Andrew's git / gitweb.git / git-cvsexportcommit.perl
?
git-cvsexportcommit.perlon commit unpack-trees: handle lstat failure for existing directory (92fda79)
   1#!/usr/bin/perl
   2
   3use 5.008;
   4use strict;
   5use warnings;
   6use Getopt::Std;
   7use File::Temp qw(tempdir);
   8use Data::Dumper;
   9use File::Basename qw(basename dirname);
  10use File::Spec;
  11use Git;
  12
  13our ($opt_h, $opt_P, $opt_p, $opt_v, $opt_c, $opt_f, $opt_a, $opt_m, $opt_d, $opt_u, $opt_w, $opt_W, $opt_k);
  14
  15getopts('uhPpvcfkam:d:w:W');
  16
  17$opt_h && usage();
  18
  19die "Need at least one commit identifier!" unless @ARGV;
  20
  21# Get git-config settings
  22my $repo = Git->repository();
  23$opt_w = $repo->config('cvsexportcommit.cvsdir') unless defined $opt_w;
  24
  25if ($opt_w || $opt_W) {
  26        # Remember where GIT_DIR is before changing to CVS checkout
  27        unless ($ENV{GIT_DIR}) {
  28                # No GIT_DIR set. Figure it out for ourselves
  29                my $gd =`git-rev-parse --git-dir`;
  30                chomp($gd);
  31                $ENV{GIT_DIR} = $gd;
  32        }
  33        # Make sure GIT_DIR is absolute
  34        $ENV{GIT_DIR} = File::Spec->rel2abs($ENV{GIT_DIR});
  35}
  36
  37if ($opt_w) {
  38        if (! -d $opt_w."/CVS" ) {
  39                die "$opt_w is not a CVS checkout";
  40        }
  41        chdir $opt_w or die "Cannot change to CVS checkout at $opt_w";
  42}
  43unless ($ENV{GIT_DIR} && -r $ENV{GIT_DIR}){
  44    die "GIT_DIR is not defined or is unreadable";
  45}
  46
  47
  48my @cvs;
  49if ($opt_d) {
  50        @cvs = ('cvs', '-d', $opt_d);
  51} else {
  52        @cvs = ('cvs');
  53}
  54
  55# resolve target commit
  56my $commit;
  57$commit = pop @ARGV;
  58$commit = safe_pipe_capture('git-rev-parse', '--verify', "$commit^0");
  59chomp $commit;
  60if ($?) {
  61    die "The commit reference $commit did not resolve!";
  62}
  63
  64# resolve what parent we want
  65my $parent;
  66if (@ARGV) {
  67    $parent = pop @ARGV;
  68    $parent =  safe_pipe_capture('git-rev-parse', '--verify', "$parent^0");
  69    chomp $parent;
  70    if ($?) {
  71        die "The parent reference did not resolve!";
  72    }
  73}
  74
  75# find parents from the commit itself
  76my @commit  = safe_pipe_capture('git-cat-file', 'commit', $commit);
  77my @parents;
  78my $committer;
  79my $author;
  80my $stage = 'headers'; # headers, msg
  81my $title;
  82my $msg = '';
  83
  84foreach my $line (@commit) {
  85    chomp $line;
  86    if ($stage eq 'headers' && $line eq '') {
  87        $stage = 'msg';
  88        next;
  89    }
  90
  91    if ($stage eq 'headers') {
  92        if ($line =~ m/^parent (\w{40})$/) { # found a parent
  93            push @parents, $1;
  94        } elsif ($line =~ m/^author (.+) \d+ [-+]\d+$/) {
  95            $author = $1;
  96        } elsif ($line =~ m/^committer (.+) \d+ [-+]\d+$/) {
  97            $committer = $1;
  98        }
  99    } else {
 100        $msg .= $line . "\n";
 101        unless ($title) {
 102            $title = $line;
 103        }
 104    }
 105}
 106
 107my $noparent = "0000000000000000000000000000000000000000";
 108if ($parent) {
 109    my $found;
 110    # double check that it's a valid parent
 111    foreach my $p (@parents) {
 112        if ($p eq $parent) {
 113            $found = 1;
 114            last;
 115        }; # found it
 116    }
 117    die "Did not find $parent in the parents for this commit!" if !$found and !$opt_P;
 118} else { # we don't have a parent from the cmdline...
 119    if (@parents == 1) { # it's safe to get it from the commit
 120        $parent = $parents[0];
 121    } elsif (@parents == 0) { # there is no parent
 122        $parent = $noparent;
 123    } else { # cannot choose automatically from multiple parents
 124        die "This commit has more than one parent -- please name the parent you want to use explicitly";
 125    }
 126}
 127
 128my $go_back_to = 0;
 129
 130if ($opt_W) {
 131    $opt_v && print "Resetting to $parent\n";
 132    $go_back_to = `git symbolic-ref HEAD 2> /dev/null ||
 133        git rev-parse HEAD` || die "Could not determine current branch";
 134    system("git checkout -q $parent^0") && die "Could not check out $parent^0";
 135}
 136
 137$opt_v && print "Applying to CVS commit $commit from parent $parent\n";
 138
 139# grab the commit message
 140open(MSG, ">.msg") or die "Cannot open .msg for writing";
 141if ($opt_m) {
 142    print MSG $opt_m;
 143}
 144print MSG $msg;
 145if ($opt_a) {
 146    print MSG "\n\nAuthor: $author\n";
 147    if ($author ne $committer) {
 148        print MSG "Committer: $committer\n";
 149    }
 150}
 151close MSG;
 152
 153if ($parent eq $noparent) {
 154    `git-diff-tree --binary -p --root $commit >.cvsexportcommit.diff`;# || die "Cannot diff";
 155} else {
 156    `git-diff-tree --binary -p $parent $commit >.cvsexportcommit.diff`;# || die "Cannot diff";
 157}
 158
 159## apply non-binary changes
 160
 161# In pedantic mode require all lines of context to match.  In normal
 162# mode, be compatible with diff/patch: assume 3 lines of context and
 163# require at least one line match, i.e. ignore at most 2 lines of
 164# context, like diff/patch do by default.
 165my $context = $opt_p ? '' : '-C1';
 166
 167print "Checking if patch will apply\n";
 168
 169my @stat;
 170open APPLY, "GIT_DIR= git-apply $context --summary --numstat<.cvsexportcommit.diff|" || die "cannot patch";
 171@stat=<APPLY>;
 172close APPLY || die "Cannot patch";
 173my (@bfiles,@files,@afiles,@dfiles);
 174chomp @stat;
 175foreach (@stat) {
 176        push (@bfiles,$1) if m/^-\t-\t(.*)$/;
 177        push (@files, $1) if m/^-\t-\t(.*)$/;
 178        push (@files, $1) if m/^\d+\t\d+\t(.*)$/;
 179        push (@afiles,$1) if m/^ create mode [0-7]+ (.*)$/;
 180        push (@dfiles,$1) if m/^ delete mode [0-7]+ (.*)$/;
 181}
 182map { s/^"(.*)"$/$1/g } @bfiles,@files;
 183map { s/\\([0-7]{3})/sprintf('%c',oct $1)/eg } @bfiles,@files;
 184
 185# check that the files are clean and up to date according to cvs
 186my $dirty;
 187my @dirs;
 188foreach my $p (@afiles) {
 189    my $path = dirname $p;
 190    while (!-d $path and ! grep { $_ eq $path } @dirs) {
 191        unshift @dirs, $path;
 192        $path = dirname $path;
 193    }
 194}
 195
 196# ... check dirs,
 197foreach my $d (@dirs) {
 198    if (-e $d) {
 199        $dirty = 1;
 200        warn "$d exists and is not a directory!\n";
 201    }
 202}
 203
 204# ... query status of all files that we have a directory for and parse output of 'cvs status' to %cvsstat.
 205my @canstatusfiles;
 206foreach my $f (@files) {
 207    my $path = dirname $f;
 208    next if (grep { $_ eq $path } @dirs);
 209    push @canstatusfiles, $f;
 210}
 211
 212my %cvsstat;
 213if (@canstatusfiles) {
 214    if ($opt_u) {
 215      my @updated = xargs_safe_pipe_capture([@cvs, 'update'], @canstatusfiles);
 216      print @updated;
 217    }
 218    # "cvs status" reorders the parameters, notably when there are multiple
 219    # arguments with the same basename.  So be precise here.
 220
 221    my %added = map { $_ => 1 } @afiles;
 222    my %todo = map { $_ => 1 } @canstatusfiles;
 223
 224    while (%todo) {
 225      my @canstatusfiles2 = ();
 226      my %fullname = ();
 227      foreach my $name (keys %todo) {
 228        my $basename = basename($name);
 229
 230        # CVS reports files that don't exist in the current revision as
 231        # "no file $basename" in its "status" output, so we should
 232        # anticipate that.  Totally unknown files will have a status
 233        # "Unknown". However, if they exist in the Attic, their status
 234        # will be "Up-to-date" (this means they were added once but have
 235        # been removed).
 236        $basename = "no file $basename" if $added{$basename};
 237
 238        $basename =~ s/^\s+//;
 239        $basename =~ s/\s+$//;
 240
 241        if (!exists($fullname{$basename})) {
 242          $fullname{$basename} = $name;
 243          push (@canstatusfiles2, $name);
 244          delete($todo{$name});
 245        }
 246      }
 247      my @cvsoutput;
 248      @cvsoutput = xargs_safe_pipe_capture([@cvs, 'status'], @canstatusfiles2);
 249      foreach my $l (@cvsoutput) {
 250        chomp $l;
 251        next unless
 252            my ($file, $status) = $l =~ /^File:\s+(.*\S)\s+Status: (.*)$/;
 253
 254        my $fullname = $fullname{$file};
 255        print STDERR "Huh? Status '$status' reported for unexpected file '$file'\n"
 256            unless defined $fullname;
 257
 258        # This response means the file does not exist except in
 259        # CVS's attic, so set the status accordingly
 260        $status = "In-attic"
 261            if $file =~ /^no file /
 262                && $status eq 'Up-to-date';
 263
 264        $cvsstat{$fullname{$file}} = $status
 265            if defined $fullname{$file};
 266      }
 267    }
 268}
 269
 270# ... Validate that new files have the correct status
 271foreach my $f (@afiles) {
 272    next unless defined(my $stat = $cvsstat{$f});
 273
 274    # This means the file has never been seen before
 275    next if $stat eq 'Unknown';
 276
 277    # This means the file has been seen before but was removed
 278    next if $stat eq 'In-attic';
 279
 280    $dirty = 1;
 281        warn "File $f is already known in your CVS checkout -- perhaps it has been added by another user. Or this may indicate that it exists on a different branch. If this is the case, use -f to force the merge.\n";
 282        warn "Status was: $cvsstat{$f}\n";
 283}
 284
 285# ... validate known files.
 286foreach my $f (@files) {
 287    next if grep { $_ eq $f } @afiles;
 288    # TODO:we need to handle removed in cvs
 289    unless (defined ($cvsstat{$f}) and $cvsstat{$f} eq "Up-to-date") {
 290        $dirty = 1;
 291        warn "File $f not up to date but has status '$cvsstat{$f}' in your CVS checkout!\n";
 292    }
 293
 294    # Depending on how your GIT tree got imported from CVS you may
 295    # have a conflict between expanded keywords in your CVS tree and
 296    # unexpanded keywords in the patch about to be applied.
 297    if ($opt_k) {
 298        my $orig_file ="$f.orig";
 299        rename $f, $orig_file;
 300        open(FILTER_IN, "<$orig_file") or die "Cannot open $orig_file\n";
 301        open(FILTER_OUT, ">$f") or die "Cannot open $f\n";
 302        while (<FILTER_IN>)
 303        {
 304            my $line = $_;
 305            $line =~ s/\$([A-Z][a-z]+):[^\$]+\$/\$$1\$/g;
 306            print FILTER_OUT $line;
 307        }
 308        close FILTER_IN;
 309        close FILTER_OUT;
 310    }
 311}
 312
 313if ($dirty) {
 314    if ($opt_f) {       warn "The tree is not clean -- forced merge\n";
 315        $dirty = 0;
 316    } else {
 317        die "Exiting: your CVS tree is not clean for this merge.";
 318    }
 319}
 320
 321print "Applying\n";
 322if ($opt_W) {
 323    system("git checkout -q $commit^0") && die "cannot patch";
 324} else {
 325    `GIT_DIR= git-apply $context --summary --numstat --apply <.cvsexportcommit.diff` || die "cannot patch";
 326}
 327
 328print "Patch applied successfully. Adding new files and directories to CVS\n";
 329my $dirtypatch = 0;
 330
 331#
 332# We have to add the directories in order otherwise we will have
 333# problems when we try and add the sub-directory of a directory we
 334# have not added yet.
 335#
 336# Luckily this is easy to deal with by sorting the directories and
 337# dealing with the shortest ones first.
 338#
 339@dirs = sort { length $a <=> length $b} @dirs;
 340
 341foreach my $d (@dirs) {
 342    if (system(@cvs,'add',$d)) {
 343        $dirtypatch = 1;
 344        warn "Failed to cvs add directory $d -- you may need to do it manually";
 345    }
 346}
 347
 348foreach my $f (@afiles) {
 349    if (grep { $_ eq $f } @bfiles) {
 350      system(@cvs, 'add','-kb',$f);
 351    } else {
 352      system(@cvs, 'add', $f);
 353    }
 354    if ($?) {
 355        $dirtypatch = 1;
 356        warn "Failed to cvs add $f -- you may need to do it manually";
 357    }
 358}
 359
 360foreach my $f (@dfiles) {
 361    system(@cvs, 'rm', '-f', $f);
 362    if ($?) {
 363        $dirtypatch = 1;
 364        warn "Failed to cvs rm -f $f -- you may need to do it manually";
 365    }
 366}
 367
 368print "Commit to CVS\n";
 369print "Patch title (first comment line): $title\n";
 370my @commitfiles = map { unless (m/\s/) { '\''.$_.'\''; } else { $_; }; } (@files);
 371my $cmd = join(' ', @cvs)." commit -F .msg @commitfiles";
 372
 373if ($dirtypatch) {
 374    print "NOTE: One or more hunks failed to apply cleanly.\n";
 375    print "You'll need to apply the patch in .cvsexportcommit.diff manually\n";
 376    print "using a patch program. After applying the patch and resolving the\n";
 377    print "problems you may commit using:";
 378    print "\n    cd \"$opt_w\"" if $opt_w;
 379    print "\n    $cmd\n";
 380    print "\n    git checkout $go_back_to\n" if $go_back_to;
 381    print "\n";
 382    exit(1);
 383}
 384
 385if ($opt_c) {
 386    print "Autocommit\n  $cmd\n";
 387    print xargs_safe_pipe_capture([@cvs, 'commit', '-F', '.msg'], @files);
 388    if ($?) {
 389        die "Exiting: The commit did not succeed";
 390    }
 391    print "Committed successfully to CVS\n";
 392    # clean up
 393    unlink(".msg");
 394} else {
 395    print "Ready for you to commit, just run:\n\n   $cmd\n";
 396}
 397
 398# clean up
 399unlink(".cvsexportcommit.diff");
 400
 401if ($opt_W) {
 402    system("git checkout $go_back_to") && die "cannot move back to $go_back_to";
 403    if (!($go_back_to =~ /^[0-9a-fA-F]{40}$/)) {
 404        system("git symbolic-ref HEAD $go_back_to") &&
 405            die "cannot move back to $go_back_to";
 406    }
 407}
 408
 409# CVS version 1.11.x and 1.12.x sleeps the wrong way to ensure the timestamp
 410# used by CVS and the one set by subsequence file modifications are different.
 411# If they are not different CVS will not detect changes.
 412sleep(1);
 413
 414sub usage {
 415        print STDERR <<END;
 416Usage: GIT_DIR=/path/to/.git git cvsexportcommit [-h] [-p] [-v] [-c] [-f] [-u] [-k] [-w cvsworkdir] [-m msgprefix] [ parent ] commit
 417END
 418        exit(1);
 419}
 420
 421# An alternative to `command` that allows input to be passed as an array
 422# to work around shell problems with weird characters in arguments
 423# if the exec returns non-zero we die
 424sub safe_pipe_capture {
 425    my @output;
 426    if (my $pid = open my $child, '-|') {
 427        @output = (<$child>);
 428        close $child or die join(' ',@_).": $! $?";
 429    } else {
 430        exec(@_) or die "$! $?"; # exec() can fail the executable can't be found
 431    }
 432    return wantarray ? @output : join('',@output);
 433}
 434
 435sub xargs_safe_pipe_capture {
 436        my $MAX_ARG_LENGTH = 65536;
 437        my $cmd = shift;
 438        my @output;
 439        my $output;
 440        while(@_) {
 441                my @args;
 442                my $length = 0;
 443                while(@_ && $length < $MAX_ARG_LENGTH) {
 444                        push @args, shift;
 445                        $length += length($args[$#args]);
 446                }
 447                if (wantarray) {
 448                        push @output, safe_pipe_capture(@$cmd, @args);
 449                }
 450                else {
 451                        $output .= safe_pipe_capture(@$cmd, @args);
 452                }
 453        }
 454        return wantarray ? @output : $output;
 455}