Andrew's git / gitweb.git / git-difftool.perl
?
git-difftool.perlon commit Merge branch 'jk/xdiff-memory-limits' into maint-2.3 (92cdfd2)
   1#!/usr/bin/perl
   2# Copyright (c) 2009, 2010 David Aguilar
   3# Copyright (c) 2012 Tim Henigan
   4#
   5# This is a wrapper around the GIT_EXTERNAL_DIFF-compatible
   6# git-difftool--helper script.
   7#
   8# This script exports GIT_EXTERNAL_DIFF and GIT_PAGER for use by git.
   9# The GIT_DIFF* variables are exported for use by git-difftool--helper.
  10#
  11# Any arguments that are unknown to this script are forwarded to 'git diff'.
  12
  13use 5.008;
  14use strict;
  15use warnings;
  16use Error qw(:try);
  17use File::Basename qw(dirname);
  18use File::Copy;
  19use File::Find;
  20use File::stat;
  21use File::Path qw(mkpath rmtree);
  22use File::Temp qw(tempdir);
  23use Getopt::Long qw(:config pass_through);
  24use Git;
  25
  26sub usage
  27{
  28        my $exitcode = shift;
  29        print << 'USAGE';
  30usage: git difftool [-t|--tool=<tool>] [--tool-help]
  31                    [-x|--extcmd=<cmd>]
  32                    [-g|--gui] [--no-gui]
  33                    [--prompt] [-y|--no-prompt]
  34                    [-d|--dir-diff]
  35                    ['git diff' options]
  36USAGE
  37        exit($exitcode);
  38}
  39
  40sub find_worktree
  41{
  42        # Git->repository->wc_path() does not honor changes to the working
  43        # tree location made by $ENV{GIT_WORK_TREE} or the 'core.worktree'
  44        # config variable.
  45        return Git::command_oneline('rev-parse', '--show-toplevel');
  46}
  47
  48sub print_tool_help
  49{
  50        # See the comment at the bottom of file_diff() for the reason behind
  51        # using system() followed by exit() instead of exec().
  52        my $rc = system(qw(git mergetool --tool-help=diff));
  53        exit($rc | ($rc >> 8));
  54}
  55
  56sub exit_cleanup
  57{
  58        my ($tmpdir, $status) = @_;
  59        my $errno = $!;
  60        rmtree($tmpdir);
  61        if ($status and $errno) {
  62                my ($package, $file, $line) = caller();
  63                warn "$file line $line: $errno\n";
  64        }
  65        exit($status | ($status >> 8));
  66}
  67
  68sub use_wt_file
  69{
  70        my ($repo, $workdir, $file, $sha1) = @_;
  71        my $null_sha1 = '0' x 40;
  72
  73        if (! -e "$workdir/$file") {
  74                # If the file doesn't exist in the working tree, we cannot
  75                # use it.
  76                return (0, $null_sha1);
  77        }
  78
  79        my $wt_sha1 = $repo->command_oneline('hash-object', "$workdir/$file");
  80        my $use = ($sha1 eq $null_sha1) || ($sha1 eq $wt_sha1);
  81        return ($use, $wt_sha1);
  82}
  83
  84sub changed_files
  85{
  86        my ($repo_path, $index, $worktree) = @_;
  87        $ENV{GIT_INDEX_FILE} = $index;
  88        $ENV{GIT_WORK_TREE} = $worktree;
  89        my $must_unset_git_dir = 0;
  90        if (not defined($ENV{GIT_DIR})) {
  91                $must_unset_git_dir = 1;
  92                $ENV{GIT_DIR} = $repo_path;
  93        }
  94
  95        my @refreshargs = qw/update-index --really-refresh -q --unmerged/;
  96        my @gitargs = qw/diff-files --name-only -z/;
  97        try {
  98                Git::command_oneline(@refreshargs);
  99        } catch Git::Error::Command with {};
 100
 101        my $line = Git::command_oneline(@gitargs);
 102        my @files;
 103        if (defined $line) {
 104                @files = split('\0', $line);
 105        } else {
 106                @files = ();
 107        }
 108
 109        delete($ENV{GIT_INDEX_FILE});
 110        delete($ENV{GIT_WORK_TREE});
 111        delete($ENV{GIT_DIR}) if ($must_unset_git_dir);
 112
 113        return map { $_ => 1 } @files;
 114}
 115
 116sub setup_dir_diff
 117{
 118        my ($repo, $workdir, $symlinks) = @_;
 119
 120        # Run the diff; exit immediately if no diff found
 121        # 'Repository' and 'WorkingCopy' must be explicitly set to insure that
 122        # if $GIT_DIR and $GIT_WORK_TREE are set in ENV, they are actually used
 123        # by Git->repository->command*.
 124        my $repo_path = $repo->repo_path();
 125        my %repo_args = (Repository => $repo_path, WorkingCopy => $workdir);
 126        my $diffrepo = Git->repository(%repo_args);
 127
 128        my @gitargs = ('diff', '--raw', '--no-abbrev', '-z', @ARGV);
 129        my $diffrtn = $diffrepo->command_oneline(@gitargs);
 130        exit(0) unless defined($diffrtn);
 131
 132        # Build index info for left and right sides of the diff
 133        my $submodule_mode = '160000';
 134        my $symlink_mode = '120000';
 135        my $null_mode = '0' x 6;
 136        my $null_sha1 = '0' x 40;
 137        my $lindex = '';
 138        my $rindex = '';
 139        my $wtindex = '';
 140        my %submodule;
 141        my %symlink;
 142        my @working_tree = ();
 143        my @rawdiff = split('\0', $diffrtn);
 144
 145        my $i = 0;
 146        while ($i < $#rawdiff) {
 147                if ($rawdiff[$i] =~ /^::/) {
 148                        warn << 'EOF';
 149Combined diff formats ('-c' and '--cc') are not supported in
 150directory diff mode ('-d' and '--dir-diff').
 151EOF
 152                        exit(1);
 153                }
 154
 155                my ($lmode, $rmode, $lsha1, $rsha1, $status) =
 156                        split(' ', substr($rawdiff[$i], 1));
 157                my $src_path = $rawdiff[$i + 1];
 158                my $dst_path;
 159
 160                if ($status =~ /^[CR]/) {
 161                        $dst_path = $rawdiff[$i + 2];
 162                        $i += 3;
 163                } else {
 164                        $dst_path = $src_path;
 165                        $i += 2;
 166                }
 167
 168                if ($lmode eq $submodule_mode or $rmode eq $submodule_mode) {
 169                        $submodule{$src_path}{left} = $lsha1;
 170                        if ($lsha1 ne $rsha1) {
 171                                $submodule{$dst_path}{right} = $rsha1;
 172                        } else {
 173                                $submodule{$dst_path}{right} = "$rsha1-dirty";
 174                        }
 175                        next;
 176                }
 177
 178                if ($lmode eq $symlink_mode) {
 179                        $symlink{$src_path}{left} =
 180                                $diffrepo->command_oneline('show', "$lsha1");
 181                }
 182
 183                if ($rmode eq $symlink_mode) {
 184                        $symlink{$dst_path}{right} =
 185                                $diffrepo->command_oneline('show', "$rsha1");
 186                }
 187
 188                if ($lmode ne $null_mode and $status !~ /^C/) {
 189                        $lindex .= "$lmode $lsha1\t$src_path\0";
 190                }
 191
 192                if ($rmode ne $null_mode) {
 193                        my ($use, $wt_sha1) = use_wt_file($repo, $workdir,
 194                                                          $dst_path, $rsha1);
 195                        if ($use) {
 196                                push @working_tree, $dst_path;
 197                                $wtindex .= "$rmode $wt_sha1\t$dst_path\0";
 198                        } else {
 199                                $rindex .= "$rmode $rsha1\t$dst_path\0";
 200                        }
 201                }
 202        }
 203
 204        # Setup temp directories
 205        my $tmpdir = tempdir('git-difftool.XXXXX', CLEANUP => 0, TMPDIR => 1);
 206        my $ldir = "$tmpdir/left";
 207        my $rdir = "$tmpdir/right";
 208        mkpath($ldir) or exit_cleanup($tmpdir, 1);
 209        mkpath($rdir) or exit_cleanup($tmpdir, 1);
 210
 211        # If $GIT_DIR is not set prior to calling 'git update-index' and
 212        # 'git checkout-index', then those commands will fail if difftool
 213        # is called from a directory other than the repo root.
 214        my $must_unset_git_dir = 0;
 215        if (not defined($ENV{GIT_DIR})) {
 216                $must_unset_git_dir = 1;
 217                $ENV{GIT_DIR} = $repo_path;
 218        }
 219
 220        # Populate the left and right directories based on each index file
 221        my ($inpipe, $ctx);
 222        $ENV{GIT_INDEX_FILE} = "$tmpdir/lindex";
 223        ($inpipe, $ctx) =
 224                $repo->command_input_pipe(qw(update-index -z --index-info));
 225        print($inpipe $lindex);
 226        $repo->command_close_pipe($inpipe, $ctx);
 227
 228        my $rc = system('git', 'checkout-index', '--all', "--prefix=$ldir/");
 229        exit_cleanup($tmpdir, $rc) if $rc != 0;
 230
 231        $ENV{GIT_INDEX_FILE} = "$tmpdir/rindex";
 232        ($inpipe, $ctx) =
 233                $repo->command_input_pipe(qw(update-index -z --index-info));
 234        print($inpipe $rindex);
 235        $repo->command_close_pipe($inpipe, $ctx);
 236
 237        $rc = system('git', 'checkout-index', '--all', "--prefix=$rdir/");
 238        exit_cleanup($tmpdir, $rc) if $rc != 0;
 239
 240        $ENV{GIT_INDEX_FILE} = "$tmpdir/wtindex";
 241        ($inpipe, $ctx) =
 242                $repo->command_input_pipe(qw(update-index --info-only -z --index-info));
 243        print($inpipe $wtindex);
 244        $repo->command_close_pipe($inpipe, $ctx);
 245
 246        # If $GIT_DIR was explicitly set just for the update/checkout
 247        # commands, then it should be unset before continuing.
 248        delete($ENV{GIT_DIR}) if ($must_unset_git_dir);
 249        delete($ENV{GIT_INDEX_FILE});
 250
 251        # Changes in the working tree need special treatment since they are
 252        # not part of the index. Remove any trailing slash from $workdir
 253        # before starting to avoid double slashes in symlink targets.
 254        $workdir =~ s|/$||;
 255        for my $file (@working_tree) {
 256                my $dir = dirname($file);
 257                unless (-d "$rdir/$dir") {
 258                        mkpath("$rdir/$dir") or
 259                        exit_cleanup($tmpdir, 1);
 260                }
 261                if ($symlinks) {
 262                        symlink("$workdir/$file", "$rdir/$file") or
 263                        exit_cleanup($tmpdir, 1);
 264                } else {
 265                        copy("$workdir/$file", "$rdir/$file") or
 266                        exit_cleanup($tmpdir, 1);
 267
 268                        my $mode = stat("$workdir/$file")->mode;
 269                        chmod($mode, "$rdir/$file") or
 270                        exit_cleanup($tmpdir, 1);
 271                }
 272        }
 273
 274        # Changes to submodules require special treatment. This loop writes a
 275        # temporary file to both the left and right directories to show the
 276        # change in the recorded SHA1 for the submodule.
 277        for my $path (keys %submodule) {
 278                my $ok;
 279                if (defined($submodule{$path}{left})) {
 280                        $ok = write_to_file("$ldir/$path",
 281                                "Subproject commit $submodule{$path}{left}");
 282                }
 283                if (defined($submodule{$path}{right})) {
 284                        $ok = write_to_file("$rdir/$path",
 285                                "Subproject commit $submodule{$path}{right}");
 286                }
 287                exit_cleanup($tmpdir, 1) if not $ok;
 288        }
 289
 290        # Symbolic links require special treatment. The standard "git diff"
 291        # shows only the link itself, not the contents of the link target.
 292        # This loop replicates that behavior.
 293        for my $path (keys %symlink) {
 294                my $ok;
 295                if (defined($symlink{$path}{left})) {
 296                        $ok = write_to_file("$ldir/$path",
 297                                        $symlink{$path}{left});
 298                }
 299                if (defined($symlink{$path}{right})) {
 300                        $ok = write_to_file("$rdir/$path",
 301                                        $symlink{$path}{right});
 302                }
 303                exit_cleanup($tmpdir, 1) if not $ok;
 304        }
 305
 306        return ($ldir, $rdir, $tmpdir, @working_tree);
 307}
 308
 309sub write_to_file
 310{
 311        my $path = shift;
 312        my $value = shift;
 313
 314        # Make sure the path to the file exists
 315        my $dir = dirname($path);
 316        unless (-d "$dir") {
 317                mkpath("$dir") or return 0;
 318        }
 319
 320        # If the file already exists in that location, delete it.  This
 321        # is required in the case of symbolic links.
 322        unlink($path);
 323
 324        open(my $fh, '>', $path) or return 0;
 325        print($fh $value);
 326        close($fh);
 327
 328        return 1;
 329}
 330
 331sub main
 332{
 333        # parse command-line options. all unrecognized options and arguments
 334        # are passed through to the 'git diff' command.
 335        my %opts = (
 336                difftool_cmd => undef,
 337                dirdiff => undef,
 338                extcmd => undef,
 339                gui => undef,
 340                help => undef,
 341                prompt => undef,
 342                symlinks => $^O ne 'cygwin' &&
 343                                $^O ne 'MSWin32' && $^O ne 'msys',
 344                tool_help => undef,
 345                trust_exit_code => undef,
 346        );
 347        GetOptions('g|gui!' => \$opts{gui},
 348                'd|dir-diff' => \$opts{dirdiff},
 349                'h' => \$opts{help},
 350                'prompt!' => \$opts{prompt},
 351                'y' => sub { $opts{prompt} = 0; },
 352                'symlinks' => \$opts{symlinks},
 353                'no-symlinks' => sub { $opts{symlinks} = 0; },
 354                't|tool:s' => \$opts{difftool_cmd},
 355                'tool-help' => \$opts{tool_help},
 356                'trust-exit-code' => \$opts{trust_exit_code},
 357                'no-trust-exit-code' => sub { $opts{trust_exit_code} = 0; },
 358                'x|extcmd:s' => \$opts{extcmd});
 359
 360        if (defined($opts{help})) {
 361                usage(0);
 362        }
 363        if (defined($opts{tool_help})) {
 364                print_tool_help();
 365        }
 366        if (defined($opts{difftool_cmd})) {
 367                if (length($opts{difftool_cmd}) > 0) {
 368                        $ENV{GIT_DIFF_TOOL} = $opts{difftool_cmd};
 369                } else {
 370                        print "No <tool> given for --tool=<tool>\n";
 371                        usage(1);
 372                }
 373        }
 374        if (defined($opts{extcmd})) {
 375                if (length($opts{extcmd}) > 0) {
 376                        $ENV{GIT_DIFFTOOL_EXTCMD} = $opts{extcmd};
 377                } else {
 378                        print "No <cmd> given for --extcmd=<cmd>\n";
 379                        usage(1);
 380                }
 381        }
 382        if ($opts{gui}) {
 383                my $guitool = Git::config('diff.guitool');
 384                if (defined($guitool) && length($guitool) > 0) {
 385                        $ENV{GIT_DIFF_TOOL} = $guitool;
 386                }
 387        }
 388
 389        if (!defined $opts{trust_exit_code}) {
 390                $opts{trust_exit_code} = Git::config_bool('difftool.trustExitCode');
 391        }
 392        if ($opts{trust_exit_code}) {
 393                $ENV{GIT_DIFFTOOL_TRUST_EXIT_CODE} = 'true';
 394        } else {
 395                $ENV{GIT_DIFFTOOL_TRUST_EXIT_CODE} = 'false';
 396        }
 397
 398        # In directory diff mode, 'git-difftool--helper' is called once
 399        # to compare the a/b directories.  In file diff mode, 'git diff'
 400        # will invoke a separate instance of 'git-difftool--helper' for
 401        # each file that changed.
 402        if (defined($opts{dirdiff})) {
 403                dir_diff($opts{extcmd}, $opts{symlinks});
 404        } else {
 405                file_diff($opts{prompt});
 406        }
 407}
 408
 409sub dir_diff
 410{
 411        my ($extcmd, $symlinks) = @_;
 412        my $rc;
 413        my $error = 0;
 414        my $repo = Git->repository();
 415        my $workdir = find_worktree();
 416        my ($a, $b, $tmpdir, @worktree) =
 417                setup_dir_diff($repo, $workdir, $symlinks);
 418
 419        if (defined($extcmd)) {
 420                $rc = system($extcmd, $a, $b);
 421        } else {
 422                $ENV{GIT_DIFFTOOL_DIRDIFF} = 'true';
 423                $rc = system('git', 'difftool--helper', $a, $b);
 424        }
 425        # If the diff including working copy files and those
 426        # files were modified during the diff, then the changes
 427        # should be copied back to the working tree.
 428        # Do not copy back files when symlinks are used and the
 429        # external tool did not replace the original link with a file.
 430        #
 431        # These hashes are loaded lazily since they aren't needed
 432        # in the common case of --symlinks and the difftool updating
 433        # files through the symlink.
 434        my %wt_modified;
 435        my %tmp_modified;
 436        my $indices_loaded = 0;
 437
 438        for my $file (@worktree) {
 439                next if $symlinks && -l "$b/$file";
 440                next if ! -f "$b/$file";
 441
 442                if (!$indices_loaded) {
 443                        %wt_modified = changed_files($repo->repo_path(),
 444                                "$tmpdir/wtindex", "$workdir");
 445                        %tmp_modified = changed_files($repo->repo_path(),
 446                                "$tmpdir/wtindex", "$b");
 447                        $indices_loaded = 1;
 448                }
 449
 450                if (exists $wt_modified{$file} and exists $tmp_modified{$file}) {
 451                        my $errmsg = "warning: Both files modified: ";
 452                        $errmsg .= "'$workdir/$file' and '$b/$file'.\n";
 453                        $errmsg .= "warning: Working tree file has been left.\n";
 454                        $errmsg .= "warning:\n";
 455                        warn $errmsg;
 456                        $error = 1;
 457                } elsif (exists $tmp_modified{$file}) {
 458                        my $mode = stat("$b/$file")->mode;
 459                        copy("$b/$file", "$workdir/$file") or
 460                        exit_cleanup($tmpdir, 1);
 461
 462                        chmod($mode, "$workdir/$file") or
 463                        exit_cleanup($tmpdir, 1);
 464                }
 465        }
 466        if ($error) {
 467                warn "warning: Temporary files exist in '$tmpdir'.\n";
 468                warn "warning: You may want to cleanup or recover these.\n";
 469                exit(1);
 470        } else {
 471                exit_cleanup($tmpdir, $rc);
 472        }
 473}
 474
 475sub file_diff
 476{
 477        my ($prompt) = @_;
 478
 479        if (defined($prompt)) {
 480                if ($prompt) {
 481                        $ENV{GIT_DIFFTOOL_PROMPT} = 'true';
 482                } else {
 483                        $ENV{GIT_DIFFTOOL_NO_PROMPT} = 'true';
 484                }
 485        }
 486
 487        $ENV{GIT_PAGER} = '';
 488        $ENV{GIT_EXTERNAL_DIFF} = 'git-difftool--helper';
 489
 490        # ActiveState Perl for Win32 does not implement POSIX semantics of
 491        # exec* system call. It just spawns the given executable and finishes
 492        # the starting program, exiting with code 0.
 493        # system will at least catch the errors returned by git diff,
 494        # allowing the caller of git difftool better handling of failures.
 495        my $rc = system('git', 'diff', @ARGV);
 496        exit($rc | ($rc >> 8));
 497}
 498
 499main();