On Tue, Feb 6, 2018 at 7:07 PM, Achim Gratz wrote: > Erik Bray writes: >> Assuming this looks good (feedback welcome) it might be nice to have >> this patch included in the next release of the official OpenBLAS >> package for Cygwin since its incompatibility with fork() has caused >> problems in the past [1]. > > It would be vastly preferrable if OPenBLAS ditched the (unfortunately > quite common) notion that "Cygwin is some sort of Windows" for "Cygwin > is some sort of Linux". I've patched out quite some bit of conditionals > like that in some other packages and it was almost always for the > better. The worst ones are those that go into the Windows conditional > branch and then on to "oh wait, but for Cygwin we need something else".
Yup--OpenBLAS treats Cygwin as "some sort of Windows" and you can see in my pull requests that there are some conditions like "if WINDOWS && !CYGWIN". I will probably try another patch to remove that notion in general. I've had a little other experience with patching OpenBLAS for Cygwin and it's one of those cases where upstream is, fortunately, receptive to supporting it as a platform. Their existing Cygwin support has just been a bit janky (a previous patch I submitted installed the DLL as cygopenblas.dll instead of libopenblas.dll, for example). Best, E -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple