[prev in list] [next in list] [prev in thread] [next in thread] 

List:       mandrake-cooker
Subject:    Re: [Cooker] dkms and update-alternatives
From:       Vincent Panel <cooker.vincent.panel () chello ! be>
Date:       2005-05-31 16:58:32
Message-ID: 200505311858.32358.cooker.vincent.panel () chello ! be
[Download RAW message or body]

Le Mardi 31 Mai 2005 15:29, Vincent Panel a écrit :
> Le Mardi 31 Mai 2005 15:18, Buchan Milne a écrit :
> > Vincent Panel wrote:
> > > Dkms should ignore your gcc choice made with update-alternatives and
> > > only use the same compiler as the one used to compile the kernel.
> > > Because if it doesn't module won't compile / won't load / will crash.
> >
> > Actually, I think you should rather file this against kernel-source.
> >
> > kernel-source should be installed in such a way that modules built
> > against it are guaranteed to have the right version magic.
> >
> > How the user triggers it (via dkms or via the nvidia installer script,
> > or via a manual build) is irrelevant to how it should be fixed.
> >
> > Regards,
> > Buchan
>
> I'm don't really understand all the C gory-details (and runtime problems),
> but don't you think a kernel compiled with gcc 4.0 might have difficulties
> to load a driver compiled with gcc 3.4 even if we trick it with the good
> version magic ? Could you guarantee it won't crash ?

Ok it looks like I misunderstood your point, I have read your update on 
bugzilla.


[prev in list] [next in list] [prev in thread] [next in thread]