From: Daniel T. <da...@re...> - 2008-03-24 17:24:55
|
On Sun, 2008-03-23 at 13:59 -0400, Andrew Kohlsmith (lists) wrote: > Is there a better (cleaner) way of developing against CVS HEAD aside from > copying the tree on top of a clean kernel? I've always run a for each on the checkout (something like the following): <untested_code> cd checked_out_directory for i in `find . ! -type d` do diff -Nau ../linux/$i $i done > gc-linux.patch </untested_code> I've then manage apply the patch to the pristine tree with quilt. This is quite a good working model if you don't have CVS write access since you just keep the master patch at the bottom of the stack and add you own changes as seperate patches higher up the stack (ready to be posted to the list). I assume the 'real' developers may have a more sophisticated approach that makes CVS commits easy as well. -- Daniel Thompson <da...@re...> |