Should I add rest parameters? If so, should I rely on there being a slice method like in
#1473 (comment)

🥳

Parser test parser/continue-on-error.ts was failing, now fixed

That's weird, it seems the features test failed because of the nvm GitHub action.
cc: @dcodeIO

please sync with main

@MaxGraey,
Are you talking to me? Is there any trouble with merging this? GitHub is telling me that there are no conflicts :)

Yeah, no need, that's why I hid that message. By the way, have you seen the other dcodeIO suggestions?
#2008 (comment)
#2008 (comment)

Thank you! I completely missed those. I'll implement those tomorrow

Could you also add destructing tests for typed and static arrays?

let arr = new Int32Array(3);
arr[0] = 1;
arr[1] = 2;
arr[2] = 3;
let [a, b, c] = arr;
assert(a == 1);
assert(b == 2);
assert(c == 3);

BTW, for the next PR I think I'm going to replace binding patterns with array literals for more flexible parsing and for normal assignments and declarations to have more similar code

@romdotdog It seems some tests have failed

I like the idea of reusing array literal expressions here, would be cool if it could be done right away :)

I like the idea of reusing array literal expressions here, would be cool if it could be done right away :)

It's an "issue" that's only worth fixing when further developing array destructuring; I don't think this base PR is worth delaying more for something that can be included in the next PR, if that makes sense (:

BTW, I couldn't reproduce the failed compiler test on my machine.
Not to mention that it failed without me making any changes to the compiler test or the compiler itself.. Maybe someone could point me in the right direction?

EDIT: Thanks to MaxGraey, I was able to reproduce and make new fixtures after pulling from upstream.

Had some extra time, pulled from upstream and squashed the resulting bugs.
I also removed the custom BindingPatternExpression and replaced it with array and object literals, as requested.

I hope this can be merged soon!

Closing for reference after the talk on Discord about the .identifierNode 'hack' used throughout this PR, view by joining the Discord server