[infinispan-dev] Raison de etre of TestObjectStreamMarshaller

[Galder Zamarreno]

Manik Surtani wrote:
> 
> On 4 Jun 2009, at 08:40, Galder Zamarreno wrote:
> 
>>
>>
>> Manik Surtani wrote:
>>> Yes, it was for ease of debugging, and also to isolate the cause of 
>>> problems.  I think we should stick with the Test marshaller for such 
>>> purposes, even if purely for the sake of component isolation during 
>>> testing.
>>
>> Ok.
>>
>>> As for the coverage issues re: the VAM, there should be a separate 
>>> set of unit tests for the VAM to ensure every known type is properly 
>>> marshalled and unmarshalled by the VAM.
>>
>> I built a similar test for JBoss Marshaller called 
>> JBossMarshallerTest. It know uses VAM rather than JBossMarshaller, so 
>> that's effectively it. I need to double check whether all types are 
>> now covered and I'll refactor it to better show it's job now.
> 
> Great!

I added an extra set of tests for each cache store where VAM is used as 
marshaller and it was well worth it. As a result of this, I found and 
resolved 3 bugs and by having separate tests that use VAM, we can 
isolated issues related to VAM vs issues related to the cache store 
implementation more easily.

> 
> Cheers
> Manik
> 
> -- 
> Manik Surtani
> manik at jboss.org
> Lead, Infinispan
> Lead, JBoss Cache
> http://www.infinispan.org
> http://www.jbosscache.org
> 
> 
> 
> 

-- 
Galder Zamarreño
Sr. Software Engineer
Infinispan, JBoss Cache