On Thu, Sep 25, 2008 at 4:20 PM, mabshoff <[EMAIL PROTECTED]> wrote: > > > > On Sep 25, 4:13 pm, "William Stein" <[EMAIL PROTECTED]> wrote: >> On Thu, Sep 25, 2008 at 4:10 PM, William Stein <[EMAIL PROTECTED]> wrote: >> > On Thu, Sep 25, 2008 at 3:40 PM, mabshoff <[EMAIL PROTECTED]> wrote: > > <SNIP> > >> >> Yes, that is really the odd thing. IIRC last time we fixed this we >> >> turned some weak references into references and the problem magically >> >> went away. Unfortunately this bug seems to be a Zombie and I have no >> >> clue how to debug this :( >> >> > It is definitely a heisenbug. I just tried the test on the same machine >> > again >> > 10 times and it passes every time: >> >> > [EMAIL PROTECTED] sage-3.1.2]$ ./sage -t --long >> > devel/sage/sage/rings/finite_field_ntl_gf2e.pyx >> > sage -t --long devel/sage/sage/rings/finite_field_ntl_gf2e.pyx >> > [1.2 s] >> >> > ---------------------------------------------------------------------- >> > All tests passed! >> > Total time for all tests: 1.2 seconds >> >> WAIT! I was just testing 3.1.2, which is behind the times. >> With 3.1.3.alpha1 it fails every time the same way (see below). >> I then turned *off* caching again for finite fields and the problem >> goes away. SO, somebody turned caching from off back to >> on in a patch that goes from 3.1.2 to 3.1.3.alpha1. And turning >> caching on not surprisingly breaks things. > > The only file directly touching that file is alpha1/trac_4175-no- > impl.patch. It looks harmless, i.e. turning a bunch of cdef into cpdef > functions. So this must be somewhat deeper.
In 3.1.2, which works, caching is off for random moduli but on otherwise. In 3.1.3.alpha1 things are broken, probably because of 4175 (??), but if I turn off caching completely -- even for non-random moduli, then finite fields work again. Another option -- leave caching on but don't use weakref -- also works fine. I've posted a patch to ticket #4197 that does that, since at least it is a (temporary?) option. -- William --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
