[usocket-devel] Re: UDP support in usocket, ideas explained

[Erik Huelsmann]

On 2/18/08, Chun Tian (binghe) <binghe.lisp at gmail.com> wrote:
> >
> > On Feb 16, 2008 8:51 PM, Erik Huelsmann <ehuels at gmail.com> wrote:
> >>> OK, so if I can see some interface function (like SOCKET-RECEIVE) in
> >>> svn trunk, I'll start to learn them and write/test some backend code
> >>> as your design and commit to you, and I'll start to make my package
> >>> depends on usocket to get networking portability.
> >>
> >> I'm sorry you didn't see anything yet. The commits you've seen until
> >> today were about me cleaning up loose ends and implementing ideas
> >> floating in my head for months (as I could not commit them - my hd
> >> had
> >> crashed). I need to get rid of that mental state to be able to start
> >> the UDP work, so, that's why. I hope to be able to do some interface
> >> definition work today.
> >
> > I was going to work with CLISP to do some exploration on this front,
> > but it turns out that the default Etch CLISP package doesn't include
> > the RAWSOCK module which I'll need to implement this feature... So,
> > instead of designing/working on a prototype, I ended up trying to
> > compile my own CLISP.
>
> Debian's CLISP package is a bit strange, which didn't compile many
> module which CLISP supported.
> Maybe you should submit this to Debian as a wishlist.

I found that this wasn't really a problem in my CLISP at all. After I
found out that the (CLISP provided) Win32 package had the same
problem, I discovered the -Kfull option I should have provided at
startup. After doing so, CLISP turned out to support RAWSOCK afterall!

> The past two weeks is Spring Festival - a very big holiday for all
> Chinese, and I didn't do anything on Lisp. My current plan is more
> developments on my SNMP package, and start a new project: IPMI[1] for
> Lisp (which still depends on a UDP support for at least LispWorks).
> I'll continue watching usocket's change and do my job as we talk before.

Great! I don't really know yet how to handle UDP socket creation, but
I'll continue defining other functions with fewer "problems" asap.

Bye,


Erik.