On Fri, Aug 12, 2016 at 1:07 PM, Thomas Schwinge <tho...@codesourcery.com> wrote: > Hi! > > On Thu, 11 Aug 2016 12:01:59 +0300, Janne Blomqvist > <blomqvist.ja...@gmail.com> wrote: >> committed a slightly modified patch as r239356. [...] > > This breaks the build for nvptx
Sorry about that! > (but it may actually be a problem > specific to the nvptx-newlib); I filed <https://gcc.gnu.org/PR74755> > "libgfortran: build breaks if localtime_r prototype is present, but > definition is not". Hmm, yeah, sounds like something nvptx-newlib should fix? More generally, however, now that you mention it; With the new PRNG we optionally seed it from /dev/urandom, and if that is not found, with gettimeofday() + getpid() (which is why it now includes time_1.h which causes this failure). But if I understand it correctly, none of these approaches work on nvtpx. So is there some sane way to seed the PRNG in an, err, random way? -- Janne Blomqvist
