https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93567
jvdelisle at charter dot net changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |jvdelisle at charter dot net
--- Comment #6 from jvdelisle at charter dot net ---
(In reply to Dominique d'Humieres from comment #4)
> The patch in comment 3 broke gfortran.dg/round_3.f08.
>
> The following should work
>
> --- ../_clean/libgfortran/io/write_float.def 2020-06-13 03:11:55.000000000
> +0200
> +++ libgfortran/io/write_float.def 2020-07-21 00:37:24.000000000 +0200
> @@ -1004,7 +1004,11 @@ determine_en_precision (st_parameter_dt
> exp_d = calculate_exp_ ## x (d);\
> r_sc = (1 - r / exp_d);\
> temp = 0.1 * r_sc;\
> - if ((m > 0.0 && ((m < temp) || (r >= (exp_d - m))))\
> + if ((m > 0.0 && ((m < temp) || (r == 0.5 && r >= (exp_d - m))\
> + || (r == 1 && !sign_bit && 1 > (exp_d - m))\
> + || (r == 1 && sign_bit && 0 >= (exp_d - m))\
> + || (r == 0 && !sign_bit && 0 >= (exp_d - m))\
> + || (r ==0 && sign_bit && 1 > (exp_d - m))))\
^
> || ((m == 0.0) && !(compile_options.allow_std\
> & (GFC_STD_F2003 | GFC_STD_F2008)))\
> || d == 0)\
It gets complicated quickly with all the corner cases. Lat line of the added
logic needs a space before the 0