perl / Git / SVN / Editor.pmon commit Merge branch 'jk/diff-algo' into maint (b28325d)
   1package Git::SVN::Editor;
   2use vars qw/@ISA $_rmdir $_cp_similarity $_find_copies_harder $_rename_limit/;
   3use strict;
   4use warnings;
   5use SVN::Core;
   6use SVN::Delta;
   7use Carp qw/croak/;
   8use IO::File;
   9use Git qw/command command_oneline command_noisy command_output_pipe
  10           command_input_pipe command_close_pipe
  11           command_bidi_pipe command_close_bidi_pipe/;
  12BEGIN {
  13        @ISA = qw(SVN::Delta::Editor);
  14}
  15
  16sub new {
  17        my ($class, $opts) = @_;
  18        foreach (qw/svn_path r ra tree_a tree_b log editor_cb/) {
  19                die "$_ required!\n" unless (defined $opts->{$_});
  20        }
  21
  22        my $pool = SVN::Pool->new;
  23        my $mods = generate_diff($opts->{tree_a}, $opts->{tree_b});
  24        my $types = check_diff_paths($opts->{ra}, $opts->{svn_path},
  25                                     $opts->{r}, $mods);
  26
  27        # $opts->{ra} functions should not be used after this:
  28        my @ce  = $opts->{ra}->get_commit_editor($opts->{log},
  29                                                $opts->{editor_cb}, $pool);
  30        my $self = SVN::Delta::Editor->new(@ce, $pool);
  31        bless $self, $class;
  32        foreach (qw/svn_path r tree_a tree_b/) {
  33                $self->{$_} = $opts->{$_};
  34        }
  35        $self->{url} = $opts->{ra}->{url};
  36        $self->{mods} = $mods;
  37        $self->{types} = $types;
  38        $self->{pool} = $pool;
  39        $self->{bat} = { '' => $self->open_root($self->{r}, $self->{pool}) };
  40        $self->{rm} = { };
  41        $self->{path_prefix} = length $self->{svn_path} ?
  42                               "$self->{svn_path}/" : '';
  43        $self->{config} = $opts->{config};
  44        $self->{mergeinfo} = $opts->{mergeinfo};
  45        return $self;
  46}
  47
  48sub generate_diff {
  49        my ($tree_a, $tree_b) = @_;
  50        my @diff_tree = qw(diff-tree -z -r);
  51        if ($_cp_similarity) {
  52                push @diff_tree, "-C$_cp_similarity";
  53        } else {
  54                push @diff_tree, '-C';
  55        }
  56        push @diff_tree, '--find-copies-harder' if $_find_copies_harder;
  57        push @diff_tree, "-l$_rename_limit" if defined $_rename_limit;
  58        push @diff_tree, $tree_a, $tree_b;
  59        my ($diff_fh, $ctx) = command_output_pipe(@diff_tree);
  60        local $/ = "\0";
  61        my $state = 'meta';
  62        my @mods;
  63        while (<$diff_fh>) {
  64                chomp $_; # this gets rid of the trailing "\0"
  65                if ($state eq 'meta' && /^:(\d{6})\s(\d{6})\s
  66                                        ($::sha1)\s($::sha1)\s
  67                                        ([MTCRAD])\d*$/xo) {
  68                        push @mods, {   mode_a => $1, mode_b => $2,
  69                                        sha1_a => $3, sha1_b => $4,
  70                                        chg => $5 };
  71                        if ($5 =~ /^(?:C|R)$/) {
  72                                $state = 'file_a';
  73                        } else {
  74                                $state = 'file_b';
  75                        }
  76                } elsif ($state eq 'file_a') {
  77                        my $x = $mods[$#mods] or croak "Empty array\n";
  78                        if ($x->{chg} !~ /^(?:C|R)$/) {
  79                                croak "Error parsing $_, $x->{chg}\n";
  80                        }
  81                        $x->{file_a} = $_;
  82                        $state = 'file_b';
  83                } elsif ($state eq 'file_b') {
  84                        my $x = $mods[$#mods] or croak "Empty array\n";
  85                        if (exists $x->{file_a} && $x->{chg} !~ /^(?:C|R)$/) {
  86                                croak "Error parsing $_, $x->{chg}\n";
  87                        }
  88                        if (!exists $x->{file_a} && $x->{chg} =~ /^(?:C|R)$/) {
  89                                croak "Error parsing $_, $x->{chg}\n";
  90                        }
  91                        $x->{file_b} = $_;
  92                        $state = 'meta';
  93                } else {
  94                        croak "Error parsing $_\n";
  95                }
  96        }
  97        command_close_pipe($diff_fh, $ctx);
  98        \@mods;
  99}
 100
 101sub check_diff_paths {
 102        my ($ra, $pfx, $rev, $mods) = @_;
 103        my %types;
 104        $pfx .= '/' if length $pfx;
 105
 106        sub type_diff_paths {
 107                my ($ra, $types, $path, $rev) = @_;
 108                my @p = split m#/+#, $path;
 109                my $c = shift @p;
 110                unless (defined $types->{$c}) {
 111                        $types->{$c} = $ra->check_path($c, $rev);
 112                }
 113                while (@p) {
 114                        $c .= '/' . shift @p;
 115                        next if defined $types->{$c};
 116                        $types->{$c} = $ra->check_path($c, $rev);
 117                }
 118        }
 119
 120        foreach my $m (@$mods) {
 121                foreach my $f (qw/file_a file_b/) {
 122                        next unless defined $m->{$f};
 123                        my ($dir) = ($m->{$f} =~ m#^(.*?)/?(?:[^/]+)$#);
 124                        if (length $pfx.$dir && ! defined $types{$dir}) {
 125                                type_diff_paths($ra, \%types, $pfx.$dir, $rev);
 126                        }
 127                }
 128        }
 129        \%types;
 130}
 131
 132sub split_path {
 133        return ($_[0] =~ m#^(.*?)/?([^/]+)$#);
 134}
 135
 136sub repo_path {
 137        my ($self, $path) = @_;
 138        if (my $enc = $self->{pathnameencoding}) {
 139                require Encode;
 140                Encode::from_to($path, $enc, 'UTF-8');
 141        }
 142        $self->{path_prefix}.(defined $path ? $path : '');
 143}
 144
 145sub url_path {
 146        my ($self, $path) = @_;
 147        if ($self->{url} =~ m#^https?://#) {
 148                # characters are taken from subversion/libsvn_subr/path.c
 149                $path =~ s#([^~a-zA-Z0-9_./!$&'()*+,-])#sprintf("%%%02X",ord($1))#eg;
 150        }
 151        $self->{url} . '/' . $self->repo_path($path);
 152}
 153
 154sub rmdirs {
 155        my ($self) = @_;
 156        my $rm = $self->{rm};
 157        delete $rm->{''}; # we never delete the url we're tracking
 158        return unless %$rm;
 159
 160        foreach (keys %$rm) {
 161                my @d = split m#/#, $_;
 162                my $c = shift @d;
 163                $rm->{$c} = 1;
 164                while (@d) {
 165                        $c .= '/' . shift @d;
 166                        $rm->{$c} = 1;
 167                }
 168        }
 169        delete $rm->{$self->{svn_path}};
 170        delete $rm->{''}; # we never delete the url we're tracking
 171        return unless %$rm;
 172
 173        my ($fh, $ctx) = command_output_pipe(qw/ls-tree --name-only -r -z/,
 174                                             $self->{tree_b});
 175        local $/ = "\0";
 176        while (<$fh>) {
 177                chomp;
 178                my @dn = split m#/#, $_;
 179                while (pop @dn) {
 180                        delete $rm->{join '/', @dn};
 181                }
 182                unless (%$rm) {
 183                        close $fh;
 184                        return;
 185                }
 186        }
 187        command_close_pipe($fh, $ctx);
 188
 189        my ($r, $p, $bat) = ($self->{r}, $self->{pool}, $self->{bat});
 190        foreach my $d (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$rm) {
 191                $self->close_directory($bat->{$d}, $p);
 192                my ($dn) = ($d =~ m#^(.*?)/?(?:[^/]+)$#);
 193                print "\tD+\t$d/\n" unless $::_q;
 194                $self->SUPER::delete_entry($d, $r, $bat->{$dn}, $p);
 195                delete $bat->{$d};
 196        }
 197}
 198
 199sub open_or_add_dir {
 200        my ($self, $full_path, $baton, $deletions) = @_;
 201        my $t = $self->{types}->{$full_path};
 202        if (!defined $t) {
 203                die "$full_path not known in r$self->{r} or we have a bug!\n";
 204        }
 205        {
 206                no warnings 'once';
 207                # SVN::Node::none and SVN::Node::file are used only once,
 208                # so we're shutting up Perl's warnings about them.
 209                if ($t == $SVN::Node::none || defined($deletions->{$full_path})) {
 210                        return $self->add_directory($full_path, $baton,
 211                            undef, -1, $self->{pool});
 212                } elsif ($t == $SVN::Node::dir) {
 213                        return $self->open_directory($full_path, $baton,
 214                            $self->{r}, $self->{pool});
 215                } # no warnings 'once'
 216                print STDERR "$full_path already exists in repository at ",
 217                    "r$self->{r} and it is not a directory (",
 218                    ($t == $SVN::Node::file ? 'file' : 'unknown'),"/$t)\n";
 219        } # no warnings 'once'
 220        exit 1;
 221}
 222
 223sub ensure_path {
 224        my ($self, $path, $deletions) = @_;
 225        my $bat = $self->{bat};
 226        my $repo_path = $self->repo_path($path);
 227        return $bat->{''} unless (length $repo_path);
 228
 229        my @p = split m#/+#, $repo_path;
 230        my $c = shift @p;
 231        $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{''}, $deletions);
 232        while (@p) {
 233                my $c0 = $c;
 234                $c .= '/' . shift @p;
 235                $bat->{$c} ||= $self->open_or_add_dir($c, $bat->{$c0}, $deletions);
 236        }
 237        return $bat->{$c};
 238}
 239
 240# Subroutine to convert a globbing pattern to a regular expression.
 241# From perl cookbook.
 242sub glob2pat {
 243        my $globstr = shift;
 244        my %patmap = ('*' => '.*', '?' => '.', '[' => '[', ']' => ']');
 245        $globstr =~ s{(.)} { $patmap{$1} || "\Q$1" }ge;
 246        return '^' . $globstr . '$';
 247}
 248
 249sub check_autoprop {
 250        my ($self, $pattern, $properties, $file, $fbat) = @_;
 251        # Convert the globbing pattern to a regular expression.
 252        my $regex = glob2pat($pattern);
 253        # Check if the pattern matches the file name.
 254        if($file =~ m/($regex)/) {
 255                # Parse the list of properties to set.
 256                my @props = split(/;/, $properties);
 257                foreach my $prop (@props) {
 258                        # Parse 'name=value' syntax and set the property.
 259                        if ($prop =~ /([^=]+)=(.*)/) {
 260                                my ($n,$v) = ($1,$2);
 261                                for ($n, $v) {
 262                                        s/^\s+//; s/\s+$//;
 263                                }
 264                                $self->change_file_prop($fbat, $n, $v);
 265                        }
 266                }
 267        }
 268}
 269
 270sub apply_autoprops {
 271        my ($self, $file, $fbat) = @_;
 272        my $conf_t = ${$self->{config}}{'config'};
 273        no warnings 'once';
 274        # Check [miscellany]/enable-auto-props in svn configuration.
 275        if (SVN::_Core::svn_config_get_bool(
 276                $conf_t,
 277                $SVN::_Core::SVN_CONFIG_SECTION_MISCELLANY,
 278                $SVN::_Core::SVN_CONFIG_OPTION_ENABLE_AUTO_PROPS,
 279                0)) {
 280                # Auto-props are enabled.  Enumerate them to look for matches.
 281                my $callback = sub {
 282                        $self->check_autoprop($_[0], $_[1], $file, $fbat);
 283                };
 284                SVN::_Core::svn_config_enumerate(
 285                        $conf_t,
 286                        $SVN::_Core::SVN_CONFIG_SECTION_AUTO_PROPS,
 287                        $callback);
 288        }
 289}
 290
 291sub A {
 292        my ($self, $m, $deletions) = @_;
 293        my ($dir, $file) = split_path($m->{file_b});
 294        my $pbat = $self->ensure_path($dir, $deletions);
 295        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
 296                                        undef, -1);
 297        print "\tA\t$m->{file_b}\n" unless $::_q;
 298        $self->apply_autoprops($file, $fbat);
 299        $self->chg_file($fbat, $m);
 300        $self->close_file($fbat,undef,$self->{pool});
 301}
 302
 303sub C {
 304        my ($self, $m, $deletions) = @_;
 305        my ($dir, $file) = split_path($m->{file_b});
 306        my $pbat = $self->ensure_path($dir, $deletions);
 307        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
 308                                $self->url_path($m->{file_a}), $self->{r});
 309        print "\tC\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
 310        $self->chg_file($fbat, $m);
 311        $self->close_file($fbat,undef,$self->{pool});
 312}
 313
 314sub delete_entry {
 315        my ($self, $path, $pbat) = @_;
 316        my $rpath = $self->repo_path($path);
 317        my ($dir, $file) = split_path($rpath);
 318        $self->{rm}->{$dir} = 1;
 319        $self->SUPER::delete_entry($rpath, $self->{r}, $pbat, $self->{pool});
 320}
 321
 322sub R {
 323        my ($self, $m, $deletions) = @_;
 324        my ($dir, $file) = split_path($m->{file_b});
 325        my $pbat = $self->ensure_path($dir, $deletions);
 326        my $fbat = $self->add_file($self->repo_path($m->{file_b}), $pbat,
 327                                $self->url_path($m->{file_a}), $self->{r});
 328        print "\tR\t$m->{file_a} => $m->{file_b}\n" unless $::_q;
 329        $self->apply_autoprops($file, $fbat);
 330        $self->chg_file($fbat, $m);
 331        $self->close_file($fbat,undef,$self->{pool});
 332
 333        ($dir, $file) = split_path($m->{file_a});
 334        $pbat = $self->ensure_path($dir, $deletions);
 335        $self->delete_entry($m->{file_a}, $pbat);
 336}
 337
 338sub M {
 339        my ($self, $m, $deletions) = @_;
 340        my ($dir, $file) = split_path($m->{file_b});
 341        my $pbat = $self->ensure_path($dir, $deletions);
 342        my $fbat = $self->open_file($self->repo_path($m->{file_b}),
 343                                $pbat,$self->{r},$self->{pool});
 344        print "\t$m->{chg}\t$m->{file_b}\n" unless $::_q;
 345        $self->chg_file($fbat, $m);
 346        $self->close_file($fbat,undef,$self->{pool});
 347}
 348
 349sub T {
 350        my ($self, $m, $deletions) = @_;
 351
 352        # Work around subversion issue 4091: toggling the "is a
 353        # symlink" property requires removing and re-adding a
 354        # file or else "svn up" on affected clients trips an
 355        # assertion and aborts.
 356        if (($m->{mode_b} =~ /^120/ && $m->{mode_a} !~ /^120/) ||
 357            ($m->{mode_b} !~ /^120/ && $m->{mode_a} =~ /^120/)) {
 358                $self->D({
 359                        mode_a => $m->{mode_a}, mode_b => '000000',
 360                        sha1_a => $m->{sha1_a}, sha1_b => '0' x 40,
 361                        chg => 'D', file_b => $m->{file_b}
 362                }, $deletions);
 363                $self->A({
 364                        mode_a => '000000', mode_b => $m->{mode_b},
 365                        sha1_a => '0' x 40, sha1_b => $m->{sha1_b},
 366                        chg => 'A', file_b => $m->{file_b}
 367                }, $deletions);
 368                return;
 369        }
 370
 371        $self->M($m, $deletions);
 372}
 373
 374sub change_file_prop {
 375        my ($self, $fbat, $pname, $pval) = @_;
 376        $self->SUPER::change_file_prop($fbat, $pname, $pval, $self->{pool});
 377}
 378
 379sub change_dir_prop {
 380        my ($self, $pbat, $pname, $pval) = @_;
 381        $self->SUPER::change_dir_prop($pbat, $pname, $pval, $self->{pool});
 382}
 383
 384sub _chg_file_get_blob ($$$$) {
 385        my ($self, $fbat, $m, $which) = @_;
 386        my $fh = $::_repository->temp_acquire("git_blob_$which");
 387        if ($m->{"mode_$which"} =~ /^120/) {
 388                print $fh 'link ' or croak $!;
 389                $self->change_file_prop($fbat,'svn:special','*');
 390        } elsif ($m->{mode_a} =~ /^120/ && $m->{"mode_$which"} !~ /^120/) {
 391                $self->change_file_prop($fbat,'svn:special',undef);
 392        }
 393        my $blob = $m->{"sha1_$which"};
 394        return ($fh,) if ($blob =~ /^0{40}$/);
 395        my $size = $::_repository->cat_blob($blob, $fh);
 396        croak "Failed to read object $blob" if ($size < 0);
 397        $fh->flush == 0 or croak $!;
 398        seek $fh, 0, 0 or croak $!;
 399
 400        my $exp = ::md5sum($fh);
 401        seek $fh, 0, 0 or croak $!;
 402        return ($fh, $exp);
 403}
 404
 405sub chg_file {
 406        my ($self, $fbat, $m) = @_;
 407        if ($m->{mode_b} =~ /755$/ && $m->{mode_a} !~ /755$/) {
 408                $self->change_file_prop($fbat,'svn:executable','*');
 409        } elsif ($m->{mode_b} !~ /755$/ && $m->{mode_a} =~ /755$/) {
 410                $self->change_file_prop($fbat,'svn:executable',undef);
 411        }
 412        my ($fh_a, $exp_a) = _chg_file_get_blob $self, $fbat, $m, 'a';
 413        my ($fh_b, $exp_b) = _chg_file_get_blob $self, $fbat, $m, 'b';
 414        my $pool = SVN::Pool->new;
 415        my $atd = $self->apply_textdelta($fbat, $exp_a, $pool);
 416        if (-s $fh_a) {
 417                my $txstream = SVN::TxDelta::new ($fh_a, $fh_b, $pool);
 418                my $res = SVN::TxDelta::send_txstream($txstream, @$atd, $pool);
 419                if (defined $res) {
 420                        die "Unexpected result from send_txstream: $res\n",
 421                            "(SVN::Core::VERSION: $SVN::Core::VERSION)\n";
 422                }
 423        } else {
 424                my $got = SVN::TxDelta::send_stream($fh_b, @$atd, $pool);
 425                die "Checksum mismatch\nexpected: $exp_b\ngot: $got\n"
 426                    if ($got ne $exp_b);
 427        }
 428        Git::temp_release($fh_b, 1);
 429        Git::temp_release($fh_a, 1);
 430        $pool->clear;
 431}
 432
 433sub D {
 434        my ($self, $m, $deletions) = @_;
 435        my ($dir, $file) = split_path($m->{file_b});
 436        my $pbat = $self->ensure_path($dir, $deletions);
 437        print "\tD\t$m->{file_b}\n" unless $::_q;
 438        $self->delete_entry($m->{file_b}, $pbat);
 439}
 440
 441sub close_edit {
 442        my ($self) = @_;
 443        my ($p,$bat) = ($self->{pool}, $self->{bat});
 444        foreach (sort { $b =~ tr#/#/# <=> $a =~ tr#/#/# } keys %$bat) {
 445                next if $_ eq '';
 446                $self->close_directory($bat->{$_}, $p);
 447        }
 448        $self->close_directory($bat->{''}, $p);
 449        $self->SUPER::close_edit($p);
 450        $p->clear;
 451}
 452
 453sub abort_edit {
 454        my ($self) = @_;
 455        $self->SUPER::abort_edit($self->{pool});
 456}
 457
 458sub DESTROY {
 459        my $self = shift;
 460        $self->SUPER::DESTROY(@_);
 461        $self->{pool}->clear;
 462}
 463
 464# this drives the editor
 465sub apply_diff {
 466        my ($self) = @_;
 467        my $mods = $self->{mods};
 468        my %o = ( D => 0, C => 1, R => 2, A => 3, M => 4, T => 5 );
 469        my %deletions;
 470
 471        foreach my $m (@$mods) {
 472                if ($m->{chg} eq "D") {
 473                        $deletions{$m->{file_b}} = 1;
 474                }
 475        }
 476
 477        foreach my $m (sort { $o{$a->{chg}} <=> $o{$b->{chg}} } @$mods) {
 478                my $f = $m->{chg};
 479                if (defined $o{$f}) {
 480                        $self->$f($m, \%deletions);
 481                } else {
 482                        fatal("Invalid change type: $f");
 483                }
 484        }
 485
 486        if (defined($self->{mergeinfo})) {
 487                $self->change_dir_prop($self->{bat}{''}, "svn:mergeinfo",
 488                                       $self->{mergeinfo});
 489        }
 490        $self->rmdirs if $_rmdir;
 491        if (@$mods == 0 && !defined($self->{mergeinfo})) {
 492                $self->abort_edit;
 493        } else {
 494                $self->close_edit;
 495        }
 496        return scalar @$mods;
 497}
 498
 4991;
 500__END__
 501
 502=head1 NAME
 503
 504Git::SVN::Editor - commit driver for "git svn set-tree" and dcommit
 505
 506=head1 SYNOPSIS
 507
 508        use Git::SVN::Editor;
 509        use Git::SVN::Ra;
 510
 511        my $ra = Git::SVN::Ra->new($url);
 512        my %opts = (
 513                r => 19,
 514                log => "log message",
 515                ra => $ra,
 516                config => SVN::Core::config_get_config($svn_config_dir),
 517                tree_a => "$commit^",
 518                tree_b => "$commit",
 519                editor_cb => sub { print "Committed r$_[0]\n"; },
 520                mergeinfo => "/branches/foo:1-10",
 521                svn_path => "trunk"
 522        );
 523        Git::SVN::Editor->new(\%opts)->apply_diff or print "No changes\n";
 524
 525        my $re = Git::SVN::Editor::glob2pat("trunk/*");
 526        if ($branchname =~ /$re/) {
 527                print "matched!\n";
 528        }
 529
 530=head1 DESCRIPTION
 531
 532This module is an implementation detail of the "git svn" command.
 533Do not use it unless you are developing git-svn.
 534
 535This module adapts the C<SVN::Delta::Editor> object returned by
 536C<SVN::Delta::get_commit_editor> and drives it to convey the
 537difference between two git tree objects to a remote Subversion
 538repository.
 539
 540The interface will change as git-svn evolves.
 541
 542=head1 DEPENDENCIES
 543
 544Subversion perl bindings,
 545the core L<Carp> and L<IO::File> modules,
 546and git's L<Git> helper module.
 547
 548C<Git::SVN::Editor> has not been tested using callers other than
 549B<git-svn> itself.
 550
 551=head1 SEE ALSO
 552
 553L<SVN::Delta>,
 554L<Git::SVN::Fetcher>.
 555
 556=head1 INCOMPATIBILITIES
 557
 558None reported.
 559
 560=head1 BUGS
 561
 562None.