[fpc-other] Microsoft to ban Memcpy() :)
Marco van de Voort
marcov at stack.nl
Sat May 16 12:33:49 CEST 2009
In our previous episode, "Vinzent Höfler" said:
> > And of course, the move() procedure in FPC/TP/Delphi has exactly the
> > same problems as memcpy() as far as bounds-unsafety is concerned (it's
> > similar with fillchar/memset, indexbyte/memchr, etc).
>
> True. But in contrast to C an average programmer doesn't need it very often. In Pascal there are easier ways to assign arrays to each other than doing a raw byte copy. ;)
Less, yes.
> But I still fail to see the advantage of
>
> void * memcpy_s ( void * destination,
> size_t dest_size,
> const void * source,
> size_t num );
>
> vs. the original:
>
> void * memcpy ( void * destination,
> const void * source,
> size_t num );
>
> Time will tell, if memcpy_s() is actually "safer". If the programmer didn't think about the destination buffer's size before, why should he now? :->
That's the feeling I had too. It will only matter in a very small number of
items, where the numbers are actually calculated on the spot in different
ways.
More information about the fpc-other
mailing list