Nick Sieger wrote:
I tried to improve the test by introducing a countdown latch like
Mental suggested on IRC, but still no go, at least on Solaris where CI
is running. See the latest version of the test here:
http://pastie.caboo.se/110419
The issue is that the thread raise line 19 is being executed before
the server goes fully into #accept on line 9.
Yeah, I ran into this on other tests. The problem is that there's no
real way to determine whether a given thread is blocking on IO (that I
know of) so we can't ever be sure the thread has called accept.
I'd love to have a good solution.
- Charlie
---------------------------------------------------------------------
To unsubscribe from this list please visit:
http://xircles.codehaus.org/manage_email
