Re: [Cdk-devel] Problems in CDK trunk

[Egon W. <ego...@gm...>]

On Wednesday 27 June 2007, Thomas Kuhn wrote:
> after an update of my cdk version to the current trunk I recognized an
> increase of the number of tests which throw errors within the cdk-taverna
> project. (From 3 to 19 errors). Has anybody an idea what this increase
> caused?

Yes, the 'upgrade' of the core classes, to work with Class-es instead of 
natives. This give a lot of NullPointerExceptions, where CDK code does not 
realize that fields (like implicit hydrogen count) might be UNSET. Those bits 
of code have now to test if the field == NULL (aka UNSET) and then make a 
reasonable assumption, or deal with it otherwise.

> I found errors while using the following classes:
>
> org.openscience.cdk.qsar.descriptors.atomic.PartialTChargePEOEDescriptor
> org.openscience.cdk.qsar.descriptors.molecular.IPMolecularDescriptor
> org.openscience.cdk.qsar.descriptors.bond.ResonancePositiveChargeDescriptor
> org.openscience.cdk.qsar.descriptors.bond.BondPartialTChargeDescriptor
>
> The error messages are all nearly identical and like:
>
> ---------------------------------------------------------------------------
>---- Test set:
> org.openscience.cdk.applications.taverna.qsar.descriptors.molecular.IPMolec
>ularTest
> ---------------------------------------------------------------------------
>---- Tests run: 1, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 0.03
> sec <<< FAILURE!
> test_LocalWorkerIPMolecular(org.openscience.cdk.applications.taverna.qsar.d
>escriptors.molecular.IPMolecularTest)  Time elapsed: 0.02 sec  <<< ERROR!
> java.lang.NoSuchMethodError:
> org.openscience.cdk.interfaces.IAtom.getFormalCharge()I at

Indeed, the method getFormalCharge() that returns an 'int' no longer exists, 
and it now returns an Integer.

Egon

-- 
ego...@gm...
Blog: http://chem-bla-ics.blogspot.com/
GPG: 1024D/D6336BA6