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

Re: 2.6.19, kernel-package problems and what are our plans for etch ...

On Mon, 18 Dec 2006 11:48:22 +0100, Bastian Blank <waldi@debian.org> said: 

> On Mon, Dec 18, 2006 at 09:31:22AM +0100, Goswin von Brederlow wrote:
>> And Bastian is decidetly anti make-kpkg and wants to remove all
>> make-kpkg use from linux-2.6 as stated several times now. You can
>> see beginings of that in the xen kernels.

> That happens if the same problems happens over and over again: kpkg
> fails silent instead of producing errors. The following reasons was
> identified, some of them fixed. This errors are now catched by the
> abi check because the complete output is missing:
> - More than one entry in the Architecture line, the code just
>   explicitely ignored anything after the first entry.

        What architecture line are we talking about here? Is there a
 bug report number I can refer to to refresh my memory on this issue?


        Again, what is broken about EXTRAVERSION? Which bug reports
 are we talking about?

> - 2.6.19-rcX failed on the autobuilders and for Sven, not for me. So
>   it failes if the environment is not as it wants.

        Has the bit about the environment that it wanted been
 identified? Is it not true that the failure was not in make-kpkg by
 itself, but only in linux-2.6, which leads one to conclude that the
 problem might have been in the so called infrastructure code?

> There are other problems like:
> * It is possible to interfere with the build from a user config.

        What do you mean, interfere? Can the same config build a
 kernel without using make-kpkg?  The only issue I can see might be
 version numbers being changed with the ser config, and yes, changing
 version numbers like that is not supported.

> * Build time raises by about 30% at worst.
>   - build target needs 30 minutes on ppc.
>   - binary-arch target needs another 30 minutes. My own routine
>     needs less than 10.

        I'd be happy to look at why bits of the build process take
 longer than a raw build without make-kpkg -- which bug number was
 this reported under?

We are what we are.
Manoj Srivastava <srivasta@acm.org> <http://www.golden-gryphon.com/>
1024D/BF24424C print 4966 F272 D093 B493 410B  924B 21BA DABB BF24 424C

Reply to: