M2 - let us try another "final" build

[Robert Greig <ro...@gmail.com>]

Hi all & Rajith in particular,

I have merged Martin's QPID-531 fix into the M2 branch.

That means we now have fixes for the three outstanding blocker Jiras
to proceed with a final build.

Rajith - can you do another build of the Java tree for M2 and we can
hopefully get this voted on at the end of the week?

Thanks,

Robert

Re: M2 - let us try another "final" build

[Robert Greig <ro...@gmail.com>]

On 18/09/2007, Carl Trieloff <cc...@redhat.com> wrote:
>
> Running org.apache.qpid.test.unit.topic.TopicSessionTest
> Tests run: 8, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 4.351
> sec <<< FAILURE!
> Results :
>
> ... is the culprit.

I have now seen this failure on our continuous build machine too.

RG

Re: M2 - let us try another "final" build

[Rafael Schloming <ra...@redhat.com>]

Carl Trieloff wrote:
> Carl Trieloff wrote:
>> Robert Greig wrote:
>>> On 18/09/2007, Rajith Attapattu <ra...@gmail.com> wrote:
>>>
>>>  
>>>> Has this been tested ?
>>>>     
>>>
>>> Well we are running a continuous build so it does build and the tests 
>>> pass.
>>>
>>> Of the three issues, QPID-531 was fixed by Martin (and includes
>>> updated test case), QPID-573 has been fixed by Rafi and QPID-572 has a
>>> fix for the most serious aspect (leaving a less critical ordering
>>> issue QPID-601 when using selectors).
>>>
>>> From the discussion on the other thread I think there is agreement we
>>> can go ahead with M2 and work on the message ordering issue when using
>>> selectors as a separate issue (to be scoped for M2.1 or later).
>>>
>>> RG
>>>   
>>
>> I see our build server is currently showing failures on java-qpid-M2. 
>> Give us a few to investigate as
>> I will base my vote from that.
>>
>>
> 
> Running org.apache.qpid.test.unit.topic.TopicSessionTest
> Tests run: 8, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 4.351 
> sec <<< FAILURE!
> Results :
> 
> ... is the culprit.

Assuming that Martin's fix is in place on M2, then this may be the 
QPID-601 race condition.

FWIW after looking at the code more, the impact of this race condition 
is actually to strand certain messages inside the broker, not to deliver 
them out of order.

--Rafael

Re: M2 - let us try another "final" build

[Robert Greig <ro...@gmail.com>]

On 19/09/2007, Carl Trieloff <cc...@redhat.com> wrote:

> We test on list of platforms but it fails consistently (every time) on
> the quad core RHEL5 64bit
> Woodcrest machine

I looked at the svn logs more carefully and there was a subsequent
change by Martin on the M2.1 branch for QPID-572. I merged that this
morning but on our continuous build machine this fails too.

We are actually seeing some odd failures so we are actively
investigating if there may be multiple problems.

RG

Re: M2 - let us try another "final" build

[Carl Trieloff <cc...@redhat.com>]

Robert Greig wrote:
> On 18/09/2007, Martin Ritchie <ri...@apache.org> wrote:
>
>   
>> Would be good to get more details.. .even the output of the surefire
>> report (while being quite useless most of the time) may provide some
>> insights.
>>
>> I haven't seen TST fail since I applied the fix.
>>     
>
> Which machines have you been testing on?
>
> RG
>   

We test on list of platforms but it fails consistently (every time) on 
the quad core RHEL5 64bit
Woodcrest machine
Carl.

Re: M2 - let us try another "final" build

[Robert Greig <ro...@gmail.com>]

On 18/09/2007, Martin Ritchie <ri...@apache.org> wrote:

> Would be good to get more details.. .even the output of the surefire
> report (while being quite useless most of the time) may provide some
> insights.
>
> I haven't seen TST fail since I applied the fix.

Which machines have you been testing on?

RG

Re: M2 - let us try another "final" build

[Martin Ritchie <ri...@apache.org>]

On 18/09/2007, Carl Trieloff <cc...@redhat.com> wrote:
> Carl Trieloff wrote:
> > Robert Greig wrote:
> >> On 18/09/2007, Rajith Attapattu <ra...@gmail.com> wrote:
> >>
> >>
> >>> Has this been tested ?
> >>>
> >>
> >> Well we are running a continuous build so it does build and the tests
> >> pass.
> >>
> >> Of the three issues, QPID-531 was fixed by Martin (and includes
> >> updated test case), QPID-573 has been fixed by Rafi and QPID-572 has a
> >> fix for the most serious aspect (leaving a less critical ordering
> >> issue QPID-601 when using selectors).
> >>
> >> From the discussion on the other thread I think there is agreement we
> >> can go ahead with M2 and work on the message ordering issue when using
> >> selectors as a separate issue (to be scoped for M2.1 or later).
> >>
> >> RG
> >>
> >
> > I see our build server is currently showing failures on java-qpid-M2.
> > Give us a few to investigate as
> > I will base my vote from that.
> >
> >
>
> Running org.apache.qpid.test.unit.topic.TopicSessionTest
> Tests run: 8, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 4.351
> sec <<< FAILURE!
> Results :
>
> ... is the culprit.

Would be good to get more details.. .even the output of the surefire
report (while being quite useless most of the time) may provide some
insights.

I haven't seen TST fail since I applied the fix.

-- 
Martin Ritchie

Re: M2 - let us try another "final" build

[Carl Trieloff <cc...@redhat.com>]

Carl Trieloff wrote:
> Robert Greig wrote:
>> On 18/09/2007, Rajith Attapattu <ra...@gmail.com> wrote:
>>
>>  
>>> Has this been tested ?
>>>     
>>
>> Well we are running a continuous build so it does build and the tests 
>> pass.
>>
>> Of the three issues, QPID-531 was fixed by Martin (and includes
>> updated test case), QPID-573 has been fixed by Rafi and QPID-572 has a
>> fix for the most serious aspect (leaving a less critical ordering
>> issue QPID-601 when using selectors).
>>
>> From the discussion on the other thread I think there is agreement we
>> can go ahead with M2 and work on the message ordering issue when using
>> selectors as a separate issue (to be scoped for M2.1 or later).
>>
>> RG
>>   
>
> I see our build server is currently showing failures on java-qpid-M2. 
> Give us a few to investigate as
> I will base my vote from that.
>
>

Running org.apache.qpid.test.unit.topic.TopicSessionTest
Tests run: 8, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 4.351 
sec <<< FAILURE!
Results :

... is the culprit.

Re: M2 - let us try another "final" build

[Carl Trieloff <cc...@redhat.com>]

Robert Greig wrote:
> On 18/09/2007, Rajith Attapattu <ra...@gmail.com> wrote:
>
>   
>> Has this been tested ?
>>     
>
> Well we are running a continuous build so it does build and the tests pass.
>
> Of the three issues, QPID-531 was fixed by Martin (and includes
> updated test case), QPID-573 has been fixed by Rafi and QPID-572 has a
> fix for the most serious aspect (leaving a less critical ordering
> issue QPID-601 when using selectors).
>
> From the discussion on the other thread I think there is agreement we
> can go ahead with M2 and work on the message ordering issue when using
> selectors as a separate issue (to be scoped for M2.1 or later).
>
> RG
>   

I see our build server is currently showing failures on java-qpid-M2. 
Give us a few to investigate as
I will base my vote from that.

Carl.

Re: M2 - let us try another "final" build

[Robert Greig <ro...@gmail.com>]

On 18/09/2007, Rajith Attapattu <ra...@gmail.com> wrote:

> Has this been tested ?

Well we are running a continuous build so it does build and the tests pass.

Of the three issues, QPID-531 was fixed by Martin (and includes
updated test case), QPID-573 has been fixed by Rafi and QPID-572 has a
fix for the most serious aspect (leaving a less critical ordering
issue QPID-601 when using selectors).

>From the discussion on the other thread I think there is agreement we
can go ahead with M2 and work on the message ordering issue when using
selectors as a separate issue (to be scoped for M2.1 or later).

RG

Re: M2 - let us try another "final" build

[Rajith Attapattu <ra...@gmail.com>]

Robert,

Has this been tested ?
What does others feel?
If people are happy I can make another build

Regards,

Rajith

On 9/18/07, Robert Greig <ro...@gmail.com> wrote:
>
> Hi all & Rajith in particular,
>
> I have merged Martin's QPID-531 fix into the M2 branch.
>
> That means we now have fixes for the three outstanding blocker Jiras
> to proceed with a final build.
>
> Rajith - can you do another build of the Java tree for M2 and we can
> hopefully get this voted on at the end of the week?
>
> Thanks,
>
> Robert
>