Use %autochangelog and %autorelease macros
This change introduces the [`%autochangelog`](1) and [`%autorelease`](2) macros.
Everytime the `Version:` tag is increased, the `Release:` tag will be
reset to `1`.
What do we get out of this change?
----------------------------------
I've had a long discussion with Zbigniew Jedrzejewski Szmek <zjedrzej@redhat.com> about
this. The outcome is as follows:
1. A packager no longer has to manually increase the number in the
`Release:` tag or reset it. Increasing is always done when a new commit is
made. Resetting happens automatically when you bump the `Version:`
tag from 13.0.0 to `14.0.0` for example.
2. A package cosumer is supposed to look at the output of
`rpm -q --changelog <PACKAGE>`. Packagers themselves only need to
care about the first line of their next commit when authoring
patches.
3. The newly added `changelog` file in this patch contains the old
content of the `%changelog` section in the spec file.
I was at first confused for why we not simply keep the section as
is and have the first line of the new commits appear in the final
spec file. But I was told that having a very clean spec file makes
cherry picking patches much more simple. I can see why that is by
looking at my own history.
This patch only makes sense, if we use %autochangelog and %autorelease
in the rest of the branches for each package as well. Then we can more easily
backport patches for example.
[1]: https://docs.pagure.org/fedora-infra.rpmautospec/autochangelog.html
[2]: https://docs.pagure.org/fedora-infra.rpmautospec/autorelease.html#using-autorelease