Re: [boost] Feature Request
At 22:40 2005-11-03, Abajyan, Hambardzum wrote:
Hi (I've already been subscribed my address as you told me to do.)
I would like to now why there isn't an implementation of TCP(UDP) Sockets ? I think it's possible to write a good platform-independent architecture based on Berkeley socket interface (supporting asynchronous (non-blocking) sockets), because both Linux and Windows systems support that architecture. There are many libraries like that: e.g. Boost.Thread library which is also platform-independent. In my opinion, this feature is very usefull as with it programmers can develop network applications on Boost. Thanks
I like the approach these guys are taking: http://www.dtilib.dk/ that actually would be 3 libraries, but I like them all.
Best Regards Hambik
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Victor A. Wagner Jr. http://rudbek.com The five most dangerous words in the English language: "There oughta be a law"
On 11/4/05, Victor A. Wagner Jr. <vawjr@rudbek.com> wrote:
At 22:40 2005-11-03, Abajyan, Hambardzum wrote:
I would like to now why there isn't an implementation of TCP(UDP)
Sockets ?
A very good one (ASIO) is in the review queue, so sit tight! I like the approach these guys are taking: http://www.dtilib.dk/
that actually would be 3 libraries, but I like them all.
This project includes a C++ socket library, a library for doing async I/O operations with IOStreams and another for database access. The designs appear to be clean and simple. The socket library doesn't seem to support non-blocking sockets or any sort of I/O multiplexing on its own, so it is not nearly as complete as ASIO. As far as sockets are concerned, layering on asynchronous operations at the IOStreams level is a mistake IMHO. These concepts need to be supported at the lowest levels of the library because some applications require them. Looking forward to the ASIO review, -- Caleb Epstein caleb dot epstein at gmail dot com
On 11/4/05, Caleb Epstein <caleb.epstein@gmail.com> wrote:
On 11/4/05, Victor A. Wagner Jr. <vawjr@rudbek.com> wrote: I like the approach these guys are taking: http://www.dtilib.dk/
that actually would be 3 libraries, but I like them all.
This project includes a C++ socket library, a library for doing async I/O operations with IOStreams and another for database access. The designs appear to be clean and simple.
The socket library doesn't seem to support non-blocking sockets or any sort of I/O multiplexing on its own, so it is not nearly as complete as ASIO. As far as sockets are concerned, layering on asynchronous operations at the IOStreams level is a mistake IMHO. These concepts need to be supported at the lowest levels of the library because some applications require them.
Agree with all these comments. I came across dtilib via the article in the October issue of CUJ. I think ASIO provides the full-blown async solution but there could be a need for an iostreams network solution (but not async?) which main benefit would be the ease of use (not performance). One thread today proposed such library.
participants (3)
-
Caleb Epstein -
Jose -
Victor A. Wagner Jr.