--82I3+IH0IqGh5yIs
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable
On 01/12, Sandro Bonazzola wrote:
Il 09/01/2015 23:48, David Caro Estevez ha scritto:
> I really discourage having separated repos for specs, scripts or any 'c=
ode' used for building any official packages.
=20
> That complicates (to the
point of making it not feasible) a continuous =
build approach, and introduces a lot
of complexity and limitations to the c=
ontinuous integration of the project.
=20
I'm not realy sure to follow. What's the difference between Juan's reques=
t and what we have already in place for jasper-report-server or
ovirt-engine-jboss-as? We have specs and sources in different
locations t=
here too.
The difference is that we control the source code, we are not doing
any continuous integration of those projects, as we don't control the
source reposiotries, if we did, it would be complicated to manage as
it's split in two repos (one for the code and another for the
specs). That makes it really hard to create a continuous deployment
pipeline as you have to keep track for each source code commit, which
commit of the spec file is the one that can build it. And we don't do
that, so we can't build any point in th esource code, once a
non-backwards compatible change is made to the spec all the previous
code becomes obsolete.
It also affects to the ability to keep a semantic versioning of the
product, as you'll have to manage generating a version from multiple
repositories (and keep the version<->hashes map and the ordering). As
a change on the spec should be reflected on the version of the
generated package.
=20
=20
> If we can avoid it, I really discourage it. Can we
talk about it and fi=
nd a more friendly solution?
=20
I've no strong feelings for existing or proposed solution.
I just think that from a packaging for distribution point of view, the ko=
ji way
of separating specs and sources is fine.
=20
The problem is in the continuos integration and build point of
view. If you are not aiming to build countinuously it does not matte.
>
=20
> > David Caro
>
=20
> > On Jan 9, 2015 4:27 PM, Sandro Bonazzola
<sbonazzo(a)redhat.com> wrote:
> >>
> >> Il 09/01/2015 15:13, Juan Hern=E1ndez ha scritto:
> >>> Hello,
> >>>
> >>> Please create the following new git repIl 09/01/2015 15:13, Juan Hern=
=E1ndez ha scritto:
>> Hello,
>>
>> Please create the following new git repositories:
>>
>> ovirt-engine-sdk-python-specs
>> ovirt-engine-sdk-java-specs
>> ovirt-engine-cli-specs
>>
>> They will be used to manage the RPM spec files (also scripts, patches,
>> and in general anything needed for building RPMs) of the Python SDK, t=
he
> >> Java SDK, and the CLI. Each repository will have branches for each pair
> >> of oVirt major release and supported distribution:
> >>
>
=20
> > Any reason for not having something like
"ovirt-specs"
> > with suggested branches:
>
=20
>
=20
>
>> ovirt-3.5-el6
> >> ovirt-3.5-el7
> >> ovirt-3.5-fc20
> >> ovirt-3.5-fc21
> >> ovirt-3.6-el6
> >> ...
>
=20
>
=20
>
> and having there
>
=20
> >
ovirt-engine-sdk-python/ovirt-engine-sdk-python.spec
> > ovirt-engine-sdk-python/0001-your-friendly-patch.patch
> > ovirt-engine-sdk-java/ovirt-engine-sdk-java.spec
> > ovirt-engine-cli/ovirt-engine-cli.spec
> > ...
>
=20
> >>
> >> Initially I will be the maintainer of these repositories.
> >>
> >> Once these repos are created the specs will be added to them and
> >> eventually removed completely from the main repositories of the SDKs a=
nd
> >> the CLI.
> >>
> >> Thanks in advance,
> >> Juan Hern=E1ndez
> >>
>
=20
>
=20
=20
=20
> --=20
> Sandro Bonazzola
> Better technology. Faster innovation. Powered by community collaboration.
> See how it works at
redhat.com
--=20
David Caro
Red Hat S.L.
Continuous Integration Engineer - EMEA ENG Virtualization R&D
Tel.: +420 532 294 605
Email: dcaro(a)redhat.com
Web:
www.redhat.com
RHT Global #: 82-62605
--82I3+IH0IqGh5yIs
Content-Type: application/pgp-signature
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAEBAgAGBQJUs4DFAAoJEEBxx+HSYmnDjLsH/3UtWBx6krCiX8g8PXhWrI+V
3T0kgRBLUhOZcrAWJDLey8aB798bjc65G+pY6b2u2GyS8CU8jT8KRJ8Rrtlofd6H
z5gmo7G5uCiE7VG5JpMnkgb2vJUZ+ps+xL6cdzAUld8P8tGY8mt8Fd/QXhvGsAXL
y3qtYWhfvecl3Fbl+C0TOir/VNqhPWFNGrF4c8EAH9NC4blsMjojUQOc9ID+t7At
kpgDSAfkBFDc27IIpIvYYsWh9DUCtQTBqYJgVN9ttzVAHacH7W2Ag8bh/jX3xhNx
7ySH0LDb7cYJlxYFHYvj4yuz2rP+WjqbDv3HYjQ42OVNucqAk5aFgB3IK+nYKqQ=
=Pxc6
-----END PGP SIGNATURE-----
--82I3+IH0IqGh5yIs--