On Wed, Jun 13, 2001 at 01:35:18AM -0400, Bill Stoddard wrote:
> I need to catch up on the other messages in this thread
Yes, please do!
The last message I posted on this thread contains even some
important info for apr_socket_t (which gets near the NAL,
but not as far). I even pointed out some problems in
apr_socket_t.
I haven't gotten a response to this posting till now (I'm
not on the list!)
> so I conceed you may be right,
>From a performance point of view, filters are probably the
right way to do this.
We (the Samba TNG people) asked for NAL, because we
currently don't need performance but _extremely clean_
design and abstraction.
> but one thing I -think- this proposal helps us avoid is
> replicating a lot of code in the core_filters (in and
> out) just to access a different network i/o primitive.
> Mucking with the core filters seems a clunky way to do
> what the NAL does.
Cool... so someone sees a use outside our goals! :)
> Bill
>
> ----- Original Message -----
> From: <[EMAIL PROTECTED]>
[...]
Elrond
