https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123702
--- Comment #4 from Coarse Rosinflower <coarse.d1rtxxx at gmail dot com> --- Okay, I figured out what that comment meant. ASCII control characters and things like spaces make the assembler complain. However even what could be considered egregious characters to put in an identifier--non-breaking spaces, word joiners, Arabic-Indic numerals, the Right-to-left Overrides, etc.--are still completely fine to put into the assembler. All the check needs to do is make sure that *if it's an ASCII character*, it is a valid C ID character. If it's not an ASCII character (if the byte is > 0x80), it can be passed along perfectly fine.
