Convert AssignOutParametersAboveReturnTests to the new test framework

[sharwell]

No description provided.

[mavasani]

Interesting, duplicate diagnostics need nested annotation?

[CyrusNajmabadi]

Interested in hearing what this is about as well.

[sharwell]

All diagnostics must be declared. There is no such thing as a "nested" annotation; this is just a visual artifact caused by one diagnostic's span being a superset of another.

[CyrusNajmabadi]

one diagnostic's span being a superset of another.

but in this case, the spans are identical.

1. does this indicate an issue with the analyzer (producing duplicate diagnostics)
2. if they're identical, does the test need to declare both?

[sharwell]

There is no analyzer here. The compiler reported both diagnostics.

All diagnostics must be declared. It is common for distinct diagnostics to appear in the same location, e.g. when the message differs.

[mavasani]

Tagging @CyrusNajmabadi for the product change, as I believe he added these code fixes.

[sharwell]

The results here were totally wrong. The code fix made a change in an unrelated method, with two impacts:

1. The original diagnostic was still reported in the same place, with a code fix offered
2. The iterative application of code fixes was an infinite loop, adding more incorrect code with each iteration

There was no way to port the test of local functions without this change.

[CyrusNajmabadi]

yikes!

[mavasani]

Test changes LGTM

[CyrusNajmabadi]

@sharwell because of the largely mechanical nature of the test changes, it's somewhat tedious to review. If possible, would you be willing to go through and call out anything of particular interest (in the test changes) to make the review more expedient? That way we can focus on the important bits, skim the rest, and get through a lot faster.

Thanks!

[sharwell]

📝 Prior to my change, the code fix for this produced this method:

void X()
{
    i = 0;
    char D(out int i) => 'a';
    D(out _);
}

[sharwell]

@sharwell because of the largely mechanical nature of the test changes, it's somewhat tedious to review. If possible, would you be willing to go through and call out anything of particular interest (in the test changes) to make the review more expedient? That way we can focus on the important bits, skim the rest, and get through a lot faster.

@CyrusNajmabadi I've now commented on the PRs sent to date. For future changes, I'll submit the PR as a draft, and include the annotations before marking it ready for review.