Return-Path: X-Original-To: apmail-hbase-dev-archive@www.apache.org Delivered-To: apmail-hbase-dev-archive@www.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id 9336817F70 for ; Thu, 5 Mar 2015 17:44:56 +0000 (UTC) Received: (qmail 95199 invoked by uid 500); 5 Mar 2015 17:44:21 -0000 Delivered-To: apmail-hbase-dev-archive@hbase.apache.org Received: (qmail 95119 invoked by uid 500); 5 Mar 2015 17:44:21 -0000 Mailing-List: contact dev-help@hbase.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: dev@hbase.apache.org Delivered-To: mailing list dev@hbase.apache.org Received: (qmail 95108 invoked by uid 99); 5 Mar 2015 17:44:21 -0000 Received: from mail-relay.apache.org (HELO mail-relay.apache.org) (140.211.11.15) by apache.org (qpsmtpd/0.29) with ESMTP; Thu, 05 Mar 2015 17:44:21 +0000 Received: from mail-lb0-f179.google.com (mail-lb0-f179.google.com [209.85.217.179]) by mail-relay.apache.org (ASF Mail Server at mail-relay.apache.org) with ESMTPSA id 1451A1A01E0 for ; Thu, 5 Mar 2015 17:44:21 +0000 (UTC) Received: by lbiv13 with SMTP id v13so34229701lbi.11 for ; Thu, 05 Mar 2015 09:44:19 -0800 (PST) X-Received: by 10.152.116.43 with SMTP id jt11mr9073674lab.30.1425577459617; Thu, 05 Mar 2015 09:44:19 -0800 (PST) MIME-Version: 1.0 Received: by 10.25.87.20 with HTTP; Thu, 5 Mar 2015 09:43:39 -0800 (PST) In-Reply-To: References: <369520731.4581185.1425529245583.JavaMail.yahoo@mail.yahoo.com> From: Andrew Purtell Date: Thu, 5 Mar 2015 09:43:39 -0800 Message-ID: Subject: Re: Client-Server wire compatibility? To: "dev@hbase.apache.org" Cc: lars hofhansl Content-Type: multipart/alternative; boundary=001a11c3656814957005108e1e07 --001a11c3656814957005108e1e07 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable On Thu, Mar 5, 2015 at 12:10 AM, Matteo Bertozzi wrote: > On Thu, Mar 5, 2015 at 4:20 AM, lars hofhansl wrote: > > > The idea actually was that a new client can never be 100% supported, > since > > a user could use it accessing new features that the server does not > > understand. > > > I think that it is fine if we have exceptions not handled when a client > tries to ask a new features on an old server. > but my question was about old operations. e.g. Assuming that you do a > create table with all the arguments/configuration that the old server was > able to understand, is it fair to have a 1.1 client that throws an > exception against a 1.0 server because in the new 1.1 the create logic is > not compatible? > =E2=80=8B=E2=80=8BThat question hinges on why the create logic is not compa= tible. =E2=80=8BWe have guidelines that allow us flexibility to break things at clearly defined points, but that doesn't mean we shouldn't avoid doing so wherever possible. Why can't create remain wire compatible in this case? Is there a JIRA? --001a11c3656814957005108e1e07--