git-cvsexportcommit.perl

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