Kernel module packages (was - Re: Pre-Review: Asterisk)

[Thorsten Leemhuis]

Am Sonntag, den 03.04.2005, 11:32 -0500 schrieb Tom 'spot' Callaway:
> Two replies, one email. What a deal!

I found it a bit confusing ;-) Anyway:

> On Sun, 2005-04-03 at 14:21 +0200, Thorsten Leemhuis wrote:
> > I don't really like this idea. The version of the kernel-module needs
> to
> > be somewhere because it is of interest sometimes 
[...]
> OK, I'm not entirely sure how:
> rpm -q kernel-module-nvidia-glx --provides | grep kernel-module-nvidia-glx-version
> is all that confusing. :)

I said it should be somewhere. Somewhere also meant a virtual
provides ;-)

> However, if the consensus is that the version number needs to appear in the n-v-r, we could 
> overload the %{release} field with that information.
> 
> So:
> Name: kernel-module-foo
> Version: 2.6.10_3smp
> Release: 1.0_7174.1

I would prefer something like that. (The reason is found below). I'm
still unsure if it should be 

Release: 1.0_7174.1
or
Release: 1.1.0_7174

I think i would prefer the later.

[...]
> So, here's my handy checklist:
> 
> - Does the end user care about the module subversion? Probably not. (And
> if they do, its easy to document how to find it for any kernel-module-*,
> we can standardize a "kernel-module-%{name}-version" provides as a
> mandatory spec item.)

If it is in n-v-r it has one major benefit: It is in the bugzilla report
*if* the reporter specifies the n-v-r of the component he reports the
bug against (as he should)

> - Can we track the module subversion using internal provides? Yes.
> Unlike your statement above, a driver package rebuild would not be
> required on each kernel-update, since the Requires is a >=. If the
> driver needs "at least" that version of a module, the >= works.

Yes, as long as we have the
Provides:kernel-module-nvidia-glx-version 

(the rebuild would be needed if we would not have it -- maybe that was
not entirely clear in my mail; I just wanted you to know about the
problem in general)

> - Then, does the kernel module version need to show up in the n-v-r? I
> say no.

I never requested it should be there ;-) But yes, I think it could be
"Release" without doing harm. 

> But again, if all the people packaging kernel modules say yes, we can
> overload the %{release} field.

Other opinions? Maybe we could make this optional?! 

>  It just won't help your dependency
> checks.

Yeah, that's right.
-- 
Thorsten Leemhuis <fedora at leemhuis.info>