Re: Windows 64 port

[Paul Eggert]

On 02/29/2012 01:24 PM, Eli Zaretskii wrote:
> I see many changes that qualify existing
> declarations with `const', which cannot possibly be wrong.

It's not an issue of whether the patch is *wrong*.
It's whether the patch is *needed*.  I can think of
hundreds of places where I could put 'const' into the
Emacs code, without breaking Emacs.  But these changes
are irrelevant to the Windows 64 port, and we should
consider these changes independently of the Windows 64
changes.

It's a significant alteration to the Emacs internal
programming style to use "const" when possible
(instead of when it's required), and any such
change should be considered on its own merits.
I happen to prefer "const" myself, but I realize that
many folks consider it to be unnecessary verbiage,
and we shouldn't introduce "const" under the argument
that it's needed for Windows 64 when in fact it's
not needed for that.

> I also see
> replacements of `unsigned long' with a `size_t', which cannot be
> wrong, either.

Same thing.  Some of those replacements are needed for Windows
64, but many (most?) are not.  I addressed this point in more
detail in my previous message that I sent out a few minutes ago.