Community
Participate
Working Groups
I setup oomph and provisioned code recommenders from the Eclipse.org catalog. That adds the following setting to my eclispe.ini: pre. -Xrunjdwp:transport=dt_socket,server=y,suspend=n,address=8001 If I now try to start Eclipse more than once (with different workspaces) it crashes with an unexpected error on Mac. This is output on the console (which users don't necessarily see): pre. ERROR: transport error 202: bind failed: Address already in use ERROR: JDWP Transport dt_socket failed to initialize, TRANSPORT_INIT(510) JDWP exit error AGENT_ERROR_TRANSPORT_INIT(197): No transports initialized [debugInit.c:750] FATAL ERROR in native method: JDWP No transports initialized, jvmtiError=AGENT_ERROR_TRANSPORT_INIT(197) Abort trap: 6 I would consider removing the eclipse.ini setting and leave that to debug launch configurations rather than the VM instance that runs Eclipse.
Is the use case to start the "Eclipse IDE for Code Recommenders Developers" twice a common use case? We added that setting b/c we experienced some strange behavior in the 'working' Eclipse which we weren't able to reproduce in debug mode. With that option set by default we were able to start another Eclipse instance and start debugging. If this is considered a severe limitation we can consider removing this option. But if it's only a nice-to-have we may stick to it. Maybe provisioning a seconds Eclipse IDE for Code Recommenders with a different name would reduce mitigate the problem to some extent? Assigning the bug to recommenders-inbox since we are responsible for / maintaining the profile ourselves.
This problem annoyed us, too, but fortunately, there was an easy fix. During the provisioning, you can specify a different debug port for each instance. Make sure to display all variables. Assign a different port to each instance you want to run at the same time and they will get along nicely.
This is a known issue that happens (at least under Mac OS, not sure about other OSes) whenever one tries to restart an Eclipse that has debugging enabled. See Bug 444690.
Waiting for a response before closing as "invalid". (btw: it would rather be wontfix or worksforme).
(In reply to comment #1) > Is the use case to start the "Eclipse IDE for Code Recommenders Developers" > twice a common use case? I can't say if it is common but I commonly start the same instance with different workspaces. > We added that setting b/c we experienced some strange behavior in the 'working' > Eclipse which we weren't able to reproduce in debug mode. With that option set > by default we were able to start another Eclipse instance and start debugging. That makes sense for advanced Eclipse developers but I wonder if that should be the default in the recommenders setup script. > If this is considered a severe limitation we can consider removing this option. > But if it's only a nice-to-have we may stick to it. Considering the side effects on Mac I would tend to remove the setting but I'm fine with whatever solution you feel makes the most sense for the project. > Maybe provisioning a seconds Eclipse IDE for Code Recommenders with a different > name would reduce mitigate the problem to some extent? I'm not entirely sure how that would fix it. I simply removed the option from the eclipse.ini
(In reply to Steffen Pingel from comment #5) > (In reply to comment #1) > > Is the use case to start the "Eclipse IDE for Code Recommenders Developers" > > twice a common use case? > > I can't say if it is common but I commonly start the same instance with > different workspaces. I did this a lot before the advent of Oomph, too. With Oomph, however, I have the feeling that the recommended workflow is to provision a "new" Eclipse for each project. (Thanks to the shared bundle pool, the Eclipse will be "recycled" rather than "new" most of the time.)
