[kepler-dev] updating Manager

Daniel Crawl crawl at sdsc.edu
Wed Apr 14 14:59:29 PDT 2010


Hi Christopher,

Thanks for folding the updates into Ptolemy. The name
changes sound good. A test case to illustrate the problem
would be nice, but I'll have to think more about how to
create one.

   --dan


On 4/14/10 12:36 PM, Christopher Brooks wrote:
> Thanks, I updated the javadocs.
>
> Is it ok if I change getThrowableExecutionId() to getExecutionId()?
> The reason is because we have setExecutionId() and the getter and
> setter should match.  Also, I'm not that excited about "Id", so
> I changed it to Identifier.
>
> I'd be happier if I had a test case that illustrates the problem,
> but understand why fabricating one would be difficult.
> http://bugzilla.ecoinformatics.org/show_bug.cgi?id=4230
> sums up the problem.
>
> I checked in an updated version to the ptII head and the
> rel-8-0-beta-2 branch.  I also updated the
> ptolemy*/module-info/revision.txt files.
>
> _Christopher
>
>
>
>
>
> On 4/14/10 11:23 AM, Daniel Crawl wrote:
>>
>> Hi Christopher,
>>
>> Thanks for taking a look.
>>
>> An executionID is an object given to the Manager via
>> setExecutionId(), and cleared during preinitialize.
>> If the Manager notifies a listener of a model error,
>> getExecutionId() can be used to map that error back
>> to the executionID object.
>>
>> --dan
>>
>>
>> On 4/14/10 7:17 AM, Christopher Brooks wrote:
>>> Hi Dan,
>>> Looks good. Can you give me a definition of an executionID
>>> that makes sense in the context of Ptolemy? In otherwords,
>>> if I'm a Ptolemy developer, and I stumble across
>>> setExecutionId(), then what is an executionID? How
>>> could I use it?
>>>
>>> Thanks,
>>>
>>> _Christopher
>>>
>>> On 4/13/10 6:36 PM, Daniel Crawl wrote:
>>>>
>>>> Hi Christopher,
>>>>
>>>> I'm working on bug 4230. The issue is that the Manager notifies
>>>> listeners of model errors asynchronously to avoid deadlocks.
>>>> However, this makes it impossible for a listener to associate
>>>> a specific execution cycle with a given error.
>>>>
>>>> The attached patch updates the Manager so that listeners can
>>>> make this association. I'd appreciate it if you could take a
>>>> look for any obvious problems. I've run the ptolemy test suite
>>>> and no new failures occurred.
>>>>
>>>> Thanks,
>>>>
>>>> --dan
>>>
>>
>



More information about the Kepler-dev mailing list