Re: [Boost-build] Boost.bjam

Subject: Re: [Boost-build] Boost.bjam
From: John Bito (jwbito_at_[hidden])
Date: 2009-08-11 14:13:13

Boost.Build is actively maintained and bona fide bugs are addressed
swiftly.  It's likely the easiest builder to get working reliably for
a non-trivial C/C++ project (possibly also other languages - I don't
have any experience with the others that appear in the set of tools),
and almost certainly the easiest way to create such a build that will
port without change to a variety of platforms.

The drawback is that it's not easy to build things that aren't
libraries (from compiled sources) or executables because the language
and runtime are arcane.  The language issue is being addressed via a
rewrite in Python.  I'm not sure what could be done to make the
runtime (generators, types, targets, toolsets) more accessible, though
diagnostics (like a visualization of the dependency digraph) might
make a big difference.

One other drawback is that it can be slower than a conventional make
on a large project because it evaluates include dependencies
dynamically.  Consequently the build is more reliable than makefiles
that don't check for dependency changes, but it's also not optimal
because it determines the dependencies every time and not only when
the files change.

My recommendation is to look at the tools that are included and if
your project doesn't need anything else, go for it.  If you'll be
doing something like Java, talk to me in a week or so.

Good luck!
John

On Tue, Aug 11, 2009 at 10:46, Anant Rao<arao_at_[hidden]> wrote:
> Hi,
>
> On
> http://en.wikipedia.org/wiki/Perforce_Jam
>
> I see this line
>
> "... Jambase is notorious among Jam users for its bugs and the infrequency with which fixes are integrated into the distribution, though few bugs are critical...."
>
> Does this affect boost.bjam?
>
> Thanks,
> _______________________________________________
> Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost-build
>