From mboxrd@z Thu Jan  1 00:00:00 1970
From: apbianco@redhat.com
To: java-gnats@sourceware.cygnus.com
Subject: libgcj/1222: Testing framework: enforce failure results correctness.
Date: Wed, 20 Dec 2000 12:22:00 -0000
Message-id: <20000322225538.14196.qmail@sourceware.cygnus.com>
X-SW-Source: 2000-q4/msg00999.html
List-Id: <gcc-prs.sourceware.org>

>Number:         1222
>Category:       libgcj
>Synopsis:       Testing framework: enforce failure results correctness.
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    tromey
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Dec 20 12:17:53 PST 2000
>Closed-Date:    
>Last-Modified:  
>Originator:     Alex Petit-Bianco
>Release:        unknown-1.0
>Organization:
>Environment:
All.
>Description:
The current testing framework already knows about tests
that are meant to fail. In order to fully test regressions,
it would be desirable to compare the output of a meant to
fail test case with a known to be correct entry and make 
sure that the compiler fails in an identical way.

I've sometimes seen gcj regressing on the way it reports
errors -- it's not something desirable.
>How-To-Repeat:

>Fix:

>Release-Note:
>Audit-Trail:

Formerly PR libgcj/180

>Unformatted: