Newsgroups: php.internals Path: news.php.net Xref: news.php.net php.internals:30355 Return-Path: Mailing-List: contact internals-help@lists.php.net; run by ezmlm Delivered-To: mailing list internals@lists.php.net Received: (qmail 43973 invoked by uid 1010); 29 Jun 2007 00:20:25 -0000 Delivered-To: ezmlm-scan-internals@lists.php.net Delivered-To: ezmlm-internals@lists.php.net Received: (qmail 43958 invoked from network); 29 Jun 2007 00:20:25 -0000 Received: from unknown (HELO lists.php.net) (127.0.0.1) by localhost with SMTP; 29 Jun 2007 00:20:25 -0000 Authentication-Results: pb1.pair.com header.from=giunta.gaetano@gmail.com; sender-id=pass; domainkeys=bad Authentication-Results: pb1.pair.com smtp.mail=giunta.gaetano@gmail.com; spf=pass; sender-id=pass Received-SPF: pass (pb1.pair.com: domain gmail.com designates 66.249.92.168 as permitted sender) DomainKey-Status: bad X-DomainKeys: Ecelerity dk_validate implementing draft-delany-domainkeys-base-01 X-PHP-List-Original-Sender: giunta.gaetano@gmail.com X-Host-Fingerprint: 66.249.92.168 ug-out-1314.google.com Received: from [66.249.92.168] ([66.249.92.168:9190] helo=ug-out-1314.google.com) by pb1.pair.com (ecelerity 2.1.1.9-wez r(12769M)) with ESMTP id AB/39-21924-84054864 for ; Thu, 28 Jun 2007 20:20:24 -0400 Received: by ug-out-1314.google.com with SMTP id c2so522931ugf for ; Thu, 28 Jun 2007 17:20:21 -0700 (PDT) DKIM-Signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:user-agent:mime-version:to:subject:references:in-reply-to:content-type:content-transfer-encoding; b=nDej1Ee4Cc8W7PxNiMQdSu/NGQmOTxUQT1fum5uZFaJzzdtKlg4DnTdnE/z8VDrahrVuz/AWMJ1aB5tN1SC2nX1tBUGWqv7Oni8Ld91R+goRKLimIjpnVty2+FPaB+Li/Cjz3HZNwodO/5+AK1JTj5i8Q0un2jwfIGTq6Rob3jw= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:user-agent:mime-version:to:subject:references:in-reply-to:content-type:content-transfer-encoding; b=GNt9dj/B7+z1/oH/x2Is6jNzuWT3HKspAaJ9FL+JyaxyvCQTgrrWEOtTQS0YY+vwxb9CBCSoOsKdVwdZEYogTYREZIRwz2lvza4a/Q/b0sZm/jLey6xle6gEY2WO+VsV1Eec8mH6KxrTTqkJW33aRQeFKeP5K5YgjBjX8TffvYk= Received: by 10.66.243.2 with SMTP id q2mr1282523ugh.1183076421736; Thu, 28 Jun 2007 17:20:21 -0700 (PDT) Received: from ?192.168.1.2? ( [151.66.60.36]) by mx.google.com with ESMTP id o55sm6959155uga.2007.06.28.17.20.20 (version=TLSv1/SSLv3 cipher=RC4-MD5); Thu, 28 Jun 2007 17:20:20 -0700 (PDT) Message-ID: <4684500A.6050805@gmail.com> Date: Fri, 29 Jun 2007 02:19:22 +0200 User-Agent: Thunderbird 2.0.0.4 (Windows/20070604) MIME-Version: 1.0 To: Greg Beaver , internals@lists.php.net References: <4682970D.5030607@gmail.com> <4682F736.3030103@gmail.com> <46831765.6000506@php.net> In-Reply-To: <46831765.6000506@php.net> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Subject: Re: SNAPS+SNAPSPecl different to Pecl4Win From: giunta.gaetano@gmail.com (Gaetano Giunta) Thanks for your encouraging words. I ran a "complete" build today against php 5.2.3, and these are the stats: - 134 packaged extensions downloaded from pecl website (sniffed the list out of cvs.php.net, then got the latest existing package for every one) of those: only 61 have a config.w32 file, while 27 are part of the php source distribution - there are 79 folders in the /ext dir of the source distribution, but only 45 dlls in the windows distribution - pecl snapshot contains 75 extensions, none of which is part of the 45 dlls in the windows distribution - pecl4win snapshot contains 94 extensions, 13 of which are distributed as dlls in the windows distribution and 75 of which are part of the pecl snapshot ==> pecl snapshot is a subset of pecl4win! I used the zip.zip file from Edin instead of as a complete build environment instead of win32build from php.net, and vc++ 2005 express (so no luck in testing the produced dlls, as they are not compatible with the official distribution). Caveats: - for packages existing both in pecl and in the base distribution, I compiled the version coming with the distribution - bitset and blenc have errors in config.w32 that prevent building a correct Makefile. Both have been fixed in cvs, but not (yet) in a release - same goes for cairo_wrapper. Strangely enough, the cvs fix to the config file dates from november, while the package is from January. Yes, I am looking at you, Hartmut! ;[ - extensions depeneding on external libs mdbtools, ekhtml and freeimage have not been built because the libs are missing. BTW: I have downloaded from sf.net a binary version of freeimage wit .h, .lib and .dll, so that might (?) be easily solved Results: 61 dll files built, 43 of which are part of the standard win32 distribution and 29 of which are already in pecl4win Compared to pecl4win build log, I have to fix a couple of includes (svn, fbcaccess), plus I lost somewhere all the win32* extensions, so the final count might be up two or four extensions. All in all, not a huge sucess, but better than nothing... Next steps: + try running the whole process again with php 4.4 + make sure that the db and gui can accomodate both a cvs version and many packaged versions + ... + profit! > Hi Gaetano, > > The flaw is not in your system, but in PECL itself. The best approach > here is for authors to ensure their packages are up to date (I don't > intend to single out any folks here, it's not easy to do what I'm > saying). Extensions that don't build simply won't be there. This is > how it works now, except that it is based on CVS rather than on releases. > No problem in singling out folks: I can do it myself! > One can't expect that all old releases will build, there are several > examples of packages at pear.php.net that are so ancient, there was no > proper validation of the package.xml file and they won't install with > newer modern versions of the PEAR installer. This is perfectly natural. > > Also, generally there is a lag between CVS and releases for good reason > - further testing of CVS is needed to ensure it doesn't break things. > If there is a pecl4win that shows broken releases, then it puts a useful > pressure on devs to release a new package version. I see no downsides > to your approach. > > What may be the best solution is to continue to build from CVS as well, > so we have "snapshots" in a similar manner to PHP snapshots, and releases. > > Thanks, > Greg > >