StubToolkit must go!

[Richard Bair]

Or something is wrong with the tests where they aren't cleaning up after themselves or leaving some static state in a state that other test code wasn't expecting, which should also be fixed. most of our tests are of the variety "create an object, test the object", where such things shouldn't really be a problem.

Richard

On Feb 6, 2013, at 4:53 PM, Jim Graham <james.graham at oracle.com> wrote:

> 
> 
> On 2/6/13 8:30 AM, Richard Bair wrote:
>> So my thought here is that if we can run 30,000 tests in < 5 min by allowing nearly all of them to be run in the same VM, then that is probably better than running 3,000 tests in separate VMs for the sake of our pre-commit testing. For the nightly run, we can certainly run them all in their same VM so that the analysis phase of failures is much simpler.
> 
> For reliability testing running them all in the same VM would actually improve the coverage of our tests.  If there is a snowball failure, then an option to run them in separate VMs to sanity check if a single test is causing the problem would be nice, but if we can't survive running them all in the same VM then something is wrong with the product and we should definitely know that...
> 
> 			...jim