On the BKsys webpage you can read now: I was then encouraged in 2005 to start porting kdelibs to bksys (the project has about 1500 cpp files). Then the following issues became obvious:
* Scons does not scale for large projects - startup time used to take more than 50 seconds on fast computers for only 1/3rd of kdelibs. And that amount of time becomes only bigger as source files are added, showing a nasty non-linear trend.
This is exactly the problem that we encounter on WengoPhone using SCons.
Finally The kde developers were in a hurry and decided in February to use Cmake instead of bksys/scons. [...]
Last November, I have decided to start an experimental branch of scons called "waf" [...] I ended up rewriting scons from the ground, starting a brand new branch.
[...]waf should be ready for production in mid-2006
# Explications de la part du developpeur de BKsys
Posté par tanguy_k (site web personnel) . En réponse au journal CMake dans KDE. Évalué à 7.
http://www.freehackers.org/~tnagy/bksys.html
dommage le lien est down actuellement.
http://blog.openwengo.com/index.php?/archives/15-SCons,-the-(...)
On the BKsys webpage you can read now:
I was then encouraged in 2005 to start porting kdelibs to bksys (the project has about 1500 cpp files). Then the following issues became obvious:
* Scons does not scale for large projects - startup time used to take more than 50 seconds on fast computers for only 1/3rd of kdelibs. And that amount of time becomes only bigger as source files are added, showing a nasty non-linear trend.
This is exactly the problem that we encounter on WengoPhone using SCons.
Finally The kde developers were in a hurry and decided in February to use Cmake instead of bksys/scons. [...]
Last November, I have decided to start an experimental branch of scons called "waf" [...] I ended up rewriting scons from the ground, starting a brand new branch.
[...]waf should be ready for production in mid-2006