[Q] Task Types

[Q] Task Types

[John McCabe]

Is there a means of providing an incomplete declaration of a Task type?

TIA

-- 
Best Regards
John McCabe

=====================================================================
Not necessarily my company or service providers opinions.
=====================================================================

Re: [Q] Task Types

[Tucker Taft]

John McCabe (jgm@hayling.cwmbran.gecm.com) wrote:

: Is there a means of providing an incomplete declaration of a Task type?

Yes.  An incomplete type declaration can be completed by *any*
kind of type, including a task or protected type.

Hence, the following is legal:

    type T;
  ....

    task type T is ...

: TIA

YW

: -- 
: Best Regards
: John McCabe
: =====================================================================
: Not necessarily my company or service providers opinions.
: =====================================================================

--
-Tucker Taft   stt@inmet.com   http://www.inmet.com/~stt/
Intermetrics, Inc.  Burlington, MA  USA

Re: [Q] Task Types

[John McCabe]

stt@houdini.camb.inmet.com (Tucker Taft) wrote:
>John McCabe (jgm@hayling.cwmbran.gecm.com) wrote:
>
>: Is there a means of providing an incomplete declaration of a Task type?
>
>Yes.  An incomplete type declaration can be completed by *any*
>kind of type, including a task or protected type.
>
>Hence, the following is legal:
>
>    type T;
>  ....
>
>    task type T is ...
>

Brilliant. I'll try that. The problem I was having was that I was trying 
to use the syntax:

	task type T;
	   :
	   :
	task type T is.....


-- 
Best Regards
John McCabe

=====================================================================
Not necessarily my company or service providers opinions.
=====================================================================

Re: [Q] Task Types

[Matthew Heaney]

John McCabe <jgm@hayling.cwmbran.gecm.com> writes:

> Is there a means of providing an incomplete declaration of a Task type?

The following code compiles cleanly using gnat v3.09.


procedure Test_Task is

   type T;
   type TA is access T;

   task type T is
      entry E (O : TA);
   end T;

   task body T is
   begin
      accept E (O : TA);
   end T;

begin

   null;

end Test_Task;


You'd often use access discriminants for this sort of thing:

type RT;

task type TT (RO : access RT) is
   entry E;
end TT;

type RT is
   limited record
      TO : TT (RT'Access);
      I  : Integer;
   end record;

task body TT is
begin
   accept E do <whatever> end E;

   ... RO.I ...

end TT;

Access discriminants give task type TT access to all the components in
the record.