Hi Andrea,
On Mon, 2014-01-13 at 22:03 +0000, Andrea Faulds wrote:
>
> On 13/01/14 21:57, Yasuo Ohgaki wrote:
> > What's the reason not to introduce larger int for 32bit platforms?
> > Too much performance impact?
>
> I've wondered this myself. It makes me also wonder if it's the case
> using 64-bit on 32-bit platforms is a little too slow, couldn't you just
> transparently store 32-bit and then switch to 64-bit if necessary?
>
thanks for the idea, but it might be much more complex as one thinks. If
one has to switch between 32 vs 64 bit in zval, it screams for API
breach. For example with functions accepting integers by reference.
Anyway, just a speculation right now :)
Regards
Anatol