Newsgroups: php.internals Path: news.php.net Xref: news.php.net php.internals:14419 Return-Path: Mailing-List: contact internals-help@lists.php.net; run by ezmlm Delivered-To: mailing list internals@lists.php.net Received: (qmail 34102 invoked by uid 1010); 23 Jan 2005 22:05:28 -0000 Delivered-To: ezmlm-scan-internals@lists.php.net Delivered-To: ezmlm-internals@lists.php.net Received: (qmail 34087 invoked from network); 23 Jan 2005 22:05:28 -0000 Received: from unknown (HELO lists.php.net) (127.0.0.1) by localhost with SMTP; 23 Jan 2005 22:05:28 -0000 X-Host-Fingerprint: 211.9.230.193 mgkyb1.nw.wakwak.com FreeBSD 4.6-4.9 Received: from ([211.9.230.193:4818] helo=mgkyb1.nw.wakwak.com) by pb1.pair.com (ecelerity HEAD (r4105:4106)) with SMTP id 62/02-29890-6AF14F14 for ; Sun, 23 Jan 2005 17:05:27 -0500 Received: from vckyb2.nw.wakwak.com (vckyb2.nw.wakwak.com [211.9.230.145]) by mgkyb1.nw.wakwak.com (8.13.3/8.13.3/2005-01-20) with SMTP id j0NM5L0p035250; Mon, 24 Jan 2005 07:05:21 +0900 (JST) (envelope-from moriyoshi@at.wakwak.com) Received: from at.wakwak.com (at.wakwak.com [211.9.230.135]) by vckyb2.nw.wakwak.com (Postfix) with ESMTP id C25C940002; Mon, 24 Jan 2005 07:05:20 +0900 (JST) Received: from [192.168.0.2] (newcycle.as.wakwak.ne.jp [218.225.209.145]) by at.wakwak.com (8.13.2/8.13.2/2005-01-05) with ESMTP/inet id j0NM5Ks3086940; Mon, 24 Jan 2005 07:05:20 +0900 (JST) (envelope-from moriyoshi@at.wakwak.com) Mime-Version: 1.0 (Apple Message framework v619) Content-Type: text/plain; charset=US-ASCII; format=flowed Message-ID: Content-Transfer-Encoding: 7bit Cc: Rui Hirokawa , "'internals List Developers' Mailing" Date: Mon, 24 Jan 2005 07:04:31 +0900 To: Ilia Alshanetsky X-Mailer: Apple Mail (2.619) Subject: Reversion of patch for bug #31398 From: moriyoshi@at.wakwak.com (Moriyoshi Koizumi) Ilia, While your opinion that we shouldn't count on the "filename" attribute in multiparted queries is quite to the point, your patch totally disregards the old behaviour. We must neither use php_basename() for pathes of a spec different from that defined by the platform which is determined in compile time. That most likely leads to inconsistent behaviour between platforms. In addition, you had removed the portion Rui added in the past that allowed multibyte-encoded filenames to get parsed properly without any notice. Actually the function in question doesn't work at all with such filenames at this momemnt. Due to those issues, I'm going to revert your patch and recommit less radical one. Regards, Moriyoshi