Re: Basic questions about wordings in “Interfacing Pragmas”

["Yannick Duchêne (Hibou57)" <yannick_duchene@yahoo.fr>]

Le Tue, 22 Nov 2011 20:10:01 +0100, AdaMagica  
<christ-usch.grein@t-online.de> a écrit:

> On 22 Nov., 19:15, Yannick Duchêne (Hibou57)
> <yannick_duch...@yahoo.fr> wrote:
>> > 9.1/1 For pragmas Import and Export, the argument for Link_Name  
>> shallnot
>> > be given without the pragma_argument_identifier unlessthe argument for
>> > External_Name is given.
>>
>> What is that pragma_argument_identifier ? The Convention ?
>
> pragma Pragma_Name (pragma_argument_identifier => Value, ...);
OK, it refers to 2.8. Still feel the wording could be more readable here.

>> Why no similar wording for imported subprograms ? Was just forget or is
>> this missing on purpose ? (or I missed it?)
>
> Don't understand. This is a rule for a type T and not for subprograms.
It's not about type, its about signature: imported subprogram has a  
convention too, and it should have the same legality requirement as access  
to subprogram has on the signature.

Later after a silly error in a text, I get a compiler warning about an  
imported C function returning… a whole tagged type record (oops). I wish I  
did not get a warning, but an error instead. It seems access to subprogram  
with a Convention have this requirement, I would expect the same for a  
imported subprogram with the same Convention too.

That's why I had a feeling something is missing.

-- 
“Syntactic sugar causes cancer of the semi-colons.”  [Epigrams on  
Programming — Alan J. — P. Yale University]
“Structured Programming supports the law of the excluded muddle.” [Idem]