Then fix them, or tell people who are in the know to do that. It's usually
rather easy to do.
[...]
> If there is a reason why non-compiling code has to be in the kernel
> sources (I would be hard-pressed to find one), then at least there should
> be a file "LATEST_COMPILING_<whatever>_IS_<version number>" in the kernel
> source directory.
Great. For 2.1.44 I count 162 CONFIG_<whatever> lines in
linux/arch/i386/defconfig, most of them would be tristate by now, but let's
assume the are just booleans. 2**162 kernels to compile... give some factor
of a bilion for comnbinations that don't make any sense, its 2**100, at the
very least.
Or do what somebody (I forget his name, sorry) used to do: Turn
_everything_ on and try to compile the beast. Then post the results. Do the
same with everything made module.
> I don't think it is a good service to the advancement of Linux if released
> kernels do not compile.
Then fix them. Or help somebody to fix them.
> Like I said, I would volunteer to test-compile 2.1.X kernels.
See above.
> If kernels are not tested in any way, they should be in a different
> directory and be called pre-<Version>.
"Not tested in any way"... Linus runs his machines with the beta kernels
before they hit the street. There are lots of people who download and
compile the beta kernels as soon as they appear. Many people try out the
patches that show up on linuxhq, &c.
Besides, who is going to do the testing if not people like you and me?
There is no budget for machines and people for any pre-release SQA. It is
done anyway, with pre-patches and beta kernels for the daring to run on
their machines and report back.
If you don't like how it works, propose something better. Or better yet,
make it work and propose something better by example.
-- Dr. Horst H. von Brand mailto:vonbrand@inf.utfsm.cl Departamento de Informatica Fono: +56 32 654431 Universidad Tecnica Federico Santa Maria +56 32 654239 Casilla 110-V, Valparaiso, Chile Fax: +56 32 797513