From: "J. Bruce Fields" Subject: Re: [PATCH] NFS: rename idmapper components Date: Tue, 26 Oct 2010 13:30:04 -0400 Message-ID: <20101026173003.GC526@fieldses.org> References: <4CC6D4F9.3040402@netapp.com> <1288100354.9435.5.camel@heimdal.trondhjem.org> <20101026172016.GA526@fieldses.org> <1288114038.12370.8.camel@heimdal.trondhjem.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Bryan Schumaker , "linux-nfs@vger.kernel.org" To: Trond Myklebust Return-path: Received: from fieldses.org ([174.143.236.118]:35010 "EHLO fieldses.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752407Ab0JZRaG (ORCPT ); Tue, 26 Oct 2010 13:30:06 -0400 In-Reply-To: <1288114038.12370.8.camel-rJ7iovZKK19ZJLDQqaL3InhyD016LWXt@public.gmane.org> Sender: linux-nfs-owner@vger.kernel.org List-ID: On Tue, Oct 26, 2010 at 01:27:18PM -0400, Trond Myklebust wrote: > On Tue, 2010-10-26 at 13:20 -0400, J. Bruce Fields wrote: > > On Tue, Oct 26, 2010 at 09:39:14AM -0400, Trond Myklebust wrote: > > > On Tue, 2010-10-26 at 09:17 -0400, Bryan Schumaker wrote: > > > > This patch renames the idmapper upcall program from nfs.upcall to nfs.idmap. > > > > It also changes the key type from id_resolver to nfs_idmap. > > ... > > > We should _not_ be doing this kind of user space api change after the > > > code has been merged by Linus. > > > > > > I don't mind a patch to the documentation that changes 'nfs.upcall' to > > > match the name given in nfs-utils, but it is too late to change the name > > > 'id_resolver': that should have been done before the merge window. > > > > There's precedent for fixing new interfaces before the final release, so > > it seems unlikely to be a huge problem before -rc1. > > > > Though agreed that there should still be a higher standard for changes > > made after the moment the interface is merged--and a slightly > > inconsistent name may not meet the standard.... > > The name isn't really inconsistent: we use "dns_resolver" for the DNS > upcall types, so "id_resolver" shouldn't really be a problem... OK, makes senses. --b.