cmake hype?

[Benjamin Reed]

On 2/3/06, David Faure <faure at kde.org> wrote:

> > Personally, I'd much rather we concentrated our efforts on scons/bksys.
>
> Why? I gave up when I couldn't write tests properly, and when kconfig_compiler wouldn't
> run out of the builddir (on mac os x) because apparently bksys doesn't have support for
> running uninstalled binaries (cmake does).
> The fact that any change we make now to bksys is very likely to be thrown away
> by ita's "current rewrite of everything" is also not very encouraging.

Yeah.  SCons got us 90% there on Mac OS X, but the bits that still
needed doing would require *major* crazy changes to SCons to make them
work well, from what I could tell.  SCons purports to have OSX
support, but it's really only useful in the most basic ways, KDE tends
to exercise a build system in ways that 99% of projects will never
see.

CMake is already past most of the hurdles, and looking at what needs
to be done, it appears it will be reasonably easy to extend the
existing CMake support for Mac OS X to handle the bits that are left.

If nothing else, the thing that strikes me most is, in bksys,
everything we've had to do to make things work on alternate platforms
has been layering hack upon hack upon hack.  In CMake, the path of
least resistance has generally been the *correct* way to do it in the
first place.  That gives it a huge advantage IMHO.