static libs ... again
Denis Leroy
denis at poolshark.org
Fri Feb 17 21:32:26 UTC 2006
Ralf Corsepius wrote:
> On Fri, 2006-02-17 at 14:04 -0500, Konstantin Ryabitsev wrote:
>
>>Quentin Spencer wrote:
>>
>>>So, not too long ago someone asked why I still had static libs in one of
>>>my packages since they are "banned" or at least strongly discouraged, so
>>>I started removing them from my packages. All of the libraries I
>>>maintain are math libraries, so security concerns are a non-issue. After
>>>removing the static libs from fftw-devel, it took less than 24 hours to
>>>get a bug report asking for them back. See
>>>https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=181897 for the
>>>reasoning.
>
>
>>It seems theirs is a rather specific and unique case. Should we change
>>policy to accommodate corner cases like that?
>
> Why? Because a user says he can't build some applications statically?
>
> What kind of rationale is this? Where is the technical explanation?
I've also had a request for reinstating static libs :
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=178853
I can see how profiling is important here, especially libsigc++ which is
a low-level library that is typically used all over C++ code...
-denis
More information about the fedora-extras-list
mailing list