Received: from carter-zimmerman.mit.edu (carter-zimmerman.suchdamage.org [69.25.196.178]) by krbdev.mit.edu (8.9.3p2) with ESMTP id WAA05000; Tue, 13 Sep 2005 22:33:31 -0400 (EDT) Received: by carter-zimmerman.mit.edu (Postfix, from userid 8042) id 37AFAE0049; Tue, 13 Sep 2005 22:33:14 -0400 (EDT) To: rt@krbdev.mit.edu Subject: Re: [krbdev.mit.edu #3176] References: From: Sam Hartman Date: Tue, 13 Sep 2005 22:33:14 -0400 In-Reply-To: (Marc Aurele La France's message of "Tue, 13 Sep 2005 20:17:21 -0600 (MDT)") Message-Id: User-Agent: Gnus/5.1006 (Gnus v5.10.6) Emacs/21.3 (gnu/linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii RT-Send-Cc: X-RT-Original-Encoding: us-ascii Content-Length: 1006 >>>>> "Marc" == Marc Aurele La France writes: Marc> On Tue, 13 Sep 2005, Sam Hartman via RT wrote: >> Note that the patch to shlib.conf breaks our ABI on AIX. Marc> How so? How can wrapping, or not, a shared object into an Marc> archive affect a programming interface? And why is this Marc> wrapping preferable to producing dlopen'able objects? Because it changes the name that appears in the liblist section of the xcoff object. Marc> Quite frankly, I find it odd that this wrapping is only Marc> being done for AIX, and only by mit-krb5. Even GNU doesn't Marc> do this. Is this a remnant of the historical Marc> misundertanding of how AIX shared objects are supposed to Marc> work? No. It's because I actually understood conventions used for the C libraries on AIX 3.1, 3.2 and 4. AIX 4.3 did add optional conventions for non-wrapped objects, although it was not clear the linker did a good job of finding them when first introduced.