Skip to main content

Please review bugfix for 6243851: DTF: J2SEMsgServiceProxy assumes that only one compontent would use it

2 replies [Last post]
Anonymous

Alexander, team,

Please review the following bugfix for the
CR #6243851: DTF: J2SEMsgServiceProxy assumes that only one remote
component would use it.

http://fisheye4.cenqua.com/changelog/cqme/branches/users/dtrounine/testE...

Basically, if j2mefw_jt.jar is in the JTharness' classpath, then
different distributed tests can interfere with each other, due to
shared usage of J2SEMsgServiceProxy, which was a singleton. So,
different tests could configure the proxy and could close the proxy,
leading to non-deterministic errors.

One of the licensees reported this problem (they misconfigured the
JTharness, there were no need to put j2mefw_jt.jar into the
classpath). Now, in FW 1.2 we actually can encounter this issue too,
since we now put j2mefw_jt.jar in the classpath too, in some cases,
like starting the JTHarness from NetBeans.

The fix is rather simple, we just eleminate "sigleton" property of
J2SEMsgServiceProxy, and create it every time when requested. Then,
different tests would use different proxies, even in single
classloader, and there will be no interference.

Thanks,
--Vladimir

---------------------------------------------------------------------
To unsubscribe, e-mail: meframework-unsubscribe@cqme.dev.java.net
For additional commands, e-mail: meframework-help@cqme.dev.java.net

Reply viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
Alexander Alexeev

Hi Vladimir,

the changes look good, just one comment - the constructor's comment
seems obsolete:
// Singleton's private constructor

Thanks,
Alexander

Vladimir Sizikov wrote:
> Alexander, team,
>
> Please review the following bugfix for the
> CR #6243851: DTF: J2SEMsgServiceProxy assumes that only one remote
> component would use it.
>
> http://fisheye4.cenqua.com/changelog/cqme/branches/users/dtrounine/testE...
>
> Basically, if j2mefw_jt.jar is in the JTharness' classpath, then
> different distributed tests can interfere with each other, due to
> shared usage of J2SEMsgServiceProxy, which was a singleton. So,
> different tests could configure the proxy and could close the proxy,
> leading to non-deterministic errors.
>
> One of the licensees reported this problem (they misconfigured the
> JTharness, there were no need to put j2mefw_jt.jar into the
> classpath). Now, in FW 1.2 we actually can encounter this issue too,
> since we now put j2mefw_jt.jar in the classpath too, in some cases,
> like starting the JTHarness from NetBeans.
>
> The fix is rather simple, we just eleminate "sigleton" property of
> J2SEMsgServiceProxy, and create it every time when requested. Then,
> different tests would use different proxies, even in single
> classloader, and there will be no interference.
>
> Thanks,
> --Vladimir
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: meframework-unsubscribe@cqme.dev.java.net
> For additional commands, e-mail: meframework-help@cqme.dev.java.net
>

---------------------------------------------------------------------
To unsubscribe, e-mail: meframework-unsubscribe@cqme.dev.java.net
For additional commands, e-mail: meframework-help@cqme.dev.java.net

Vladimir Sizikov

Alexander,

On Fri, Apr 27, 2007 at 09:29:13PM +0400, Alexander Alexeev wrote:
> Hi Vladimir,
>
> the changes look good, just one comment - the constructor's comment
> seems obsolete:
> // Singleton's private constructor

Thanks! I've fixed it and committed the changes.

--Vladimir

---------------------------------------------------------------------
To unsubscribe, e-mail: meframework-unsubscribe@cqme.dev.java.net
For additional commands, e-mail: meframework-help@cqme.dev.java.net