Athlon Incompatible Packages

[Mike A. Harris]

On Tue, 24 Feb 2004, Ivan Gyurdiev wrote:

>>>I think you're avoiding my question, which is whether this is a bug or 
>>>not. It seems to me though, that this is a bug... and all bugs have to 
>>>be fixed, correct? As far as the real issues go, I'll get to those 
>>>too... - I have 67 additional packages that will not build on fedora.
>> 
>> 
>> It is not a bug.  rpmbuild --rebuild --target=i386.
>
>But shouldn't the athlon arch be compatible with 386?
>In which case shouldn't the packages also allow athlon?
>Excuse my total ignorance on this... If in fact, the packages in 
>question will not work correctly on athlon, and fedora is unwilling to
>support athlon, then I can see how this is not a bug.

Exactly..  That's the difference.  If the rpm package is wrongly 
excluding athlon, it should be fixed.  However, if the rpm 
package allows rpm recompilation with athlon and the compile 
fails, or even if it succeeds and then fails at runtime, then 
that is up to some volunteer to fix really as it isn't supported.  

Bugs being reported that end up being something unsupported, 
however in which the bug reporter supplies a patch to fix the 
bug, which is clean and correct, and isn't likely to regress 
builds for official architectures, should certainly be considered 
by package maintainers at least.

I know if someone filed a "XFree86 wont recompile with
--target==athlon" bug report to me, I would probably try to
rebuild it myself, just to see what the failure is and wether it
is a simple fix or not.  If it required more involvement, I'd 
probably add a comment to the report something like:

"I've reproduced that, and at a first non-detailed examination,
it appears to require more attention than I'm willing to give it
since that is not something that we officially support.  However
if you or someone else is willing to investigate the matter and
fix it voluntarily, and supply a patch, I'd be more than happy to
review your patch and consider it for inclusion in future
builds."

Just because we (Red Hat) might not support something does not
mean we should outright reject things that volunteers might be
willing to fix on their own IMHO.  Remember, it's a community 
project, so external contributions are welcome.  ;o)



-- 
Mike A. Harris     ftp://people.redhat.com/mharris
OS Systems Engineer - XFree86 maintainer - Red Hat