Re: [PATCH v5] float128: Add wrappers to override ldbl-128 as float128.

["Gabriel F. T. Gomes" <gftg at linux dot vnet dot ibm dot com>]

On Wed, 24 May 2017 16:32:20 +0000
Joseph Myers <joseph@codesourcery.com> wrote:

> On Fri, 19 May 2017, Gabriel F. T. Gomes wrote:
> 
> > +#include <math.h>
> > +#undef HUGE_VALL  
> 
> > +/* float.h constants.  */
> > +#include <float.h>
> > +#undef LDBL_DIG
> > +#undef LDBL_EPSILON
> > +#undef LDBL_MANT_DIG
> > +#undef LDBL_MAX
> > +#undef LDBL_MAX_10_EXP
> > +#undef LDBL_MAX_EXP
> > +#undef LDBL_MIN
> > +#undef LDBL_MIN_10_EXP
> > +#undef LDBL_MIN_EXP
> > +#undef LDBL_TRUE_MIN
> > +#define HUGE_VALL HUGE_VAL_F128  
> 
> HUGE_VALL is not a float.h constant; it would seem better to redefine it 
> immediately after its #undef.
> 
> > diff --git a/sysdeps/ieee754/float128/k_rem_pio2f128.c b/sysdeps/ieee754/float128/k_rem_pio2f128.c
> > new file mode 100644
> > index 0000000..5140266
> > --- /dev/null
> > +++ b/sysdeps/ieee754/float128/k_rem_pio2f128.c
> > @@ -0,0 +1,2 @@
> > +#include <float128_private.h>
> > +#include <math/k_rem_pio2l.c>  
> 
> This file is not needed, and math/k_rem_pio2l.c does not exist, as 
> k_rem_pio2 is only built for the types that need it.  So remove this file.
> 
> OK with those changes.
> 

Thanks.  Pushed with the changes.