Re: WIP: About CMake v2

On 29 August 2015 at 02:04, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> Christopher Browne <cbbrowne(at)gmail(dot)com> writes:
> > (Does CMake run on a VAX 11/780?? :-))
>
> Yeah. I see the two major risks as being:
>
> 1. We limit ourselves to platforms that cmake works on.
>
> 2. We lose the ability to handle weird special-case tests that are
> possible (if not necessarily pleasant) with autoconf/gmake.
>
> I might be overly worried about #2 --- but the impression I have of cmake
> is that they've spent lots of time on "make easy cases easy" and maybe not
> enough on "make hard cases possible".
>

I would agree with that, and even go further to "make easy but common cases
ugly anyway".

I converted Scribus to CMake years ago as part of porting to Windows. It
promised to be a cleaner, nicer build system. It delivered in some ways and
was rather ugly in others.

Its configuration/scripting language feels like someone thought "hey, M4 is
horrible, lets make sure we keep that it-hurts-so-good feeling in our build
system too".

Many prewritten CMake modules fail to follow basic practices that make them
work anywhere but where the author wanted/needed. I rewrote pretty much
every CMake module I used in the build. That was years ago, mind, but I
haven't seen CMake exactly taking over the world since then.

For me CMake saved lots of pain on Windows, but added pain on Linux distros
and other platforms where I had to deal more directly with package naming
differences, path differences, etc. Maybe I was doing it wrong, or maybe
it's improved since.

To me, CMake is just another build system. It has some great features and
some truly awful ones, just like most others. It does have the major
advantage of working on Windows, but I'd want to know just how much of a
hit we were going to take on other platforms. If it breaks m68k netbsd I
really, really don't care, but platforms a sane person would use with a
modern Pg shouldn't be hurt in the name of making Windows nicer if it
doesn't actually gain real functionality.

A possible appeal of CMake to me is that we could more easily support
PGXS-like functionality sanely on Windows. It *really* sucks that we don't
have anything like that now, and how hard extensions are to build on
Windows. OTOH I think it's reasonably practical to use CMake to build a
"PGXS-alike" for windows that works with existing unmodified Pg installs of
existing versions... so the win isn't big.

Anyway, we won't know unless somebody tries it. There will certainly be
> some pluses and some minuses, and we'll have to decide whether the pluses
> outweigh the minuses. I don't think we can tell that without a fairly
> detailed attempt at making it work.
>

I agree.

It's pretty much a case of "show us how it looks and we'll decide whether
to use your hours and hours of work or throw it away". Don't forget the
docs builds... unless of course the CMake conversion is also proposed to
come with a full conversion of the docs to a less prehistoric,
easier-to-work-with format. In which case you can quadruple the hours
you're gambling too.

--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services