>>13C standard
Nobody cares about your primitive caveman bytefuck languages where you can't even use things before defining them.
Besides, you haven't even read what is linked in
>>8, otherwise you would have concurred that I'm right:
Tokenization
In the third phase of translation, the preprocessor tokenizes the source file into preprocessing tokens and sequences of whitespace characters. Comments are placed by a single whitespace character
Tokenization is another word for lexing (do I need to explain that more or can you follow at least this far?). Thus, comments are stripped out only at the lexing phase. Which is a part of compilation (because without lexing there would be no parsing and thus the compiler wouldn't "see" any actual code - do I need to explain that more, or is your pea-brain powerful enough to understand English?