Compiler: Parse infix bitwise operators

[ospencer]

Even if they aren't defined yet, we should still parse &, |, ^ (tbd), <<, >>, and >>> so they can be used.

[bmakuh]

@ospencer you cool if I take this one?

[bmakuh]

Also, what's the tbd on ^?

[ospencer]

@bmakuh Absolutely!

[ospencer]

The tbd on ^ is that we currently use it as the unboxing operator. We have let mut now, so it's a little less necessary that we have syntax for that operator.

Now that I think about it though, we can totally unambiguously parse when it's an infix operator versus when it's a prefix operator.

[ospencer]

If you feel like giving that a shot, definitely give it a go. It may be a little annoying depending on your experience with parser generators (and more specifically dypgen), but we could also pair on it if it gives you a headache!

[ospencer]

Here's where we keep all of the operator precedences: https://github.com/grain-lang/grain/blob/master/docs/contributor/operator_precedence.md

[bmakuh]

@ospencer sorry just finally getting around to this. For these operators, is this what you were intending?

sigil	operator
&	AND
|	OR
^	XOR
<<	left shift
>>	arithmetic right shift
>>>	logical right shift

[bmakuh]

Also, do we want a NOT operator like ~?

[ospencer]

Yup, >> is the arithmetic (signed) right shift operator, and >>> is the logical (unsigned) right shift operator.

And yeah, we don't want ~ right now since we might be using ~ for something else 🙂

[bmakuh]

Ok, thanks. And is there any danger of ambiguity between an infix bitwise pipe operator and other uses of the pipe? I guess we're not using it for match anymore, but it came to mind after you mentioned the unboxing operator

[ospencer]

There shouldn't be! It's only used for || now (and pipeline in the future).

[phated]

This was completed by #425 🎉