Re: [MiNT] XaAES modifications and clarification

[Paul Wratt <paul.wratt@gmail.com>]

2009/12/11 Vincent Rivière <vincent.riviere@freesbee.fr>:
> Miro Kropacek wrote:
>>
>> Not true. At this time, compiler is already aware of the structure
>> declaration. That declaration statement is useless.
>
> And furthermore, it works also for interdependent structures !
>
> /* This is valid C */
> struct A
> {
>        struct B* b;
> };
>
> struct B
> {
>        struct A* a;
> };
>
> However, I finally found the case where the forward declaration is
> mandatory. It is in C++, when a struct is referred to without the struct
> keyword. Without the forward declaration, the compiler can't know where that
> type come from.
>
> // This is valid C++
> struct B;
>
> struct A
> {
>        B* b; // This requires the forward declaration of struct B
> };
>
> struct B
> {
>        A* a;
> };
>
> --
> Vincent Rivière
>
Thanks, that seems to have cleared up a few others things also, nice
to know (the C++ bit as well)

Johan - thanks for the test, this will help when I get to some
HighWire code which can compile with GCC & PureC (AHCC). I cant find
your post regarding slowness of GCC, if I had dedicated hardware as
you have, I would take the following approach:

PC box (no screen or keyboard) via network, using either FTP or shared
filesystem, and shell into it to do cross-compiling.
All the editing is done local, but the back breaking is done remotely

I have finally got emails down to "between builds" (mostly) so
compiling on ARAnyM is not so bad..

An update on the enum, it appears it is only the comments which are
out of since, I think the H set are hilighted, and that that is a
separate attribute on the object, it may be something to do with the
"caret"

Cheers

Paul