Mixed EOL significance

[MihaMarkic]

I have to parse code where in some rules EOL matters but in most of them does not.

i.e.

lda #99 // EOL matters here
lda #100 
// also here 
#define xxx

// but it does not matter here
.for (
;
;
) {
}

so I don't hide EOLs.

The above is an excerpt from kickassembler wher code and statements are mixed.

The obvious solution would be to populate parser rules with EOL* where it does not matter, but I'd like to avoid it if possible:
for: FOR EOL* OPEN_PARENS EOL* forInit? EOL* SEMICOLON EOL* condition? EOL* SEMICOLON EOL* expression? EOL* CLOSE_PARENS EOL* unit?;
Having zillion of EOL*˛ all over parser rules doesn't look too apealling to me.

Is there a better solution?

[MihaMarkic]

This answer looks promising https://stackoverflow.com/questions/41667217/antlr4-how-can-i-both-hide-and-use-tokens-in-a-grammar