Print (foo^)##bar with the right precedence

[anmonteiro]

fixes #1554

[IwanKaramazow]

method simple_enough_to_be_lhs_dot_send ?(except_unary_postfix=false)
The except_unary_postfix name arg feels very specific/specialized. Can we pass the relevant expression down to make it a bit more generic/extendable in the future? There might be other cases.
I might be totally wrong though, this is just an idea.

[anmonteiro]

I agree it feels too specialized. I'm definitely open to other ideas, but the point here is that in the presence of a SHARPOP we want to formatPrecedence of UnaryPostfix.

Maybe the optional argument is fine and we just need to change its name to sharpop?

[IwanKaramazow]

random idea, what about passing ?infixOp as some kind of context?

[anmonteiro]

@IwanKaramazow check my latest commit. I replaced except_unary_postfix with an infix_context that we can pattern match against.

I liked your idea, and I believe this might prove to be more extensible in the future if we wanna tweak other infix ops.

What do you think?

[IwanKaramazow]

Feels like a nice step into the right direction :)
If you want to avoid the extra optional allocation, you can use the empty string as default value ?(except_unary_postfix=""). The current approach is fine, won't make much of a difference here.

@chenglou ready for me, merge approved.

[anmonteiro]

I thought about the empty string but it didn't feel very nice to me. I'll change it if you'd like though

[IwanKaramazow]

No, not necessary, option makes sense here

[chenglou]

Thanks