Closed gcmalloc closed 7 years ago
So, this fixes a test failure where iptables complains that it doesn't have the lock. However, we're supposedly doing the locking ourselves. Doesn't this just mean there is a bug in go-iptables' locking code, or am I totally confused?
according to the changelog, iptables 1.6.0 replaced the old unix socket locked by a proper flock.
iptables: use flock() instead of abstract unix sockets
So the locking of go-iptables in the test is redundant with the one done by iptables.
Reworked and resubmitted as #38 - thanks!
mostly based on #24 with addition of a flag for version < 1.6.0 and modification for the test matrix