From: Dirk Griffioen (dirk_at_[hidden])
Date: 2007-05-15 02:29:15
I just read (a bit of) the discussion on cmake - to me it's ominous
boost is even considering this.
For me there were 2 reasons to start using boost in 2001, the libs *and*
the build system. NO more vc projects file, make files, generated
makefiles etc. But an elegant tool with a simple syntax building all my
plaforms. (And please know, at some point we wrote our own makefile
generator, but I can admit, bjam is just a better tool).
After that I started using bjam (bbv2) as my main buildtool and been
happy with it ever since:
* I used it for over 4 years in all my work/commercial/consulting and
* 'it just works'
* it forces you to structure your code well, think 'mountpoint = project'
* target definition is simple and elegant
* it nicely intergrates in a buildsystem like cerberus
(http://cerberus.rubyforge.org), on all platforms I use(d)
Some thing that worry me in the discussion:
We expend a lot of effort maintaining Boost.Build, much of which
could be avoided if we bought into an existing system rather than
rolling our own.
This is marketing speak! (And has nothing to do with how things work in
And secondly, I believe tools like VC are not in the best interest of
developers, so remarks like
CMake is about integrating well with the way the user develops code.
That might be makefiles, VC++ projects, XCode projects, whatever.
Having used CMake, bjam, and autotools, I've found CMake to be by far
the easiest system for *users* to get up and running with.
may sound reasonable, they are in fact a corruption - I apologize if
this sounds offending, but I have never seen anything good coming from
have to wade through a sea of
projectfiles,makefiles,makefile-alike-syntax,the code itself *and* a
directory structure below it all: too many universes to juggle around.
It's complex enough even if you have a 1-on-1 mapping between
buildsystem, code and directory layout; so in my view there's nothing
wrong with imposing structure on users, they might even benefit from it
(I know Id did ...).
To me it's feel backwards to try and get some 'old school' stuff like
cmake into boost. Sure, bjam documentation isnt' all great and if
there's only one guy doing the work, maybe that's the key to a solution
then: write some docs, and get Mr Prus some help ...
And I may offend people here, but everyone I ever heard or saw
complaining about bjam's difficulty (on windows or ...), well, they're
just lazy. It's not that hard to build and I think some of the remarks
by Mr Caisse, Mr Rivera, Mr Rozental and others demonstrate this.
Well, there you have it -
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk