fix Oscar #4170 - too low precision in factoring

[thofma]

No description provided.

[thofma]

#1639 without the other commit

[thofma]

@fieker do you have the condensed example?

[fieker]

On Wed, Oct 09, 2024 at 03:08:16AM -0700, Tommy Hofmann wrote: @fieker do you have the condensed example?

Its still the degree 96 over quad. Don't know if we want to add it

…

-- Reply to this email directly or view it on GitHub: #1640 (comment) You are receiving this because you were mentioned. Message ID: ***@***.***>

[thofma]

The following is hanging for me with this change:

Zx, x = FlintZZ["x"]
k, a = number_field(swinnerton_dyer(3, x))
kt, t = k["t"]
g = swinnerton_dyer(8, x)
length(factor(change_base_ring(k, g))) == 8 # van Hoeij

P.S.: It is an example from the tests test/NfAbs/Elem.jl, which are currently hanging.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 75.85%. Comparing base (137e629) to head (2407ea8).
Report is 15 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1640      +/-   ##
==========================================
+ Coverage   75.80%   75.85%   +0.04%     
==========================================
  Files         361      361              
  Lines      113700   113720      +20     
==========================================
+ Hits        86192    86263      +71     
+ Misses      27508    27457      -51     

Files with missing lines	Coverage Δ
src/NumField/NfAbs/Poly.jl	44.41% <100.00%> (ø)
src/NumField/NfAbs/PolyFact.jl	86.22% <100.00%> (+8.62%)	⬆️

... and 41 files with indirect coverage changes

[thofma]

The NfAbs/Elem.jl test finished, but they took 1700 seconds compared to 230 seconds without the changes here.

[fieker]

I checked out the th/facccc and it terminates on my machine in 150sec - now twice

[fieker]

OK, 13sec in the previous version. I see the problem