Add handling of params collections for natural delegate types

[AlekseyTs]

No description provided.

[AlekseyTs]

@333fred, @dotnet/roslyn-compiler Please review.

[333fred]

Done review pass.

[AlekseyTs]

@333fred Please take another look.

[AlekseyTs]

@333fred, @dotnet/roslyn-compiler Please review.

[333fred]

Minor grammar comment, otherwise LGTM.

[AlekseyTs]

@dotnet/roslyn-compiler For the second review.

[AlekseyTs]

@dotnet/roslyn-compiler For the second review.

[AlekseyTs]

@dotnet/roslyn-compiler For the second review.

[AlekseyTs]

@dotnet/roslyn-compiler For the second review.

[AlekseyTs]

@dotnet/roslyn-compiler For the second review.

[AlekseyTs]

@RikkiGibson, @dotnet/roslyn-compiler For the second review

[jaredpar]

Is this meant to address #71333?

[AlekseyTs]

Is this meant to address #71333?

No. This is an unrelated change

[AlekseyTs]

@RikkiGibson, @dotnet/roslyn-compiler For the second review

[RikkiGibson]

Do we have any test for a scenario like the following? Would such a test be useful? No need to take any action in this PR, just want to have the test plan item written down if applicable.

// comp1
public delegate void Del(params IEnumerable<int> e);
public static void Test1(Del del) { }

// comp2
public static void Main()
{
    Test1(e => { });
}
``` #Pending

[AlekseyTs]

Would such a test be useful?

During an offline conversation we arrived to the answer that it would be interesting to verify whether any language version diagnostics going to be reported on the Test1 invocation in the second compilation. Since the scenario is not primarily related to the changes made in this PR, I am not going to add a test like this here. But, I will incorporate it to one of my next PRs.