From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 31402 invoked by alias); 24 Dec 2004 20:26:43 -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 31364 invoked by uid 48); 24 Dec 2004 20:26:38 -0000 Date: Fri, 24 Dec 2004 20:26:00 -0000 Message-ID: <20041224202638.31363.qmail@sourceware.org> From: "mmitchel at gcc dot gnu dot org" To: gcc-bugs@gcc.gnu.org In-Reply-To: <20041224172149.19148.bugzilla-gcc@thewrittenword.com> References: <20041224172149.19148.bugzilla-gcc@thewrittenword.com> Reply-To: gcc-bugzilla@gcc.gnu.org Subject: [Bug c++/19148] [4.0 Regression] ICE: gimplification failed X-Bugzilla-Reason: CC X-SW-Source: 2004-12/txt/msg03488.txt.bz2 List-Id: ------- Additional Comments From mmitchel at gcc dot gnu dot org 2004-12-24 20:26 ------- The front end is generating as the RHS: (INDIRECT_REF (NOP_EXPR (ADDR_EXPR (TARGET_EXPR (COND_EXPR ...))) which seems reasonable. The LHS is: (INDIRECT_REF (SAVE_EXPR (POSTINCREMENT_EXPR (PARM_DECL)))) which also seems reasonable. Due to my patch, we now call gimplify_cond_expr on the RHS, with the "target" parameter non-NULL. That causes us to call gimplify_expr on the LHS, with fb_lvalue, which aborts, because the gimplifier doesn't have a good place to put the post-increment. Adding a post_p parameter to gimplify_cond_expr fixes the problem (and passing it down to gimplify_expr) fixes the problem. Is there any reason that's not a bad idea? -- What |Removed |Added ---------------------------------------------------------------------------- CC| |jason at redhat dot com, rth | |at redhat dot com http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19148