(retracted by #16258)
authorJarkko Hietaniemi <jhi@iki.fi>
Sun, 28 Apr 2002 16:13:34 +0000 (16:13 +0000)
committerJarkko Hietaniemi <jhi@iki.fi>
Sun, 28 Apr 2002 16:13:34 +0000 (16:13 +0000)
commitf1f3224a6db45e106d059fb8564b0c9f6dc649b4
tree8023452237f7774b7a020da0e7ae061fb9897b70
parentcd5e0f48cb0779f625af5d8146ee1d6c56f31dcb
(retracted by #16258)

Fix thread tests not to depend on sleep() as
a scheduling aid.  In two tests (basic and list)
I had to change from sleep() hack to another hack...
basically, using the filesystem as a semaphore.
The assumption made is that rmdir() is atomic.
(The sleep() scheduling assumption broke with the
GNU pth in NetBSD.)

(the cond.t part retracted by #16250)

p4raw-id: //depot/perl@16249
ext/threads/shared/t/cond.t
ext/threads/t/basic.t
ext/threads/t/list.t
ext/threads/t/thread.t