cvs diff howto?

Daniel Jacobowitz drow@mvista.com
Mon Aug 26 14:17:00 GMT 2002


On Mon, Aug 26, 2002 at 10:57:07PM +0200, Svein E. Seldal wrote:
> Daniel Jacobowitz wrote:
> >Generally, if you use 'cvs add' and then cvs diff (or maybe cvs diff
> >-N? I don't remember), this works.  However, you might need to have
> >commit access to do it that way; CVS is awkward about that.  I have a
> >little script that just looks for the "? Filename" lines and replaces
> >them with the appropriate diff.
> >
> 
> My problem is that I do not have write access to the CVS so I cant add 
> the new files, hence I cannot use cvs diff for these files.
> 
> What would be the simplest way for me to submit such patches that 
> contains both modified and new files? I'm tempted to submit an ordinary 
> patch for the modified files and a tgz for the new ones. Will that do?

Please don't; instead, you can use such a script, or add dummy entries
to CVS/Entries by hand.  They look like:

/filename/0/Initial filename//

> This script you were talking about sounds interesting. Would it be 
> possble to get it?

It's pretty trivial.  Here it is, usage "fillpatch <patchfilename>":

#!/usr/bin/perl

use strict;

my $file = shift @ARGV;
my $newfile = $file . ".new";

open (FILE, $file);
open (OUT, "> $newfile");

while (<FILE>)
{
  if (/^\? (.*)$/) {
    print OUT "Index: $1\n";
    print OUT "===================================================================\n";
    print OUT "RCS file: N/A\n";
    print OUT "diff -u /dev/null $1\n";
    close OUT;
    system ("diff -u /dev/null \"$1\" >> $newfile");
    open (OUT, ">> $newfile");
  } else {
    print OUT $_;
  }
}

close OUT;


-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer



More information about the Binutils mailing list