Re: [announcement] SYSAPI and SYSSVC for Windows

[Dave Thompson <david.thompson1@worldnet.att.net>]

On Sun, 21 Dec 2003 14:42:09 +0100, "Dmitry A. Kazakov"
<mailbox@dmitry-kazakov.de> wrote:

> Ekkehard Morgenstern wrote:
<snip>
> >> A *number of* standards, fascinating! (:-))
> > 
Aside from the "development under actual use" period, C has had two
full standards (89/90 and 99) and one Normative Amendment (95),
compared to two for Ada. Not *that* different.

> > For Ada, there's also an '83 and a '95 rationale. ;)
> 
> Ada 95 practically includes Ada 83, with very minor exceptions. Rationale is
> not the standard, it is rather comments to the standard. BTW, there is also
> AARM - Annotated Ada Reference Manual.
> 
There was a rationale in C89 (the ANSI-only standard), and is a draft
still being worked on last I looked for C99, but neither of them as
complete or IMO clear, much less integrated, as the ARM annotations.

> >> Come on! C++ does not have:
> >> 
> >> 1. multiple dispatch;
> > 
> > True, except for constructors and destructors. They're always called for
> > all derived classes of an object.
> 
> This is not multiple dispatch. MD is when, for example:
> 
> type Matrix is tagged ...;
> function "+" (X, Y : Matrix) return Matrix;
>    -- This is dispatching in X, Y and the result
> 
Or more generally types not in the same hierarchy like the "standard"
  procedure Eat (X: Animal; Y: Food);

> In C++ only one (hidden) parameter can be dispatching. All others are
> contravariant, which leads to numerous nasty pitfalls
> 
No, in C++ child method parameters are the same type. In fact, trying
to make an overrider covariant instead hides the parent method, which
seems to be a fairly common bug. The *return* type (which in C++ does
not contribute to dispatching) can be covariant, but only as a pointer
or reference; as you note there are no polymorphic/classwide *values*
in C++. The only *contra*variance is in pointer-to-member conversions,
to which Ada has no direct analog I know of.

<snip>
> >> 5. functions returning class-wide objects on the stack;
> > 
> > True, but you can return a reference to an object. (I'm not sure if
> > passing a reference to a temporary object would work)
> 
> This won't help, because of 4. The same type name is sometimes denotes
> class-wide, sometimes specific in C++. In declarations it is always
> specific, so you cannot do:
> 
For an actual object it is always specific; for a pointer or reference
always polymorphic/classwide iff the class has any 'virtual' methods
(i.e. has a vtable = is tagged and not just a record/struct/data).

> declare
>    X : Object'Class := Get_From_File; -- Any descendant of Object
> 
> To achive same effect, you have to use the heap:
> 
Or other programmer-managed storage like static, a private pool, etc.
But not automatic, which is the point.

>    X_Ptr * Object = Get_From_File (); // internally calls new
>    ...
>    free (X_Ptr); // do not forget to release it
> 
You presumably meant free(Object). In fact, officially you must use
delete iff it was allocated with new* and free() iff it was allocated
with C-style malloc-or-friends; implementations of these are not
required to be interchangeable, though they usually if not always are.
(* and even worse delete [] if the new was actually of an array, even
though that may be hidden inside the other module. Which is a good
reason to provide and use an e.g. do_delete method instead.)

- David.Thompson1 at worldnet.att.net