Re: [Nautilus-list] Solaris memory bug

[Alex Larsson <alexl redhat com>]

On 11 Mar 2002, Michael Meeks wrote:

> Hi Havoc,
> 
> On Mon, 2002-03-11 at 14:18, Havoc Pennington wrote:
> > Maybe this is useful - running "valgrind" on gnome-terminal
> > (requires "chmod 700 /lib/i686" to get rid of thread stuff), 
> > I get warnings about uninitialized memory passed to writev().
> 
> 	I don't believe so - I think that's simply an artefact of the fact that
> CORBA doesn't specify what data is in the padding between requests, and
> as such we don't spend time initializing it.

But valgrind won't report it unless we actually use the uninitialized 
data. It actually knows of the validity of each *bit*, so even partially 
initialized bitfields should work.

And using the uninitialized padding would be bad.

-- 
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 Alexander Larsson                                            Red Hat, Inc 
                   alexl redhat com    alla lysator liu se 
He's a short-sighted small-town messiah with acid for blood. She's a beautiful 
renegade queen of the dead from Mars. They fight crime!