Return-Path: linux-nfs-owner@vger.kernel.org Received: from nm37-vm5.bullet.mail.bf1.yahoo.com ([72.30.238.205]:48592 "EHLO nm37-vm5.bullet.mail.bf1.yahoo.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754155Ab3C3L1t convert rfc822-to-8bit (ORCPT ); Sat, 30 Mar 2013 07:27:49 -0400 References: <1364550512-14832-1-git-send-email-oakad@yahoo.com> <1364564090.2660.303.camel@willson.li.ssimo.org> Message-ID: <1364642465.72286.YahooMailNeo@web161301.mail.bf1.yahoo.com> Date: Sat, 30 Mar 2013 04:21:05 -0700 (PDT) From: Alex Dubov Reply-To: Alex Dubov Subject: Re: [PATCH] Fix compilation against Heimdal kerberos implementation To: Simo Sorce Cc: "steved@redhat.com" , "linux-nfs@vger.kernel.org" In-Reply-To: <1364564090.2660.303.camel@willson.li.ssimo.org> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Sender: linux-nfs-owner@vger.kernel.org List-ID: ----- Original Message ----- > From: Simo Sorce > > > krb5_cc_retrieve_cred is available in MIT too, I would try to avoid > using ifedfs here and come up with common code that works for both to > avoid bugs that may cause different behavior in parallel code paths > depending on which library is used. > > This looks like just gratuitous re-indentation, I would drop it it's not > needed and makes the patch bigger. I did it on purpose and said so in the patch description header, because I knew you are going to optimize it out anyway. I don't think that iterating through every credential was ever considered a proper way to implement check_for_taregt function. > > MIT doesn't have krb5_get_err_text at all so it shouldn't be tried out > at all if MIT is available. > > I think you should ratrher add a keberos.m4 macro to check for > krb5_get_err_text then change this to something like: If it's not available on mit-krb5 and is deprecated on Heimdal, why bother with krb5_get_err_text at all?