[clang] [Clang] Treat address-of template substitution as SFINAE (PR #208062)

[Aaron Ballman via cfe-commits]

AaronBallman wrote:

> I was not aware there were concerns about including internal issue IDs. This was just a one-off, not an AMD-wide push. I agree it is better to avoid setting that precedent here, so I will drop it.

Thank you for understanding! If something changes in the future, we can definitely discuss more at that point.

https://github.com/llvm/llvm-project/pull/208062