I will keep going with this on Monday.... It was my intent to make the test fail last night on platforms where we could not determine the processor count, so we'd get a good sense of how much work remains for the platforms not yet accounted for. But I messed up the test a bit.
I'll correct that, and apply this patch on Monday and we'll keep moving forward. Hopefully other folks can chime in after that if there are still more platforms where we need to use other techniques to come up with the result. Thanks for the patch! More next week, David On Sat, Nov 6, 2010 at 6:38 AM, Rolf Eike Beer <e...@sf-mail.de> wrote: > I feel like I'm taking the position the great Greg KH has in Linux kernel > development: the maintainer of crap. You write it "QNX" but you speak it > "crap". Don't get me wrong, I hate this stuff. But I have to deal with it so I > want CMake work there properly to reduce my pain. > > So here is a fix for the ProcessorCount.cmake module to work properly there. > This also makes it possible on all platforms that have getconf to detect if > _NPROCESSORS_ONLN is not supported. > > I would love to see if someone of Kitware could get into contact with the guys > at QNX. AFAIK there is sort of a free partner program where you get developer > licenses of their OS. I have two virtual machines (Linux KVM) running with > 6.4.1 and 6.5.0 which works (for some values of works). If you would go and > throw something like 6.3.2, 6.4.1, and 6.5.0 in virtual machines you could > make them build cmake nightly and test all this stuff. Sadly I can't use my > work machines for that (for different reasons). > > Eike > > _______________________________________________ > cmake-developers mailing list > cmake-developers@cmake.org > http://public.kitware.com/cgi-bin/mailman/listinfo/cmake-developers > > _______________________________________________ cmake-developers mailing list cmake-developers@cmake.org http://public.kitware.com/cgi-bin/mailman/listinfo/cmake-developers
