Small fix of is_divisible (and divexact) + tests

[SirToby25]

Given an instance A of a subtype of Hecke.AbstractAssociativeAlgebra{T}, trying

divexact(zero(A), zero(A))

would result in a BoundsError instead of the usual ErrorException, due to r == 0 in the Elem.jl.

This commit fixes this an also adds two small tests.

[thofma]

Hm, I think the answer for is_divisible(zero(A), zero(A)) should be true and divexact(zero(A), zero(A)) should not throw.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 75.99%. Comparing base (83f49f5) to head (6f53a9b).
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1713      +/-   ##
==========================================
- Coverage   76.03%   75.99%   -0.05%     
==========================================
  Files         361      361              
  Lines      114377   114377              
==========================================
- Hits        86963    86917      -46     
- Misses      27414    27460      +46     

Files with missing lines	Coverage Δ
src/AlgAss/Elem.jl	88.65% <100.00%> (+0.29%)	⬆️

... and 29 files with indirect coverage changes

[SirToby25]

I agree that is_divisible(zero(A), zero(A)) should return true.

However, the following already throw a DivideError, so I think divexact(zero(A), zero(A)) throwing would be preferable for consistency.

divexact(zero(ZZ), zero(ZZ))
divexact(zero(QQ), zero(QQ))

[thofma]

Right, let's keep it consistent for now.