qpid-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Robbie Gemmell <robbie.gemm...@gmail.com>
Subject Re: [NOTICE/DISCUSS] Making proton-j and proton-c more independent
Date Mon, 09 Jan 2017 15:27:32 GMT
On 6 January 2017 at 21:54, Robbie Gemmell <robbie.gemmell@gmail.com> wrote:
> On 6 January 2017 at 20:32, Andrew Stitcher <astitcher@redhat.com> wrote:
>> On Fri, 2017-01-06 at 19:28 +0000, Robbie Gemmell wrote:
>>> On 18 December 2016 at 23:17, Robbie Gemmell <robbie.gemmell@gmail.co
>>> m> wrote:
>>> > On 16 December 2016 at 18:11, Robbie Gemmell <robbie.gemmell@gmail.
>>> > com> wrote:
>>> > > On 15 December 2016 at 18:05, Robbie Gemmell <robbie.gemmell@gmai
>>> > > l.com> wrote:
>>> > > > On 23 November 2016 at 12:40, Robbie Gemmell <robbie.gemmell@gm
>>> > > > ail.com> wrote:
>>> > > > > I'd like to call this out more specifically following
>>> > > > > discussion in a
>>> > > > > couple of recent threads, to ensure folks are clear on things
>>> > > > > before
>>> > > > > they proceed. If not, shout.
>>> > > > >
>>> > > > > Following completion of the 0.16.0 release, I will look to
>>> > > > > proceed
>>> > > > > with making proton-j and proton-c (+bindings) more
>>> > > > > independent to
>>> > > > > allow them to each be released individually and more easily
>>> > > > > differ
>>> > > > > where appropriate in order to best serve users needs. As
part
>>> > > > > of this
>>> > > > > I will look to:
>>> > > > >
>>> > > > > - Have infra create a separate repository for proton-j,
>>> > > > > likely "qpid-proton-j".
>>> > > > > - Remove Messenger from proton-j.
>>> > > > >
>>> > > > > On the former, I think just cloning the existing repo and
>>> > > > > trimming
>>> > > > > things to leave the bits of interest is probably the easiest
>>> > > > > way to
>>> > > > > go, with a reverse trim happening for the existing
>>> > > > > repository. Any
>>> > > > > other thoughts on that?
>>> > > > >
>>> > > > > On the latter, I still need to look at the overall impact,
>>> > > > > since some
>>> > > > > of the python tests do use Messenger and I believe not always
>>> > > > > to test
>>> > > > > Messenger-specific things. If there are tests using messenger
>>> > > > > that are
>>> > > > > of obvious interest which can't quickly be replaced, then
I
>>> > > > > would move
>>> > > > > the messenger impl to just being part of the tests to at
the
>>> > > > > very
>>> > > > > least remove it from the main component until they can be
>>> > > > > replaced.
>>> > > > > I'd look to keep as many of the non-messenger tests as we
can
>>> > > > > initially, replacing them over time where appropriate.
>>> > > > >
>>> > > > > We should start by marking proton-j messenger as deprecated
>>> > > > > in 0.16.0,
>>> > > > > I will do that shortly. The rest will begin after 0.16.0
is
>>> > > > > complete.
>>> > > > >
>>> > > > > Robbie
>>> > > >
>>> > > > As an update on this, I raised
>>> > > > https://issues.apache.org/jira/browse/INFRA-13080 earlier in
>>> > > > the week,
>>> > > > and we now have a new repo in place for proton-j,
>>> > > > https://git-wip-us.apache.org/repos/asf?p=qpid-proton-j.git
>>> > > >
>>> > > > The GitHub mirror should go live during the next 24hrs, theres
>>> > > > a daily
>>> > > > sync point for the process.
>>> > >
>>> > >
>>> > > The GitHub mirror is now also in place:
>>> > > https://github.com/apache/qpid-proton-j
>>> > >
>>> > > I didn't get a chance to start scrubbing things yet, but I may
>>> > > get
>>> > > bored during my holidays...which start right about now :)
>>> > >
>>> > > Robbie
>>> >
>>> > I decided not to leave this hanging entirely until the new year and
>>> > so
>>> > made an initial stab at carving proton-j out on its own and then
>>> > removing Messenger. It still needs more cleanup before I land it in
>>> > the new year but it is well on the way, current state is viewable
>>> > at
>>> > https://github.com/gemmellr/qpid-proton-j/tree/independent-proton-j
>>> > for anyone wanting to check it out before that. I haven't looked at
>>> > the reverse case of removing proton-j from the original repo, I
>>> > will
>>> > leave that until after the holidays are over.
>>> >
>>> > Robbie
>>>
>>> I've been working on the reverse case of removing proton-j from the
>>> original qpid-proton repo now that its new repo has been created.
>>>
>>> The current state is viewable on this branch:
>>> https://github.com/gemmellr/qpid-proton/tree/independent-proton-c
>>>
>>> Or as a diff from the current master, look at
>>> https://github.com/gemmellr/qpid-proton/compare/master...gemmellr:ind
>>> ependent-proton-c
>>>
>>> I think this side of things is about ready to go, I plan to land it
>>> on
>>> master next week.
>>
>> From the pov of changes to the proton-c build system this looks right -
>> and if not then I'm sure the fixes will be small.
>>
>> Andrew
>>
>
> Great, thanks for looking it over.

I've made some more changes, and pushed things to master. I updated
the ASF CI job configs earlier, and the Travis / Appveyor jobs seem
happy following the change but shout if you spot issues. The proton-c
job on the ASF CI is not happy, but that seems unrelated, and on
checking it wasnt happy before the change either.

Robbie

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@qpid.apache.org
For additional commands, e-mail: users-help@qpid.apache.org


Mime
View raw message