testing get_git_username using git-testtools for #700

[atluft]

Trying to add a test for a function that uses a git repo created specifically to test the function. In this case the repo has a local configuration w/ the "committer.name" and "committer.email" set to known values.

Other typical values are added to the repo too. Like "user.name" and "user.email" and "author.name" and "author.email" to ensure those aren't selected as the correct values, although there are cases where those would be correct.

Interesting discovery is that "committer.name" won't be selected if "committer.email" isn't present too. Not sure that is an issue to pursue or not.

Testing:

cargo test -- --nocapture

[atluft]

Hi @o2sh and @spenserblack,
Please advise about the code coverage failures below, is the new test failing or is it that code coverage went negative by adding more code into measurement, or something else :-) ?

[o2sh]

I suspect a recent update of tarpaulin to be the cause for the drop:

Indeed, we always pull the latest version when generating the codecov report.

- name: Install cargo-tarpaulin
  uses: actions-rs/install@v0.1
  with:
    crate: cargo-tarpaulin
    version: latest 

This is just a supposition, I haven't tested it locally.

[spenserblack]

🤔 Looks like this file was (correctly) converted to LF line endings.
For the sake of a reasonable diff, I think we should convert all files to LF so that it's easier to compare this and #813

[atluft]

I see working on converting in #815
Oops, did I do that? Shoot, I really was trying to keep it CR LF, I guess 2 wrongs do make a right!
It is crazy that in 2022 we have to think about line endings, maybe in another 100 years this will be resolved.

[spenserblack]

Oops, did I do that?

You correctly did that (See .editorconfig file) 😆

I incorrectly set up an EditorConfig and never enforced its settings in a workflow.

[spenserblack]

Thanks @atluft! If you feel comfortable rebasing onto main then can you please do so?
There's a good chance that this might be squash-merged, so a normal merge commit is fine, too.

git checkout --ours and git checkout --theirs are useful when resolving these full-file conflicts.

[atluft]

@spenserblack
I think this PR is in good shape now.
git rebase main, correct conflicts, continue, push force at end.

squash is a wonderful feature!

[atluft]

Discovered the "Ready for Review" button!

[spenserblack]

LGTM, thank you for helping us with testing! Leaving one minor nitpick.