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

Re: Kernel compiling....



On Sat, 7 May 2005 11:36:08 +0200, maximilian attems <debian@sternwelten.at> said: 

> On Sat, 07 May 2005, Manoj Srivastava wrote:
>> On Fri, 6 May 2005 18:11:27 +0200, maximilian attems
>> <debian@sternwelten.at> said:
snipp> 
>> > the strings out of kernel-package are sometimes quite offensive.
>> 
>> Ah. Drive by criticism.

> no picking up the subject for discussions.

> "# work around idiocy in recent kernel versions" i find that output
> quite wierd. localversion was meant for distributions and is much
> easier to use than to change EXTRAVERSION in the Makefile.

        You are just displaying your ignorance. That message is for a
 workaround in kernel scripts, where they just delete ./debian on
 clean (like, when dpkg-buildpackage is run) -- whether or not they
 created the dir.

>> > also the ouput isn't up to the nice Kbuild of the 2.6, which will
>> > be the common case (afair you also support older non maintained
>> > trees).
>> 
>> The output from kernel-package is mostly the output from the
>> underlying kernel build system, so I am intrigued by this comment.

> well most of the time yes, but not always. for example call

> ~/src/linux-2.6.11-rc4$ make config
>   CLEAN scripts/basic CLEAN scripts/kconfig
> ..

> ~/src/linux-2.6.11-rc4$ fakroot make-kpkg clean /usr/bin/make -f
> /usr/share/kernel-package/rules real_stamp_clean make[1]: Entering
> directory `/home/max/src/linux-2.6.11-rc4' test ! -f .config || cp
> -pf .config config.precious test -f Makefile && \ /usr/bin/make
> ARCH=i386 distclean make[2]: Entering directory
> `/home/max/src/linux-2.6.11-rc4' ..

> but i agree that this is an implementation detail.

	And the difference in putput style, which kernel-package has
 always had, is somehow technocally releveant?

        If this is the tenor of the complaints you have, I think your
 case is pretty darned weak.

snipp> 
>> >> For starters, it would be nice if the configs used in official
>> >> kernels were available in the configs directory of
>> >> kernel-package,so that user of stand alone kernels could use
>> >> those as a guide.
>> 
>> > that doesn't make that match sense because of the modular debian
>> > patchw.  make defconfig comes to mind.
>> 
>> Parse error. A use gets kernels from kernel.org, does not have a
>> .config, and kernel-package provides one. Would it not make sense
>> for the config provided to be as close to the config the official
>> packages use? make defconfig does not, as far as I know, give you
>> the config that debian uses -- it just produces a .config from the
>> default values of config elements, which is nowhere close to the
>> same thing.

> yup you are right, responded to quickly, make defconfig doesn't make
> sense in that case.  but that would make sense for volatile as an
> config for 2.6.8 is not complete for A, who is taking 2.6.11.8 from
> kernel.org and it will get worser.  the main difference for using a
> config out of the svn are the modular ide and fb patches by xu,
> which aren't mainline.

	Frankly, starting with the 2.6.8 config is way better than
 starting with defconfig, or nothing.

	manoj
-- 
Batteries not included.
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: