[M3devel] M3 concerns

[Tony Hosking]

On Jan 5, 2008, at 8:59 PM, Mika Nystrom wrote:

> "Randy Coleburn" writes:
>>
>>
>>>>
>>>> LONGINT = INTEGER on Windows, which is fine as far as it goes
>>>> (everything builds appropriately), but not particularly useful.
>>>
>>> I know that it should run this way, but pickles for example wouldn't
>>> be exchangeable then, would they? Has anybody tested pickle support
>>> for LONGINT yet?
>

Sorry, yes, I was incorrect in saying LONGINT=INTEGER on WIN32.  In  
fact, they are distinct types that just happen to have the same  
representation, so the backend can handle them.

> Well, I think this depends on exactly what you (not sure who said
> this originally) mean by "LONGINT = INTEGER".  What you say is true
> if indeed LONGINT and INTEGER are the same type.  If, however, they
> are different types (as they should be), which just happen to be
> of the same size (on Windows/i386), there should be no problem.
> The Pickle library (at least ver. 2) transparently performs endianness
> reordering and word-size adjustments.  (Pickles *should* be
> exchangeable between 32- and 64-bit architectures, for example, and
> between big-endian sparcs and little-endian i386s and alphas---this
> latter property I use to exchange pickles between PPC_DARWIN and
> FreeBSD4.)  Of course, have all the proper hooks been called to
> enable this?  I don't know, I just know it's definitely possible
> to set things up so that Pickles are exchangeable even under these
> conditions.
>
>      Mika