Re: Array of records with default values not propagating to array

[Simon Wright <simon@pushface.org>]

Niklas Holsti <niklas.holsti@tidorum.invalid> writes:

> On 18-02-03 17:26 , Simon Wright wrote:
>
>> I find that
>>
>>    type Card_Info (Trump : Boolean := False) is record
>>       Name : Ada.Strings.Unbounded.Unbounded_String;
>>       Divinatory_Meaning : Ada.Strings.Unbounded.Unbounded_String;
>>       ...
>>
>>          TDeck (Index) :=
>>            (Trump       => True,
>>             Trump_Value => Index_1,
>>             others      => <>);
>>          ...
>>
>>             TDeck (Index) :=
>>               (Trump      => False,
>>                Card_Value => Card_Index,
>>                Suit_Value => Suit_Index,
>>                others     => <>);
>>
>> compiles OK and seems to run perfectly well. Presumably this is OK
>> because the initial value for an otherwise-uninitialized unbounded
>> string is the empty string. Or maybe it's GNAT being over-helpful, since
>> it seems not to obey ARM 4.3.1(17.1)[1].
>>
>> [1] http://www.ada-auth.org/standards/rm12_w_tc1/html/RM-4-3-1.html#p17.1
>
> I don't think that ARM paragraph applies to the "others" in this
> code. The paragraph says:
>
> "A record_component_association for a discriminant without a
> default_expression shall have an expression rather than <>"
>
> but the components covered by the "others" in this code are not
> discriminants. The discriminant is the Trump component, and the
> associations for Trump have an expression (True or False), not <>.

Oh, yes.

I should have looked a bit further down: (19.1) says

   "... For a record_component_association with <>, if the
   component_declaration has a default_expression, that
   default_expression defines the value for the associated component(s);
   otherwise, the associated component(s) are initialized by default as
   for a stand-alone object of the component subtype (see 3.3.1)."

which is exactly what OP needs.