Simpler package layout for natvis files

[kennykerr]

I don't think we need the wordier file paths.

[tim-weis]

Seems reasonable.

I initially renamed the .natvis files expecting them to carry over into the PDBs and showing up in .nvlist. For some reason that didn't happen. (I'll sink some more time into investigating this when I can.)

The rationale behind moving them into a folder was to declutter the crate root. Doing that felt naive and I think leaving them in the crate root and marking them "hidden" (by convention) does a better job at that.

The only concern I have is whether this affects publishing. "Hidden" files (i.e., files starting with a .) follow special rules. Can you --dry-run a cargo publish to make sure they aren't getting lost? Not that the visualizers are stable enough to be of much help at this time, but things are moving.

[kennykerr]

I just published an update:

https://github.com/microsoft/windows-rs/releases/tag/0.53.0

Please do kick the tires and let me know if you find any issues. Thanks again for the help!

[tim-weis]

Running .nvlist (WinDbg Preview) on a re-targeted binary displays both .natvis files (named <crate>-<index>.natvis), dumping the contents produces the expected files, and dx-ing the respective expressions delivers the desired results. (VS 2022 Preview still fails to expand the expressions.)

LGTM.