[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: gcc compilation error with abs() affects sarge, etch, lenny, sid

On Thu, Nov 22, 2007 at 08:52:09AM +0300, Nikita V. Youshchenko wrote:
> > Today I saw a thread on the linux-kernel list [1] about a gcc compilation
> > issue that affects _all_ gcc versions in Debian after gcc-2.95.
> > I've filed bugs reports:
> > #452108 (gcc-4.2); #452113 (gcc-4.1); #452114 (gcc-3.3); #452115 (gcc-3.4)
> > 
> > The issue is that in some cases a negative const before an abs() function
> > will be brought _inside_ the abs() function and will thus be converted to
> > a positive.
> I think this should be fixed in stable as well. And probably even in
> oldstable.
> Miscompilation is BAD thing ...

You sound like this would be the first miscompilation bug in gcc in a
long time. I think we actually have several of them in each gcc version
at any time. They are usually not architecture-independent like this one

Frank Lichtenheld <djpig@debian.org>
www: http://www.djpig.de/

Reply to: