From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 19341 invoked by alias); 21 Aug 2003 04:28:24 -0000 Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org Received: (qmail 19326 invoked by uid 48); 21 Aug 2003 04:28:23 -0000 Date: Thu, 21 Aug 2003 04:28:00 -0000 Message-ID: <20030821042823.19324.qmail@sources.redhat.com> From: "nbryant at allegientsystems dot com" To: gcc-bugs@gcc.gnu.org In-Reply-To: <20030821012448.12001.nbryant@allegientsystems.com> References: <20030821012448.12001.nbryant@allegientsystems.com> Reply-To: gcc-bugzilla@gcc.gnu.org Subject: [Bug libgcj/12001] _Jv_FindClass dumps core when Oracle's classes12.zip in $CLASSPATH X-Bugzilla-Reason: CC X-SW-Source: 2003-08/txt/msg02259.txt.bz2 List-Id: PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org. http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12001 ------- Additional Comments From nbryant at allegientsystems dot com 2003-08-21 04:28 ------- I thought about this (making a parent class loader that handles the binary classes, and making a child that handles the interpreted ones) but I wasn't sure what the semantic consequences would be. So, I am currently testing a patch to URLClassloader which delays the initialization of the protocol handlers until after the class loader is completely constructed. this has the side effect of moving the mutex locking to runtime instead of init time, but it shouldn't be too bad.