You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@qpid.apache.org by Alan Conway <ac...@redhat.com> on 2007/09/25 14:55:57 UTC
Weekly plans [was Re: C++ code generator]
On Tue, 2007-09-25 at 08:26 -0400, Carl Trieloff wrote:
> Suggestion:
> ---------------
> One thing that might help is if we pick a day of the week and everyone
> on the project posts what they are
> going to do that week, or at the end of the week post everything that
> they have done. That way it is easy
> for everyone on the project to be able to ask about any work being done.
>
> Thoughts..... would everyone support such a process?
> Carl.
I'm in favour!
My plan this week:
Finish C++ session support: suspend, resume and session level acks in C
++ client & broker. Possibly some refactoring of client handler chains
to be more consistent with broker to allow re-use of a common session
ack handler.
Start on Queue proxies & queue replication for C++ broker.
Re: Weekly plans
Posted by Gordon Sim <gs...@redhat.com>.
I'm mainly going to be working on additional tests for the 0-10 c++ code
this next week (and fixing bugs that come up). (Not very worthy of
announcement I know...)
Re: Weekly plans [was Re: C++ code generator]
Posted by Gordon Sim <gs...@redhat.com>.
Alan Conway wrote:
> On Tue, 2007-09-25 at 08:26 -0400, Carl Trieloff wrote:
>> Suggestion:
>> ---------------
>> One thing that might help is if we pick a day of the week and everyone
>> on the project posts what they are
>> going to do that week, or at the end of the week post everything that
>> they have done. That way it is easy
>> for everyone on the project to be able to ask about any work being done.
>>
>> Thoughts..... would everyone support such a process?
>> Carl.
>
> I'm in favour!
Me too... lets hope I get into the habit!
Over the next week I plan to add support for 'packed' structs, include
the execution header in all L4 method segments and handle the sync bit
in that header correctly.
Re: Weekly plans
Posted by Andrew Stitcher <as...@redhat.com>.
This week I plan to finish 0-10 field tables for C++ broker/client API,
probably including a rewrite of the C++ client API to be nicer to use.
Work on allowing multiple 0-10 frames in a segment.
Andrew
Re: Weekly plans
Posted by Marnie McCormack <ma...@googlemail.com>.
Hi All,
Great idea. This week I'm catching up with the mailing lists and M2. Over
the next few weeks I plan to start easing back into the world of Qpid,
probably on the doc side, time permitting.
I'll be back in anger after Christmas.
Bfn,
Marnie
On 9/25/07, Nuno Santos <ns...@redhat.com> wrote:
>
> On Tue, 2007-09-25 at 08:26 -0400, Carl Trieloff wrote:
> > Suggestion: One thing that might help is if we pick a day of the week
> > and everyone on the project posts what they are going to do that
> > week, or at the end of the week post everything that they have done.
> > That way it is easy for everyone on the project to be able to ask
> > about any work being done.
>
> My plan for this week is to work with the C++ and Java guys to automate
> the cross-language tests as much as possible, with the end goal of
> integrating them into our continuous build system.
>
> Nuno
>
Re: Weekly plans
Posted by Nuno Santos <ns...@redhat.com>.
Nuno Santos wrote:
> My plan for this week is to work with the C++ and Java guys to automate
> the cross-language tests as much as possible, with the end goal of
> integrating them into our continuous build system.
The above has spilled over into this week... I also plan to start
working on the management console, namely going over the current
requirements, doing a gap-analysis versus the existing eclipse plugin,
and if time allows, start a mock-up of possible user interactions.
Nuno
Re: Weekly plans
Posted by Nuno Santos <ns...@redhat.com>.
On Tue, 2007-09-25 at 08:26 -0400, Carl Trieloff wrote:
> Suggestion: One thing that might help is if we pick a day of the week
> and everyone on the project posts what they are going to do that
> week, or at the end of the week post everything that they have done.
> That way it is easy for everyone on the project to be able to ask
> about any work being done.
My plan for this week is to work with the C++ and Java guys to automate
the cross-language tests as much as possible, with the end goal of
integrating them into our continuous build system.
Nuno
Re: Weekly plans [was Re: C++ code generator]
Posted by Arnaud Simon <as...@redhat.com>.
I am working on
- Getting java systests and interoptests running against a 0_10 C++
broker
- Adding dtx (crash recovery) tests
Re: Weekly plans [was Re: C++ code generator]
Posted by Carl Trieloff <cc...@redhat.com>.
I am working on
- async persistence integration for C++
- performance (work waiting for interop issues to complete)
Re: Status dates
Posted by Alan Conway <ac...@redhat.com>.
On Tue, 2007-09-25 at 10:17 -0400, Carl Trieloff wrote:
> Let's plan to send out the status updates every Tuesday.
>
> Carl.
+1
Status dates
Posted by Carl Trieloff <cc...@redhat.com>.
Let's plan to send out the status updates every Tuesday.
Carl.
Re: Weekly plans [was Re: C++ code generator]
Posted by Martin Ritchie <ri...@apache.org>.
My plan is ensuring all the outstanding M2 issues are resolved then I
hope to find some time to look at reinstating our Mina changes.
On 25/09/2007, Robert Greig <ro...@gmail.com> wrote:
> On 25/09/2007, Alan Conway <ac...@redhat.com> wrote:
>
> > I'm in favour!
>
> Yep this is a good idea.
>
> My plan is to keep trying to get M2 Java into a state where it
> actually passes all the tests consistently so that we might actually
> be able to release it.
>
> RG
>
--
Martin Ritchie
Re: Weekly plans [was Re: C++ code generator]
Posted by Robert Greig <ro...@gmail.com>.
On 25/09/2007, Alan Conway <ac...@redhat.com> wrote:
> I'm in favour!
Yep this is a good idea.
My plan is to keep trying to get M2 Java into a state where it
actually passes all the tests consistently so that we might actually
be able to release it.
RG
Re: Weekly plans [was Re: C++ code generator]
Posted by Robert Greig <ro...@gmail.com>.
On 25/09/2007, Robert Godfrey <ro...@gmail.com> wrote:
> Nope - I've had a request to get all the AMQP implementations to
> interoperate (Qpid / Rabbit / OpenAMQ) ... and after discussions with those
> guys we decided it was best to aim for interoperability at 0-9 (non-WIP).
> So I'm adding some code to the Java side so it can speak 0-8 and 0-9, as
> well as fixing up a few other compliance issues (like frame size
> negotiation).
>
> I was *sure* I'd mentioned this on the list somewhere :-)
We did discuss this - under an M2.1 subject.
The idea, in summary, is to have an M2.1 release for at least the Java
components (there didn't appear to be much interest from others) that
is very targetted and limited in scope providing interop at the 0-9
level (to allow interop with OpenAMQ and Rabbit in particular) and
fixing only key issues.
Obviously M2 has dragged on a bit due to issues found in testing the Java.
RG
Re: Weekly plans [was Re: C++ code generator]
Posted by Alan Conway <ac...@redhat.com>.
On Tue, 2007-09-25 at 15:30 +0100, Robert Godfrey wrote:
> On 25/09/2007, Alan Conway <ac...@redhat.com> wrote:
> >
> > On Tue, 2007-09-25 at 14:05 +0100, Robert Godfrey wrote:
> > > On Qpid my main plan is to try to get back to putting 0-9 support into
> > > M2.1...
> > 0-9??? Is that a typo?
>
> Nope - I've had a request to get all the AMQP implementations to
> interoperate (Qpid / Rabbit / OpenAMQ) ... and after discussions with those
> guys we decided it was best to aim for interoperability at 0-9 (non-WIP).
> So I'm adding some code to the Java side so it can speak 0-8 and 0-9, as
> well as fixing up a few other compliance issues (like frame size
> negotiation).
>
> I was *sure* I'd mentioned this on the list somewhere :-)
>
I'm sure you did and I wasn't paying attention. I'm a bit perplexed but
no need to re-open the discussion, I'm fully occupied with 0.10 these
days.
Re: Weekly plans [was Re: C++ code generator]
Posted by Robert Godfrey <ro...@gmail.com>.
On 25/09/2007, Alan Conway <ac...@redhat.com> wrote:
>
> On Tue, 2007-09-25 at 14:05 +0100, Robert Godfrey wrote:
> > On Qpid my main plan is to try to get back to putting 0-9 support into
> > M2.1...
> 0-9??? Is that a typo?
Nope - I've had a request to get all the AMQP implementations to
interoperate (Qpid / Rabbit / OpenAMQ) ... and after discussions with those
guys we decided it was best to aim for interoperability at 0-9 (non-WIP).
So I'm adding some code to the Java side so it can speak 0-8 and 0-9, as
well as fixing up a few other compliance issues (like frame size
negotiation).
I was *sure* I'd mentioned this on the list somewhere :-)
-- Rob
> Alan - is there a design doc somewhere about how you are approaching
> > replication / clustering in the C++ broker?
>
> http://cwiki.apache.org/qpid/clusteringha.html was written in the 0-9
> days, the implementation will use 0-10 session acks for failover and
> AMQP-95 for notification, so it will work with any 0-10 client.
>
>
>
>
Re: Weekly plans [was Re: C++ code generator]
Posted by Alan Conway <ac...@redhat.com>.
On Tue, 2007-09-25 at 14:05 +0100, Robert Godfrey wrote:
> On Qpid my main plan is to try to get back to putting 0-9 support into
> M2.1...
0-9??? Is that a typo?
> Alan - is there a design doc somewhere about how you are approaching
> replication / clustering in the C++ broker?
http://cwiki.apache.org/qpid/clusteringha.html was written in the 0-9
days, the implementation will use 0-10 session acks for failover and
AMQP-95 for notification, so it will work with any 0-10 client.
Re: Weekly plans [was Re: C++ code generator]
Posted by Robert Godfrey <ro...@gmail.com>.
On 25/09/2007, Alan Conway <ac...@redhat.com> wrote:
>
> On Tue, 2007-09-25 at 08:26 -0400, Carl Trieloff wrote:
> > Suggestion:
> > ---------------
> > One thing that might help is if we pick a day of the week and everyone
> > on the project posts what they are
> > going to do that week, or at the end of the week post everything that
> > they have done. That way it is easy
> > for everyone on the project to be able to ask about any work being done.
> >
> > Thoughts..... would everyone support such a process?
> > Carl.
>
> I'm in favour!
+1 from me too...
On Qpid my main plan is to try to get back to putting 0-9 support into
M2.1... Also, after issues we've been having today, butting
re-fragmentation of
messages where the max-frame-size is not 64Kb into the Java broker.
My plan this week:
>
> Finish C++ session support: suspend, resume and session level acks in C
> ++ client & broker. Possibly some refactoring of client handler chains
> to be more consistent with broker to allow re-use of a common session
> ack handler.
>
> Start on Queue proxies & queue replication for C++ broker.
Alan - is there a design doc somewhere about how you are approaching
replication / clustering in the C++ broker?
Thanks,
Rob