[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Chicken-users] chickenlib
From: |
tonyg |
Subject: |
Re: [Chicken-users] chickenlib |
Date: |
Fri, 9 Aug 2002 16:23:02 +0100 |
User-agent: |
Mutt/1.2.5i |
Hi...
> 1) I think "bind-unix", "bind-inet" etc. are mistakes. In the manner
> of Scsh's interface, things that map more straightforwardly into
> the underlying Unix calls are likely of use. See:
> http://www.scsh.net/docu/html/man-Z-H-5.html#%_chap_4
> I think that some of the design choices in the Scsh interface are a
> mistake, but having a straightforward bind etc. is not.
You're right. Okay, then: users of socket.scm are hereby warned that
it is about to be gutted and replaced :-) and to use plt-net.scm if
they'd rather have a stable interface!
> 2) It would be good to have abstract types that encapsulate a sockaddr
> style structure and functions to manipulate that. Similarly for
> things we might dub "inaddr4", "inaddr6", and "localaddr" (though
> perhaps we don't need the last as it is really just a string.)
> By doing this, we can handle the range of available calls more
> nicely. Scsh goes some of the way in this direction but not far
> enough.
Right. Okay.
Maybe I'll change it to have (define-record)s for sockaddr_in- and
sockaddr_un-equivalents, with conversion procedures
(inaddr4->sockaddr), (unix-addr->sockaddr) etc. to convert to a
bytevector suitable for use with a lowlevel call like bind?
I'll think about this some more, and with it a bit over the weekend.
> 3) I think the names should probably be simply the names of the
> underlying Unix system calls, which is mostly what you do, rather
> than the altered names in scsh.
Okay. Good idea not to gratuitously rename things I suppose.
Thanks for the suggestions!
Tony
--
Monkeys high on math -- some of the best comedy on earth
- Tom Lord, regarding comp.lang.scheme