On Tue, 2018-01-23 at 07:36 +0100, Igor Gnatenko wrote:
On Mon, 2018-01-22 at 14:33 -0500, Matthew Miller wrote:
> On Sat, Jan 20, 2018 at 02:23:16PM +0100, Igor Gnatenko wrote:
> > What I'm trying to say here is that each time we want to
> > implement
> > some feature in Fedora, we either need to have some replacement
> > in
> > EPEL or diverge Fedora branches from EPEL branches. Having
> > replacement is not always possible, especially if we start
> > utilizing
> > new (actually 8 years old) features of RPM.
>
> I'd really like to see us tend towards coming up with macros that
> provide elegant fallbacks on EPEL. (The %license macro is a good
> example.)
There is no fallback for rich dependencies. There is no fallback for
filetriggers.
> I get the frustration with older stuff holding us back, but we
> really
> have a _lot_ of users who get value from doing so.
>
https://twitter.com/mattdm/status/936243506355621888
I'm definitely not against supporting EPEL, but right now it is
hurting me that
much that I don't do that.
AWS (Amazon web services) technologies are based on Centos 6 and 7 ,
the recente announced "AWS Cloud9 a cloud IDE for writing, running, and
debugging code" is running in a Centos 6 ! and use epel repos ! maybe
that is why I more motivate to packaging for EPEL 6 .
TL;DR;
Can't we fix things on EPEL, to speed up Fedora devel ?
another story that is bugging me is python2 packages for example [1]
if we want call it python2-foo , so we have to guarantee that rule also
applies on RHEL(s) / Centos .
[1]
%if 0%{?fedora}
BuildRequires: python2-six
%else
BuildRequires: python-six
%endif
https://src.fedoraproject.org/rpms/wammu/blob/master/f/wammu.spec#_18
--
Sérgio M. B.