This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] [BZ #15381] Initialize wide struct info.
- From: Rich Felker <dalias at aerifal dot cx>
- To: OndÅej BÃlka <neleai at seznam dot cz>
- Cc: Julius Plenz <julius at plenz dot com>, Carlos O'Donell <carlos at redhat dot com>, Andreas Schwab <schwab at suse dot de>, libc-alpha at sourceware dot org
- Date: Tue, 14 May 2013 10:13:33 -0400
- Subject: Re: [PATCH] [BZ #15381] Initialize wide struct info.
- References: <mvmmwryyq0g dot fsf at hawking dot suse dot de> <20130514115913 dot GF20282 at plenz dot com> <20130514123120 dot GA18821 at domone dot kolej dot mff dot cuni dot cz>
On Tue, May 14, 2013 at 02:31:20PM +0200, OndÅej BÃlka wrote:
> On Tue, May 14, 2013 at 01:59:13PM +0200, Julius Plenz wrote:
> > Hi,
> >
> > * Andreas Schwab <schwab@suse.de>:
> > > > Are you opposed to initializing _wide_data to NULL?
> > >
> > > I would prefer -1 so that broken uses crash predictably. :-)
> >
> > Since undefined behaviour can be anything you like, both initializing
> > with NULL and with -1 is equally "valid".
> >
> OK, here is version with -1 (or 0xdeadbeef if you like it more)
-1 will reliably crash on any real-world target. 0xdeadbeef is likely
only to crash on small processes or machines that enforce alignment,
and would otherwise potentially corrupt application memory badly.
Rich