You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@ode.apache.org by Tammo van Lessen <tv...@gmail.com> on 2007/05/23 21:31:27 UTC

BPELTestAbstract

Hi there,

I've two questions regarding the BPELTestAbstract class:

a) Would it be possible to move it from bpel-test:src/test to
src/main? This would allow external developers to write tests in
separate modules (to not tainting the ode working copy) as this class
is then included in the ode-bpel-test jar.

b) Most of the bpel-test currently fail in line 559 because finalstat
is null. Some debugging unveiled that it is null because line 544
returns null (If I'm not mistaken because the status gets removed once
the MEX is completed). However, replacing finalstat with
mex.getStatus() in this block helps, but of course I have no idea
whether this breaks the actual intension of this code - haven't dive
into too much yet. Maybe you can shed some light?

Best regards,
  Tammo

-- 
Tammo van Lessen - tvanlessen@gmail.com - http://www.taval.de

Re: BPELTestAbstract

Posted by Matthieu Riou <ma...@gmail.com>.
Hi Tammo,

a) Would it be possible to move it from bpel-test:src/test to
> src/main? This would allow external developers to write tests in
> separate modules (to not tainting the ode working copy) as this class
> is then included in the ode-bpel-test jar.


Yep, good point. That's fixed.

b) Most of the bpel-test currently fail in line 559 because finalstat
> is null. Some debugging unveiled that it is null because line 544
> returns null (If I'm not mistaken because the status gets removed once
> the MEX is completed). However, replacing finalstat with
> mex.getStatus() in this block helps, but of course I have no idea
> whether this breaks the actual intension of this code - haven't dive
> into too much yet. Maybe you can shed some light?


Yeah, actually there was a couple of wrong assumption in that test. First
that the future was returning the mex status and second that this could be
done outside of a transaction. I've replaced it by an explicit status
interrogation done inside a transaction, which fixed some of those tests.
Thanks a lot for the debugging!

Matthieu

Best regards,
>   Tammo
>
> --
> Tammo van Lessen - tvanlessen@gmail.com - http://www.taval.de
>