kernel headers weirdness.

Dheart · 2009-12-22 07:17:57

I used to testing for a while to install the 2.6.32 kernel and I installed the kernel26-headers package.
Today I upgraded (testing repo off) only to find that there's a kernel-headers package which installs to /usr/include/
as opposed to the kernel26-headers (from testing (and previous kernel-headers)) which install to /usr/src/linux-*
Furthermore the kernel-headers package from core comes with version 2.6.32!? But the kernel in core is only 2.6.31
?!?

Allan · 2009-12-22 07:36:39

kernel-headers are the sanitised userspace headers that are part of the toolchain (glibc, gcc, binutils).

kernel26-headers are for building kernel modules. They were split from the kernel26 package. You of course know this given you subscribe to arch-dev-public for the important announcements for [testing] packages...

ngoonee · 2009-12-22 09:46:33

I've seen that last sentence many times now, and wonder whether anyone actually uses [testing] without subscribing there. It would be insane, like deciding to upgrade your sedan to a monster truck and sliming the entire windscreen with egg.

Oh, and with the windows wound up and stuck closed.

Dheart · 2009-12-22 10:37:03

I used testing a approximately 2 weeks ago and back then the packages weren't split (At least I didn't have two different kernel-headers packages on my system).
And regarding the kernel-headers 2.6.32 package in core while the kernel itself is 2.6.31? Why's that?

wonder · 2009-12-22 12:33:44

Dheart wrote:

And regarding the kernel-headers 2.6.32 package in core while the kernel itself is 2.6.31? Why's that?

because of what allan said and doesn't need to follow your kernel version.

Last edited by wonder (2009-12-22 12:34:40)

Mr.Elendig · 2009-12-22 14:20:56

@ Dheart: A somewhat related thread: http://bbs.archlinux.org/viewtopic.php?id=49105

Dheart · 2009-12-22 17:24:24

I know that they don't need to match by version, just that usually the kernel is the first thing upgraded and after that come the headers, not the other way around...

wonder · 2009-12-22 18:04:57

Dheart wrote:

I know that they don't need to match by version, just that usually the kernel is the first thing upgraded and after that come the headers, not the other way around...

usually is like that but now there are a few bugs concerning kernel 2.6.32 kernel and can't be moved from testing until those are fixed.

lucke · 2009-12-22 18:24:55

ngoonee wrote:

I've seen that last sentence many times now, and wonder whether anyone actually uses [testing] without subscribing there. It would be insane, like deciding to upgrade your sedan to a monster truck and sliming the entire windscreen with egg.

I'm guilty of using [testing] without being suscribed to ML for over 5 years (dev ML hasn't been public for that long, IIRC). I'm alive.

Allan · 2009-12-22 19:43:33

lucke wrote:

ngoonee wrote:
I've seen that last sentence many times now, and wonder whether anyone actually uses [testing] without subscribing there. It would be insane, like deciding to upgrade your sedan to a monster truck and sliming the entire windscreen with egg.
I'm guilty of using [testing] without being suscribed to ML for over 5 years (dev ML hasn't been public for that long, IIRC). I'm alive.

Well, lets call that luck... I frequently post a message saying "now is probably not a good time to update" message during big rebuilds.

Arch Linux

#1 2009-12-22 07:17:57

kernel headers weirdness.

#2 2009-12-22 07:36:39

Re: kernel headers weirdness.

#3 2009-12-22 09:46:33

Re: kernel headers weirdness.

#4 2009-12-22 10:37:03

Re: kernel headers weirdness.

#5 2009-12-22 12:33:44

Re: kernel headers weirdness.

#6 2009-12-22 14:20:56

Re: kernel headers weirdness.

#7 2009-12-22 17:24:24

Re: kernel headers weirdness.

#8 2009-12-22 18:04:57

Re: kernel headers weirdness.

#9 2009-12-22 18:24:55

Re: kernel headers weirdness.

#10 2009-12-22 19:43:33

Re: kernel headers weirdness.

Board footer