[PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

[Tamar Christina]

Hi Joseph,

I have updated the patch with the changes,
w.r.t to the formatting, there are tabs there that seem to be rendered
at 4 spaces wide. In my editor setup at 8 spaces they are correct.

Kind Regards,
Tamar
________________________________________
From: Joseph Myers <joseph@codesourcery.com>
Sent: Thursday, November 24, 2016 6:28:18 PM
To: Tamar Christina
Cc: GCC Patches; Wilco Dijkstra; rguenther@suse.de; law@redhat.com; Michael Meissner; nd
Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On Thu, 24 Nov 2016, Tamar Christina wrote:

> @@ -11499,6 +11503,53 @@ to classify.  GCC treats the last argument as type-generic, which
>  means it does not do default promotion from float to double.
>  @end deftypefn
>
> +@deftypefn {Built-in Function} int __builtin_isnan (...)
> +This built-in implements the C99 isnan functionality which checks if
> +the given argument represents a NaN.  The return value of the
> +function will either be a 0 (false) or a 1 (true).
> +On most systems, when an IEEE 754 floating point is used this
> +built-in does not produce a signal when a signaling NaN is used.

"an IEEE 754 floating point" should probably be "an IEEE 754
floating-point type" or similar.

> +GCC treats the argument as type-generic, which means it does
> +not do default promotion from float to double.

I think type names such as float and double should use @code{} in the
manual.

> +the given argument represents an Infinite number.  The return

Infinite should not have a capital letter there.

> +@deftypefn {Built-in Function} int __builtin_iszero (...)
> +This built-in implements the C99 iszero functionality which checks if

This isn't C99, it's TS 18661-1:2014.

> +the given argument represents the number 0.  The return

0 or -0.

> +@deftypefn {Built-in Function} int __builtin_issubnormal (...)
> +This built-in implements the C99 issubnormal functionality which checks if

Again, TS 18661-1.

> +the given argument represents a sub-normal number.  The return

Do not hyphenate subnormal.

> +         switch (DECL_FUNCTION_CODE (decl))
> +         {
> +                 case BUILT_IN_SETJMP:
> +                     lower_builtin_setjmp (gsi);
> +                     data->cannot_fallthru = false;
> +                     return;

The indentation in this whole block of code (not all quoted) is wrong.

> +    real_inf(&rinf);

Missing space before '('.

> +  emit_tree_cond (&body, dest, done_label,
> +               is_normal(&body, arg, loc), fp_normal);
> +  emit_tree_cond (&body, dest, done_label,
> +               is_zero(&body, arg, loc), fp_zero);
> +  emit_tree_cond (&body, dest, done_label,
> +               is_nan(&body, arg, loc), fp_nan);
> +  emit_tree_cond (&body, dest, done_label,
> +               is_infinity(&body, arg, loc), fp_infinite);

Likewise.

> +               fndecl(&body, arg, loc), t_true);

Likewise.

--
Joseph S. Myers
joseph@codesourcery.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: fp-classify-new2.patch
Type: text/x-patch
Size: 65047 bytes
Desc: fp-classify-new2.patch
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20161125/4fc20bc7/attachment.bin>