[Squeakfoundation] Re: Shrinking alpha image (was Re: Proposal to get to the triad)

[Daniel Vainsencher]

[Tests are already failing, is it relevant to harvesting?]
No. Solutions are harvested, not problems. BUT - tests are supposed to
be relevant earlier in the process - to what people work on, and submit
fixes for, and claim is [su] compliant. But to help with that, the tests
have to be more visible.

[Testing server would add visibility and help us USE our measures]
I agree. I think such a server would be a great opportunity. Note that
one you have the technology of a server that brings up images,
configures them, and runs tests, you can also use it to do many other
things:
- SLint reports on the image, or other metrics
- running graphs of results over updates.

Anyway, something that would be able to just:
- Get a latest image
- Load updates
- load test packages
- run them
- email results to list

Would be great. I think a first iteration could be in the form of an SM
package, that once loaded, performs the rest of the process. It could be
platform specific and use OSProcess to start the other image. The first
release needn't even be completely automatic - it could helps automate
this process for a person that runs it manually every week, and adds
comments on the results.

This would already make the tests much more visible.

Daniel

Marcus Denker <marcus at ira.uka.de> wrote:
> Another Question is what to do about failing Testcases... the baseimage
> testing package allready accumulated 16(!) failing tests. 
> 
> Will this be of any relevance to the release process?
> 
> Somehow I have the impression that a central testing server would help
> a lot: If we'd get an email every week (or even day), we would have
> much more pressure to actually harvest the existing bugfixes.
> 
> 
>        Marcus
> 
> 
> -- 
> Marcus Denker marcus at ira.uka.de  -- Squeak! http://squeak.de
> 
> _______________________________________________
> Squeakfoundation mailing list
> Squeakfoundation at lists.squeakfoundation.org
> http://lists.squeakfoundation.org/listinfo/squeakfoundation