DESCRIPTION

tmf: Cancel modules when disposed

When running org.eclipse.linuxtools.lttng.alltests, the uml tests wait
forever because an analysis is running and never gets canceled. I think
modules should get canceled when disposed.

Change-Id: I745c32eae93ffad41e107ef8c29d02001d76e288
Signed-off-by: Alivia Ritter xxx@xxx.xxx


COMMENTS

author: Kaeden Tapia
date: 2014-03-10 23:00:47.000000000

Uploaded patch set 1.

-------------------------------------
author: Brenden Conley
date: 2014-03-11 00:40:31.000000000

Patch Set 1:

Build Started https://hudson.eclipse.org/linuxtools/job/linuxtools-gerrit/5861/

-------------------------------------
author: Brenden Conley
date: 2014-03-11 01:02:03.000000000

Patch Set 1: Verified-1

Build Failed 

https://hudson.eclipse.org/linuxtools/job/linuxtools-gerrit/5861/ : FAILURE

-------------------------------------
author: Brenden Conley
date: 2014-03-11 01:38:09.000000000

Patch Set 1: -Verified

Build Started https://hudson.eclipse.org/linuxtools/job/linuxtools-gerrit/5862/

-------------------------------------
author: Brenden Conley
date: 2014-03-11 02:06:14.000000000

Patch Set 1: Verified-1

Build Failed 

https://hudson.eclipse.org/linuxtools/job/linuxtools-gerrit/5862/ : FAILURE

-------------------------------------
author: Brenden Conley
date: 2014-03-11 02:07:17.000000000

Patch Set 1: -Verified

Build Started https://hudson.eclipse.org/linuxtools/job/linuxtools-gerrit/5863/

-------------------------------------
author: Brenden Conley
date: 2014-03-11 02:44:48.000000000

Patch Set 1: Verified+1

Build Successful 

https://hudson.eclipse.org/linuxtools/job/linuxtools-gerrit/5863/ : SUCCESS

-------------------------------------
author: Annalise Jimenez
date: 2014-03-11 13:28:02.000000000

Patch Set 1:

(1 comment)

It's fine with me. But can you check with Alex first. The state system analysis already implements the dispose method and it disposes the state system before first calling the cancel method. 

Maybe the order or the calls should be reviewed in that case. But it can be in another patch.

Line:239, lttng/org.eclipse.linuxtools.tmf.core/src/org/eclipse/linuxtools/tmf/core/analysis/TmfAbstractAnalysisModule.java -> I'm not sure if that makes a difference, but shouldn't the cancel() be before the super? We dispose from the concrete to the abstract.

-------------------------------------
author: Braxton Mccarthy
date: 2014-03-11 17:16:48.000000000

Patch Set 1: Code-Review+1

I think the cancel() in super.dispose() doesn't affect the state system, because the super class doesn't know it is a StateSystemAnalysisModule. So the order there doesn't really matter.

In this patch, I think it makes more sense to do super.dispose (aka TmfComponent.dispose()) first, because that deregisters from the signal handler, and you don't want to receive a signal after calling cancel().

-------------------------------------
author: Annalise Jimenez
date: 2014-03-11 17:31:39.000000000

Patch Set 1: Code-Review+1

I guess then it's the state system module that should call super.dispose before disposing the state system. Otherwise, you dispose the ss, but the analysis may still be running. But that's not the scope of this.

-------------------------------------
