From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 3837 invoked by alias); 3 Apr 2006 14:04:06 -0000 Received: (qmail 3785 invoked by uid 22791); 3 Apr 2006 14:04:05 -0000 X-Spam-Check-By: sourceware.org Received: from mta08-winn.ispmail.ntl.com (HELO mtaout02-winn.ispmail.ntl.com) (81.103.221.48) by sourceware.org (qpsmtpd/0.31) with ESMTP; Mon, 03 Apr 2006 14:04:04 +0000 Received: from aamtaout04-winn.ispmail.ntl.com ([81.103.221.35]) by mtaout02-winn.ispmail.ntl.com with ESMTP id <20060403140401.FEHT29040.mtaout02-winn.ispmail.ntl.com@aamtaout04-winn.ispmail.ntl.com>; Mon, 3 Apr 2006 15:04:01 +0100 Received: from zapata.pink ([82.6.101.32]) by aamtaout04-winn.ispmail.ntl.com with ESMTP id <20060403140400.DNBK16086.aamtaout04-winn.ispmail.ntl.com@zapata.pink>; Mon, 3 Apr 2006 15:04:00 +0100 Received: from zapata.pink (localhost.localdomain [127.0.0.1]) by zapata.pink (8.13.1/8.13.1) with ESMTP id k33E3ud0008460; Mon, 3 Apr 2006 15:03:56 +0100 Received: (from aph@localhost) by zapata.pink (8.13.1/8.13.1/Submit) id k33E3uCp008457; Mon, 3 Apr 2006 15:03:56 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <17457.11083.990954.190967@zapata.pink> Date: Mon, 03 Apr 2006 14:04:00 -0000 From: Andrew Haley To: gcc-patches@gcc.gnu.org, java-patches@gcc.gnu.org Subject: Re: PR java/26858: NullPointerException not generated for large classes... In-Reply-To: <17452.246.662800.831324@zapata.pink> References: <17452.246.662800.831324@zapata.pink> X-IsSubscribed: yes Mailing-List: contact java-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: java-patches-owner@gcc.gnu.org X-SW-Source: 2006-q2/txt/msg00011.txt.bz2 This causes an ICE when building some codebases because we generate a check when we shouldn't. Fixed thusly. Thanks to MJW for finding this. Andrew. 2006-04-03 Andrew Haley PR java/26858 * expr.c (build_field_ref): Don't check the field offset if flag_syntax_only. Index: expr.c =================================================================== --- expr.c (revision 112638) +++ expr.c (working copy) @@ -1696,7 +1696,8 @@ memory may be allocated from any other page, so only field offsets < pagesize are guaratneed to trap. We also assume the smallest page size we'll encounter is 4k bytes. */ - if (check && ! flag_check_references && ! flag_indirect_dispatch) + if (! flag_syntax_only && check && ! flag_check_references + && ! flag_indirect_dispatch) { tree field_offset = byte_position (field_decl); if (! page_size)