From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 15071 invoked by alias); 23 Sep 2004 07:50:13 -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 15015 invoked by uid 48); 23 Sep 2004 07:50:08 -0000 Date: Thu, 23 Sep 2004 07:50:00 -0000 Message-ID: <20040923075008.15014.qmail@sourceware.org> From: "aph at gcc dot gnu dot org" To: gcc-bugs@gcc.gnu.org In-Reply-To: <20020807153600.7532.jmr@fulcrummicro.com> References: <20020807153600.7532.jmr@fulcrummicro.com> Reply-To: gcc-bugzilla@gcc.gnu.org Subject: [Bug libgcj/7532] shutdown hooks not run upon abnormal termination X-Bugzilla-Reason: CC X-SW-Source: 2004-09/txt/msg02708.txt.bz2 List-Id: ------- Additional Comments From aph at gcc dot gnu dot org 2004-09-23 07:49 ------- The signal handler (here called "shutdown()") can't legally be unwound if ::java::lang::System::exit throws an exception. Studying the libgcj runtime code, I'm not sure if this might happen. The Java class lib spec says "This method never returns normally", suggesting that an exceptional return is possible. We could simply catch everything in shutdown() and then call exit(). With that and the removal of ::gcj::reduceSignalUsage this patch would be fine for POSIXy systems. -- What |Removed |Added ---------------------------------------------------------------------------- Last reconfirmed|2004-08-27 04:05:12 |2004-09-23 07:50:02 date| | http://gcc.gnu.org/bugzilla/show_bug.cgi?id=7532