Add iroot and check parameter to root for Integer.

[wbhart]

Fixes #556

Also see Nemocas/Nemo.jl#862

This PR is the AbstractAlgebra counterpart to Nemocas/Nemo.jl#1118 however I believe the two PR's to be completely independent.

It's technically breaking as it changes the semantics of root, but should not change any current use cases, so it is weakly breaking in that sense.

[codecov]

Codecov Report

Merging #939 (4014c06) into master (39aadab) will decrease coverage by 0.31%.
The diff coverage is 86.48%.

@@            Coverage Diff             @@
##           master     #939      +/-   ##
==========================================
- Coverage   82.57%   82.25%   -0.32%     
==========================================
  Files          86       86              
  Lines       20058    20275     +217     
==========================================
+ Hits        16562    16678     +116     
- Misses       3496     3597     +101     

Impacted Files	Coverage Δ
src/AbstractAlgebra.jl	41.50% <ø> (ø)
src/julia/Integer.jl	82.22% <86.48%> (+2.97%)	⬆️
src/AbsSeries.jl	90.34% <0.00%> (-5.27%)	⬇️
src/RelSeries.jl	90.16% <0.00%> (-4.73%)	⬇️
src/generic/RelSeries.jl	95.02% <0.00%> (-2.45%)	⬇️
src/generic/AbsSeries.jl	94.11% <0.00%> (-2.44%)	⬇️
src/Rings.jl	67.85% <0.00%> (-1.84%)	⬇️
src/generic/SparsePoly.jl	25.85% <0.00%> (-1.19%)	⬇️
src/generic/LaurentSeries.jl	93.29% <0.00%> (-0.20%)	⬇️
src/Matrix.jl	91.54% <0.00%> (-0.07%)	⬇️
... and 6 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 39aadab...4014c06. Read the comment docs.

[fieker]

On Fri, Jul 16, 2021 at 07:22:05AM -0700, wbhart wrote: Fixes #556 Also see Nemocas/Nemo.jl#862 This PR is the AbstractAlgebra counterpart to Nemocas/Nemo.jl#1118 however I believe the two PR's to be completely independent. It's technically breaking as it changes the semantics of root, but should not change any current use cases, so it is weakly breaking in that sense. You can view, comment on, or merge this pull request online at: #939 -- Commit Summary -- * Add iroot and check parameter to root for Integer.

It is breaking as the current semantics is used to compute bounds... So, once merged this will need Hecke and Oscar updating. (I am not unhappy with the suggestion) Am I missing s.th.? But this defines it for BigInt, not fmpz? And the semantic for BigInt will be different than for fmpz?

…

-- File Changes -- M docs/src/integer.md (5) M src/julia/Integer.jl (68) M test/julia/Integers-test.jl (44) -- Patch Links -- https://github.com/Nemocas/AbstractAlgebra.jl/pull/939.patch https://github.com/Nemocas/AbstractAlgebra.jl/pull/939.diff -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: #939

[wbhart]

The new function does exactly the same thing as the old for any conceivable call. There was no check flag previously, and the default here is false, so.....

As for fmpz, that is handled in the Nemo PR linked above. And as you can see from the docstrings, they have identical semantics.

The difference I am talking about is a technical difference with the original semantics which were to compute the truncated integer part of the root. That's no longer part of the definition of the root function (but still happens in practice anyway).

[wbhart]

Note that the Nemo iroot branch is not expected to pass until flint-2.9 is out (not 2.8), hence the failure on the iroot CI.

[thofma]

Does the corresponding Nemo PR need flint 2.9 or is 2.8 sufficient?

[wbhart]

2.8 only.