Re: [Boost-build] BjamC++ state

From: Steven Knight (knight_at_[hidden])
Date: 2006-11-08 14:30:10

Hi Alex--

[I overlooked the original post.]

>> From: boost-build-bounces_at_[hidden]
>> [mailto:boost-build-bounces_at_[hidden]] On Behalf Of
>> Alex Besogonov
>>
>> This is a fork of SCons with MUCH better speed and some unique
>> capabilities. Waf is _very_ fast - it requires just 5 seconds to do a
>> no-op build for infamous
>> http://www.gamesfromwithin.com/articles/0506/000092.html test
>> (pure Jam
>> requires 1.5 seconds).

I want to correct one misimpression:  Waf isn't really a fork of SCons,
at least the way the term is usually defined.  "Fork" implies starting
from the same code base and creating a variant, or at least an initial
goal of creating a compatible rewrite (a la the FSF's initial creation
of GNU tools as rewrites of classic UNIX utilities).

Waf is really a complete blank-sheet-of-paper, newly-written build tool.
It started with none of the SCons code, and its configuration is not
at all compatible with SCons (and, according to Thomas, not possible
to make compatible).  Waf has some similar concepts because, like most
good software engineers, Thomas built on what came before by trying to
re-use ideas that seemed to work well and improve others that didn't.
But beyond that sort of shared inspiration, I seen no basis for the
claim that Waf is derived from or a "fork" of SCons.

Waf may or may not be a vast improvement, especially for certain
configurations, but that's for the users to decide.

My two cents,

          --SK