[LLVMdev] [RFC] Raise minimum required CMake version to 3.0

[Rafael Espíndola]

> I don't know that it is special.
>
> I don't work on OS X at all, and I've not heard complaints from users on OS
> X about the CMake versions or problems with getting or installing LLVM. I
> have no idea what the problems are there, but if a particular version of
> CMake makes it harder to grab LLVM on OS X and start hacking on it, we
> *should* consider than when raising the minimum version.
>
> Same thing for Windows. One interesting aspect though is that Windows
> systems don't generally come with *any* version of CMake (or configure, or
> autotools or....). So there is no way to make grabbing LLVM and building it
> avoid grabbing a version of CMake. We should probably still give some
> temporal buffer so folks aren't having to constantly update CMake, etc.
>
> If Linux is special in any regard it is special in that it has a preset
> system for managing installed software which does provide CMake and would be
> more convenient to use than a manually installed CMake. That said, clearly
> any inconvenience here only impacts some of the users, not all of them.
>
>
> However, everyone seems to think I'm advocating we never move the CMake
> version forward. That isn't what I'm saying at all. What I am saying is that
> moving the CMake version forward has a cost. Not a huge insurmountable cost,
> but non-zero and I suspect non-trivial cost. As a consequence, I'm
> suggesting we do so *once we have a use case* (and I don't mean a
> hypothetical use case, but patches or planned patches) and when the merits
> of that use case make it worthwhile (I suspect they will be).

Well, there is a clear proposal for what it would be used. It seems
counter productive to ask Chris to implement a patch using 3.0
features first and then get it rejected because we decided that we
don't want to move to 3.0 after all.

It is OK to decide to move once the patch is ready, but I don't see
the point of delaying the decision.

As to way it looks to me linux is being given special treatment:

In the compiler case, we selected minimum versions of different
compilers that are easy to get on each system. In the cmake (and
python) case, things are simpler, there is more or less the same
difficulty in installing a recent version on OS X, Linux and Windows.

Linux is special in that unlike OS X or windows it is even easier to
install old version of cmake (and python), but why do have to support
that extra simplicity? If cmake gets bundled with some future OS X or
Windows toolkit, would that make the situation worse for us?

Cheers,
Rafael