limited with

limited with

[Lucretia]

Hi,

I finally got an example of mutually recursive "with'ing" working using
GCC-4.2-CVS and the -gnat05 switch.

Now my question is this; is it possible to do this without having to
use access types for the limited with'd type? I did try to make one of
my types a tagged type and remove the aliased and the 'access
attribute, but that didn't work.

Thanks,
Luke.

Re: limited with

[Randy Brukardt]

"Lucretia" <lucretia9@lycos.co.uk> wrote in message
news:1128604768.018159.141700@g44g2000cwa.googlegroups.com...
> Hi,
>
> I finally got an example of mutually recursive "with'ing" working using
> GCC-4.2-CVS and the -gnat05 switch.
>
> Now my question is this; is it possible to do this without having to
> use access types for the limited with'd type? I did try to make one of
> my types a tagged type and remove the aliased and the 'access
> attribute, but that didn't work.

You can avoid using access types in the visible part of your specification
as long as your types are tagged. (I highly recommend this for all tagged
types, period. Leave the memory allocation up to your cleints.) But you will
have to use an access type for any objects/components of the type, because
you can't declare objects in the scope of a limited with. Essentially, in
any such case, you'll need a level of indirection somewhere (the objects
being declared/allocated in the body or in a called routine). If you can
avoid objects (which is unlikely), then you don't need any access types.

                             Randy.

Re: limited with

[Anh Vo]

It is interesting to the fact that gcc-4.1 exists while gcc-4.1.0 has
not been released yet. It is currently at stage 3. Did you mean
gcc-4.0.2?

AV

Re: limited with

[Anh Vo]

I guess made a typo, too. I meant .. that gcc-4.2 exists ..

Anh Vo wrote:
> It is interesting to the fact that gcc-4.1 exists while gcc-4.1.0 has
> not been released yet. It is currently at stage 3. Did you mean
> gcc-4.0.2?
> 
> AV