[Openais] defect 791 - return codes not returned for both synchronizeasync and syncasync callback

Steven Dake sdake at mvista.com
Thu Jul 28 11:27:25 PDT 2005


On Thu, 2005-07-28 at 11:06 -0700, Mark Haverkamp wrote:
> On Thu, 2005-07-28 at 10:50 -0700, Steven Dake wrote:
> > On Thu, 2005-07-28 at 07:40 -0700, Mark Haverkamp wrote:
> > > On Wed, 2005-07-27 at 16:11 -0700, Steven Dake wrote:
> > > > this is unfortunately required by the saftest project in the B.01.01
> > > > specification.  patch attached to fix.
> > > 
> > > Are you sure that we want to modify our code just to satisfy saftest?
> > > This sounds like a case that I have in the event service where there is
> > > an inconsistency in the specification in that it looks possible to get
> > > both an error code from the async open and the open callback.  It seems
> > > wrong to allow an error code to be returned from a call and also execute
> > > a callback. The people on the saf-open list seemed to think so too.  I
> > > think that the tests shouldn't be the ones interpreting the
> > > specification, but the people who wrote it.
> > > 
> > > Mark.
> > >  
> > 
> > Unfortunately at this time the saf-test test cases are the only
> > interpretation that matters for certification :(.  So with broken tests
> > we are somewhat stuck unless we can get them changed.  And conformance
> > is pretty important for us..
> 
> Is the saf-test suite the official conformance test for SAF?
> 

yes so any failing tests we should get worked out either in our tree or
in saftest tree by December when saf conformance for ais goes live..

> 
> > 
> > I will submit a bug to saftest for this issue.  If they fix it, then
> > I'll not commit this patch.  If they won't fix it, then I'll commit the
> > patch.
> > 
> > Can you submit a bug/patch to saftest for any evt tests with related
> > behavior?
> 
> I will do that.
> 

thx

> > 
> > Regards
> > -steve




More information about the Openais mailing list