From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 15934 invoked by alias); 10 Jan 2002 19:59:22 -0000 Mailing-List: contact gdb-patches-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sources.redhat.com Received: (qmail 15868 invoked from network); 10 Jan 2002 19:59:18 -0000 Received: from unknown (HELO nevyn.them.org) (128.2.145.6) by sources.redhat.com with SMTP; 10 Jan 2002 19:59:18 -0000 Received: from drow by nevyn.them.org with local (Exim 3.33 #1 (Debian)) id 16OlMn-0002kr-00; Thu, 10 Jan 2002 15:00:01 -0500 Date: Thu, 10 Jan 2002 11:59:00 -0000 From: Daniel Jacobowitz To: Michael Elizabeth Chastain Cc: fnf@redhat.com, gdb-patches@sources.redhat.com Subject: Re: [RFC] const qualifiers in gdb.c++/method.exp Message-ID: <20020110150000.D10383@nevyn.them.org> Mail-Followup-To: Michael Elizabeth Chastain , fnf@redhat.com, gdb-patches@sources.redhat.com References: <200201101507.JAA01023@duracef.shout.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <200201101507.JAA01023@duracef.shout.net> User-Agent: Mutt/1.3.23i X-SW-Source: 2002-01/txt/msg00248.txt.bz2 On Thu, Jan 10, 2002 at 09:07:18AM -0600, Michael Elizabeth Chastain wrote: > Daniel Jacobowitz writes: > > First of all, as best as I can tell, GCC simply doesn't support the > > stabs extensions for const and volatile. Rather than accepting this, > > we should detect it and mark them XFAILS. I would prefer to XFAIL > > based on debug info type and compiler rather than on result - if we lose the > > DWARF-2 const tag in a bug some day, it should start FAILing, not > > XFAILing! > > I would prefer to FAIL rather than XFAIL. However, I accept XFAIL in > this situation, because the bug does not hurt users very much and our > chances of getting the compiler fixed are low. Not that low; it's pretty simplistic. > setup_xfail_format is the function to use here. See gdb.c++/templates.exp > for an example. Note that the caller must call get_debug_format first. > > mec> "A * const" > mec> native i686-pc-linux-gnu, gcc 3.0.2, -gdwarf-2 > mec> native i686-pc-linux-gnu, gcc HEAD, -gdwarf-2 > mec> native i686-pc-linux-gnu, gcc gcc-3_0-branch, -gdwarf-2 > mec> > mec> "const A * const" > mec> native i686-pc-linux-gnu, gcc 2.95.3, -gdwarf-2 > mec> > mec> The first seven of these are wrong. The type needs to have a "const A *" > mec> in it, rather than an "A *". > > drow> That we lost this const is very interesting. It is present in the > drow> dwarf-2 information without a doubt. What version of GDB are you > drow> actually using to test this? > > This is a recent CVS pull of gdb HEAD: > > Checkout begin: Sat Dec 22 18:12:43 PST 2001 > Checkout end: Sat Dec 22 20:08:47 PST 2001 > > Hmmm, that's not all that recent. :( Mind trying again? I think Fred fixed these - or someone else did... -- Daniel Jacobowitz Carnegie Mellon University MontaVista Software Debian GNU/Linux Developer