@cmuratori @sssmcgrath Like for a long time I didn't know it worked with sockets, until I found out you could cast sockets to handles.
-
-
Replying to @cmuratori
@cmuratori it's underlying socket implementation is basically the same as bsd's kqueue, but it applies to files properly, kqueue doesn't =(2 replies 0 retweets 1 like -
Replying to @sssmcgrath
@sssmcgrath All of those systems _were based on IOCP_. Unix didn't have shit for this until after IOCP, and then sun cloned it, IIRC.2 replies 0 retweets 0 likes -
Replying to @cmuratori
@cmuratori I wasn't sure who had it first, sunos and aix had it around the same time, then free bsd, but that's cool if it was msft first3 replies 0 retweets 0 likes -
Replying to @sssmcgrath
@sssmcgrath Yep, it was Cutler and Co. It was patented, believe it or not, but MSFT happily never sued. Patent's expired now, I believe...1 reply 0 retweets 0 likes -
Replying to @cmuratori
@cmuratori ahh.. I wonder if that's why linux went with epoll: to avoid patents.. sad =(1 reply 0 retweets 1 like -
Replying to @sssmcgrath
@sssmcgrath Doubtful, since epoll would presumably be considered to violate the patent, assuming the patent was even found valid.1 reply 0 retweets 0 likes -
Replying to @cmuratori
@cmuratori@sssmcgrath epoll is just a really shitty version of IOCP :)3 replies 1 retweet 1 like -
Replying to @cmuratori
@cmuratori@sssmcgrath Not really. IOCP tells you X is done; epoll just tells you X is _ready_ to _be_done_. Async I/O on Linux is dumb.1 reply 0 retweets 0 likes -
Replying to @jwatte
@jwatte@sssmcgrath Technically IOCP tells you either. For example, if you recv with a 0 buffer, IOCP will tell you when you _could_ recv.1 reply 0 retweets 1 like
@cmuratori @jwatte @sssmcgrath This is why I say epoll is a really shitty version of IOCP :) That said, I haven't worked much with it.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.