Qpid C++ status

[Alan Conway <ac...@redhat.com>]

The Qpid C++ code is in flux, I did some restructuring yesterday and
will do some more today. If you're working on anything and need help
merging give me a shout. 

Yesterday I renamed the broker-side Session classes (excluding
SessionContext, will rename today) to use the more AMQP-standard name
"Connection" and separated the "adapter" layer of the broker from the
Connection core class.

Today I will move most of the Connection logic to the Channel classes
(client & broker) since most processing happens on a channel and in
particular request/response ID handling has to happen in the context of
a channel. Stuff for channel 0 will be handled by Channe(0) since this
too must have proper request-response ID handling.

Cheers,
Alan.

Re: Qpid C++ status

[Marnie McCormack <ma...@googlemail.com>]

Thanks :-)

Marnie


On 1/18/07, Alan Conway <ac...@redhat.com> wrote:
>
> On Thu, 2007-01-18 at 09:35 +0000, Marnie McCormack wrote:
> > Alan,
> >
> > Just to be clear - you're talking about the branch here, rather than
> trunk ?
>
> Yes, I'll put [c++ 0-9 branch] in the subject in future to make that
> clear. Rest assured I would never do a commit that put the trunk in a
> state like the 0-9 branch is in now (at least not on purpose...;)
>
> It'll all be better soon...
>
> >
> > Thanks & Regards,
> > Marnie
> >
> >
> > On 1/17/07, Alan Conway <ac...@redhat.com> wrote:
> > >
> > > The Qpid C++ code is in flux, I did some restructuring yesterday and
> > > will do some more today. If you're working on anything and need help
> > > merging give me a shout.
> > >
> > > Yesterday I renamed the broker-side Session classes (excluding
> > > SessionContext, will rename today) to use the more AMQP-standard name
> > > "Connection" and separated the "adapter" layer of the broker from the
> > > Connection core class.
> > >
> > > Today I will move most of the Connection logic to the Channel classes
> > > (client & broker) since most processing happens on a channel and in
> > > particular request/response ID handling has to happen in the context
> of
> > > a channel. Stuff for channel 0 will be handled by Channe(0) since this
> > > too must have proper request-response ID handling.
> > >
> > > Cheers,
> > > Alan.
> > >
> > >
>
>

Re: Qpid C++ status

[Alan Conway <ac...@redhat.com>]

On Thu, 2007-01-18 at 09:35 +0000, Marnie McCormack wrote:
> Alan,
> 
> Just to be clear - you're talking about the branch here, rather than trunk ?

Yes, I'll put [c++ 0-9 branch] in the subject in future to make that
clear. Rest assured I would never do a commit that put the trunk in a
state like the 0-9 branch is in now (at least not on purpose...;)

It'll all be better soon...

> 
> Thanks & Regards,
> Marnie
> 
> 
> On 1/17/07, Alan Conway <ac...@redhat.com> wrote:
> >
> > The Qpid C++ code is in flux, I did some restructuring yesterday and
> > will do some more today. If you're working on anything and need help
> > merging give me a shout.
> >
> > Yesterday I renamed the broker-side Session classes (excluding
> > SessionContext, will rename today) to use the more AMQP-standard name
> > "Connection" and separated the "adapter" layer of the broker from the
> > Connection core class.
> >
> > Today I will move most of the Connection logic to the Channel classes
> > (client & broker) since most processing happens on a channel and in
> > particular request/response ID handling has to happen in the context of
> > a channel. Stuff for channel 0 will be handled by Channe(0) since this
> > too must have proper request-response ID handling.
> >
> > Cheers,
> > Alan.
> >
> >

Re: Qpid C++ status

[Marnie McCormack <ma...@googlemail.com>]

Alan,

Just to be clear - you're talking about the branch here, rather than trunk ?

Thanks & Regards,
Marnie


On 1/17/07, Alan Conway <ac...@redhat.com> wrote:
>
> The Qpid C++ code is in flux, I did some restructuring yesterday and
> will do some more today. If you're working on anything and need help
> merging give me a shout.
>
> Yesterday I renamed the broker-side Session classes (excluding
> SessionContext, will rename today) to use the more AMQP-standard name
> "Connection" and separated the "adapter" layer of the broker from the
> Connection core class.
>
> Today I will move most of the Connection logic to the Channel classes
> (client & broker) since most processing happens on a channel and in
> particular request/response ID handling has to happen in the context of
> a channel. Stuff for channel 0 will be handled by Channe(0) since this
> too must have proper request-response ID handling.
>
> Cheers,
> Alan.
>
>