Daniel P. Berrange wrote:
> 2) Are ifdef an acceptable solution to epoll avaliability
issues?
There are quite a large number of ifdefs in your patch below.
I think it could be worth having a separate vireventepoll.c
for the epoll impl. Then choose to compile vireventpoll.c vs
vireventepoll.c at configure time. We could have vireventcommon.c
if there is some portion of code that can be shared.
If we're going this route, would it make sense to use an existing
abstraction layer like libevent?
(Asking out of curiosity, not from experience)
Roman Bogorodskiy