I added a null-catcher to the Form 6 deserializer code - that particular value
should never be null; it prints out the typecode that can't be converted into a
type.  -M


On 10/9/2017 10:35 AM, Marshall Schor wrote:
> looking at this stack trace, NPE on line 1839 implies tgtType is null.
> Looking up at previous lines, the initPrevIntValue(tgtType) would have thrown 
> a
> NPE if it was executed, so therefore "storeIt" must be false.
> Which means srcType == null.
> If lenient is being used, this probably means the src and target (the
> deserialized form) type systems are different.
>
> So, as far as I can tell, this means that there's a type code in the form 
> being
> deserialized which is a number not in the target type.
> Because this is unlikely, it's more likely something else went amiss, and
> there's some garbage being read for the type code.
>
> I think this will need a test case - can you provide one?
>
> -Marshall
>
>
> On 10/3/2017 6:14 PM, Richard Eckart de Castilho wrote:
>> On 02.10.2017, at 19:33, Marshall Schor <m...@schor.com> wrote:
>>> this should now be fixed.  Can you update to head and retest?  Thanks! 
>>> -Marshall
>> From DKPro Core: a considerable number of tests have been fixed. Several 
>> ones need to be updated again due to changes in FS.toString(). However, it 
>> seems there are still some issues. Before looking into these, I want to fix 
>> the more trivial things first though.
>>
>> That said, here is an example of one of the potentially remaining issues: a 
>> stacktrace (completely undiagnosed) that I still get when deserializing a 
>> form 6 binary CAS in lenient mode:
>>
>> java.lang.NullPointerException
>>      at 
>> org.apache.uima.cas.impl.BinaryCasSerDes6.deserializeAfterVersion(BinaryCasSerDes6.java:1839)
>>      at 
>> org.apache.uima.cas.impl.BinaryCasSerDes.reinit(BinaryCasSerDes.java:594)
>>      at org.apache.uima.util.CasIOUtils.load(CasIOUtils.java:382)
>>      at org.apache.uima.util.CasIOUtils.load(CasIOUtils.java:344)
>>      at 
>> de.tudarmstadt.ukp.dkpro.core.io.bincas.BinaryCasReader.getNext(BinaryCasReader.java:194)
>>      at 
>> de.tudarmstadt.ukp.dkpro.core.io.bincas.BinaryCasWriterReaderTest.read(BinaryCasWriterReaderTest.java:521)
>>      at 
>> de.tudarmstadt.ukp.dkpro.core.io.bincas.BinaryCasWriterReaderTest.test6Lenient(BinaryCasWriterReaderTest.java:160)
>>
>> Cheers,
>>
>> -- Richard
>

Reply via email to