glibc-headers RPM bug? (was: Re: Kernel dependency issues)

[Panu Matilainen]

Apologies if you receive this twice but I've a feeling the original reply 
ended up somewhere in the bit-heaven of /dev/null...

 	- Panu -

On Fri, 12 Dec 2008, Panu Matilainen wrote:

> On Fri, 12 Dec 2008, Seth Vidal wrote:
>> On Fri, 12 Dec 2008, Michael Schwendt wrote:
>> 
>>> On Fri, 12 Dec 2008 12:13:21 -0500, Christopher wrote:
>>> 
>>>>> I'm surprised one can "rpm -e glibc-headers" while keeping glibc-devel
>>>>> That looks a bit strange.
>>>> 
>>>> I'll leave you to file a bug on that if you think it's wrong.
>>> 
>>> If it were that easy:
>>>
>>>  $ rpm -qR glibc-devel|grep ^glib
>>>  glibc = 2.9-2
>>>  glibc-headers = 2.9-2
>>> 
>>> So, glibc-devel requires glibc-headers.
>>>
>>>  $ rpm -q --provides glibc-headers
>>>  glibc-headers(i386)
>>>  glibc-headers = 2.9-2
>>>  glibc-headers(x86-32) = 2.9-2
>>>
>>>  $ rpm -q --whatprovides glibc-headers
>>>  glibc-headers-2.9-2.i386
>>>
>>>  $ sudo rpm -e glibc-headers
>>> 
>>> Uh? That should not have been possible, because now:
>>>
>>>  $ rpm -V glibc-devel
>>>  Unsatisfied dependencies for glibc-devel-2.9-2.i386:
>>>          glibc-headers = 2.9-2 is needed by glibc-devel-2.9-2.i386
>>> 
>>> Can anyone reproduce this? Surely I cannot file a bug about glibc. This
>>> looks more like a bug in RPM. F10 here, btw.
>> 
>> it's a Requires(Pre) thats allowing it to happen.
>
> Yup, just checked (from F-10 glibc.spec):
>
> %package devel
> Summary: Object files for development using standard C libraries.
> Group: Development/Libraries
> Requires(pre): /sbin/install-info
> Requires(pre): %{name}-headers = %{version}-%{release}
>
>> 
>> Try it with xorg-x11-filesystem
>> 
>> rpm -e xorg-x11-filesystem
>> rpm -Va --nofiles --nomd5
>> 
>> notice the broken deps :(
>
> It's actually the verification which is showing bogus results here 
> (https://bugzilla.redhat.com/show_bug.cgi?id=223642) wrt rpm semantics: 
> "Requires(pre): pkg1" means "pkg1" is required to be present during execution 
> of %pre scriptlet of the package and nothing else. It does not imply 
> "Requires: pkg1", that would have to be separately added if pkg1 is needed 
> during installation *and* runtime.
>
> Whether this is sane behavior is debatable, and has been debated several 
> times in various places...
>
> 	- Panu -
>