[kepler-users] Problem with multiple plays of same KAR.

[Tom Parris]

Dan,

Fix confirmed.  Thanks!

-- Tom

On Tue, Jan 15, 2013 at 12:34 PM, Daniel Crawl <danielcrawl at gmail.com> wrote:
>
> Hi Tom,
>
> I fixed the problem.
>
>   --dan
>
>
>
> On 1/9/13 1:33 PM, Tom Parris wrote:
>>
>> Dan,
>>
>> Yes.  That's what we are using and that is the observed behavior.  I just
>> assumed the problem originated with our own code (which is generally a
>> safe
>> assumption).
>>
>> We would be most grateful if you could look into.  But please realize this
>> is not currently a show stopper for us, just an annoyance.
>>
>> -- Tom
>>
>> On Wed, Jan 9, 2013 at 4:23 PM, Daniel Crawl <danielcrawl at gmail.com>
>> wrote:
>>
>>>
>>> Hi Tom,
>>>
>>> Are you using the MatrixViewer actor to display the matrix tokens?
>>>
>>> I recently reported a similar problem with this actor:
>>>
>>> The MatrixViewer actor pops up a window displaying a matrix. If the
>>> window is closed, it does not reappear after running the model again
>>> (unlike the Display or plotting actors).
>>>
>>>
>>> http://bugzilla.**ecoinformatics.org/show_bug.**cgi?id=5759<http://bugzilla.ecoinformatics.org/show_bug.cgi?id=5759>
>>>
>>>
>>> I can look into fixing MatrixViewer.
>>>
>>>    --dan
>>>
>>>
>>>
>>> On 1/9/13 9:27 AM, Tom Parris wrote:
>>>
>>>> Dear Kepler-Users:
>>>>
>>>> I am hoping you can help point me in the right direction to track down
>>>> an
>>>> odd problem we are having (with Kepler trunk on Windows 7 x64 pro, but
>>>> has
>>>> been happening for a couple of years through multiple versions of the
>>>> trunk
>>>> and on x86 XP systems as well)
>>>>
>>>> 1. It involves a simple test KAR for a self-developed actor that uses
>>>> GDAL
>>>> to read a matrix and produce two tokens (the matrix and the metadata for
>>>> the matrix).  I know it is customary to provide an example for
>>>> replication
>>>> purposes and I am willing to do so, but it is non trivial to get it
>>>> installed correctly.  I'm also willing to share the code of the actor
>>>> for
>>>> review (with the usual caveats that we know it is not pretty).
>>>>
>>>> 2. The KAR opens and runs once without difficulty.  Three display
>>>> windows
>>>> are generated.  One with the name of the input file, one with the
>>>> metadata,
>>>> and one with the matrix.  If one closes these three windows and presses
>>>> the
>>>> "play" button a second time, only the first two windows are produced.
>>>> The
>>>> display window with the matrix is not produced.
>>>>
>>>> 3. Closing Kepler and relaunching solves the problem.
>>>>
>>>> 4. After pressing "play" a second time, closing the KAR and re-opening
>>>> fails (due to an issue resolving a simple jython class and instance for
>>>> an
>>>> actor that subsets the matrix).
>>>>
>>>> 5. If instead of pressing "play" a second time, I close the KAR and
>>>> re-open
>>>> it, then it opens successfully and runs once, but not a second time.
>>>>
>>>> 6. We have not tracked this problem down previously because in "real"
>>>> workflows we don't seem to encounter this problem (though we typically
>>>> only
>>>> run once per open).  These workflows typically make extensive use of
>>>> this
>>>> actor, with multiple instances and multiple firings per workflow.
>>>>
>>>> I suspect this is a problem in our code (either the GDAL actor or the
>>>> jython actor) understanding differences in the roles of initalize(),
>>>> prefire(), and postfire().
>>>>
>>>> One oddity is that self developed GDAL actor outputs the metadata token
>>>> after it outputs the matrix token.  So, it is odd that the metadata
>>>> token
>>>> is displayed on multiple runs, but not the matrix token.
>>>>
>>>> -- Tom
>>>>
>>>>
>>>>
>>>> ______________________________**_________________
>>>> Kepler-users mailing list
>>>> Kepler-users at kepler-project.**org <Kepler-users at kepler-project.org>
>>>>
>>>> http://lists.nceas.ucsb.edu/**kepler/mailman/listinfo/**kepler-users<http://lists.nceas.ucsb.edu/kepler/mailman/listinfo/kepler-users>
>>>>
>>>>
>>>
>>
>