On 28/08/07, Carl Trieloff <[EMAIL PROTECTED]> wrote:
> Gordon Sim wrote:
> > Martin Ritchie wrote:
> >> Hi,
> >>
> >> Can someone explain to me the warnings I'm seeing from the python tests
> >> ([java][OUT] has been prepended by maven)
> >>
> >>      [java] [OUT]Using specification from: ../specs\amqp.0-8.xml
> >>      [java] [OUT]Warning: duplicate id: Constant(name=xa_rbrollback,
> >> id=1)
> >>      [java] [OUT]Warning: duplicate id: Constant(name=xa_rbtimeout,
> >> id=2)
> >>      [java] [OUT]Warning: duplicate id: Constant(name=xa_heurhaz, id=3)
> >>      [java] [OUT]Warning: duplicate id: Constant(name=xa_rdonly, id=7)
> >>      [java] [OUT]Warning: duplicate id: Constant(name=xa_rbrollback,
> >> id=1)
> >>      [java] [OUT]Warning: duplicate id: Constant(name=xa_rbtimeout,
> >> id=2)
> >>      [java] [OUT]Warning: duplicate id: Constant(name=xa_heurhaz, id=3)
> >>      [java] [OUT]Warning: duplicate id: Constant(name=xa_rdonly, id=7)
> >
> > The warnings are emitted when the spec file loaded has constants with
> > duplicate values (the python client keys constants by value as well as
> > by name in order to use the frame type constants in decoding).
> >
> > However for the 0-8 spec file there are no duplicates. I suspect that
> > some recent tests and test changes that are now loading all the spec
> > versions are causing these errors to be output though you are actually
> > using 0-8.
> >
> >> also there are two tests that are currently failing:
> >>
> >>      [java] [ERR]test_auto_rollback (tests_0-8.tx.TxTests) ... ERROR
> >>      [java] [ERR]test_rollback (tests_0-8.tx.TxTests) ... ERROR
> >>
> >> I shall wind back trunk to find the point at which they started to fail.
> >
> > There have been a fair amount of changes to the python client so I
> > can't rule that out as a cause. However as mentioned in my mail of 2nd
> > August, the python client on trunk passed those two tests when I ran
> > it against the M2 java broker so my initial suspicion would be that it
> > was a java broker issue.
>
> If it helps, these tests currently pass on the C++ broker - so the tests
> are working
> Carl.

Thanks good to know they work on the C++ broker. Still trying to
narrow down when the failures started.

-- 
Martin Ritchie

Reply via email to