Avoid using SIZEOF_VOID_P, remove unsused SYS_IS_CYGWIN32

[fingolfin]

This allows future GAP versions to drop SIZEOF_VOID_P (although 4.12
definitely will support it).

[fingolfin]

I should point out that this check (and also the one it replaces) does not seem to fit the code it wraps 100%: I think what you really need to check is whether sizeof(unsigned long long) == 8, and then use unsigned long long instead of uin64_t below (as uint64_t may be long, not long long, which is technically a different type, even if it has the same size as long). Anyway, it probably doesn't matter in practice on most systems, and I didn't want to make a semantic change.

[james-d-mitchell]

I see what you mean, I'd be happy to accept a PR that made this change, or to make this change myself.

[fingolfin]

Those CI failures don't seem to be caused by this PR?

E: The repository 'http://www.math.uiuc.edu/Macaulay2/Repositories/Ubuntu bionic Release' no longer has a Release file.

[wilfwilson]

Thanks for the PR @fingolfin and the comment, yes the CI failure is surely unrelated to your PR.

[james-d-mitchell]

Thanks for the PR @fingolfin, much appreciated. I'm happy enough to merge this, and agree it doesn't seem to be the fault of this PR that the CI is failing.

[james-d-mitchell]

Those CI failures don't seem to be caused by this PR?

E: The repository 'http://www.math.uiuc.edu/Macaulay2/Repositories/Ubuntu bionic Release' no longer has a Release file.

Looking at this a bit further, I think this is caused by the following line:

https://github.com/gap-system/gap-docker-base/blob/f18e5899d0106e0575a7e37ee823b6f4d0196a81/Dockerfile#L130

in the gap-system/gap-docker-base repo. The failing CI job for Digraphs uses the gap-docker-master docker container. I'm not sure if this is a bug in the gap-docker-base, or not, probably I should report it there? Any advice? @fingolfin @wilfwilson

[wilfwilson]

@james-d-mitchell I really don't know the first thing about Docker, sorry. I think opening the issue that you did is your best bet. For now I'd be happy for us to merge PRs where the 64-bit/GAP master/latest packages job is the only failing one.