Newsgroups: php.internals Path: news.php.net Xref: news.php.net php.internals:47139 Return-Path: Mailing-List: contact internals-help@lists.php.net; run by ezmlm Delivered-To: mailing list internals@lists.php.net Received: (qmail 32486 invoked from network); 11 Mar 2010 22:55:04 -0000 Received: from unknown (HELO lists.php.net) (127.0.0.1) by localhost with SMTP; 11 Mar 2010 22:55:04 -0000 Authentication-Results: pb1.pair.com header.from=johannes@php.net; sender-id=unknown Authentication-Results: pb1.pair.com smtp.mail=johannes@php.net; spf=unknown; sender-id=unknown Received-SPF: unknown (pb1.pair.com: domain php.net does not designate 83.243.58.134 as permitted sender) X-PHP-List-Original-Sender: johannes@php.net X-Host-Fingerprint: 83.243.58.134 mailout2.netbeat.de Linux 2.6 Received: from [83.243.58.134] ([83.243.58.134:48825] helo=mailout2.netbeat.de) by pb1.pair.com (ecelerity 2.1.1.9-wez r(12769M)) with ESMTP id 93/2A-18254-5C4799B4 for ; Thu, 11 Mar 2010 17:55:04 -0500 Received: (qmail 26004 invoked by uid 89); 11 Mar 2010 22:49:19 -0000 Received: from unknown (HELO ?192.168.1.28?) (postmaster%schlueters.de@93.104.127.222) by mailout2.netbeat.de with ESMTPA; 11 Mar 2010 22:49:19 -0000 X-Originator: 9e51b244e0a38413ab6a9876e36ba9df To: Ilia Alshanetsky Cc: internals@lists.php.net In-Reply-To: <4B997230.3070801@prohost.org> References: <4B9926E8.4080202@lerdorf.com> <4B995F83.1000605@prohost.org> <4B99639B.9030405@oracle.com> <4B996975.9070003@oracle.com> <4B997230.3070801@prohost.org> Content-Type: text/plain; charset="UTF-8" Organization: php.net Date: Thu, 11 Mar 2010 23:54:50 +0100 Message-ID: <1268348090.3959.52.camel@guybrush> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Subject: Re: [PHP-DEV] PHP 6 From: johannes@php.net (Johannes =?ISO-8859-1?Q?Schl=FCter?=) On Thu, 2010-03-11 at 17:44 -0500, Ilia Alshanetsky wrote: > Even though the 5.2 code base is fairly mature, it is far from being bug > free. Unit tests are often a good way to identify corner cases that may > not be handled properly even in the stable branches, so more tests IMHO > is a good thing. Until the decision is made to discontinue the 5.2 > branch, which should be done eventually, maybe even later this year, we > should not neglect it from any standpoint, especially tests. I don't think we should force merging in 5.2, Chris madea good point about proper testing of tests with external systems/libs. On the other hand merging tests to5.2 and 5.3 means that we can find new BC breaks we had overseen and either fix them or document them properly. So I won't "waste" too much time but not forget about 5.2. johannes > On 10-03-11 5:06 PM, Christopher Jones wrote: > > > > > > Eric Stewart wrote: > > > > > Focusing TestFest on 5.3 was not how I had envisioned TestFest 2010, > > I was > > > operating under the assumption we would be committing all applicable > > tests > > > to 5.2, 5.3 and 6.0 as was done last year. Focusing strictly on 5.3 > > > certainly makes my job (and others working on TestFest 2010) easier > > but I > > > wasn't aware that anyone was even entertaining this. > > > > > > This is obviously not strictly on point with this discussion and I > > don't > > > think a divergence from the main topic is appropriate at this > > moment. But it > > > underscores the importance that we will be much better off hashing > > this out > > > sooner rather than later as things in the near future will certainly > > bear > > > significant effects. > > > > > > As this thread winds down and consensus is reached, I'll bring this > > > particular topic up again on the QA List with a discussion based on > > results > > > of late events. > > > > > > Eric Lee Stewart > > > > > > > I think they should be merged to 5.3 and trunk (which is the "branch" > > I really meant in my previous email, not 5.4). > > > > Looking at e.g. the DBA tests, I am building with 3 different versions > > of the Berkeley DB client library. With OCI8 tests I build with three > > different client libraries and then test against three different DB > > versions. Repeating this kind of effort for 5.2 at this mature stage > > of its life isn't worthwhile use of resources IMHO (when no code has > > changed). In a perfect world we'd do it, but where we are now I think > > we need to focus the efforts of testers on 5.3+. > > > > Chris > > >