Fix arb.neg()

[Maegereg]

Fixes #335, and adds tests for arb.neg()

[oscarbenjamin]

It looks like you have run a formatter such as black or ruff on the test code. Those formatting changes should be reverted and if we do want to use a formatter then that should be done in a separate PR.

[Maegereg]

Oops, sorry about that, I have black and isort set to run automatically on save in my editor. Fixed.

[oscarbenjamin]

Looks good. Thanks.

Can you add a release note with your name here in the README:

python-flint/README.md

Lines 162 to 175 in e8e716b

	Next release (0.9.0)...
	-----------------------
	Contributors (0.9.0):
	- Rémy Oudompheng (RO)
	- Agriya Khetarpal (AK)
	- Oscar Benjamin (OB)
	Changes (0.9.0):
	- [gh-322](https://github.com/flintlib/python-flint/pull/322),
	Add `mul_low` and `pow_trunc` methods to `fmpz_poly`, `fmpq_poly` and
	`nmod_poly`. (RO)

[Maegereg]

Can you add a release note with your name here in the README:

Done!

[oscarbenjamin]

Looks good. Thanks!

[oscarbenjamin]

It would be good to have some more testing for arb so any contribution is welcome. Mostly it is only covered by doctests and those are also quite flakey for arb given that in different FLINT versions the exact ball boundaries are not stable.

[oscarbenjamin]

For testing purposes now that this PR is merged you can install the nightly wheel with these changes from here:
https://anaconda.org/scientific-python-nightly-wheels/python-flint

If you want to add something in CI to test with these then that is a good option.