Support BigFloat polynomials

[michakraus]

Add GenericLinearAlgebra in order to support BigFloat polynomials, e.g.,

> roots(derivative(Polynomial(BigFloat[0, 1, -1])^(k-1), k-2))
5-element Array{Complex{BigFloat},1}:
 0.1726731646460114281008537718765708222153959588022877212423398482923665410323274 + 0.0im
 0.4999999999999999999999999999999999999999999999999999999999999999999999999998834 + 0.0im
 0.8273268353539885718991462281234291777846040411977122787576601517076334589676855 - 0.0im
  1.000000000000000000000000000000000000000000000000000000000000000000000000000086 + 0.0im
                                                                               0.0 + 0.0im

Currently roots cannot be computed for BigFloat polynomials as LinearAlgebra.eigvals does not support BigFloat.

[codecov]

Codecov Report

Merging #297 (8b6a3d2) into master (33a45ae) will decrease coverage by 0.64%.
The diff coverage is 79.05%.

@@            Coverage Diff             @@
##           master     #297      +/-   ##
==========================================
- Coverage   88.49%   87.85%   -0.65%     
==========================================
  Files          16       16              
  Lines        1617     1663      +46     
==========================================
+ Hits         1431     1461      +30     
- Misses        186      202      +16     

Impacted Files	Coverage Δ
src/Polynomials.jl	100.00% <ø> (ø)
src/abstract.jl	23.52% <0.00%> (-1.48%)	⬇️
src/polynomials/ngcd.jl	87.13% <ø> (ø)
src/polynomials/SparsePolynomial.jl	90.43% <62.50%> (-3.03%)	⬇️
src/polynomials/LaurentPolynomial.jl	82.32% <70.58%> (+0.91%)	⬆️
src/polynomials/Polynomial.jl	95.23% <71.42%> (-4.77%)	⬇️
src/show.jl	75.75% <80.00%> (-0.29%)	⬇️
src/polynomials/ImmutablePolynomial.jl	87.24% <84.52%> (-5.56%)	⬇️
src/common.jl	91.97% <100.00%> (+1.17%)	⬆️
src/polynomials/ChebyshevT.jl	96.64% <100.00%> (+0.06%)	⬆️
... and 2 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c16d49f...8b6a3d2. Read the comment docs.

[jverzani]

Thanks, this seems quite reasonable.

[michakraus]

Would it be possible to trigger a new release soon? This would be very much appreciated as this issue is preventing me from releasing the latest developments in RungeKutta.jl and GeometricIntegrators.jl... thanks!

[jverzani]

Sure, I can get to this this afternoon.

…

On Wed, Dec 2, 2020 at 3:40 PM Michael Kraus ***@***.***> wrote: Would it be possible to trigger a new release soon? This would be very much appreciated as this issue is preventing me from releasing the latest developments in RungeKutta.jl and GeometricIntegrators.jl... thanks! — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#297 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AADG6TEQ5PNFUBWCDZ5YISLSS2Q4PANCNFSM4UA6YSUQ> .

-- John Verzani Department of Mathematics College of Staten Island, CUNY [email protected]

[jverzani]

Sorry @michakraus . This fix is going to have to be shelved, as surprisingly it causes issues with many packages when a combination of packages including this one is loaded. I'll open an issue so that we can reinstall this once that dust settles.

[michakraus]

No problem, I saw the discussion, makes sense. I can work with v1.1.13 for now and will investigate other options.

[jverzani]

I think you only need to do: using Polynomials, GenericLinearAlgebra and v1.2.0 will be like v1.1.3. At least, when I tried something with roots it all worked as desired.

[michakraus]

Yes, indeed. I updated my packages and this works everywhere. I guess we can dump this PR as there is no real need for it anymore.