Am 19.07.2012 09:26, schrieb ArtemGr:
> Hi, Tommi!
>
> Is there a way to extend the TCP/IP socket lifetime past the
> Worker::processRequest in order to send the response later
> while not keeping the Tntnet thread pinned?
>
> If I'm not mistaken, the socket which is left after the 
> HttpReply::setDirectMode
> has a chance of not being destructed immediately
> because Tntnet would wait for further KeepAlive requests,
> but that doesn't seem reliable.
> Would it be hard to implement a method in Tntnet which takes the socket
> out of the Tntnet loop?
>
>
>
Hi Artem,

it is not just a small change to implement that. It needs really 
rethinking, how the main loop works.

Another feature request is spdy, which is a extension to http from 
google. This also needs rethinking.

So we have now 2 reasons to do it.

I feel, that it is time to make a larger change. I have already thought 
a little about replacing the whole event handling with the cxxtools 
event handling, so that tntnet can run in the same event loop like the 
rpc servers in cxxtools. But this is really a larger project, which 
needs some time.

The goal is to keep tntnet source compatible, so that no user changes 
are needed.


Tommi

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Tntnet-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/tntnet-general

Reply via email to