Possible Bug, Need Second Opinion

Possible Bug, Need Second Opinion

[ANH_VO]

I am nearly 100% certain that the codes below should compile fine. But, I would
like to have second opinion possibly more before reporting it.

with Interfaces;

package Global is

   type Size_T is new Interfaces.Unsigned_32;

end Global;


with Global;
with Interfaces.C.Pointers;

package Index_Rejection is

   use Global;
      
   type Data_Array is array (Size_T range <>) of aliased Integer;

   package Data_Pointers is
      new Interfaces.C.Pointers
             (Index => Size_T,
              Element => Integer,
              Element_Array => Data_Array,
              Default_Terminator => 0);
                                              
end Index_Rejection;

However, when compiled using GNAT 3.13p on Windows 95, the compiler complains
with an error message of  'index_rejection.ads:14:32: index types of actual do
not match those of formal "Element_Array"'. On the other hand, if Size_T on
lines 8 and 12 is fully qualified as shown below, the compiler was happy. 

with Global;
with Interfaces.C.Pointers;

package Index_Rejection is

   use Global;
      
   type Data_Array is array (Global.Size_T range <>) of aliased Integer;

   package Data_Pointers is
      new Interfaces.C.Pointers
             (Index => Global.Size_T,
              Element => Integer,
              Element_Array => Data_Array,
              Default_Terminator => 0);
                                              
end Index_Rejection;

Am I correct to conclude that there is a bug in the compiler?

Anh Vo

Re: Possible Bug, Need Second Opinion

[Florian Weimer]

ANH_VO@udlp.com writes:

> However, when compiled using GNAT 3.13p on Windows 95, the compiler
> complains with an error message of 'index_rejection.ads:14:32: index
> types of actual do not match those of formal "Element_Array"'.

Further experiments suggest that GNAT performs the name resolution for
generic actual parameters in the context of the generic, and not in
instantiating context.  (Sorry, the terminology is probably
non-standard.)

I don't know if this is a property of the Ada language, but I would be
surprised if it was.

Re: Possible Bug, Need Second Opinion

[Ted Dennison]

In article <mailman.1001275761.28439.comp.lang.ada@ada.eu.org>, ANH_VO@udlp.com
says...
>
>I am nearly 100% certain that the codes below should compile fine. But, I would
>like to have second opinion possibly more before reporting it.
..
>not match those of formal "Element_Array"'. On the other hand, if Size_T on
>lines 8 and 12 is fully qualified as shown below, the compiler was happy. 

Its tempting to term anything that encourages full qualification as a "feature".
:-)

---
T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html
          home email - mailto:dennison@telepath.com

Re: Possible Bug, Need Second Opinion

[Simon Wright]

ANH_VO@udlp.com writes:

> I am nearly 100% certain that the codes below should compile
> fine. But, I would like to have second opinion possibly more before
> reporting it.

This compiles fine on GNAT 3.14a1.

-Simon