river-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Peter Firmstone <j...@zeus.net.au>
Subject Re: Build failed in Hudson: River-trunk #163
Date Fri, 24 Apr 2009 08:06:22 GMT
It appears to me that I should perhaps be writing jtreg style unit tests 
for ClassDep rather than JUnit?

3 different test toolkits might be going overboard, especially since 
there appears to be a good deal of overlap between jtreg and junit.  I 
don't want to get into a philosophical idealistic comparison between the 
two tools as I personally prefer junit; only because I'm familiar with 
it, it just seems like the most commonsense approach.

Cheers,

Peter Firmstone.

Peter Jones wrote:
> On Wed, Apr 22, 2009 at 08:37:58PM +0200, Jonathan Costers wrote:
>   
>> Waw, thanks a lot for your explanation, Peter. It clears up a lot of
>> things for me :-)
>>     
>
> You're welcome.
>
>   
>> So, do you think it would be useful to enable these tests in River?
>>     
>
> I think that they provide valuable coverage.
>
>   
>> Are they overlapping with the QA suite?
>>     
>
> I don't think that there's much overlap.
>
>   
>> Should we choose one or the other framework? Or should we keep both?
>>     
>
> Well, jtreg is in no position to handle the main Jini QA suite, so it
> wouldn't be the one chosen.  As mentioned, the jtreg model is simple
> enough that it's conceivable that support for it could be added to
> another harness, like the Jini QA harness.  But if the OpenJDK's jtreg
> implementation can be used where necessary, doing that might not seem
> worth the effort.
>
> Another option of course would be to port these jtreg-style tests to
> some other test framework, but that might require a good deal of
> effort (and for regression tests, some risk whether the regression
> condition is still being effectively tested).  I should disclose some
> bias: as a JDK old-timer, I am fond of jtreg, for what it's good at.
>
> -- Peter
>
>   


Mime
View raw message