Re: Guile 64-bit Windows support, redux

[Mike Gran]

> (The following response is a bit muddled -- it initially says I don't
> follow but later I will.  Still, there is some stuff of which I really 
> disagree with the implementation.)

> I don't quite follow why the change from 'long' to 'intptr_t'.  I mean,
> reading the commit message you linked to:

Hello Maxime,

Thank you for the review, and I'll go through at some point, but, for now
I've restarting a new patch set based on what I've learned before and
with a higher level of formality. I'm starting
with the easy parts and will leave this numerical part until the very end.

So it may take some weeks to form a proper reply.

Fundamentally, Guile has a habit of putting pointers into longs, so that's
really all that needs to be fixed.  But is ends up being more complicated
that it sounds.

But when I get back to it, I'll try to be more precise
1. long or long long - an integer that is "big" in some general sense
2. int64_t - an integer that needs to be 64-bit because may hold a large integer
3. intptr_t - an integer that holds a pointer

I saw Ludo in chat and he said that it was important to not change the ABI,
so I'll need to create some tooling to ensure that everything stays
consistent as I progress.

>> 
>> [1] Last two times I volunteered myself for something on Guile/Guix,
>> I got quite ill. Let's not hope for a repeat.

> Ill, like in, bad social interactions, or like in independent medical 
> issues that happened at the same time as working on Guile/Guix stuff?
>I don't need to know per se but this footnote is ambiguous.

Oh sorry, I didn't imagine it could be read that way. Medically ill.
I recommend against getting old, if you can avoid it.

Regards,
Mike Gran