Raise CMake minimum version to 3.16, CMake fixes, update Homepage

[a17r]

Raise CMake minimum version to 3.16

CMake 3.31 warns about compat. for <3.10 being removed in the future. Qt6
CMake modules already require 3.16 though, so it is a good base line.

CMake 3.16 was released in 2019.

Do not expand variables beforehand

See also:
https://invent.kde.org/plasma/libplasma/-/merge_requests/267
https://cmake.org/cmake/help/latest/command/if.html#variable-expansion
https://cmake.org/cmake/help/latest/policy/CMP0054.html#policy:CMP0054

Update Homepage throughout repository (misc)

[u-235]

These changes break the build.

[a17r]

These changes break the build.

I've checked all build logs and none of them are one of them is related to my changes fwics. 1) I have not initialised bundled chmlib in my fork [I hope I've found the reason there] 2) the libzip version (1.7.3 is very old) that is being attempted to be installed for macOS is too old for >=CMake-4 which seems to be part of your CI setup now (1.11.4 would fix that).

Do you have different build.log for me?

[u-235]

I agree that the CMake version needs to be bumped up, especially since it differs in various places. But I don't see the point of commit 3755074, which changes working code. I read the explanation in your links, but I think the problem is exaggerated.

[a17r]

Why is this closed? I'm just away from git access for a couple of days before I can address your comments.

[u-235]

Why is this closed?

I removed KDE support and this makes bug 964965 irrelevant.

[a17r]

This PR is unrelated to KDE support.

[u-235]

Do you still need this PR?

[a17r]

I'll get to it and rebase later.

[a17r]

But I don't see the point of commit 3755074, which changes working code. I read the explanation in your links, but I think the problem is exaggerated.

I've kept this for now, but as last commit. If you do not like it at all, you can pick the first two commits only, alternatively I can drop it in another force-push.

I have had to fix nasty issues related to the CMake variables in if() in the past, so it is definitely a thing. Sure the code works now, but this way of doing it is prone to subtle breakage when it suddenly doesn't work anymore. (this is also more in line with 99% of other CMake build systems out there, of which I have seen probably thousands)