I looked in the log you posted and noticed that the build of Numpy actually succeeds, despite showing an error about xlocale.h not found. So this error was ignored, as expected. It's actually pretty common that a configuration script tests things and shows errors, only to ignore them - that's by design.
HTH Dima On Sat, May 30, 2020 at 9:07 AM Dima Pasechnik <dimp...@gmail.com> wrote: > > Note that Numpy 1.16.6 has in numpy/core/setup_common.py > xlocale.h listed in > > OPTIONAL_HEADERS = [ > # sse headers only enabled automatically on amd64/x32 builds > "xmmintrin.h", # SSE > "emmintrin.h", # SSE2 > "features.h", # for glibc version linux > "xlocale.h", # see GH#8367 > "dlfcn.h", # dladdr > "sys/mman.h", #madvise > ] > > (and the xlocale.h entry there is commented pointing at the GitHub > issue I mentioned) > I imagine that OPTIONAL_HEADERS are meant to be optional, absence of > any of them should not > create a problem. > E.g. on Debian 10 there is no xlocale.h, but Numpy builds just fine. > > I'd say it's an upstream bug/feature, not Sage's bug - but I'd be glad > to be proved wrong. > It appears one needs to debug this on cygwin, so that's your turn... > > HTH > Dima > > > > On Sat, May 30, 2020 at 4:54 AM ilyub <ilyubar...@gmail.com> wrote: > > > > Attached is the log. There are two fatal errors there: one is: > > [numpy-1.16.6] _configtest.c:1:10: fatal error: xlocale.h: No such file or > > directory > > [numpy-1.16.6] 1 | #include <xlocale.h> > > [numpy-1.16.6] | ^~~~~~~~~~~ > > [numpy-1.16.6] compilation terminated. > > > > On Friday, May 29, 2020 at 1:38:39 PM UTC-4, Matthias Koeppe wrote: > >> > >> > >> > >> On Thursday, May 28, 2020 at 2:13:30 PM UTC-7, Dima Pasechnik wrote: > >>> > >>> On Thu, May 28, 2020 at 7:10 PM Matthias Koeppe > >>> <matthia...@gmail.com> wrote: > >>> > > >>> > On Thursday, May 28, 2020 at 7:21:35 AM UTC-7, ilyub wrote: > >>> >> > >>> >> The cygwin people seemed to have removed the physical file (xlocale.h) > >>> >> from the cygwin distribution starting with version 3.1.1-1 > >>> >> https://cygwin.com/packages/summary/cygwin-devel.html. > >>> >> Because the include file is missing, sage build fails because the > >>> >> numpy module has an include of xlocale.h . > >>> >>>> > >>> >>>> > >>> > > >>> > Is upstream numpy aware of this issue? > >>> > >>> It has been done and dusted in numpy: > >>> https://github.com/numpy/numpy/pull/8367 > >>> > >>> they took Sage's patch from embray, and then Sage's patch was removed. > >>> I have no idea why this pops up again. > >>> Please post the logs. > >>> > >> > >> Indeed we need to see the logs to know what is happening on the reporter's > >> machine. > >> > >> Numpy within Sage 9.1 builds just fine on current Cygwin (with 3.1.4) -- > >> see https://github.com/sagemath/sage/runs/694846429?check_suite_focus=true > >> > >> > > -- > > You received this message because you are subscribed to the Google Groups > > "sage-support" group. > > To unsubscribe from this group and stop receiving emails from it, send an > > email to sage-support+unsubscr...@googlegroups.com. > > To view this discussion on the web visit > > https://groups.google.com/d/msgid/sage-support/f9d09f43-d327-4c78-bac4-b705848d74c4%40googlegroups.com. -- You received this message because you are subscribed to the Google Groups "sage-support" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-support+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-support/CAAWYfq0PedSS6YwYEyBBBbjjfyk-nUxdrqnrR6OyVvtGKNRFzg%40mail.gmail.com.
