[kepler-dev] Datasource changes

Kevin Ruland kruland at ku.edu
Fri Feb 3 14:10:03 PST 2006


Dan Higgins wrote:
> Kevin,
>    Nice job on changing the EML2DataSource to use initialize to wait 
> for data to be loaded. Now, if I click on 'Run' before data is 
> finished loading, the workflows will just wait (inside 'initialize') 
> until data is ready, rather than throwing an error. (at least the ones 
> I looked at did when run from within Vergil).
>
thanks.  I guess from the paragraph below that it almost works.  BTW - I 
have committed essentially this code.  But now the line numbers have 
changed.  I'll run ptexecute and see what I get.
>    However, I decided to test execution using the ptexecute macro. 
> [You can use the ant target(s) 'ptexecute' or 'ptexecute-dev', or 
> generate a batch file with 'build_ptexecute'.] ptexecute will run the 
> workflow 'workflows/eco/IPCC_Base_Layers.xml' OK if the data is in the 
> cache; however, if the data is NOT in the cache (i.e. needs to be 
> downloaded), I get a null pointer exception
> Caused by: java.lang.NullPointerException
>        at 
> org.ecoinformatics.seek.datasource.eml.eml2.Eml200DataSource.initiali
> ze(Eml200DataSource.java:355)
>
>    Another problem is that I cannot run multiple 'ptexecute' process 
> simultaneously nor can I run kelper while a long ptexecute process is 
> running. I suspect this is the same problem we saw before with 
> multiple Kepler processes, but I am somewhat surprized that it occurs 
> with ptexecute (This is with workflows that have datasources).
>

This is exactly the same problem as trying to run two kepler instances 
simultaneously.  Maybe we need to have an "execute" method which simply 
messages a running kepler instance to load and execute another 
workflow?  this would be the same as using "file->open" twice to get two 
different workflows running.

Kevin



More information about the Kepler-dev mailing list