On Sun, Apr 29, 2018 at 2:29 PM, Timo Kaufmann <[email protected]> wrote: > Hi, > > I'm currently in the process of upgrading nixos's sage package to 8.2. While > doing > that, https://trac.sagemath.org/ticket/18919 resurfaced: > > ``` > alarm(9/11); (2^100).binomial(2^22, algorithm='pari') > ``` > > sometimes hangs. "sometimes" seems to be whenever the system is under load. > > This behaviour is tested in src/sage/rings/integer.pyx. This also happens on > 8.1, > so it shouldn't be cause by the update. Its probably a packaging issue. > > However I'm hoping that somebody more familiar with sage development than me > has > an idea what might be causing this before I spend more time blindly > debugging it. > > The bug occurs with cysignals 1.7.1 (and also 1.7.0) and the same pari git > revision that the sage spkg uses. The versions of the other direct > dependencies > are here: > > https://gist.github.com/timokau/a4ca039811de43e849b1e147e43ef2f3. > > Does anybody have any pointers to potential causes?
Hi Timo, I haven't tried this test yet myself and am not sure about how it's mis-interacting with the binomial method or with PARI. But it might be something similar to this: https://trac.sagemath.org/ticket/24986 , Not the same, but still a case of a signal being handled at some point that causes some memory assignment to go awry and data structures to be corrupted. Just as shot in the dark though. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
