Bill,
Such was my frustration for weeks at a time last fall. Very annoying
when you have code you'd like to commit!
The truth is that the tests are:
1) single threaded (in the sense of using only one core, single-
threaded client)
2) don't cover significant areas (no AMX coverage for example).
3) don't test real-world loads
Lloyd
On Jun 15, 2007, at 2:11 PM, Bill Shannon wrote:
> After failing three nights in a row, I assumed the failures were real.
> I just ran the tests *again*, without updating my workspace, and this
> time they passed!
>
> Grumble, why aren't these tests more reliable?
>
> If the normal behavior is that sometimes these tests fail and
> sometimes
> they pass, what do we really know when they pass?
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>