Oddity with function returning image of fixed point type

Oddity with function returning image of fixed point type

[Jesper Quorning]

Is something odd going on here?
I did not expect Image_Odd1/2 to return floating point images.


with Ada.Text_Io;     use Ada.Text_Io;

procedure Fpt_Last is

   type Fpt is delta 0.01 digits 4;

   Image_Last : constant String := Fpt'Image (Fpt'Last);

   function Image_Ok return String is
   begin
      return Fpt'Last'Image;
      --  return Fpt'Image (Fpt'Last); -- Also ok
   end Image_Ok;

   Last : constant Fpt := Fpt'Last;

   function Image_Odd_1 return String is (Fpt'Last'Img);
   function Image_Odd_2 return String is (Fpt'Last'Image);
   function Image_Ok_2  return String is (Fpt'Image (FPT'Last));
   function Image_Ok_3  return String is (Last'Image);

begin
   Put_Line ("Image_Last  : " & Image_Last);
   Put_Line ("Image_Ok    : " & Image_Ok);
   Put_Line ("Image_Odd_1 : " & Image_Odd_1);
   Put_Line ("Image_Odd_2 : " & Image_Odd_2);
   Put_Line ("Image_Ok_2  : " & Image_Ok_2);
   Put_Line ("Image_Ok_3  : " & Image_Ok_3);
end Fpt_Last;


Output:

Image_Last  :  99.99
Image_Ok    :  99.99
Image_Odd_1 :  9.99900000000000000E+01
Image_Odd_2 :  9.99900000000000000E+01
Image_Ok_2  :  99.99
Image_Ok_3  :  99.99

Compiled with gnatmake version 10.3.0 or CE 2020 on macOS 10.13.6.


/Jesper

Re: Oddity with function returning image of fixed point type

[Stephen Leake]

Jesper Quorning <jesper.quorning@gmail.com> writes:

> Is something odd going on here?
> I did not expect Image_Odd1/2 to return floating point images.

Right, this looks like a compiler bug. LRM 3.5(13):

S'Last
               S'Last denotes the upper bound of the range of S. The
               value of this attribute is of the type of S.

But this is acting like Fpt'Last is universal_real.

-- 
-- Stephe

Re: Oddity with function returning image of fixed point type

[Dmitry A. Kazakov]

On 2021-09-02 19:08, Stephen Leake wrote:
> Jesper Quorning <jesper.quorning@gmail.com> writes:
> 
>> Is something odd going on here?
>> I did not expect Image_Odd1/2 to return floating point images.
> 
> Right, this looks like a compiler bug. LRM 3.5(13):
> 
> S'Last
>                 S'Last denotes the upper bound of the range of S. The
>                 value of this attribute is of the type of S.
> 
> But this is acting like Fpt'Last is universal_real.

No, it is this:

LRM 4.10 (40/5)

"X'Image denotes the result of calling function S'Image with Arg being 
X, where S is the nominal subtype of X."

The question is what is the nominal subtype of Fpt'Last.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: Oddity with function returning image of fixed point type

[Jesper Quorning]

torsdag den 2. september 2021 kl. 19.33.00 UTC+2 skrev Dmitry A. Kazakov:
> On 2021-09-02 19:08, Stephen Leake wrote: 
> > Jesper Quorning <jesper....@gmail.com> writes: 
> > 
> >> Is something odd going on here? 
> >> I did not expect Image_Odd1/2 to return floating point images. 
> > 
> > Right, this looks like a compiler bug. LRM 3.5(13): 
> > 
> > S'Last 
> > S'Last denotes the upper bound of the range of S. The 
> > value of this attribute is of the type of S. 
> > 
> > But this is acting like Fpt'Last is universal_real.

> No, it is this: 
> 
> LRM 4.10 (40/5)
Had to go back to Ada 83 LRM to find a chapter 4.10.

> "X'Image denotes the result of calling function S'Image with Arg being 
> X, where S is the nominal subtype of X." 
LRM 3.5 (35-36) says about the same.

> The question is what is the nominal subtype of Fpt'Last. 

Well Image_Ok and Image_Odd_2 should both return Fpt'Last'Image so one of them must be bad.

Found LRM 3.5 (27.7/2) describing the image of a fixed point type.

Thanks for your responses. I will report the issue.

/Jesper