On Feb 17, 2014, at 8:28 AM, Gustavo Carneiro <[email protected]> wrote:

> I wish this sort of thing were a bit more intuitive in a future revision of 
> tulip.  A synchronous close() would be helpful, I think.

Actually I think this is great.  This is reflecting a very real feature of 
networks in general: unless you receive acknowledgement from the remote 
application that your information was received, you really have no idea if the 
data was even transmitted all the way, let a lone processed.

Making it as simple as "synchronous close()" is basically facilitating a newbie 
mistake; repeating the error of HTTP/0.9's 
the-body-is-over-when-the-connection-is-closed non-deterministic framing.  You 
still need to be able to flush the transport somehow for tuning reasons, but 
it's good that people trip over this early.

-glyph

Attachment: smime.p7s
Description: S/MIME cryptographic signature

Reply via email to