[clang] Revert "[clang-format] Annotate ::operator and Foo::operator correctly" (PR #164670)

[Eli Friedman via cfe-commits]

efriedma-quic wrote:

The recent activity here is #164048, then #164670, then a948f25?

Reverting a commit from the 18th on the 22nd is within the policy, although I'd usually prefer to give the author of the patch a little more time to respond before merging.

Every commit has a goal to fix something, but the point of the revert policy is to ensure stability for people closely tracking top of tree.  Maintaining that stability is higher priority than a general bugfix, even if it's fixing an important bug.  If you're causing immediate breakage, the right response is to back out, then reland when you have a handle on the regressions.  The fact that the bugfix is fixing a regression from a year ago isn't really relevant; keeping stability is the highest priority.

----

Not sure what we want to do for the branch; we can maybe take a bit more time there to figure out the right approach, once we know what the fix for top-of-tree looks like.

https://github.com/llvm/llvm-project/pull/164670