git-cvsexportcommit.perl

   1 #!/usr/bin/perl -w
   2
   3 # Known limitations:
   4 # - does not propagate permissions
   5 # - error handling has not been extensively tested
   6 #
   7
   8 use strict;
   9 use Getopt::Std;
  10 use File::Temp qw(tempdir);
  11 use Data::Dumper;
  12 use File::Basename qw(basename dirname);
  13
  14 unless ($ENV{GIT_DIR} && -r $ENV{GIT_DIR}){
  15     die "GIT_DIR is not defined or is unreadable";
  16 }
  17
  18 our ($opt_h, $opt_P, $opt_p, $opt_v, $opt_c, $opt_f, $opt_a, $opt_m, $opt_d, $opt_u);
  19
  20 getopts('uhPpvcfam:d:');
  21
  22 $opt_h && usage();
  23
  24 die "Need at least one commit identifier!" unless @ARGV;
  25
  26 my @cvs;
  27 if ($opt_d) {
  28         @cvs = ('cvs', '-d', $opt_d);
  29 } else {
  30         @cvs = ('cvs');
  31 }
  32
  33 # resolve target commit
  34 my $commit;
  35 $commit = pop @ARGV;
  36 $commit = safe_pipe_capture('git-rev-parse', '--verify', "$commit^0");
  37 chomp $commit;
  38 if ($?) {
  39     die "The commit reference $commit did not resolve!";
  40 }
  41
  42 # resolve what parent we want
  43 my $parent;
  44 if (@ARGV) {
  45     $parent = pop @ARGV;
  46     $parent =  safe_pipe_capture('git-rev-parse', '--verify', "$parent^0");
  47     chomp $parent;
  48     if ($?) {
  49         die "The parent reference did not resolve!";
  50     }
  51 }
  52
  53 # find parents from the commit itself
  54 my @commit  = safe_pipe_capture('git-cat-file', 'commit', $commit);
  55 my @parents;
  56 my $committer;
  57 my $author;
  58 my $stage = 'headers'; # headers, msg
  59 my $title;
  60 my $msg = '';
  61
  62 foreach my $line (@commit) {
  63     chomp $line;
  64     if ($stage eq 'headers' && $line eq '') {
  65         $stage = 'msg';
  66         next;
  67     }
  68
  69     if ($stage eq 'headers') {
  70         if ($line =~ m/^parent (\w{40})$/) { # found a parent
  71             push @parents, $1;
  72         } elsif ($line =~ m/^author (.+) \d+ [-+]\d+$/) {
  73             $author = $1;
  74         } elsif ($line =~ m/^committer (.+) \d+ [-+]\d+$/) {
  75             $committer = $1;
  76         }
  77     } else {
  78         $msg .= $line . "\n";
  79         unless ($title) {
  80             $title = $line;
  81         }
  82     }
  83 }
  84
  85 my $noparent = "0000000000000000000000000000000000000000";
  86 if ($parent) {
  87     my $found;
  88     # double check that it's a valid parent
  89     foreach my $p (@parents) {
  90         if ($p eq $parent) {
  91             $found = 1;
  92             last;
  93         }; # found it
  94     }
  95     die "Did not find $parent in the parents for this commit!" if !$found and !$opt_P;
  96 } else { # we don't have a parent from the cmdline...
  97     if (@parents == 1) { # it's safe to get it from the commit
  98         $parent = $parents[0];
  99     } elsif (@parents == 0) { # there is no parent
 100         $parent = $noparent;
 101     } else { # cannot choose automatically from multiple parents
 102         die "This commit has more than one parent -- please name the parent you want to use explicitly";
 103     }
 104 }
 105
 106 $opt_v && print "Applying to CVS commit $commit from parent $parent\n";
 107
 108 # grab the commit message
 109 open(MSG, ">.msg") or die "Cannot open .msg for writing";
 110 if ($opt_m) {
 111     print MSG $opt_m;
 112 }
 113 print MSG $msg;
 114 if ($opt_a) {
 115     print MSG "\n\nAuthor: $author\n";
 116     if ($author ne $committer) {
 117         print MSG "Committer: $committer\n";
 118     }
 119 }
 120 close MSG;
 121
 122 if ($parent eq $noparent) {
 123     `git-diff-tree --binary -p --root $commit >.cvsexportcommit.diff`;# || die "Cannot diff";
 124 } else {
 125     `git-diff-tree --binary -p $parent $commit >.cvsexportcommit.diff`;# || die "Cannot diff";
 126 }
 127
 128 ## apply non-binary changes
 129
 130 # In pedantic mode require all lines of context to match.  In normal
 131 # mode, be compatible with diff/patch: assume 3 lines of context and
 132 # require at least one line match, i.e. ignore at most 2 lines of
 133 # context, like diff/patch do by default.
 134 my $context = $opt_p ? '' : '-C1';
 135
 136 print "Checking if patch will apply\n";
 137
 138 my @stat;
 139 open APPLY, "GIT_DIR= git-apply $context --summary --numstat<.cvsexportcommit.diff|" || die "cannot patch";
 140 @stat=<APPLY>;
 141 close APPLY || die "Cannot patch";
 142 my (@bfiles,@files,@afiles,@dfiles);
 143 chomp @stat;
 144 foreach (@stat) {
 145         push (@bfiles,$1) if m/^-\t-\t(.*)$/;
 146         push (@files, $1) if m/^-\t-\t(.*)$/;
 147         push (@files, $1) if m/^\d+\t\d+\t(.*)$/;
 148         push (@afiles,$1) if m/^ create mode [0-7]+ (.*)$/;
 149         push (@dfiles,$1) if m/^ delete mode [0-7]+ (.*)$/;
 150 }
 151 map { s/^"(.*)"$/$1/g } @bfiles,@files;
 152 map { s/\\([0-7]{3})/sprintf('%c',oct $1)/eg } @bfiles,@files;
 153
 154 # check that the files are clean and up to date according to cvs
 155 my $dirty;
 156 my @dirs;
 157 foreach my $p (@afiles) {
 158     my $path = dirname $p;
 159     while (!-d $path and ! grep { $_ eq $path } @dirs) {
 160         unshift @dirs, $path;
 161         $path = dirname $path;
 162     }
 163 }
 164
 165 # ... check dirs,
 166 foreach my $d (@dirs) {
 167     if (-e $d) {
 168         $dirty = 1;
 169         warn "$d exists and is not a directory!\n";
 170     }
 171 }
 172
 173 # ... query status of all files that we have a directory for and parse output of 'cvs status' to %cvsstat.
 174 my @canstatusfiles;
 175 foreach my $f (@files) {
 176     my $path = dirname $f;
 177     next if (grep { $_ eq $path } @dirs);
 178     push @canstatusfiles, $f;
 179 }
 180
 181 my %cvsstat;
 182 if (@canstatusfiles) {
 183     if ($opt_u) {
 184       my @updated = safe_pipe_capture(@cvs, 'update', @canstatusfiles);
 185       print @updated;
 186     }
 187     my @cvsoutput;
 188     @cvsoutput= safe_pipe_capture(@cvs, 'status', @canstatusfiles);
 189     my $matchcount = 0;
 190     foreach my $l (@cvsoutput) {
 191         chomp $l;
 192         if ( $l =~ /^File:/ and  $l =~ /Status: (.*)$/ ) {
 193             $cvsstat{$canstatusfiles[$matchcount]} = $1;
 194             $matchcount++;
 195         }
 196     }
 197 }
 198
 199 # ... validate new files,
 200 foreach my $f (@afiles) {
 201     if (defined ($cvsstat{$f}) and $cvsstat{$f} ne "Unknown") {
 202         $dirty = 1;
 203         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";
 204         warn "Status was: $cvsstat{$f}\n";
 205     }
 206 }
 207 # ... validate known files.
 208 foreach my $f (@files) {
 209     next if grep { $_ eq $f } @afiles;
 210     # TODO:we need to handle removed in cvs
 211     unless (defined ($cvsstat{$f}) and $cvsstat{$f} eq "Up-to-date") {
 212         $dirty = 1;
 213         warn "File $f not up to date but has status '$cvsstat{$f}' in your CVS checkout!\n";
 214     }
 215 }
 216 if ($dirty) {
 217     if ($opt_f) {       warn "The tree is not clean -- forced merge\n";
 218         $dirty = 0;
 219     } else {
 220         die "Exiting: your CVS tree is not clean for this merge.";
 221     }
 222 }
 223
 224 print "Applying\n";
 225 `GIT_DIR= git-apply $context --summary --numstat --apply <.cvsexportcommit.diff` || die "cannot patch";
 226
 227 print "Patch applied successfully. Adding new files and directories to CVS\n";
 228 my $dirtypatch = 0;
 229
 230 #
 231 # We have to add the directories in order otherwise we will have
 232 # problems when we try and add the sub-directory of a directory we
 233 # have not added yet.
 234 #
 235 # Luckily this is easy to deal with by sorting the directories and
 236 # dealing with the shortest ones first.
 237 #
 238 @dirs = sort { length $a <=> length $b} @dirs;
 239
 240 foreach my $d (@dirs) {
 241     if (system(@cvs,'add',$d)) {
 242         $dirtypatch = 1;
 243         warn "Failed to cvs add directory $d -- you may need to do it manually";
 244     }
 245 }
 246
 247 foreach my $f (@afiles) {
 248     if (grep { $_ eq $f } @bfiles) {
 249       system(@cvs, 'add','-kb',$f);
 250     } else {
 251       system(@cvs, 'add', $f);
 252     }
 253     if ($?) {
 254         $dirtypatch = 1;
 255         warn "Failed to cvs add $f -- you may need to do it manually";
 256     }
 257 }
 258
 259 foreach my $f (@dfiles) {
 260     system(@cvs, 'rm', '-f', $f);
 261     if ($?) {
 262         $dirtypatch = 1;
 263         warn "Failed to cvs rm -f $f -- you may need to do it manually";
 264     }
 265 }
 266
 267 print "Commit to CVS\n";
 268 print "Patch title (first comment line): $title\n";
 269 my @commitfiles = map { unless (m/\s/) { '\''.$_.'\''; } else { $_; }; } (@files);
 270 my $cmd = join(' ', @cvs)." commit -F .msg @commitfiles";
 271
 272 if ($dirtypatch) {
 273     print "NOTE: One or more hunks failed to apply cleanly.\n";
 274     print "You'll need to apply the patch in .cvsexportcommit.diff manually\n";
 275     print "using a patch program. After applying the patch and resolving the\n";
 276     print "problems you may commit using:";
 277     print "\n    $cmd\n\n";
 278     exit(1);
 279 }
 280
 281 if ($opt_c) {
 282     print "Autocommit\n  $cmd\n";
 283     print safe_pipe_capture(@cvs, 'commit', '-F', '.msg', @files);
 284     if ($?) {
 285         die "Exiting: The commit did not succeed";
 286     }
 287     print "Committed successfully to CVS\n";
 288     # clean up
 289     unlink(".msg");
 290 } else {
 291     print "Ready for you to commit, just run:\n\n   $cmd\n";
 292 }
 293
 294 # clean up
 295 unlink(".cvsexportcommit.diff");
 296
 297 # CVS version 1.11.x and 1.12.x sleeps the wrong way to ensure the timestamp
 298 # used by CVS and the one set by subsequence file modifications are different.
 299 # If they are not different CVS will not detect changes.
 300 sleep(1);
 301
 302 sub usage {
 303         print STDERR <<END;
 304 Usage: GIT_DIR=/path/to/.git ${\basename $0} [-h] [-p] [-v] [-c] [-f] [-m msgprefix] [ parent ] commit
 305 END
 306         exit(1);
 307 }
 308
 309 # An alternative to `command` that allows input to be passed as an array
 310 # to work around shell problems with weird characters in arguments
 311 # if the exec returns non-zero we die
 312 sub safe_pipe_capture {
 313     my @output;
 314     if (my $pid = open my $child, '-|') {
 315         @output = (<$child>);
 316         close $child or die join(' ',@_).": $! $?";
 317     } else {
 318         exec(@_) or die "$! $?"; # exec() can fail the executable can't be found
 319     }
 320     return wantarray ? @output : join('',@output);
 321 }
 322
 323 sub safe_pipe_capture_blob {
 324     my $output;
 325     if (my $pid = open my $child, '-|') {
 326         local $/;
 327         undef $/;
 328         $output = (<$child>);
 329         close $child or die join(' ',@_).": $! $?";
 330     } else {
 331         exec(@_) or die "$! $?"; # exec() can fail the executable can't be found
 332     }
 333     return $output;
 334 }