[bind10-dev] Trying to shoe-horn truncate (trac #499)

Scott Mann smann at isc.org
Sun Jan 30 20:30:00 UTC 2011


Actually, I failed to mention that the intermediate option is to just 
abstract UDPQuery into IOQuery...it would be "heavier" than if IOSocket 
were completed. But, I think it can be done without IOSocket. Maybe this 
is the way to go for now.


On 01/30/2011 10:34 AM, Scott Mann wrote:
> Currenty, a UDPQuery object is used to send and receive dns messages 
> via the resolver. While this was probably expedient at the time it was 
> written, it is a problem now.
>
> I think that the correct way to solve this problem is to complete the 
> IOSocket abstraction layer (see src/lib/asiolink/iosocket.cc) and then 
> abstract UDPQuery to IOQuery (or pick a better name) which would 
> handle either UDP or TCP communications. Another, less desirable 
> approach, would be to create a separate TCPQuery class that would 
> effectively behave like UDPQuery but for TCP. The problem here is that 
> much of the code would be duplicated (callbacks, send functions, etc) 
> and there might also be some additional code necessary for this 
> approach (that wouldn't otherwise be needed).
>
> In an effort to try to find a short-cut, I've spent a fair bit of time 
> trying to massage UDPQuery into something that could handle the TCP 
> communication and I've run into various problems. All of which has 
> left me wondering if we ought not just fix IOSocket. I think this 
> would likely add a task or two. At this point, I am unsure of the 
> impact that would have on attaining our 1st year goals on the other 
> hand, I am also concerned about the amount of refactoring that will 
> become necessary beginning in April.
>
> Thoughts or suggestions?
>
> Thank you.
> -Scott
> _______________________________________________
> bind10-dev mailing list
> bind10-dev at lists.isc.org
> https://lists.isc.org/mailman/listinfo/bind10-dev




More information about the bind10-dev mailing list