Re: bignum branch

[Eli Zaretskii]

> Cc: address@hidden, address@hidden
> From: Paul Eggert <address@hidden>
> Date: Tue, 17 Jul 2018 10:41:43 -0700
> 
> On 07/17/2018 10:38 AM, Eli Zaretskii wrote:
> > I don't understand why this is needed, given that Tom published a
> > proposed solution for the issue.  If you are bothered by possible
> > run-time penalty for LP64 systems, we can use compile-time checks
> > instead of run-time checks.
> 
> It's mostly a performance concern on 32-bit platforms.

Only on those built --with-wide-int, right?  Because otherwise, a
32-bit 'long' is just fine with GMP.

> But if nobody else cares about performance on 32-bit platforms, I'll
> stop worrying too.

I see no reason to worry about slightly less efficient bignums on some
platforms, until GMP folks get their act together.  Maybe someone
should ask the GMP developers to make those changes sooner rather than
later, to expedite the process.  After all, they should care about
Emacs, I think.