qpid-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Robert Godfrey <rob.j.godf...@gmail.com>
Subject Re: Java InVM test exclusion using Ant test profiles
Date Tue, 08 Dec 2009 12:08:34 GMT
2009/12/8 Martin Ritchie <ritchiem@apache.org>

> 2009/12/8 Robert Godfrey <rob.j.godfrey@gmail.com>:
> > 2009/12/8 Martin Ritchie <ritchiem@apache.org>
> >
> >> 2009/12/7 Rajith Attapattu <rajith77@gmail.com>:
> >> > On Mon, Dec 7, 2009 at 1:39 PM, Robbie Gemmell <
> robbie.gemmell@gmail.com>
> >> wrote:
> >> >> My understanding of it was that the default profile was for InVM
> testing
> >> and
> >> >> pulling in the global excludes, with other broker profiles (java
> >> standalone,
> >> >> cpp etc) based off that and pulling in any profile specific excludes.
> I
> >> >> think thats a god system to keep..it just isn't quite working like
> that
> >> at
> >> >> the moment.
> >> >
> >> > Yes that was the initial approach.
> >> > However I also believe we need to have excludes per protocol version
> as
> >> well.
> >> > We would need to continue to run the 0-8 test profile for quite some
> >> > time as we have those systems in production.
> >> >
> >> > As I outlined in my email I think it's reasonable to assume the
> >> > default profile is all tests that can be run with the in-vm broker.
> >> > Since the Java broker supports both 0-8 and 0-10 I assume both types
> >> > of test can be run against the broker during a single run?
> >>
> >> All protocols can be tested on a single run but InVM currently only
> >> supports 0-8/9
> >>
> >>
> > We should work out a way of enabling 0-10 inVM runs also... should
> certainly
> > be possible with a bit of work I would think...
>
> Yeah should be do able but there are now quite a few broker classes
> required, to setup the protocol, it is just much more awkward. The
> Enum for the supported versions jumps to mind, give the
> TransportConnection class is in the client package so can't directly
> reference the broker.
>
>
Would it be any easier to simply define to different InVM invocation
mechanisms - one for 0-8/0-9/0-9-1 and one for 0-10?

-- Rob


> Martin
> > -- Rob
> >
>
>
>
> --
> Martin Ritchie
>
> ---------------------------------------------------------------------
> Apache Qpid - AMQP Messaging Implementation
> Project:      http://qpid.apache.org
> Use/Interact: mailto:dev-subscribe@qpid.apache.org
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message