[kepler-dev] I broke Eml200DataSource.

[Kevin Ruland]

Dan,

Thanks.  I was going to build one myself but this will be easier.  I saw 
some obvious problem in the eml200DataSource actor including the dreaded 
busy-wait.  The obvious synchronization points result in deadlocks.  I 
will look at this very carefully and have it fixed shortly.

BTW - does this workflow have two actors using the same datasets?  That 
is another case which we must handle as well.

Kevin


Dan Higgins wrote:

> Kevin,
>     Another workflow to check is " IPCC_BASE_Layers.xml" (in 
> workflows/eco; #16 on the ENM_Workflows html page). This workflow has 
> 10 EMLDataSources. Jing and I spent a lot of time earlier this week 
> getting it to run (prior to your last set of changes) and there were 
> apparently a number of thread issues with 10 datasources opening in 
> separate threads. Jing got it working both from the cache and from 
> Ecogrid, but I haven't tried your new code (since I have been 
> preparing for a training session next week in Albuquerque). So if 
> there are thread problems, they will probably appear in this workflow!
>
> Dan
>
> Dan
>
> Kevin Ruland wrote:
>
>> I have this mostly working again.  The GDAL_h1K_NS.xml workflow works 
>> most of the time.  It will always work from cache. There are some 
>> significant threading issues in this class currently.  I am going to 
>> take a close look at them now.
>>
>> Kevin
>>
>> Kevin Ruland wrote:
>>
>>  
>>
>>> Hi all,
>>>
>>> In the recent commits, I ended up breaking the Eml200DataSource 
>>> actor.  Unfortunately, I have not tracked down the problem just 
>>> yet.  I expect  to have it repaired by tomorrow morning.
>>>
>>> Sorry for any inconvience.
>>>
>>> Kevin
>>> _______________________________________________
>>> Kepler-dev mailing list
>>> Kepler-dev at ecoinformatics.org
>>> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/kepler-dev 
>>>
>>>
>>>
>>>   
>>
>>
>> _______________________________________________
>> Kepler-dev mailing list
>> Kepler-dev at ecoinformatics.org
>> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/kepler-dev
>>  
>>
>
>