I was not really asking to validate the patches, but simply to provide a way to 
rebuild the kernel modules on the fly for these packages.  If the build bails 
then the user knows that it does not work.

In that case it would be nice if the package (eg. DIPC) provide a known working 
kernel so the user does not have to muck around with building stuff.

Also, I do not think it would be possible to use the same kernel for DIPC, 
Kerrighed, Lustre etc. - I don't think anybody would want to do that anyways... 
;-)

My $0.02.

Cheers,

Bernard 

> -----Original Message-----
> From: [EMAIL PROTECTED] 
> [mailto:[EMAIL PROTECTED] On Behalf Of 
> Geoffroy Vallée
> Sent: Friday, March 24, 2006 10:43
> To: [email protected]
> Subject: Re: [Oscar-devel] Re: Creating an RPM for DIPC
> 
> Hi Bernard,
> 
> I agree that we have to address the bigger issue of how to 
> create a package 
> for a specific version of the Linux kernel. Unfortunately, i 
> do not see any 
> solution: if you want to use different patches (e.g. DIPC + 
> Lustre) they may 
> create conflicts and unfortunately it seems that major Linux 
> distributions 
> provide deeply modified Linux kernels. 
> Therefore, i don't think it is possible to validate patches 
> for all the 
> kernels available around and so the only solution according 
> to me is to 
> identify a specific configuration, validate patches regarding 
> that and 
> provide a binary packages. If we do not do that i think it 
> will be very 
> difficult to support lustre, kerrighed or dipc via OSCAR on a 
> large set of 
> Linux distributions. But i may be wrong...
> 
> My 2 cents,
> 
> Le Vendredi 24 Mars 2006 13:19, Bernard Li a écrit :
> > Hi Kamran:
> >
> > Alternatively, you can build a SRPM including the diffs etc. (and
> > requiring the "kernel" and probably "kernel-devel" package) 
> so that the
> > user can on the fly rebuild the dipc RPM based on their kernel.  It
> > would probably be a good idea to ship one that is known to 
> work though.
> >
> > Anyways, we need to address the bigger issue here, which is 
> how we want
> > to treat packages that depend on specific kernels - eg. lustre,
> > kerrighed, dipc.
> >
> > Cheers,
> >
> > Bernard
> >
> > > -----Original Message-----
> > > From: [EMAIL PROTECTED]
> > > [mailto:[EMAIL PROTECTED] On Behalf Of
> > > Kamran Karimi
> > > Sent: Friday, March 24, 2006 9:27
> > > To: [email protected]
> > > Subject: [Oscar-devel] Re: Creating an RPM for DIPC
> > >
> > > Thanks Geoffroy. I've been looking at Kerrighed's packaging,
> > > and it seems
> > > that creating an RPM is not suitable for systems like DIPC
> > > and Kerrighed.
> > > The reason is that installing such systems requires patching
> > > the kernel. I
> > > see the following problems:
> > >
> > > 1) Including a pre-compiled kernel forces others to use a
> > > specific version
> > > of the kernel, with whatever options that were selected at
> > > the compile time.
> > > When distributing a patch, chances are users can apply it to
> > > a similar
> > > kernel. DIPC's kernel patch in particular is fairly small and
> > > non-intrusive, so chances are it can be applied to other
> > > kernel versions.
> > >
> > > 2) DIPC can work on non-x86 CPUs, so distributing an i386 RPM is
> > > restrictive.
> > >
> > > Yes, for a user-level system, an RPM distribution is
> > > excellent. I won't
> > > speak for Kerrighed, but creating an RPM for DIPC seems to
> > > complicates
> > > things more than it helps.
> > >
> > > Maybe I need to be educated on the merits of creating an RPM
> > > for a system
> > > like DIPC
> > >
> > > -Kamran
> > >
> > >
> > >
> > >
> > > -------------------------------------------------------
> > > This SF.Net email is sponsored by xPML, a groundbreaking
> > > scripting language
> > > that extends applications into web and mobile media. Attend
> > > the live webcast
> > > and join the prime developer group breaking into this new
> > > coding territory!
> > > http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&;
> > > dat=121642
> > > _______________________________________________
> > > Oscar-devel mailing list
> > > [email protected]
> > > https://lists.sourceforge.net/lists/listinfo/oscar-devel
> >
> > -------------------------------------------------------
> > This SF.Net email is sponsored by xPML, a groundbreaking 
> scripting language
> > that extends applications into web and mobile media. Attend the live
> > webcast and join the prime developer group breaking into 
> this new coding
> > territory!
> > http://sel.as-us.falkag.net/sel?cmd=lnk&kid0944&bid$1720&dat1642
> > _______________________________________________
> > Oscar-devel mailing list
> > [email protected]
> > https://lists.sourceforge.net/lists/listinfo/oscar-devel
> 
> -- 
> Geoffroy
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by xPML, a groundbreaking 
> scripting language
> that extends applications into web and mobile media. Attend 
> the live webcast
> and join the prime developer group breaking into this new 
> coding territory!
> http://sel.as-us.falkag.net/sel?cmd=k&kid0944&bid$1720&dat1642
> _______________________________________________
> Oscar-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/oscar-devel
> 


-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid0944&bid$1720&dat1642
_______________________________________________
Oscar-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/oscar-devel

Reply via email to