[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: kmail and korn don't work for me (from unstable)



El Sáb 22 Dic 2001 18:34, Bjoern Krombholz escribió:
> On Sat, Dec 22, 2001 at 04:15:25PM +0100, Pablo de Vicente wrote:
> > El Sáb 22 Dic 2001 09:47, Wolfgang Ratzka escribió:
> > > $ kmail
> > > *** KMail got signal 4 (Crashing)
>
> So the crash is caused bye SIGILL - means Illegal Instruction, CPU throws
> these signals in case it doesn't know an OP-Code. F.e. MMX-Code would give
> SIGILL on 486.
>
> >  My system is an AMD K6/450 and none of the kdenetwork applications work.
> > KMail crashes the same way you see. I purged and reinstalled KDE from
> > scratch but that did not helped. I am also using sid for KDE.
>
> Reinstalling the same binaries won't help. It doesn't look like but seems
> that kdenetwork is compiled with some special optimization flags not
> supported on AMD K6 CPUs.
>
> >   My last option was to grab the kdenetwork-2.2.2.tar.bz2 file from one
> > of the KDE ftp mirrors and compile it myself. It compiled properly. Then
> > I copied from kmail/.libs/kmail to /usr/bin/kmail and now I have a
> > working kmail. I admit this is a dirty solution but it is the only one I
> > have for the moment.
>
> Try apt-get -b source kmail. To build the binary deb yourself from the
> source packages.
>
> >  There is something wrong with the Debian kdenetwork packages and I have
> > no hint of what it is. I also have a laptop with a Pentium III and the
> > Debian packages work fine. May be the problem is related to the type of
> > CPU?.
>
> Probably. Maybe the deb is ppro optimized what wouldn't work on K6 or
> something similiar.

 I have just read another thread in this same list stating that the problem 
may arise from compilation flags which do not apply to AMD K6 CPUs. There are 
new debs for all kdenetwork packages. I have installed kmail-2.2.2-8 and 
kppp-2.2.2-8 on my AMD K6  and now they work fine. (well the latter has a 
problem that I will mention in a separate thread because I think is not 
related to the topic discussed here).

Thanks to Daniel Stone (the maintainer for these packages) for fixing this 
problem.

Pablo de Vicente.



Reply to: