Tommy Hofmann
Tommy Hofmann
My reply from slack: It is calling the wrong function. The modul homorphisms for free modules over affine algebras are not hooked up with Singular. There is a catchall implementation...
I think we can fix it. I think @tthsqe12 provided some root functionality for composite moduli. But it seems it is not properly hooked up.
Yes, is is wrapped (somewhere). The question is now whether we always want to factor the modulus?
We could throw an error and refer to the method `issquare_with_square_root(a::nmod, f::Fac{fmpz})`? I am also happy to cache the factorization.
Yes, I think this is what we want. Can you elaborate on the > So this means that many tests will break if we are not caching internally constructed rings...
I like it. But I would be in favour of keeping `rand(ZZ, -10:10)` (forever, and not just as a deprecated method).
I think we already have the sampler/iterator and I am afraid no one ever uses them. It is very cumbersome to write `r = RandomGenerator(QQ, numerator = bla, denominator =...
Just wanted to confirm what @lgoettgens observed. In most (all?) cases I saw, the last line before the stacktrace was ``` GC: pause 38.81ms. collected 269.267041MB. incr explG25_5.jlcon [7030] signal...
I think the allocations with dynamic dispatch are due to boxing of the return value (even for immutables). Last time I checked out JET, it gave waaaay too many false...