Fixed Kernel?

[Charles R. Anderson]

On Mon, Mar 01, 2004 at 03:16:24PM -1000, Warren Togami wrote:
> Note that just "running" a kernel usually wont expose less obvious 
> problems.  You folks testing the update-testing kernel should run stress 
> tests that cause CPU/disk/RAM to be heavily saturated.  If it swaps 
> heavily that is a good test.

Rebuilding the kernel src.rpm and comparing the resulting extracted
binaries would be a good stress test.  Something along these lines:

boot to updates-testing kernel

mkdir original
cd original
rpm2cpio kernel-from-updates-testing.i686.rpm | cpio -idv
cd ..

mkdir rebuild
cd rebuild
rpmbuild --rebuild --target=i686 kernel-from-updates-testing.src.rpm
rpm2cpio kernel-just-rebuilt.i686.rpm | cpio -idv
cd ..

diff -ur original rebuild