[kepler-dev] [Bug 5095] test kepler and wrp for memory leaks

[bugzilla-daemon at ecoinformatics.org]

http://bugzilla.ecoinformatics.org/show_bug.cgi?id=5095

--- Comment #22 from jianwu <jianwu at sdsc.edu> 2011-01-06 16:37:39 PST ---
My tests show Kepler has memory leaks for "workflow open, execution and close"
and "workflow open, edit and close". More detailed information can be found at 
https://kepler-project.org/developers/teams/build/dealing-with-memory-leak-in-kepler

>From the tests, I doubt some memory leaks might be from Ptolemy also. So I
tested Ptolemy itself today. Here is what I did:

1) Start Ptolemy
2) Open a new 'Graph Editor', whose title is "Unamed".
3) Open a sample workflow (constant actor + display actor in SDF)
4) Run the sample workflow
4) Close the window (now only the "Unamed" window left)

>From the jprofiler (other tools should also be good to see it), I can see there
is one instance for class "ptolemy.actor.lib.Const" and 
"ptolemy.actor.lib.gui.Display" after garbage collection. Clearly the objects
for these two actors are not released. I also saw 4 instances for
"ptolemy.gui.Top$1", which should be 2 if there is one window left, and 8
instances for "ptolemy.kernel.util.Workspace". 

>From "Heap Walker" in JProfiler, I didn't see obvious wrong references in the
GC path of "ptolemy.actor.lib.Const" and other instances. 

Christopher Brooks, do you want to verify the phenomenon?

-- 
Configure bugmail: http://bugzilla.ecoinformatics.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.