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);
  12
  13 getopts('uhPpvcfam:d: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) {
  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         if (! -d $opt_w."/CVS" ) {
  35                 die "$opt_w is not a CVS checkout";
  36         }
  37         chdir $opt_w or die "Cannot change to CVS checkout at $opt_w";
  38 }
  39 unless ($ENV{GIT_DIR} && -r $ENV{GIT_DIR}){
  40     die "GIT_DIR is not defined or is unreadable";
  41 }
  42
  43
  44 my @cvs;
  45 if ($opt_d) {
  46         @cvs = ('cvs', '-d', $opt_d);
  47 } else {
  48         @cvs = ('cvs');
  49 }
  50
  51 # resolve target commit
  52 my $commit;
  53 $commit = pop @ARGV;
  54 $commit = safe_pipe_capture('git-rev-parse', '--verify', "$commit^0");
  55 chomp $commit;
  56 if ($?) {
  57     die "The commit reference $commit did not resolve!";
  58 }
  59
  60 # resolve what parent we want
  61 my $parent;
  62 if (@ARGV) {
  63     $parent = pop @ARGV;
  64     $parent =  safe_pipe_capture('git-rev-parse', '--verify', "$parent^0");
  65     chomp $parent;
  66     if ($?) {
  67         die "The parent reference did not resolve!";
  68     }
  69 }
  70
  71 # find parents from the commit itself
  72 my @commit  = safe_pipe_capture('git-cat-file', 'commit', $commit);
  73 my @parents;
  74 my $committer;
  75 my $author;
  76 my $stage = 'headers'; # headers, msg
  77 my $title;
  78 my $msg = '';
  79
  80 foreach my $line (@commit) {
  81     chomp $line;
  82     if ($stage eq 'headers' && $line eq '') {
  83         $stage = 'msg';
  84         next;
  85     }
  86
  87     if ($stage eq 'headers') {
  88         if ($line =~ m/^parent (\w{40})$/) { # found a parent
  89             push @parents, $1;
  90         } elsif ($line =~ m/^author (.+) \d+ [-+]\d+$/) {
  91             $author = $1;
  92         } elsif ($line =~ m/^committer (.+) \d+ [-+]\d+$/) {
  93             $committer = $1;
  94         }
  95     } else {
  96         $msg .= $line . "\n";
  97         unless ($title) {
  98             $title = $line;
  99         }
 100     }
 101 }
 102
 103 my $noparent = "0000000000000000000000000000000000000000";
 104 if ($parent) {
 105     my $found;
 106     # double check that it's a valid parent
 107     foreach my $p (@parents) {
 108         if ($p eq $parent) {
 109             $found = 1;
 110             last;
 111         }; # found it
 112     }
 113     die "Did not find $parent in the parents for this commit!" if !$found and !$opt_P;
 114 } else { # we don't have a parent from the cmdline...
 115     if (@parents == 1) { # it's safe to get it from the commit
 116         $parent = $parents[0];
 117     } elsif (@parents == 0) { # there is no parent
 118         $parent = $noparent;
 119     } else { # cannot choose automatically from multiple parents
 120         die "This commit has more than one parent -- please name the parent you want to use explicitly";
 121     }
 122 }
 123
 124 $opt_v && print "Applying to CVS commit $commit from parent $parent\n";
 125
 126 # grab the commit message
 127 open(MSG, ">.msg") or die "Cannot open .msg for writing";
 128 if ($opt_m) {
 129     print MSG $opt_m;
 130 }
 131 print MSG $msg;
 132 if ($opt_a) {
 133     print MSG "\n\nAuthor: $author\n";
 134     if ($author ne $committer) {
 135         print MSG "Committer: $committer\n";
 136     }
 137 }
 138 close MSG;
 139
 140 if ($parent eq $noparent) {
 141     `git-diff-tree --binary -p --root $commit >.cvsexportcommit.diff`;# || die "Cannot diff";
 142 } else {
 143     `git-diff-tree --binary -p $parent $commit >.cvsexportcommit.diff`;# || die "Cannot diff";
 144 }
 145
 146 ## apply non-binary changes
 147
 148 # In pedantic mode require all lines of context to match.  In normal
 149 # mode, be compatible with diff/patch: assume 3 lines of context and
 150 # require at least one line match, i.e. ignore at most 2 lines of
 151 # context, like diff/patch do by default.
 152 my $context = $opt_p ? '' : '-C1';
 153
 154 print "Checking if patch will apply\n";
 155
 156 my @stat;
 157 open APPLY, "GIT_DIR= git-apply $context --summary --numstat<.cvsexportcommit.diff|" || die "cannot patch";
 158 @stat=<APPLY>;
 159 close APPLY || die "Cannot patch";
 160 my (@bfiles,@files,@afiles,@dfiles);
 161 chomp @stat;
 162 foreach (@stat) {
 163         push (@bfiles,$1) if m/^-\t-\t(.*)$/;
 164         push (@files, $1) if m/^-\t-\t(.*)$/;
 165         push (@files, $1) if m/^\d+\t\d+\t(.*)$/;
 166         push (@afiles,$1) if m/^ create mode [0-7]+ (.*)$/;
 167         push (@dfiles,$1) if m/^ delete mode [0-7]+ (.*)$/;
 168 }
 169 map { s/^"(.*)"$/$1/g } @bfiles,@files;
 170 map { s/\\([0-7]{3})/sprintf('%c',oct $1)/eg } @bfiles,@files;
 171
 172 # check that the files are clean and up to date according to cvs
 173 my $dirty;
 174 my @dirs;
 175 foreach my $p (@afiles) {
 176     my $path = dirname $p;
 177     while (!-d $path and ! grep { $_ eq $path } @dirs) {
 178         unshift @dirs, $path;
 179         $path = dirname $path;
 180     }
 181 }
 182
 183 # ... check dirs,
 184 foreach my $d (@dirs) {
 185     if (-e $d) {
 186         $dirty = 1;
 187         warn "$d exists and is not a directory!\n";
 188     }
 189 }
 190
 191 # ... query status of all files that we have a directory for and parse output of 'cvs status' to %cvsstat.
 192 my @canstatusfiles;
 193 foreach my $f (@files) {
 194     my $path = dirname $f;
 195     next if (grep { $_ eq $path } @dirs);
 196     push @canstatusfiles, $f;
 197 }
 198
 199 my %cvsstat;
 200 if (@canstatusfiles) {
 201     if ($opt_u) {
 202       my @updated = xargs_safe_pipe_capture([@cvs, 'update'], @canstatusfiles);
 203       print @updated;
 204     }
 205     # "cvs status" reorders the parameters, notably when there are multiple
 206     # arguments with the same basename.  So be precise here.
 207
 208     my %added = map { $_ => 1 } @afiles;
 209     my %todo = map { $_ => 1 } @canstatusfiles;
 210
 211     while (%todo) {
 212       my @canstatusfiles2 = ();
 213       my %fullname = ();
 214       foreach my $name (keys %todo) {
 215         my $basename = basename($name);
 216
 217         $basename = "no file " . $basename if (exists($added{$basename}));
 218         chomp($basename);
 219
 220         if (!exists($fullname{$basename})) {
 221           $fullname{$basename} = $name;
 222           push (@canstatusfiles2, $name);
 223           delete($todo{$name});
 224         }
 225       }
 226       my @cvsoutput;
 227       @cvsoutput = xargs_safe_pipe_capture([@cvs, 'status'], @canstatusfiles2);
 228       foreach my $l (@cvsoutput) {
 229         chomp $l;
 230         if ($l =~ /^File:\s+(.*\S)\s+Status: (.*)$/) {
 231           if (!exists($fullname{$1})) {
 232             print STDERR "Huh? Status reported for unexpected file '$1'\n";
 233           } else {
 234             $cvsstat{$fullname{$1}} = $2;
 235           }
 236         }
 237       }
 238     }
 239 }
 240
 241 # ... validate new files,
 242 foreach my $f (@afiles) {
 243     if (defined ($cvsstat{$f}) and $cvsstat{$f} ne "Unknown") {
 244         $dirty = 1;
 245         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";
 246         warn "Status was: $cvsstat{$f}\n";
 247     }
 248 }
 249 # ... validate known files.
 250 foreach my $f (@files) {
 251     next if grep { $_ eq $f } @afiles;
 252     # TODO:we need to handle removed in cvs
 253     unless (defined ($cvsstat{$f}) and $cvsstat{$f} eq "Up-to-date") {
 254         $dirty = 1;
 255         warn "File $f not up to date but has status '$cvsstat{$f}' in your CVS checkout!\n";
 256     }
 257 }
 258 if ($dirty) {
 259     if ($opt_f) {       warn "The tree is not clean -- forced merge\n";
 260         $dirty = 0;
 261     } else {
 262         die "Exiting: your CVS tree is not clean for this merge.";
 263     }
 264 }
 265
 266 print "Applying\n";
 267 `GIT_DIR= git-apply $context --summary --numstat --apply <.cvsexportcommit.diff` || die "cannot patch";
 268
 269 print "Patch applied successfully. Adding new files and directories to CVS\n";
 270 my $dirtypatch = 0;
 271
 272 #
 273 # We have to add the directories in order otherwise we will have
 274 # problems when we try and add the sub-directory of a directory we
 275 # have not added yet.
 276 #
 277 # Luckily this is easy to deal with by sorting the directories and
 278 # dealing with the shortest ones first.
 279 #
 280 @dirs = sort { length $a <=> length $b} @dirs;
 281
 282 foreach my $d (@dirs) {
 283     if (system(@cvs,'add',$d)) {
 284         $dirtypatch = 1;
 285         warn "Failed to cvs add directory $d -- you may need to do it manually";
 286     }
 287 }
 288
 289 foreach my $f (@afiles) {
 290     if (grep { $_ eq $f } @bfiles) {
 291       system(@cvs, 'add','-kb',$f);
 292     } else {
 293       system(@cvs, 'add', $f);
 294     }
 295     if ($?) {
 296         $dirtypatch = 1;
 297         warn "Failed to cvs add $f -- you may need to do it manually";
 298     }
 299 }
 300
 301 foreach my $f (@dfiles) {
 302     system(@cvs, 'rm', '-f', $f);
 303     if ($?) {
 304         $dirtypatch = 1;
 305         warn "Failed to cvs rm -f $f -- you may need to do it manually";
 306     }
 307 }
 308
 309 print "Commit to CVS\n";
 310 print "Patch title (first comment line): $title\n";
 311 my @commitfiles = map { unless (m/\s/) { '\''.$_.'\''; } else { $_; }; } (@files);
 312 my $cmd = join(' ', @cvs)." commit -F .msg @commitfiles";
 313
 314 if ($dirtypatch) {
 315     print "NOTE: One or more hunks failed to apply cleanly.\n";
 316     print "You'll need to apply the patch in .cvsexportcommit.diff manually\n";
 317     print "using a patch program. After applying the patch and resolving the\n";
 318     print "problems you may commit using:";
 319     print "\n    cd \"$opt_w\"" if $opt_w;
 320     print "\n    $cmd\n\n";
 321     exit(1);
 322 }
 323
 324 if ($opt_c) {
 325     print "Autocommit\n  $cmd\n";
 326     print xargs_safe_pipe_capture([@cvs, 'commit', '-F', '.msg'], @files);
 327     if ($?) {
 328         die "Exiting: The commit did not succeed";
 329     }
 330     print "Committed successfully to CVS\n";
 331     # clean up
 332     unlink(".msg");
 333 } else {
 334     print "Ready for you to commit, just run:\n\n   $cmd\n";
 335 }
 336
 337 # clean up
 338 unlink(".cvsexportcommit.diff");
 339
 340 # CVS version 1.11.x and 1.12.x sleeps the wrong way to ensure the timestamp
 341 # used by CVS and the one set by subsequence file modifications are different.
 342 # If they are not different CVS will not detect changes.
 343 sleep(1);
 344
 345 sub usage {
 346         print STDERR <<END;
 347 Usage: GIT_DIR=/path/to/.git ${\basename $0} [-h] [-p] [-v] [-c] [-f] [-u] [-w cvsworkdir] [-m msgprefix] [ parent ] commit
 348 END
 349         exit(1);
 350 }
 351
 352 # An alternative to `command` that allows input to be passed as an array
 353 # to work around shell problems with weird characters in arguments
 354 # if the exec returns non-zero we die
 355 sub safe_pipe_capture {
 356     my @output;
 357     if (my $pid = open my $child, '-|') {
 358         @output = (<$child>);
 359         close $child or die join(' ',@_).": $! $?";
 360     } else {
 361         exec(@_) or die "$! $?"; # exec() can fail the executable can't be found
 362     }
 363     return wantarray ? @output : join('',@output);
 364 }
 365
 366 sub xargs_safe_pipe_capture {
 367         my $MAX_ARG_LENGTH = 65536;
 368         my $cmd = shift;
 369         my @output;
 370         my $output;
 371         while(@_) {
 372                 my @args;
 373                 my $length = 0;
 374                 while(@_ && $length < $MAX_ARG_LENGTH) {
 375                         push @args, shift;
 376                         $length += length($args[$#args]);
 377                 }
 378                 if (wantarray) {
 379                         push @output, safe_pipe_capture(@$cmd, @args);
 380                 }
 381                 else {
 382                         $output .= safe_pipe_capture(@$cmd, @args);
 383                 }
 384         }
 385         return wantarray ? @output : $output;
 386 }