Newsgroups: php.internals Path: news.php.net Xref: news.php.net php.internals:15399 Return-Path: Mailing-List: contact internals-help@lists.php.net; run by ezmlm Delivered-To: mailing list internals@lists.php.net Received: (qmail 93310 invoked by uid 1010); 12 Mar 2005 14:17:16 -0000 Delivered-To: ezmlm-scan-internals@lists.php.net Delivered-To: ezmlm-internals@lists.php.net Received: (qmail 93295 invoked from network); 12 Mar 2005 14:17:16 -0000 Received: from unknown (HELO mailserverfree01.dynip.com) (127.0.0.1) by localhost with SMTP; 12 Mar 2005 14:17:16 -0000 X-Host-Fingerprint: 64.233.184.195 wproxy.gmail.com Linux 2.4/2.6 Received: from ([64.233.184.195:18057] helo=wproxy.gmail.com) by pb1.pair.com (ecelerity HEAD r(5124)) with SMTP id FD/7C-31540-CE9F2324 for ; Sat, 12 Mar 2005 09:17:16 -0500 Received: by wproxy.gmail.com with SMTP id 37so883275wra for ; Sat, 12 Mar 2005 06:17:14 -0800 (PST) DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:reply-to:to:subject:mime-version:content-type:content-transfer-encoding; b=abRRigfc8IDAj/YOWLvCXJO9/GDCjg2Jc2NTXT8fGjtoXxWoBDfuZdaGWvue1HUwCfQ7InFGhhBX5bAB+yjIk7I+zrGbwrDJjVNyZFysb/z1zuB2zfT5U+iozW2YoH3f8/y3eedcKrw0zwmq4J0Esygx8pglqQ7HPsQ16CIQrX0= Received: by 10.54.25.18 with SMTP id 18mr618965wry; Sat, 12 Mar 2005 06:17:14 -0800 (PST) Received: by 10.54.44.57 with HTTP; Sat, 12 Mar 2005 06:17:14 -0800 (PST) Message-ID: <4e89b42605031206177b564e50@mail.gmail.com> Date: Sat, 12 Mar 2005 09:17:14 -0500 Reply-To: Wez Furlong To: PHP Development Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Subject: Weird time-sensitive leaks / problems in the memory manager From: kingwez@gmail.com (Wez Furlong) Yesterday I saw some leaks with foreach() and custom zend_iterators. Today, the same binary doesn't have those leaks. I had another similar issue a few weeks back on a different machine, but with an overrun reported on shutdown. The next day, the same binary was fine. Unless the linux-fairy visited both those boxes overnight and magically fixed the bugs, I think we have an issue either in the memory manager, or worse, in some other piece of code that is subtly mangling the memory manager headers. (this is with HEAD). Valgrind shows no problems. Looks like it's going to be hours of "fun" trying to track this down. --Wez.