Re: [build-test alert] BUILD FAILED WinXP SP2 1CPU msvc debug: classlib

2006-11-28 Thread Vladimir Ivanov

Thanks, now all luni tests passed for me on my WinXP.

On 11/28/06, Tim Ellison <[EMAIL PROTECTED]> wrote:


Vladimir Ivanov wrote:
> On 11/28/06, Paulex Yang <[EMAIL PROTECTED]> wrote:
>>
>> Tim Ellison wrote:
>> > Vladimir Ivanov wrote:
>> >
>> >> Thanks, now the build OK.
>> >>
>> >
>> > Ok, you see no failures in LUNI tests?
>> >
>> I've reverted the patch for HARMONY-2157 and HARMONY-2270 because all
>> the patches have some issues, the build should be OK now. Sorry for the
>> inconvenience caused.
>
>
> Actually, the tests.api.java.io.FileTest failed for me again. I think it
is
> due to not updated drlvm (classlib version is r480066, drlvm is
r480039).

It was reverted in the classlib at r480075 -- so you need to update.

Regards,
Tim

--

Tim Ellison ([EMAIL PROTECTED])
IBM Java technology centre, UK.



Re: [build-test alert] BUILD FAILED WinXP SP2 1CPU msvc debug: classlib

2006-11-28 Thread Tim Ellison
Vladimir Ivanov wrote:
> On 11/28/06, Paulex Yang <[EMAIL PROTECTED]> wrote:
>>
>> Tim Ellison wrote:
>> > Vladimir Ivanov wrote:
>> >
>> >> Thanks, now the build OK.
>> >>
>> >
>> > Ok, you see no failures in LUNI tests?
>> >
>> I've reverted the patch for HARMONY-2157 and HARMONY-2270 because all
>> the patches have some issues, the build should be OK now. Sorry for the
>> inconvenience caused.
> 
> 
> Actually, the tests.api.java.io.FileTest failed for me again. I think it is
> due to not updated drlvm (classlib version is r480066, drlvm is r480039).

It was reverted in the classlib at r480075 -- so you need to update.

Regards,
Tim

-- 

Tim Ellison ([EMAIL PROTECTED])
IBM Java technology centre, UK.


Re: [build-test alert] BUILD FAILED WinXP SP2 1CPU msvc debug: classlib

2006-11-28 Thread Paulex Yang

Vladimir Ivanov wrote:

On 11/28/06, Paulex Yang <[EMAIL PROTECTED]> wrote:


Tim Ellison wrote:
> Vladimir Ivanov wrote:
>
>> Thanks, now the build OK.
>>
>
> Ok, you see no failures in LUNI tests?
>
I've reverted the patch for HARMONY-2157 and HARMONY-2270 because all
the patches have some issues, the build should be OK now. Sorry for the
inconvenience caused.



Actually, the tests.api.java.io.FileTest failed for me again. I think 
it is

due to not updated drlvm (classlib version is r480066, drlvm is r480039).

 time="

0.25">
   junit.framework.AssertionFailedError: 


mkdir 248 failed
at tests.api.java.io.FileTest.test_mkdir(FileTest.java:1812)
at java.lang.reflect.VMReflection.invokeMethod(Native Method)

 
Not surprised you still get failure...I reverted those two patches at 
revision r480070 and r480075 for tests(luni)/impl(portlib) respectively





Regards,
> Tim
>
>


--
Paulex Yang
China Software Development Lab
IBM








--
Paulex Yang
China Software Development Lab
IBM




Re: [build-test alert] BUILD FAILED WinXP SP2 1CPU msvc debug: classlib

2006-11-28 Thread Vladimir Ivanov

On 11/28/06, Paulex Yang <[EMAIL PROTECTED]> wrote:


Tim Ellison wrote:
> Vladimir Ivanov wrote:
>
>> Thanks, now the build OK.
>>
>
> Ok, you see no failures in LUNI tests?
>
I've reverted the patch for HARMONY-2157 and HARMONY-2270 because all
the patches have some issues, the build should be OK now. Sorry for the
inconvenience caused.



Actually, the tests.api.java.io.FileTest failed for me again. I think it is
due to not updated drlvm (classlib version is r480066, drlvm is r480039).

 
   junit.framework.AssertionFailedError:
mkdir 248 failed
at tests.api.java.io.FileTest.test_mkdir(FileTest.java:1812)
at java.lang.reflect.VMReflection.invokeMethod(Native Method)

 




Regards,
> Tim
>
>


--
Paulex Yang
China Software Development Lab
IBM





Re: [build-test alert] BUILD FAILED WinXP SP2 1CPU msvc debug: classlib

2006-11-28 Thread Paulex Yang

Tim Ellison wrote:

Vladimir Ivanov wrote:
  

Thanks, now the build OK.



Ok, you see no failures in LUNI tests?
  
I've reverted the patch for HARMONY-2157 and HARMONY-2270 because all 
the patches have some issues, the build should be OK now. Sorry for the 
inconvenience caused.

Regards,
Tim

  



--
Paulex Yang
China Software Development Lab
IBM




Re: [build-test alert] BUILD FAILED WinXP SP2 1CPU msvc debug: classlib

2006-11-28 Thread Tim Ellison
Vladimir Ivanov wrote:
> Thanks, now the build OK.

Ok, you see no failures in LUNI tests?

Regards,
Tim

-- 

Tim Ellison ([EMAIL PROTECTED])
IBM Java technology centre, UK.


Re: [build-test alert] BUILD FAILED WinXP SP2 1CPU msvc debug: classlib

2006-11-28 Thread Vladimir Ivanov

Thanks, now the build OK.
Thanks, Vladimir


On 11/28/06, Vladimir Ivanov <[EMAIL PROTECTED]> wrote:




On 11/28/06, Tim Ellison <[EMAIL PROTECTED] > wrote:
>
> Vladimir Ivanov wrote:
> > Should I stop the CC while the issue with test compilation will be
> > resolved?
>
> It should be resolved now.  Are there any other outstanding issues to
> get the build fixed?



Ups, my winxp was failed and status for classlib/trunk is 'obstructed' now
:(

My issue was bind with 'test' target for classlib - it reports compilation
problems.

To verify it for me I need more time.

 Thanks, Vladimir

Regards,
> Tim
>
> --
>
> Tim Ellison ( [EMAIL PROTECTED])
> IBM Java technology centre, UK.
>




Re: [build-test alert] BUILD FAILED WinXP SP2 1CPU msvc debug: classlib

2006-11-28 Thread Vladimir Ivanov

On 11/28/06, Tim Ellison <[EMAIL PROTECTED]> wrote:


Vladimir Ivanov wrote:
> Should I stop the CC while the issue with test compilation will be
> resolved?

It should be resolved now.  Are there any other outstanding issues to
get the build fixed?




Ups, my winxp was failed and status for classlib/trunk is 'obstructed' now
:(

My issue was bind with 'test' target for classlib - it reports compilation
problems.

To verify it for me I need more time.

Thanks, Vladimir

Regards,

Tim

--

Tim Ellison ([EMAIL PROTECTED])
IBM Java technology centre, UK.



Re: [build-test alert] BUILD FAILED WinXP SP2 1CPU msvc debug: classlib

2006-11-28 Thread Tim Ellison
Vladimir Ivanov wrote:
> Should I stop the CC while the issue with test compilation will be
> resolved?

It should be resolved now.  Are there any other outstanding issues to
get the build fixed?

Regards,
Tim

-- 

Tim Ellison ([EMAIL PROTECTED])
IBM Java technology centre, UK.


Re: [build-test alert] BUILD FAILED WinXP SP2 1CPU msvc debug: classlib

2006-11-28 Thread Vladimir Ivanov

Not yet

On 11/28/06, Geir Magnusson Jr. <[EMAIL PROTECTED]> wrote:


it is getting spam-like

Doesn't CC already do the "send no mail until state change"?



Not yet.
OK. I'll stop CC while this feature will be implemented.


geir



Vladimir Ivanov wrote:
> Should I stop the CC while the issue with test compilation will be
> resolved?
>
> Vladimir
>
> On 11/28/06, Geir Magnusson Jr. <[EMAIL PROTECTED]> wrote:
>>
>>
>>
>> Tim Ellison wrote:
>> > Vladimir Ivanov wrote:
>> >> On 11/28/06, Geir Magnusson Jr. <[EMAIL PROTECTED]> wrote:
>> >>> I've seen a few of these repeated - shouldn't the CI system not
send
>> the
>> >>> same message over and over?
>> >> Now the system sends notification for each failure. It can de
changed
>> to
>> >> send notification only for first failure but such system will be
>> useless in
>> >> the time of failure investigation.
>> >>
>> >> The current scenario based on the assumption that each failure is
>> critical
>> >> and should be fixed ASAP or patch should be reverted. In this case
>> next
>> CC
>> >> cycle will pass.
>> >>
>> >> Should I change the CC to send one notification for build failure?
>> >
>> > I would say so, otherwise we become blind to the repeated messages.
>> > Once the build break flag goes up people should turn their attention
to
>> > it until it is resolved.
>>
>> Exactly.  I now that I'm getting a bit jaded from the repeated stream.
>>
>> One message, and then stop. When the CI build/test run finally works
>> again, then another message saying "BUILD FIXED" (or whatever).
>>
>> geir
>>
>>
>



Re: [build-test alert] BUILD FAILED WinXP SP2 1CPU msvc debug: classlib

2006-11-28 Thread Geir Magnusson Jr.

it is getting spam-like

Doesn't CC already do the "send no mail until state change"?

geir


Vladimir Ivanov wrote:
Should I stop the CC while the issue with test compilation will be 
resolved?


Vladimir

On 11/28/06, Geir Magnusson Jr. <[EMAIL PROTECTED]> wrote:




Tim Ellison wrote:
> Vladimir Ivanov wrote:
>> On 11/28/06, Geir Magnusson Jr. <[EMAIL PROTECTED]> wrote:
>>> I've seen a few of these repeated - shouldn't the CI system not send
the
>>> same message over and over?
>> Now the system sends notification for each failure. It can de changed
to
>> send notification only for first failure but such system will be
useless in
>> the time of failure investigation.
>>
>> The current scenario based on the assumption that each failure is
critical
>> and should be fixed ASAP or patch should be reverted. In this case 
next

CC
>> cycle will pass.
>>
>> Should I change the CC to send one notification for build failure?
>
> I would say so, otherwise we become blind to the repeated messages.
> Once the build break flag goes up people should turn their attention to
> it until it is resolved.

Exactly.  I now that I'm getting a bit jaded from the repeated stream.

One message, and then stop. When the CI build/test run finally works
again, then another message saying "BUILD FIXED" (or whatever).

geir






Re: [build-test alert] BUILD FAILED WinXP SP2 1CPU msvc debug: classlib

2006-11-28 Thread Vladimir Ivanov

Should I stop the CC while the issue with test compilation will be resolved?

Vladimir

On 11/28/06, Geir Magnusson Jr. <[EMAIL PROTECTED]> wrote:




Tim Ellison wrote:
> Vladimir Ivanov wrote:
>> On 11/28/06, Geir Magnusson Jr. <[EMAIL PROTECTED]> wrote:
>>> I've seen a few of these repeated - shouldn't the CI system not send
the
>>> same message over and over?
>> Now the system sends notification for each failure. It can de changed
to
>> send notification only for first failure but such system will be
useless in
>> the time of failure investigation.
>>
>> The current scenario based on the assumption that each failure is
critical
>> and should be fixed ASAP or patch should be reverted. In this case next
CC
>> cycle will pass.
>>
>> Should I change the CC to send one notification for build failure?
>
> I would say so, otherwise we become blind to the repeated messages.
> Once the build break flag goes up people should turn their attention to
> it until it is resolved.

Exactly.  I now that I'm getting a bit jaded from the repeated stream.

One message, and then stop. When the CI build/test run finally works
again, then another message saying "BUILD FIXED" (or whatever).

geir




Re: [build-test alert] BUILD FAILED WinXP SP2 1CPU msvc debug: classlib

2006-11-28 Thread Geir Magnusson Jr.



Tim Ellison wrote:

Vladimir Ivanov wrote:

On 11/28/06, Geir Magnusson Jr. <[EMAIL PROTECTED]> wrote:

I've seen a few of these repeated - shouldn't the CI system not send the
same message over and over?

Now the system sends notification for each failure. It can de changed to
send notification only for first failure but such system will be useless in
the time of failure investigation.

The current scenario based on the assumption that each failure is critical
and should be fixed ASAP or patch should be reverted. In this case next CC
cycle will pass.

Should I change the CC to send one notification for build failure?


I would say so, otherwise we become blind to the repeated messages.
Once the build break flag goes up people should turn their attention to
it until it is resolved.


Exactly.  I now that I'm getting a bit jaded from the repeated stream.

One message, and then stop. When the CI build/test run finally works 
again, then another message saying "BUILD FIXED" (or whatever).


geir



Re: [build-test alert] BUILD FAILED WinXP SP2 1CPU msvc debug: classlib

2006-11-28 Thread Tim Ellison
Vladimir Ivanov wrote:
> On 11/28/06, Geir Magnusson Jr. <[EMAIL PROTECTED]> wrote:
>>
>> I've seen a few of these repeated - shouldn't the CI system not send the
>> same message over and over?
> 
> Now the system sends notification for each failure. It can de changed to
> send notification only for first failure but such system will be useless in
> the time of failure investigation.
> 
> The current scenario based on the assumption that each failure is critical
> and should be fixed ASAP or patch should be reverted. In this case next CC
> cycle will pass.
> 
> Should I change the CC to send one notification for build failure?

I would say so, otherwise we become blind to the repeated messages.
Once the build break flag goes up people should turn their attention to
it until it is resolved.

Regards,
Tim

-- 

Tim Ellison ([EMAIL PROTECTED])
IBM Java technology centre, UK.