On 5/19/2010 1:06 PM, Vladimir Prus wrote:
David Abrahams wrote:
It's hard to say how soon Boost will make the transition, or even if it will. I'm trying assiduously not to tie the fate of Ryppl to Boost's use thereof. However, I predict that Boost *will* make that transition
I would suggest to wait with any such predictions until you can demonstrate that whatever new setup you propose can build all of boost, run all the tests, get exactly same results on those tests and further, and is actually maintained for a few releases and is actually better.
I should remind about the boost-cmake project, which was announced with great noise, promoted to a state of an almost-official component, got mentioned at the top of release announcement for a certain release (*), got featured in marketing materials of certain company and then ended up being totally unmaintained.
It's not totally unmaintained, but your point is a fair one and I'm surprised nobody responded. Before we could consider switching to something like ryppl and cmake, it needs to be proven *and* there needs to some folks (plural) committed to maintaining it long-term. To that end, I'm throwing my hat in the ring. I'll be moving to Boston for the summer to help Dave turn ryppl/cmake into a viable alternative. Dave and I are both here for the long-term and are deeply invested in making Boost just work. Also, KitWare has donated one man-month over the next year to helping make this happen. They'll add any features to CMake/CTest/CDash/CPack we deem necessary. -- Eric Niebler BoostPro Computing http://www.boostpro.com