && $l->[($#$l - 1)] eq "\n") {
4825                        pop @$l;
4826                }
4827                $nr_line = scalar @$l;
4828                if (!$nr_line) {
4829                        print "1 line\n\n\n";
4830                } else {
4831                        if ($nr_line == 1) {
4832                                $nr_line = '1 line';
4833                        } else {
4834                                $nr_line .= ' lines';
4835                        }
4836                        print $nr_line, "\n";
4837                        show_commit_changed_paths($c);
4838                        print "\n";
4839                        print $_ foreach @$l;
4840                }
4841        } else {
4842                print "1 line\n";
4843                show_commit_changed_paths($c);
4844                print "\n";
4845
4846        }
4847        foreach my $x (qw/raw stat diff/) {
4848                if ($c->{$x}) {
4849                        print "\n";
4850                        print $_ foreach @{$c->{$x}}
4851                }
4852        }
4853}
4854
4855sub cmd_show_log {
4856        my (@args) = @_;
4857        my ($r_min, $r_max);
4858        my $r_last = -1; # prevent dupes
4859        set_local_timezone();
4860        if (defined $::_revision) {
4861                if ($::_revision =~ /^(\d+):(\d+)$/) {
4862                        ($r_min, $r_max) = ($1, $2);
4863                } elsif ($::_revision =~ /^\d+$/) {
4864                        $r_min = $r_max = $::_revision;
4865                } else {
4866                        ::fatal "-r$::_revision is not supported, use ",
4867                                "standard 'git log' arguments instead";
4868                }
4869        }
4870
4871        config_pager();
4872        @args = git_svn_log_cmd($r_min, $r_max, @args);
4873        if (!@args) {
4874                print commit_log_separator unless $incremental || $oneline;
4875                return;
4876        }
4877        my $log = command_output_pipe(@args);
4878        run_pager();
4879        my (@k, $c, $d, $stat);
4880        my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
4881        while (<$log>) {
4882                if (/^${esc_color}commit -?($::sha1_short)/o) {
4883                        my $cmt = $1;
4884                        if ($c && cmt_showable($c) && $c->{r} != $r_last) {
4885                                $r_last = $c->{r};
4886                                process_commit($c, $r_min, $r_max, \@k) or
4887                                                                goto out;
4888                        }
4889                        $d = undef;
4890                        $c = { c => $cmt };
4891                } elsif (/^${esc_color}author (.+) (\d+) ([\-\+]?\d+)$/o) {
4892                        get_author_info($c, $1, $2, $3);
4893                } elsif (/^${esc_color}(?:tree|parent|committer) /o) {
4894                        # ignore
4895                } elsif (/^${esc_color}:\d{6} \d{6} $::sha1_short/o) {
4896                        push @{$c->{raw}}, $_;
4897                } elsif (/^${esc_color}[ACRMDT]\t/) {
4898                        # we could add $SVN->{svn_path} here, but that requires
4899                        # remote access at the moment (repo_path_split)...
4900                        s#^(${esc_color})([ACRMDT])\t#$1   $2 #o;
4901                        push @{$c->{changed}}, $_;
4902                } elsif (/^${esc_color}diff /o) {
4903                        $d = 1;
4904                        push @{$c->{diff}}, $_;
4905                } elsif ($d) {
4906                        push @{$c->{diff}}, $_;
4907                } elsif (/^\ .+\ \|\s*\d+\ $esc_color[\+\-]*
4908                          $esc_color*[\+\-]*$esc_color$/x) {
4909                        $stat = 1;
4910                        push @{$c->{stat}}, $_;
4911                } elsif ($stat && /^ \d+ files changed, \d+ insertions/) {
4912                        push @{$c->{stat}}, $_;
4913                        $stat = undef;
4914                } elsif (/^${esc_color}    (git-svn-id:.+)$/o) {
4915                        ($c->{url}, $c->{r}, undef) = ::extract_metadata($1);
4916                } elsif (s/^${esc_color}    //o) {
4917                        push @{$c->{l}}, $_;
4918                }
4919        }
4920        if ($c && defined $c->{r} && $c->{r} != $r_last) {
4921                $r_last = $c->{r};
4922                process_commit($c, $r_min, $r_max, \@k);
4923        }
4924        if (@k) {
4925                ($r_min, $r_max) = ($r_max, $r_min);
4926                process_commit($_, $r_min, $r_max) foreach reverse @k;
4927        }
4928out:
4929        close $log;
4930        print commit_log_separator unless $incremental || $oneline;
4931}
4932
4933sub cmd_blame {
4934        my $path = pop;
4935
4936        config_pager();
4937        run_pager();
4938
4939        my ($fh, $ctx, $rev);
4940
4941        if ($_git_format) {
4942                ($fh, $ctx) = command_output_pipe('blame', @_, $path);
4943                while (my $line = <$fh>) {
4944                        if ($line =~ /^\^?([[:xdigit:]]+)\s/) {
4945                                # Uncommitted edits show up as a rev ID of
4946                                # all zeros, which we can't look up with
4947                                # cmt_metadata
4948                                if ($1 !~ /^0+$/) {
4949                                        (undef, $rev, undef) =
4950                                                ::cmt_metadata($1);
4951                                        $rev = '0' if (!$rev);
4952                                } else {
4953                                        $rev = '0';
4954                                }
4955                                $rev = sprintf('%-10s', $rev);
4956                                $line =~ s/^\^?[[:xdigit:]]+(\s)/$rev$1/;
4957                        }
4958                        print $line;
4959                }
4960        } else {
4961                ($fh, $ctx) = command_output_pipe('blame', '-p', @_, 'HEAD',
4962                                                  '--', $path);
4963                my ($sha1);
4964                my %authors;
4965                while (my $line = <$fh>) {
4966                        if ($line =~ /^([[:xdigit:]]{40})\s\d+\s\d+/) {
4967                                $sha1 = $1;
4968                                (undef, $rev, undef) = ::cmt_metadata($1);
4969                                $rev = '0' if (!$rev);
4970                        }
4971                        elsif ($line =~ /^author (.*)/) {
4972                                $authors{$rev} = $1;
4973                                $authors{$rev} =~ s/\s/_/g;
4974                        }
4975                        elsif ($line =~ /^\t(.*)$/) {
4976                                printf("%6s %10s %s\n", $rev, $authors{$rev}, $1);
4977                        }
4978                }
4979        }
4980        command_close_pipe($fh, $ctx);
4981}
4982
4983package Git::SVN::Migration;
4984# these version numbers do NOT correspond to actual version numbers
4985# of git nor git-svn.  They are just relative.
4986#
4987# v0 layout: .git/$id/info/url, refs/heads/$id-HEAD
4988#
4989# v1 layout: .git/$id/info/url, refs/remotes/$id
4990#
4991# v2 layout: .git/svn/$id/info/url, refs/remotes/$id
4992#
4993# v3 layout: .git/svn/$id, refs/remotes/$id
4994#            - info/url may remain for backwards compatibility
4995#            - this is what we migrate up to this layout automatically,
4996#            - this will be used by git svn init on single branches
4997# v3.1 layout (auto migrated):
4998#            - .rev_db => .rev_db.$UUID, .rev_db will remain as a symlink
4999#              for backwards compatibility
5000#
5001# v4 layout: .git/svn/$repo_id/$id, refs/remotes/$repo_id/$id
5002#            - this is only created for newly multi-init-ed
5003#              repositories.  Similar in spirit to the
5004#              --use-separate-remotes option in git-clone (now default)
5005#            - we do not automatically migrate to this (following
5006#              the example set by core git)
5007#
5008# v5 layout: .rev_db.$UUID => .rev_map.$UUID
5009#            - newer, more-efficient format that uses 24-bytes per record
5010#              with no filler space.
5011#            - use xxd -c24 < .rev_map.$UUID to view and debug
5012#            - This is a one-way migration, repositories updated to the
5013#              new format will not be able to use old git-svn without
5014#              rebuilding the .rev_db.  Rebuilding the rev_db is not
5015#              possible if noMetadata or useSvmProps are set; but should
5016#              be no problem for users that use the (sensible) defaults.
5017use strict;
5018use warnings;
5019use Carp qw/croak/;
5020use File::Path qw/mkpath/;
5021use File::Basename qw/dirname basename/;
5022use vars qw/$_minimize/;
5023
5024sub migrate_from_v0 {
5025        my $git_dir = $ENV{GIT_DIR};
5026        return undef unless -d $git_dir;
5027        my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
5028        my $migrated = 0;
5029        while (<$fh>) {
5030                chomp;
5031                my ($id, $orig_ref) = ($_, $_);
5032                next unless $id =~ s#^refs/heads/(.+)-HEAD$#$1#;
5033                next unless -f "$git_dir/$id/info/url";
5034                my $new_ref = "refs/remotes/$id";
5035                if (::verify_ref("$new_ref^0")) {
5036                        print STDERR "W: $orig_ref is probably an old ",
5037                                     "branch used by an ancient version of ",
5038                                     "git-svn.\n",
5039                                     "However, $new_ref also exists.\n",
5040                                     "We will not be able ",
5041                                     "to use this branch until this ",
5042                                     "ambiguity is resolved.\n";
5043                        next;
5044                }
5045                print STDERR "Migrating from v0 layout...\n" if !$migrated;
5046                print STDERR "Renaming ref: $orig_ref => $new_ref\n";
5047                command_noisy('update-ref', $new_ref, $orig_ref);
5048                command_noisy('update-ref', '-d', $orig_ref, $orig_ref);
5049                $migrated++;
5050        }
5051        command_close_pipe($fh, $ctx);
5052        print STDERR "Done migrating from v0 layout...\n" if $migrated;
5053        $migrated;
5054}
5055
5056sub migrate_from_v1 {
5057        my $git_dir = $ENV{GIT_DIR};
5058        my $migrated = 0;
5059        return $migrated unless -d $git_dir;
5060        my $svn_dir = "$git_dir/svn";
5061
5062        # just in case somebody used 'svn' as their $id at some point...
5063        return $migrated if -d $svn_dir && ! -f "$svn_dir/info/url";
5064
5065        print STDERR "Migrating from a git-svn v1 layout...\n";
5066        mkpath([$svn_dir]);
5067        print STDERR "Data from a previous version of git-svn exists, but\n\t",
5068                     "$svn_dir\n\t(required for this version ",
5069                     "($::VERSION) of git-svn) does not exist.\n";
5070        my ($fh, $ctx) = command_output_pipe(qw/rev-parse --symbolic --all/);
5071        while (<$fh>) {
5072                my $x = $_;
5073                next unless $x =~ s#^refs/remotes/##;
5074                chomp $x;
5075                next unless -f "$git_dir/$x/info/url";
5076                my $u = eval { ::file_to_s("$git_dir/$x/info/url") };
5077                next unless $u;
5078                my $dn = dirname("$git_dir/svn/$x");
5079                mkpath([$dn]) unless -d $dn;
5080                if ($x eq 'svn') { # they used 'svn' as GIT_SVN_ID:
5081                        mkpath(["$git_dir/svn/svn"]);
5082                        print STDERR " - $git_dir/$x/info => ",
5083                                        "$git_dir/svn/$x/info\n";
5084                        rename "$git_dir/$x/info", "$git_dir/svn/$x/info" or
5085                               croak "$!: $x";
5086                        # don't worry too much about these, they probably
5087                        # don't exist with repos this old (save for index,
5088                        # and we can easily regenerate that)
5089                        foreach my $f (qw/unhandled.log index .rev_db/) {
5090                                rename "$git_dir/$x/$f", "$git_dir/svn/$x/$f";
5091                        }
5092                } else {
5093                        print STDERR " - $git_dir/$x => $git_dir/svn/$x\n";
5094                        rename "$git_dir/$x", "$git_dir/svn/$x" or
5095                               croak "$!: $x";
5096                }
5097                $migrated++;
5098        }
5099        command_close_pipe($fh, $ctx);
5100        print STDERR "Done migrating from a git-svn v1 layout\n";
5101        $migrated;
5102}
5103
5104sub read_old_urls {
5105        my ($l_map, $pfx, $path) = @_;
5106        my @dir;
5107        foreach (<$path/*>) {
5108                if (-r "$_/info/url") {
5109                        $pfx .= '/' if $pfx && $pfx !~ m!/$!;
5110                        my $ref_id = $pfx . basename $_;
5111                        my $url = ::file_to_s("$_/info/url");
5112                        $l_map->{$ref_id} = $url;
5113                } elsif (-d $_) {
5114                        push @dir, $_;
5115                }
5116        }
5117        foreach (@dir) {
5118                my $x = $_;
5119                $x =~ s!^\Q$ENV{GIT_DIR}\E/svn/!!o;
5120                read_old_urls($l_map, $x, $_);
5121        }
5122}
5123
5124sub migrate_from_v2 {
5125        my @cfg = command(qw/config -l/);
5126        return if grep /^svn-remote\..+\.url=/, @cfg;
5127        my %l_map;
5128        read_old_urls(\%l_map, '', "$ENV{GIT_DIR}/svn");
5129        my $migrated = 0;
5130
5131        foreach my $ref_id (sort keys %l_map) {
5132                eval { Git::SVN->init($l_map{$ref_id}, '', undef, $ref_id) };
5133                if ($@) {
5134                        Git::SVN->init($l_map{$ref_id}, '', $ref_id, $ref_id);
5135                }
5136                $migrated++;
5137        }
5138        $migrated;
5139}
5140
5141sub minimize_connections {
5142        my $r = Git::SVN::read_all_remotes();
5143        my $new_urls = {};
5144        my $root_repos = {};
5145        foreach my $repo_id (keys %$r) {
5146                my $url = $r->{$repo_id}->{url} or next;
5147                my $fetch = $r->{$repo_id}->{fetch} or next;
5148                my $ra = Git::SVN::Ra->new($url);
5149
5150                # skip existing cases where we already connect to the root
5151                if (($ra->{url} eq $ra->{repos_root}) ||
5152                    ($ra->{repos_root} eq $repo_id)) {
5153                        $root_repos->{$ra->{url}} = $repo_id;
5154                        next;
5155                }
5156
5157                my $root_ra = Git::SVN::Ra->new($ra->{repos_root});
5158                my $root_path = $ra->{url};
5159                $root_path =~ s#^\Q$ra->{repos_root}\E(/|$)##;
5160                foreach my $path (keys %$fetch) {
5161                        my $ref_id = $fetch->{$path};
5162                        my $gs = Git::SVN->new($ref_id, $repo_id, $path);
5163
5164                        # make sure we can read when connecting to
5165                        # a higher level of a repository
5166                        my ($last_rev, undef) = $gs->last_rev_commit;
5167                        if (!defined $last_rev) {
5168                                $last_rev = eval {
5169                                        $root_ra->get_latest_revnum;
5170                                };
5171                                next if $@;
5172                        }
5173                        my $new = $root_path;
5174                        $new .= length $path ? "/$path" : '';
5175                        eval {
5176                                $root_ra->get_log([$new], $last_rev, $last_rev,
5177                                                  0, 0, 1, sub { });
5178                        };
5179                        next if $@;
5180                        $new_urls->{$ra->{repos_root}}->{$new} =
5181                                { ref_id => $ref_id,
5182                                  old_repo_id => $repo_id,
5183                                  old_path => $path };
5184                }
5185        }
5186
5187        my @emptied;
5188        foreach my $url (keys %$new_urls) {
5189                # see if we can re-use an existing [svn-remote "repo_id"]
5190                # instead of creating a(n ugly) new section:
5191                my $repo_id = $root_repos->{$url} || $url;
5192
5193                my $fetch = $new_urls->{$url};
5194                foreach my $path (keys %$fetch) {
5195                        my $x = $fetch->{$path};
5196                        Git::SVN->init($url, $path, $repo_id, $x->{ref_id});
5197                        my $pfx = "svn-remote.$x->{old_repo_id}";
5198
5199                        my $old_fetch = quotemeta("$x->{old_path}:".
5200                                                  "refs/remotes/$x->{ref_id}");
5201                        command_noisy(qw/config --unset/,
5202                                      "$pfx.fetch", '^'. $old_fetch . '$');
5203                        delete $r->{$x->{old_repo_id}}->
5204                               {fetch}->{$x->{old_path}};
5205                        if (!keys %{$r->{$x->{old_repo_id}}->{fetch}}) {
5206                                command_noisy(qw/config --unset/,
5207                                              "$pfx.url");
5208                                push @emptied, $x->{old_repo_id}
5209                        }
5210                }
5211        }
5212        if (@emptied) {
5213                my $file = $ENV{GIT_CONFIG} || "$ENV{GIT_DIR}/config";
5214                print STDERR <<EOF;
5215The following [svn-remote] sections in your config file ($file) are empty
5216and can be safely removed:
5217EOF
5218                print STDERR "[svn-remote \"$_\"]\n" foreach @emptied;
5219        }
5220}
5221
5222sub migration_check {
5223        migrate_from_v0();
5224        migrate_from_v1();
5225        migrate_from_v2();
5226        minimize_connections() if $_minimize;
5227}
5228
5229package Git::IndexInfo;
5230use strict;
5231use warnings;
5232use Git qw/command_input_pipe command_close_pipe/;
5233
5234sub new {
5235        my ($class) = @_;
5236        my ($gui, $ctx) = command_input_pipe(qw/update-index -z --index-info/);
5237        bless { gui => $gui, ctx => $ctx, nr => 0}, $class;
5238}
5239
5240sub remove {
5241        my ($self, $path) = @_;
5242        if (print { $self->{gui} } '0 ', 0 x 40, "\t", $path, "\0") {
5243                return ++$self->{nr};
5244        }
5245        undef;
5246}
5247
5248sub update {
5249        my ($self, $mode, $hash, $path) = @_;
5250        if (print { $self->{gui} } $mode, ' ', $hash, "\t", $path, "\0") {
5251                return ++$self->{nr};
5252        }
5253        undef;
5254}
5255
5256sub DESTROY {
5257        my ($self) = @_;
5258        command_close_pipe($self->{gui}, $self->{ctx});
5259}
5260
5261package Git::SVN::GlobSpec;
5262use strict;
5263use warnings;
5264
5265sub new {
5266        my ($class, $glob) = @_;
5267        my $re = $glob;
5268        $re =~ s!/+$!!g; # no need for trailing slashes
5269        $re =~ m!^([^*]*)(\*(?:/\*)*)([^*]*)$!;
5270        my $temp = $re;
5271        my ($left, $right) = ($1, $3);
5272        $re = $2;
5273        my $depth = $re =~ tr/*/*/;
5274        if ($depth != $temp =~ tr/*/*/) {
5275                die "Only one set of wildcard directories " .
5276                        "(e.g. '*' or '*/*/*') is supported: '$glob'\n";
5277        }
5278        if ($depth == 0) {
5279                die "One '*' is needed for glob: '$glob'\n";
5280        }
5281        $re =~ s!\*!\[^/\]*!g;
5282        $re = quotemeta($left) . "($re)" . quotemeta($right);
5283        if (length $left && !($left =~ s!/+$!!g)) {
5284                die "Missing trailing '/' on left side of: '$glob' ($left)\n";
5285        }
5286        if (length $right && !($right =~ s!^/+!!g)) {
5287                die "Missing leading '/' on right side of: '$glob' ($right)\n";
5288        }
5289        my $left_re = qr/^\/\Q$left\E(\/|$)/;
5290        bless { left => $left, right => $right, left_regex => $left_re,
5291                regex => qr/$re/, glob => $glob, depth => $depth }, $class;
5292}
5293
5294sub full_path {
5295        my ($self, $path) = @_;
5296        return (length $self->{left} ? "$self->{left}/" : '') .
5297               $path . (length $self->{right} ? "/$self->{right}" : '');
5298}
5299
5300__END__
5301
5302Data structures:
5303
5304
5305$remotes = { # returned by read_all_remotes()
5306        'svn' => {
5307                # svn-remote.svn.url=https://svn.musicpd.org
5308                url => 'https://svn.musicpd.org',
5309                # svn-remote.svn.fetch=mpd/trunk:trunk
5310                fetch => {
5311                        'mpd/trunk' => 'trunk',
5312                },
5313                # svn-remote.svn.tags=mpd/tags/*:tags/*
5314                tags => {
5315                        path => {
5316                                left => 'mpd/tags',
5317                                right => '',
5318                                regex => qr!mpd/tags/([^/]+)$!,
5319                                glob => 'tags/*',
5320                        },
5321                        ref => {
5322                                left => 'tags',
5323                                right => '',
5324                                regex => qr!tags/([^/]+)$!,
5325                                glob => 'tags/*',
5326                        },
5327                }
5328        }
5329};
5330
5331$log_entry hashref as returned by libsvn_log_entry()
5332{
5333        log => 'whitespace-formatted log entry
5334',                                              # trailing newline is preserved
5335        revision => '8',                        # integer
5336        date => '2004-02-24T17:01:44.108345Z',  # commit date
5337        author => 'committer name'
5338};
5339
5340
5341# this is generated by generate_diff();
5342@mods = array of diff-index line hashes, each element represents one line
5343        of diff-index output
5344
5345diff-index line ($m hash)
5346{
5347        mode_a => first column of diff-index output, no leading ':',
5348        mode_b => second column of diff-index output,
5349        sha1_b => sha1sum of the final blob,
5350        chg => change type [MCRADT],
5351        file_a => original file name of a file (iff chg is 'C' or 'R')
5352        file_b => new/current file name of a file (any chg)
5353}
5354;
5355
5356# retval of read_url_paths{,_all}();
5357$l_map = {
5358        # repository root url
5359        'https://svn.musicpd.org' => {
5360                # repository path               # GIT_SVN_ID
5361                'mpd/trunk'             =>      'trunk',
5362                'mpd/tags/0.11.5'       =>      'tags/0.11.5',
5363        },
5364}
5365
5366Notes:
5367        I don't trust the each() function on unless I created %hash myself
5368        because the internal iterator may not have started at base.