I noticed something while writing the last two sections on Integer literals and Floating-point literals, and I’m curious if anyone else noticed.

There’s no way to express a negative integer or floating-point literal!

If we look specifically at the definition of an decimal literal in EBNF format, we see:

decimal_lit    = "0" | ( "1" … "9" ) [ [ "_" ] decimal_digits ] .

Notably, we do not see this:

decimal_lit    = [ "+" | "-" ] "0" | ( "1" … "9" ) [ [ "_" ] decimal_digits ] .

Does this mean Go doesn’t support negative numbers? Well, obviously not. What use is a programming language without negative numbers?

Does it mean it doesn’t allow assigning negative values directly to constants or variables? No, that’s easily disproven as well.

So what gives?

The + and - signs are interpteted by Go as Unary operators (there are a few others as well).

So what that means is when you write

var myInt = -30

You have not written a negative integer literal (-30). Rather, you’ve written an integer literal (30), preceeded by the unary operator - which negates it value.

Remember that one for your next Go Pub Quiz night!