[PATCH] PR libstdc++/13045 / C++ demangler, floating values and function pointer type return type postfix fix.

[Geoff Keating]

> X-Original-To: geoffk@foam.wonderslug.com
> Date: Tue, 25 Nov 2003 03:03:50 +0100
> From: Carlo Wood <carlo@alinoe.com>
> Cc: Benjamin Kosnik <bkoz@redhat.com>, gcc-patches@gcc.gnu.org,
>    libstdc++@gcc.gnu.org
> Mail-Followup-To: Geoff Keating <geoffk@geoffk.org>,
> 	Benjamin Kosnik <bkoz@redhat.com>, gcc-patches@gcc.gnu.org,
> 	libstdc++@gcc.gnu.org
> Content-Disposition: inline
> User-Agent: Mutt/1.4.1i
> X-OriginalArrivalTime: 25 Nov 2003 01:56:48.0890 (UTC) FILETIME=[6348FDA0:01C3B2F7]
> 
> On Mon, Nov 24, 2003 at 01:36:53PM -0800, Geoff Keating wrote:
> > That might be reasonable for 'float' and 'double', but 'long double'
> > varies greatly between platforms.
> 
> sizeof(long double) is > 8 I suppose?  In that case it will
> print just the hex value: (double)[3ff999294912bc6120e95828]
> or actually, it currently would fail to demangle it,
> it doesn't support the 'long double' casts in literals.

It varies.  On some platforms, 'long double' is the same as 'double'
(possibly depending on a flag).

-- 
- Geoffrey Keating <geoffk@geoffk.org>