>
>> @Daniel:
>>
>> What do you think about it?
>>
>
> If you want a replacement for accept() then I think such a callback makes
> sense. If you want to just do a setsockopt() on the new socket then I think
> the CURL_SOCKOPTFUNCTION is the correct callback and it allows fairly easy
> extension with its 'purpose' option. I certainly wouldn't mind seeing that
> it gets used after accept() has returned successfully.
>
>
"replacement for accept()" sounds like a good idea but in the second
thought, I do not find any use case in which an application would need to
implement this callback. On the other hand, applications may need to set
socket options before data transfer kicks off. I am going to prepare a
patch for CURL_SOCKOPTFUNCTION after accept() gets called and returned
success.

@Joe: Does this change make it for you?

-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette: http://curl.haxx.se/mail/etiquette.html
Received on 2012-06-24