git-cvsexportcommit.perl

   1 #!/usr/bin/perl
   2
   3 use 5.008001;
   4 use strict;
   5 use warnings;
   6 use Getopt::Std;
   7 use File::Temp qw(tempdir);
   8 use Data::Dumper;
   9 use File::Basename qw(basename dirname);
  10 use File::Spec;
  11 use Git;
  12
  13 our ($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
  15 getopts('uhPpvcfkam:d:w:W');
  16
  17 $opt_h && usage();
  18
  19 die "Need at least one commit identifier!" unless @ARGV;
  20
  21 # Get git-config settings
  22 my $repo = Git->repository();
  23 $opt_w = $repo->config('cvsexportcommit.cvsdir') unless defined $opt_w;
  24
  25 my $tmpdir = File::Temp::tempdir(CLEANUP => 1);
  26 my $hash_algo = $repo->config('extensions.objectformat') || 'sha1';
  27 my $hexsz = $hash_algo eq 'sha256' ? 64 : 40;
  28
  29 if ($opt_w || $opt_W) {
  30         # Remember where GIT_DIR is before changing to CVS checkout
  31         unless ($ENV{GIT_DIR}) {
  32                 # No GIT_DIR set. Figure it out for ourselves
  33                 my $gd =`git rev-parse --git-dir`;
  34                 chomp($gd);
  35                 $ENV{GIT_DIR} = $gd;
  36         }
  37
  38         # On MSYS, convert a Windows-style path to an MSYS-style path
  39         # so that rel2abs() below works correctly.
  40         if ($^O eq 'msys') {
  41                 $ENV{GIT_DIR} =~ s#^([[:alpha:]]):/#/$1/#;
  42         }
  43
  44         # Make sure GIT_DIR is absolute
  45         $ENV{GIT_DIR} = File::Spec->rel2abs($ENV{GIT_DIR});
  46 }
  47
  48 if ($opt_w) {
  49         if (! -d $opt_w."/CVS" ) {
  50                 die "$opt_w is not a CVS checkout";
  51         }
  52         chdir $opt_w or die "Cannot change to CVS checkout at $opt_w";
  53 }
  54 unless ($ENV{GIT_DIR} && -r $ENV{GIT_DIR}){
  55     die "GIT_DIR is not defined or is unreadable";
  56 }
  57
  58
  59 my @cvs;
  60 if ($opt_d) {
  61         @cvs = ('cvs', '-d', $opt_d);
  62 } else {
  63         @cvs = ('cvs');
  64 }
  65
  66 # resolve target commit
  67 my $commit;
  68 $commit = pop @ARGV;
  69 $commit = safe_pipe_capture('git', 'rev-parse', '--verify', "$commit^0");
  70 chomp $commit;
  71 if ($?) {
  72     die "The commit reference $commit did not resolve!";
  73 }
  74
  75 # resolve what parent we want
  76 my $parent;
  77 if (@ARGV) {
  78     $parent = pop @ARGV;
  79     $parent =  safe_pipe_capture('git', 'rev-parse', '--verify', "$parent^0");
  80     chomp $parent;
  81     if ($?) {
  82         die "The parent reference did not resolve!";
  83     }
  84 }
  85
  86 # find parents from the commit itself
  87 my @commit  = safe_pipe_capture('git', 'cat-file', 'commit', $commit);
  88 my @parents;
  89 my $committer;
  90 my $author;
  91 my $stage = 'headers'; # headers, msg
  92 my $title;
  93 my $msg = '';
  94
  95 foreach my $line (@commit) {
  96     chomp $line;
  97     if ($stage eq 'headers' && $line eq '') {
  98         $stage = 'msg';
  99         next;
 100     }
 101
 102     if ($stage eq 'headers') {
 103         if ($line =~ m/^parent ([0-9a-f]{$hexsz})$/) { # found a parent
 104             push @parents, $1;
 105         } elsif ($line =~ m/^author (.+) \d+ [-+]\d+$/) {
 106             $author = $1;
 107         } elsif ($line =~ m/^committer (.+) \d+ [-+]\d+$/) {
 108             $committer = $1;
 109         }
 110     } else {
 111         $msg .= $line . "\n";
 112         unless ($title) {
 113             $title = $line;
 114         }
 115     }
 116 }
 117
 118 my $noparent = "0" x $hexsz;
 119 if ($parent) {
 120     my $found;
 121     # double check that it's a valid parent
 122     foreach my $p (@parents) {
 123         if ($p eq $parent) {
 124             $found = 1;
 125             last;
 126         }; # found it
 127     }
 128     die "Did not find $parent in the parents for this commit!" if !$found and !$opt_P;
 129 } else { # we don't have a parent from the cmdline...
 130     if (@parents == 1) { # it's safe to get it from the commit
 131         $parent = $parents[0];
 132     } elsif (@parents == 0) { # there is no parent
 133         $parent = $noparent;
 134     } else { # cannot choose automatically from multiple parents
 135         die "This commit has more than one parent -- please name the parent you want to use explicitly";
 136     }
 137 }
 138
 139 my $go_back_to = 0;
 140
 141 if ($opt_W) {
 142     $opt_v && print "Resetting to $parent\n";
 143     $go_back_to = `git symbolic-ref HEAD 2> /dev/null ||
 144         git rev-parse HEAD` || die "Could not determine current branch";
 145     system("git checkout -q $parent^0") && die "Could not check out $parent^0";
 146 }
 147
 148 $opt_v && print "Applying to CVS commit $commit from parent $parent\n";
 149
 150 # grab the commit message
 151 open(MSG, ">.msg") or die "Cannot open .msg for writing";
 152 if ($opt_m) {
 153     print MSG $opt_m;
 154 }
 155 print MSG $msg;
 156 if ($opt_a) {
 157     print MSG "\n\nAuthor: $author\n";
 158     if ($author ne $committer) {
 159         print MSG "Committer: $committer\n";
 160     }
 161 }
 162 close MSG;
 163
 164 if ($parent eq $noparent) {
 165     `git diff-tree --binary -p --root $commit >.cvsexportcommit.diff`;# || die "Cannot diff";
 166 } else {
 167     `git diff-tree --binary -p $parent $commit >.cvsexportcommit.diff`;# || die "Cannot diff";
 168 }
 169
 170 ## apply non-binary changes
 171
 172 # In pedantic mode require all lines of context to match.  In normal
 173 # mode, be compatible with diff/patch: assume 3 lines of context and
 174 # require at least one line match, i.e. ignore at most 2 lines of
 175 # context, like diff/patch do by default.
 176 my $context = $opt_p ? '' : '-C1';
 177
 178 print "Checking if patch will apply\n";
 179
 180 my @stat;
 181 open APPLY, "GIT_INDEX_FILE=$tmpdir/index git apply $context --summary --numstat<.cvsexportcommit.diff|" || die "cannot patch";
 182 @stat=<APPLY>;
 183 close APPLY || die "Cannot patch";
 184 my (@bfiles,@files,@afiles,@dfiles);
 185 chomp @stat;
 186 foreach (@stat) {
 187         push (@bfiles,$1) if m/^-\t-\t(.*)$/;
 188         push (@files, $1) if m/^-\t-\t(.*)$/;
 189         push (@files, $1) if m/^\d+\t\d+\t(.*)$/;
 190         push (@afiles,$1) if m/^ create mode [0-7]+ (.*)$/;
 191         push (@dfiles,$1) if m/^ delete mode [0-7]+ (.*)$/;
 192 }
 193 map { s/^"(.*)"$/$1/g } @bfiles,@files;
 194 map { s/\\([0-7]{3})/sprintf('%c',oct $1)/eg } @bfiles,@files;
 195
 196 # check that the files are clean and up to date according to cvs
 197 my $dirty;
 198 my @dirs;
 199 foreach my $p (@afiles) {
 200     my $path = dirname $p;
 201     while (!-d $path and ! grep { $_ eq $path } @dirs) {
 202         unshift @dirs, $path;
 203         $path = dirname $path;
 204     }
 205 }
 206
 207 # ... check dirs,
 208 foreach my $d (@dirs) {
 209     if (-e $d) {
 210         $dirty = 1;
 211         warn "$d exists and is not a directory!\n";
 212     }
 213 }
 214
 215 # ... query status of all files that we have a directory for and parse output of 'cvs status' to %cvsstat.
 216 my @canstatusfiles;
 217 foreach my $f (@files) {
 218     my $path = dirname $f;
 219     next if (grep { $_ eq $path } @dirs);
 220     push @canstatusfiles, $f;
 221 }
 222
 223 my %cvsstat;
 224 if (@canstatusfiles) {
 225     if ($opt_u) {
 226       my @updated = xargs_safe_pipe_capture([@cvs, 'update'], @canstatusfiles);
 227       print @updated;
 228     }
 229     # "cvs status" reorders the parameters, notably when there are multiple
 230     # arguments with the same basename.  So be precise here.
 231
 232     my %added = map { $_ => 1 } @afiles;
 233     my %todo = map { $_ => 1 } @canstatusfiles;
 234
 235     while (%todo) {
 236       my @canstatusfiles2 = ();
 237       my %fullname = ();
 238       foreach my $name (keys %todo) {
 239         my $basename = basename($name);
 240
 241         # CVS reports files that don't exist in the current revision as
 242         # "no file $basename" in its "status" output, so we should
 243         # anticipate that.  Totally unknown files will have a status
 244         # "Unknown". However, if they exist in the Attic, their status
 245         # will be "Up-to-date" (this means they were added once but have
 246         # been removed).
 247         $basename = "no file $basename" if $added{$basename};
 248
 249         $basename =~ s/^\s+//;
 250         $basename =~ s/\s+$//;
 251
 252         if (!exists($fullname{$basename})) {
 253           $fullname{$basename} = $name;
 254           push (@canstatusfiles2, $name);
 255           delete($todo{$name});
 256         }
 257       }
 258       my @cvsoutput;
 259       @cvsoutput = xargs_safe_pipe_capture([@cvs, 'status'], @canstatusfiles2);
 260       foreach my $l (@cvsoutput) {
 261         chomp $l;
 262         next unless
 263             my ($file, $status) = $l =~ /^File:\s+(.*\S)\s+Status: (.*)$/;
 264
 265         my $fullname = $fullname{$file};
 266         print STDERR "Huh? Status '$status' reported for unexpected file '$file'\n"
 267             unless defined $fullname;
 268
 269         # This response means the file does not exist except in
 270         # CVS's attic, so set the status accordingly
 271         $status = "In-attic"
 272             if $file =~ /^no file /
 273                 && $status eq 'Up-to-date';
 274
 275         $cvsstat{$fullname{$file}} = $status
 276             if defined $fullname{$file};
 277       }
 278     }
 279 }
 280
 281 # ... Validate that new files have the correct status
 282 foreach my $f (@afiles) {
 283     next unless defined(my $stat = $cvsstat{$f});
 284
 285     # This means the file has never been seen before
 286     next if $stat eq 'Unknown';
 287
 288     # This means the file has been seen before but was removed
 289     next if $stat eq 'In-attic';
 290
 291     $dirty = 1;
 292         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";
 293         warn "Status was: $cvsstat{$f}\n";
 294 }
 295
 296 # ... validate known files.
 297 foreach my $f (@files) {
 298     next if grep { $_ eq $f } @afiles;
 299     # TODO:we need to handle removed in cvs
 300     unless (defined ($cvsstat{$f}) and $cvsstat{$f} eq "Up-to-date") {
 301         $dirty = 1;
 302         warn "File $f not up to date but has status '$cvsstat{$f}' in your CVS checkout!\n";
 303     }
 304
 305     # Depending on how your GIT tree got imported from CVS you may
 306     # have a conflict between expanded keywords in your CVS tree and
 307     # unexpanded keywords in the patch about to be applied.
 308     if ($opt_k) {
 309         my $orig_file ="$f.orig";
 310         rename $f, $orig_file;
 311         open(FILTER_IN, "<$orig_file") or die "Cannot open $orig_file\n";
 312         open(FILTER_OUT, ">$f") or die "Cannot open $f\n";
 313         while (<FILTER_IN>)
 314         {
 315             my $line = $_;
 316             $line =~ s/\$([A-Z][a-z]+):[^\$]+\$/\$$1\$/g;
 317             print FILTER_OUT $line;
 318         }
 319         close FILTER_IN;
 320         close FILTER_OUT;
 321     }
 322 }
 323
 324 if ($dirty) {
 325     if ($opt_f) {       warn "The tree is not clean -- forced merge\n";
 326         $dirty = 0;
 327     } else {
 328         die "Exiting: your CVS tree is not clean for this merge.";
 329     }
 330 }
 331
 332 print "Applying\n";
 333 if ($opt_W) {
 334     system("git checkout -q $commit^0") && die "cannot patch";
 335 } else {
 336     `GIT_INDEX_FILE=$tmpdir/index git apply $context --summary --numstat --apply <.cvsexportcommit.diff` || die "cannot patch";
 337 }
 338
 339 print "Patch applied successfully. Adding new files and directories to CVS\n";
 340 my $dirtypatch = 0;
 341
 342 #
 343 # We have to add the directories in order otherwise we will have
 344 # problems when we try and add the sub-directory of a directory we
 345 # have not added yet.
 346 #
 347 # Luckily this is easy to deal with by sorting the directories and
 348 # dealing with the shortest ones first.
 349 #
 350 @dirs = sort { length $a <=> length $b} @dirs;
 351
 352 foreach my $d (@dirs) {
 353     if (system(@cvs,'add',$d)) {
 354         $dirtypatch = 1;
 355         warn "Failed to cvs add directory $d -- you may need to do it manually";
 356     }
 357 }
 358
 359 foreach my $f (@afiles) {
 360     if (grep { $_ eq $f } @bfiles) {
 361       system(@cvs, 'add','-kb',$f);
 362     } else {
 363       system(@cvs, 'add', $f);
 364     }
 365     if ($?) {
 366         $dirtypatch = 1;
 367         warn "Failed to cvs add $f -- you may need to do it manually";
 368     }
 369 }
 370
 371 foreach my $f (@dfiles) {
 372     system(@cvs, 'rm', '-f', $f);
 373     if ($?) {
 374         $dirtypatch = 1;
 375         warn "Failed to cvs rm -f $f -- you may need to do it manually";
 376     }
 377 }
 378
 379 print "Commit to CVS\n";
 380 print "Patch title (first comment line): $title\n";
 381 my @commitfiles = map { unless (m/\s/) { '\''.$_.'\''; } else { $_; }; } (@files);
 382 my $cmd = join(' ', @cvs)." commit -F .msg @commitfiles";
 383
 384 if ($dirtypatch) {
 385     print "NOTE: One or more hunks failed to apply cleanly.\n";
 386     print "You'll need to apply the patch in .cvsexportcommit.diff manually\n";
 387     print "using a patch program. After applying the patch and resolving the\n";
 388     print "problems you may commit using:";
 389     print "\n    cd \"$opt_w\"" if $opt_w;
 390     print "\n    $cmd\n";
 391     print "\n    git checkout $go_back_to\n" if $go_back_to;
 392     print "\n";
 393     exit(1);
 394 }
 395
 396 if ($opt_c) {
 397     print "Autocommit\n  $cmd\n";
 398     print xargs_safe_pipe_capture([@cvs, 'commit', '-F', '.msg'], @files);
 399     if ($?) {
 400         die "Exiting: The commit did not succeed";
 401     }
 402     print "Committed successfully to CVS\n";
 403     # clean up
 404     unlink(".msg");
 405 } else {
 406     print "Ready for you to commit, just run:\n\n   $cmd\n";
 407 }
 408
 409 # clean up
 410 unlink(".cvsexportcommit.diff");
 411
 412 if ($opt_W) {
 413     system("git checkout $go_back_to") && die "cannot move back to $go_back_to";
 414     if (!($go_back_to =~ /^[0-9a-fA-F]{$hexsz}$/)) {
 415         system("git symbolic-ref HEAD $go_back_to") &&
 416             die "cannot move back to $go_back_to";
 417     }
 418 }
 419
 420 # CVS version 1.11.x and 1.12.x sleeps the wrong way to ensure the timestamp
 421 # used by CVS and the one set by subsequence file modifications are different.
 422 # If they are not different CVS will not detect changes.
 423 sleep(1);
 424
 425 sub usage {
 426         print STDERR <<END;
 427 usage: GIT_DIR=/path/to/.git git cvsexportcommit [-h] [-p] [-v] [-c] [-f] [-u] [-k] [-w cvsworkdir] [-m msgprefix] [ parent ] commit
 428 END
 429         exit(1);
 430 }
 431
 432 # An alternative to `command` that allows input to be passed as an array
 433 # to work around shell problems with weird characters in arguments
 434 # if the exec returns non-zero we die
 435 sub safe_pipe_capture {
 436     my @output;
 437     if (my $pid = open my $child, '-|') {
 438         binmode($child, ":crlf");
 439         @output = (<$child>);
 440         close $child or die join(' ',@_).": $! $?";
 441     } else {
 442         exec(@_) or die "$! $?"; # exec() can fail the executable can't be found
 443     }
 444     return wantarray ? @output : join('',@output);
 445 }
 446
 447 sub xargs_safe_pipe_capture {
 448         my $MAX_ARG_LENGTH = 65536;
 449         my $cmd = shift;
 450         my @output;
 451         my $output;
 452         while(@_) {
 453                 my @args;
 454                 my $length = 0;
 455                 while(@_ && $length < $MAX_ARG_LENGTH) {
 456                         push @args, shift;
 457                         $length += length($args[$#args]);
 458                 }
 459                 if (wantarray) {
 460                         push @output, safe_pipe_capture(@$cmd, @args);
 461                 }
 462                 else {
 463                         $output .= safe_pipe_capture(@$cmd, @args);
 464                 }
 465         }
 466         return wantarray ? @output : $output;
 467 }