[concurrency-interest] Racy lazy initialization: blush test

thurston at nomagicsoftware.com thurston at nomagicsoftware.com
Fri Feb 22 13:57:22 EST 2013

Changing the example somewhat (just excluding registers):

if (0 == x.f)
     x.f = 1

To my mind, this code contains a data race even though the "result" 
(not precisely sure what that means frankly) is the same, i.e. it is 
"racy" even if not "buggy"

More information about the Concurrency-interest mailing list