3.6.4/issue/22692 unused import

[som-snytt]

Fixes #22692

If the explicit name is different from the symbol name, then it must have been a rename on import.

That is never satisfied by a wildcard selector.

[WojciechMazur]

Open CB stats

Scala Version	Unused Warnings
3.6.3	1079
3.6.4-RC2	1191
3.6.4-RC2 + this PR	1194
3.6.4-RC2 + revert #22314	1070
3.7.0-RC1-bin-20250227-1b0be01-NIGHTLY	7378

So it seems like it fixes the linked issue, but at the same time the total amount of unused warnings has increased since 3.6.3.
Is it a good change? Hard to say without investigating them in the context of source code.

If anyone is instrested in the details warnings I've uploaded them as gists (it contains all warnings, not only Unused Warnings)
3.6.3 - https://gist.github.com/WojciechMazur/48ff0e124ae6862a8d139e5f7592b216
3.7-nightly - https://gist.github.com/WojciechMazur/c827f11d9f72ad67c9b79c4f39ee9f3b
3.6.4-RCs - https://gist.github.com/WojciechMazur/aadb545954190b81f7340f81dc79897e :

• this PR results are stored in 3.6.4-RC2-20250304-4c0ddac.json
• the revert of 22314 in 3.6.4-RC2-20250303-fce9437.json
• 3.6.4-RC2 in 3.6.4-RC2.json

[som-snytt]

I will take a look, understanding that it may not pay off. If it is a fool's errand, I'm the fool for it.

It may not be worth the uncertainty in any case, of course.

[WojciechMazur]

We've just discussed this topic internally and decided we'd release the 3.6.4-RC2 as 3.6.4 (final) without this change or revert of the PR that caused the regression in the RC3 release.
As the discussions proved the issue is non trivial to solve, each approach has it's own pros and cons

Thank you for investigating the issue and creating a PR!