On 18/04/12 03:39, Steven Chamberlain wrote:
> Strange, the workaround hasn't really changed anything for me on
> kfreebsd-i386
Sorry that was wrong, I had dry-run the patch without actually applying
it :(
With it, I hit a tst-timer5 regression during build.
But if I keep the timeout at its default 2000ms and:
On 18/04/12 07:02, Petr Salinger wrote:
- if (terminated_children || main_thread_exiting) {
+ if (1 || terminated_children || main_thread_exiting) {
Then I see 100/100 passes of test_socket. That used to hang always.
There are still occasional hangs in test_threading (about 20% of the
time) and also very rarely in test_signal (seen 1% of the time).
And in the ruby1.9.1 test suite, test_io looks like it might be fixed by
this, but test_thread still hangs sometimes.
Regards,
--
Steven Chamberlain
[email protected]
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
