Fedora Issues

[Michael K. Johnson]

On Tue, Oct 21, 2003 at 07:52:27AM -0700, Jesse Keating wrote:
> > 2.) why 3.3 was used if it doesn't work and 
> 
> Because waiting for the huge kernel to come together is just silly.  Moving 
> forward and putting even more pressure on the kernel folks is a bit better.  
> No sense in holding back the rest of the class because of one slow student.

FYI, the kernel has basically always been sensitive to new gcc versions;
very few even minor versions of gcc haven't required some kernel
cleanup, and more often than not the problems haven't been obvious.
It's real convenient when the problems are "doesn't compile any more"
but more often they have been more like "assembly interfacing very
subtly with C has its operating assumptions broken" and the results
are more often subtle and difficult-to-debug race conditions.  Heck,
we've seen even kernel *bugfixes* cause problems in the kernel's
restricted operating environment!

Therefore, when a new compiler version comes out, we'd wait a bit
even if it did happen to compile the kernel just fine.

I wouldn't call the kernel a "slow student" here -- it's generally the
fact that it's trying to do clever things that gets it into trouble
here.

michaelkjohnson

 "He that composes himself is wiser than he that composes a book."
 Linux Application Development                     -- Ben Franklin
 http://people.redhat.com/johnsonm/lad/