refactor: unify internal naming of allow_invalid_certificates

[CommanderStorm]

Which issue does this PR close?

None, because should not be mentioned in a changelog.
Is internal-only

Rationale for this change

Inconsistent between external api and internal api.

What changes are included in this PR?

Rename internal variable

Are there any user-facing changes?

No user facing or breaking changes

[kevinjqliu]

LGTM

make sense to align with the builder method with_allow_invalid_certificates

btw ci is failing due to formatting issue

[kevinjqliu]

LGTM

[CommanderStorm]

🎉

[kylebarron]

This may have broken CI on main 🥲 https://github.com/apache/arrow-rs-object-store/actions/runs/27298445894

[kylebarron]

I re-ran the CI and it passed the second time, so it must just be a flakey test

[kevinjqliu]

its green! looks like it might be an intermittent issue

[CommanderStorm]

The way in which that failed does not make sense to me as far as I undestand the rules of cargo test.

It is as if cargo test narrowed the feature set to what is actually being used in the testcase..
I checked release notes and commit logs and did not find any smoking guns of mayne a new version rolling out or something. strange.

[kylebarron]

🤷‍♂️ we'll see if it happens again! Let's just say it's GitHub's fault haha