From cvs-return-2842-apmail-apr-cvs-archive=apr.apache.org@apr.apache.org Thu Jan 31 16:02:42 2002
Return-Path: <cvs-return-2842-apmail-apr-cvs-archive=apr.apache.org@apr.apache.org>
Delivered-To: apmail-apr-cvs-archive@apr.apache.org
Received: (qmail 69230 invoked by uid 500); 31 Jan 2002 16:02:42 -0000
Mailing-List: contact cvs-help@apr.apache.org; run by ezmlm
Precedence: bulk
List-Post: <mailto:cvs@apr.apache.org>
List-Help: <mailto:cvs-help@apr.apache.org>
List-Unsubscribe: <mailto:cvs-unsubscribe@apr.apache.org>
List-Subscribe: <mailto:cvs-subscribe@apr.apache.org>
Reply-To: dev@apr.apache.org
Delivered-To: mailing list cvs@apr.apache.org
Received: (qmail 69219 invoked from network); 31 Jan 2002 16:02:41 -0000
X-Authentication-Warning: rdu163-40-092.nc.rr.com: trawick set sender to trawick@attglobal.net using -f
Sender: trawick@rdu163-40-092.nc.rr.com
To: dev@apr.apache.org
Cc: apr-cvs@apache.org
Subject: Re: cvs commit: apr/network_io/unix sa_common.c
References: <20020131132132.57067.qmail@icarus.apache.org>
From: Jeff Trawick <trawick@attglobal.net>
Date: 31 Jan 2002 11:00:43 -0500
In-Reply-To: <20020131132132.57067.qmail@icarus.apache.org>
Message-ID: <m3n0yuh5n8.fsf@rdu163-40-092.nc.rr.com>
Lines: 34
User-Agent: Gnus/5.0808 (Gnus v5.8.8) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Spam-Rating: daedalus.apache.org 1.6.2 0/1000/N

martin@apache.org writes:

> martin      02/01/31 05:21:32
> 
>   Modified:    network_io/unix sa_common.c
>   Log:
>   Be a bit more conservative regarding all those broken implementations out
>   there. Not everyone has FreeBSD or Linux.

broken?  whatever...  AFAICT, FreeBSD is "good" and Linux (or at least
some very popular levels of glibc) is "bad"

The unfortunate fact is that getnameinfo() and getaddrinfo() are in
general problematic right now when you look at the install base of
glibc, AIX, Solaris, et al.  Things like APR don't like the
getaddrinfo() on older AIX.  Things like lynx don't like the
getaddrinfo() on Solaris.  The informational RFC unfortunately
neglected to specify some of the behaviors that apps look for (and
sometimes don't find).  Spelling out how getaddrinfo() reports such
problems is just one example of that.

>   However, and because h_errno has been zeroed out before the call,
>   I added another test for nonzero h_errno. Everone happy?

It looks okay to me.

When there is some breathing room, I'll pursue some autodetection of
how getnameinfo() reports feedback so that clumsy run-time checks like
I had originally and like you restored are not necessary.

-- 
Jeff Trawick | trawick@attglobal.net | PGP public key at web site:
       http://www.geocities.com/SiliconValley/Park/9289/
             Born in Roswell... married an alien...