Yes, it's in the forecast since I bought A7 in summer 2007 (together with enums and unions). I guess a major compiler update will take at least another three years?

I can understand that you don't want to touch the compilers code because it's a really critical part and adding things in one place can destroy completely other places.
But couldn't you create a fork of the current compiler and create an experimental branch to test things? Users could then decide which compiler they want to use to test if the experimental compiler breaks the compatibility with the existing one. Looks like a win-win situation for me, you don't lose compatibility and new compiler features could be deployed faster.

Oh and about the __FILE__ and __LINE__ macros, do they also need a major compiler update? Because imo it's just a search and replace operation and could be done in the compiler frontend