From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 26863 invoked by alias); 27 Mar 2006 17:09:49 -0000 Received: (qmail 26811 invoked by uid 48); 27 Mar 2006 17:09:16 -0000 Date: Mon, 27 Mar 2006 17:09:00 -0000 Message-ID: <20060327170916.26810.qmail@sourceware.org> X-Bugzilla-Reason: CC References: Subject: [Bug tree-optimization/26797] [4.2 Regression] ACATS c35507m cd2a23e cxh1001 failures In-Reply-To: Reply-To: gcc-bugzilla@gcc.gnu.org To: gcc-bugs@gcc.gnu.org From: "law at redhat dot com" 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 X-SW-Source: 2006-03/txt/msg02655.txt.bz2 List-Id: ------- Comment #4 from law at redhat dot com 2006-03-27 17:09 ------- It seems to me that the fundamental problem is that the testcase manages to stuff the value "3" into an object which is supposed to only have the range [4, 5]. ie, if you put a breakpoint in C35507M.NEWCHARBRP and examine the first parameter (ie A) you'll find that on the first call and second calls "A" has the value 4 and 5 respectively. However, on the 3rd call it has the value 3, which is clearly out of "A"'s range [3, 4]. I suspect there are similar issues with C35507M.CHARRP since it has similar looking code. Someone with a better knowledge of the Ada front-end, the langauge and the testsuite is going to need to chime in at this point. From the data I see, VRP is doing exactly what we want and expect it to do. -- law at redhat dot com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |kenner at vlsi1 dot ultra | |dot nyu dot edu http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26797