On 01/07/2014 02:47 PM, Christophe Fergeau wrote:
Hey,
On Mon, Jan 06, 2014 at 03:35:24PM +0200, Uri Lublin wrote:
On 01/06/2014 01:22 PM, Christophe Fergeau wrote:
find_model_params() is first dereferencing nbuckets when setting
it to 0, and then it checks it for NULL. The NULL-check will nev
On Tue, Jan 07, 2014 at 01:47:12PM +0100, Christophe Fergeau wrote:
> Hey,
>
> On Mon, Jan 06, 2014 at 03:35:24PM +0200, Uri Lublin wrote:
> > On 01/06/2014 01:22 PM, Christophe Fergeau wrote:
> > >find_model_params() is first dereferencing nbuckets when setting
> > >it to 0, and then it checks it
Hey,
On Mon, Jan 06, 2014 at 03:35:24PM +0200, Uri Lublin wrote:
> On 01/06/2014 01:22 PM, Christophe Fergeau wrote:
> >find_model_params() is first dereferencing nbuckets when setting
> >it to 0, and then it checks it for NULL. The NULL-check will never trigger
> >as if the pointer was NULL, we'd
On 01/06/2014 01:22 PM, Christophe Fergeau wrote:
find_model_params() is first dereferencing nbuckets when setting
it to 0, and then it checks it for NULL. The NULL-check will never trigger
as if the pointer was NULL, we'd crash when we dereference it.
This commit removes the redundant test, but
find_model_params() is first dereferencing nbuckets when setting
it to 0, and then it checks it for NULL. The NULL-check will never trigger
as if the pointer was NULL, we'd crash when we dereference it.
This commit removes the redundant test, but adds an assert so that we
catch this condition if it
