[boost] Re: Extensible polymorphic pointer

From: David B. Held (dheld_at_[hidden])
Date: 2002-08-07 18:10:43

"Philippe A. Bouchard" <philippeb_at_[hidden]> wrote in message
news:ais7bn$gl6$1_at_main.gmane.org...
> [...]
> get_type_id() gets the s_id associated with the target; it's a distinct
> number in the range of [0, no of pointed classes].  It is transformed
> into an index for lookups.

Isn't it true that get_type_id() make assumptions about the binary layout
of the class?  That is undefined, let alone non-portable.

"Philippe A. Bouchard" <philippeb_at_[hidden]> wrote in message
news:ais5tv$d00$1_at_main.gmane.org...
> [...]
> Up to now I've just tested ptr.h with pointer.cpp; only one source file.
I
> think gcc is duplicating templates & static members everywhere but
> can be prevented with #pragma interface macros.

Is that portable?

> [...]
> Let's not forget that (int *) 0 + sizeof(double) will be equal to 0x20
since
> the addition multiplies sizeof(int) sizeof(double) times.  It is so much
> easier with offsets.

I see that you might have the desired semantics with int* arithmetic, but
you convert to char* anyway, so why not use sizeof() directly with char*?

"Philippe A. Bouchard" <philippeb_at_[hidden]> wrote in message
news:ais1eq$aj$1_at_main.gmane.org...
> [...]
> If I use resize() instead, will it shrink down the allocated block
(exactly
> what I don't want) if the new size is smaller than the previous one?

Yes, it will.  Which is why the correct solution is to put a check to see if
the new size is larger than the existing size, push_back until it is the
right size.  And yes, that is probably just as expensive as it sounds.

> [...]
> ptrdiff_t is a macro, is therefore not flexible.  offset was created to
> simplify pointer arithmetics and to remove some implicit headaches...
> [...]

How is ptrdiff_t being a macro "not flexible"?  It should be defined as
the difference type for pointer arithmetic on any given platform, which
is what offset does.

Dave