You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The problem is that __dealloc__ can happen at "random" times (whenever the garbage collector happens to trigger), so it must not have global side-effects.
To the reviewer: Note that the NTL documentation explicitly says you don't need to have the correct context when you destroy an object:
Note, however, that if a GF2E object is created under one modulus
and then used in any way (except destroyed) under another,
program behavior is not predictable.
Essentially identical language occurs in the documentation for lzz_pE, lzz_p, ZZ_pE, and ZZ_p.
I fixed 9 potential instances of the problem, but only added a doctest for one of them; you'll understand why when you see how hard it is to doctest.
All doctests pass.
This is based on sage-3.3 + ReST patches, but I think it would probably apply without the ReST patches just fine.
Positive review. I did valgrind all of the Sage 3.3 doctests + this patch (while testing gsw's libSingular work) and no issue popped up. It also passes doctests on top of the ReST patches.
I should have a fix for this very soon.
Component: number theory
Issue created by migration from https://trac.sagemath.org/ticket/5340
The text was updated successfully, but these errors were encountered: