[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 #30 from Aaron <aschultz at nceas.ucsb.edu> 2011-03-30 19:51:29 PDT ---
Continuing to make progress on this bug.

There are 3 levels I am working at, the ptolemy level, the kepler core level
and the sensor-view level.  The ultimate goal is to free up as many resources
as possible whenever windows are closed so memory usage does not build up over
time as windows are repeatedly opened and closed.  

Unfortunately it is difficult to declare that a class is free from leaks since
it can easily depend on other classes in different situations.  For example, I
could say KeplerGraphFrame was free from leaks because when you open a new
workflow then immediately close it all the memory for the frame gets garbage
collected.  However, in a more complex scenario, if I open a KeplerGraphFrame
and build a workflow containing a Display actor, run it, then close the window,
KeplerGraphFrame will not be garbage collected, so the Display actor leak
causes KeplerGraphFrame to leak (and the leak in the Display actor may in turn
come from a leak in the TextEditor class, and so on).  So my point is that
characterizing leaks can be difficult (not to mention there are over 300 actors
in our library that can contribute to leaky graph frames).

Memory improvements have been made on leaks found in TextEditor, Display actor,
WindowPropertiesAttribute, ActorGraphFrame (and all super classes),
EditParametersDialog, FigureAction
At the Kepler level improvements have been made in KeplerGraphFrame
At sensor-view level improvements have been made in WorkflowManager

As leaks are discovered and addressed at the lower levels it becomes easier to
identify and address leaks at the higher levels.

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