Subject: re-designing libwrap's API (was: CVS commit: src)
To: None <>
From: Greg A. Woods <>
List: tech-userlevel
Date: 01/18/1999 16:37:16
[ On Mon, January 18, 1999 at 11:47:36 (-0800), Matthias Drochner wrote: ]
> Subject: CVS commit: src
> Needless to say that libwrap's interface sucks.

Badly.  On several fronts.

The patch I made to TCP Wrappers to introduce RBL support, and recently
submitted to NetBSD (PR#6837, already closed, and I think applied to
-current), would greatly benefit from being able to return more complex
information to the caller (such as the line number of the match, the
actual value that matched and the option, the DNS records associated
with the RBL entry, etc.).

Some of my concerns are alleviated due to the fact that the mailer I
intend to use with RBL domains in the future doesn't use libwrap, but
there may still be some call to cleaning up libwrap's API and extending
it for such purposes.

I sent e-mail to libwrap's author, Wietse Venema, at the time I created
the RBL patch, but I said it wasn't an high priority thing and
definitely not as important as his work on Postfix.

Give the code I've read in Postfix I think he could easily cook up a
much better replacement for TCP Wrappers in no time flat, but I still
don't know if he's interested in doing so.  I'm pretty sure I could
borrow from the code in Postfix and libwrap and come up with something
pretty good too.  That kind of "derivative" might even qualify as being
a unique new work under most copyright laws....  ;-)

I guess the reason for this message is to find out if anyone else thinks
this would be a high-demand thing or whether we can just live with what
we've got.

							Greg A. Woods

+1 416 218-0098      VE3TCP      <>      <robohack!woods>
Planix, Inc. <>; Secrets of the Weird <>