Add and use zero!, one!, neg! methods

[fingolfin]

Add a bunch of one! and neg! methods. Also used zero!, one!, neg! in a few more spots. And placed these function in a uniform way into the code based (i.e. these are now always the first three unsafe operations).

This is incomplete in so far as there are still types missing a "native" neg! (and in some very few cases also one!).

Unfortunately this function is not covered by the ring conformance test suite -- we should change that (in AA).

[codecov]

Codecov Report

Attention: Patch coverage is 66.00000% with 119 lines in your changes missing coverage. Please review.

Project coverage is 87.37%. Comparing base (ef20920) to head (ab457ad).
Report is 3 commits behind head on master.

Files with missing lines	Patch %	Lines
src/flint/fmpz_mod.jl	0.00%	8 Missing ⚠️
src/flint/fmpq_mat.jl	40.00%	6 Missing ⚠️
src/flint/fq_default_mat.jl	40.00%	6 Missing ⚠️
src/flint/fq_default_mpoly.jl	0.00%	6 Missing ⚠️
src/flint/fq_mat.jl	40.00%	6 Missing ⚠️
src/flint/fq_nmod_mat.jl	40.00%	6 Missing ⚠️
src/flint/gfp_fmpz_elem.jl	0.00%	6 Missing ⚠️
src/flint/fmpz_mod_rel_series.jl	0.00%	5 Missing ⚠️
src/flint/fq_default_rel_series.jl	58.33%	5 Missing ⚠️
src/flint/fq_rel_series.jl	54.54%	5 Missing ⚠️
... and 18 more

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1869      +/-   ##
==========================================
+ Coverage   87.32%   87.37%   +0.04%     
==========================================
  Files          97       97              
  Lines       35658    36561     +903     
==========================================
+ Hits        31139    31945     +806     
- Misses       4519     4616      +97     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[thofma]

Seems to be ok. The unfortunate thing is that we would not notice if a method went missing.

[lgoettgens]

found some things. for the valuation issue, can you please check all other relative series types as well (I might have forgotten a comment somewhere).

Overall, I would really like to have Nemocas/AbstractAlgebra.jl#1814 in, but let's merge this once the comments are resolved and fix any issues later

[lgoettgens]

no idea if this is correct (seems non-trivial enough at least to deserve a comment)

[fingolfin]

Thanks for the careful feedback, @lgoettgens . I am busy with other things right now, but I plan to go over it all, and maybe use this as incentive to add a few more ring interface tests which hopefully will detect all my mistakes here, to increase confidence in this and further work.

[fingolfin]

I think this is ready in so far as that I addressed review feedback by @lgoettgens and am not aware of other issues right now. But I'd still feel better if we first had the conformance tests in place to properly check all of these. Thankfully @lgoettgens is working on that.

[fingolfin]

I made sure this branch passes the new conformance tests in PR #1880 (I.e. I did run the Nemo tests suite with this branch plus the one in PR #1880 merged, against the code in Nemocas/AbstractAlgebra.jl#1827 and it passed).

As such we could also just merge it now.

[fingolfin]

I've restarted the CI tests to verify the conformance tests pass. They do. I don't see a point in waiting for OscarCI to pass again, as they passed before and nothing changed in the code, just in the tests we run.