git-cvsexportcommit.perl

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