[PATCH] Avoid signed overflow in num_get::_M_extract_int (PR libstdc++/67214)

[Xi Ruoyao]

On 2017-05-19 15:38 +0100, Jonathan Wakely wrote:
> On 18/05/17 19:10 +0800, Xi Ruoyao wrote:
> > This UB has been hiding so long...
> 
> Indeed! Thanks for the patch.
> 
> > 2017-03-11  Xi Ruoyao  <ryxi@stu.xidian.edu.cn>
> > 
> > 	PR libstdc++/67214
> > 	* include/bits/locale_facets.tcc (_M_extract_int):
> > 	Â Â Add explicit conversion to avoid signed overflow.
> > ---
> > Â libstdc++-v3/include/bits/locale_facets.tcc | 3 ++-
> > Â 1 file changed, 2 insertions(+), 1 deletion(-)
> > 
> > diff --git a/libstdc++-v3/include/bits/locale_facets.tcc b/libstdc++-v3/include/bits/locale_facets.tcc
> > index 351190c..5f85d15 100644
> > --- a/libstdc++-v3/include/bits/locale_facets.tcc
> > +++ b/libstdc++-v3/include/bits/locale_facets.tcc
> > @@ -470,7 +470,8 @@ _GLIBCXX_BEGIN_NAMESPACE_LDBL
> > Â 	bool __testoverflow = false;
> > Â 	const __unsigned_type __max =
> > Â 	Â Â (__negative && __gnu_cxx::__numeric_traits<_ValueT>::__is_signed)
> > -	Â Â ? -__gnu_cxx::__numeric_traits<_ValueT>::__min
> > +	Â Â ? -static_cast<__unsigned_type>(__gnu_cxx::
> > +	Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â __numeric_traits<_ValueT>::__min)
> 
> Do we need to keep the negation, or can we just cast to
> __unsigned_type?

For 2's complement we can just cast to __unsigned_type. Â But for
clarity and other strange architectures I think we should keep
the negation.
-- 
Xi Ruoyao <ryxi@stu.xidian.edu.cn>
School of Aerospace Science and Technology, Xidian University