RFR 8049696: com/sun/jdi/RunToExit fails with "ConnectException: Connection refused"

Staffan Larsen staffan.larsen at oracle.com
Tue Mar 10 08:04:32 UTC 2015


When catching the ConnectException, can you print it out as well? It may contain something that is interesting. With that change it looks good to me - no need for an updated review.

Thanks,
/Staffan

> On 9 mar 2015, at 19:40, Jaroslav Bachorik <jaroslav.bachorik at oracle.com> wrote:
> 
> Please, review the following test change
> 
> Issue : https://bugs.openjdk.java.net/browse/JDK-8049696
> Webrev: http://cr.openjdk.java.net/~jbachorik/8049696/webrev.00
> 
> The test seems to fail due to a premature attempt to connect the debugee. The test is trying to guard against this by monitoring the debugee's stdout and allowing to proceed to connect only if something has already arrived on stdout (presumably the "Listening for transport dt_socket" message). However, it does not wait for the whole message to be present in stdout and this might contribute to the instability.
> 
> The proposed fix replaces the arbitrary code used to launch the debugee with the test library ProcessTools.startProcess() and waits for the whole message being read from stdout before proceeding. If, for any reason, synchronizing on the presence of this message is still not enough, the call to method AttachingConnector.attach() is retried if it has lead to ConnectException (probably not initialized debugee).
> 
> Thanks,
> 
> -JB-



More information about the serviceability-dev mailing list