Fix zsuper with args named with js reserved word

Fixes #2361

GHA: Test head Ruby

Revert "Fix defining multiple methods with the same block"

Fix multiline x-string in a case assignment

This reverts commit 3eacd508eb55daa34198330bf6820579428d845d.

Docs & opal/platform fixes

Also remove requiring "native" which is not used here.

Also require nodejs/file so that IO.read and friends won't break.

It's been deprecated for a while.

It's very outdated and not working, no need for a deprecation cycle.

Opal v1.4 alpha release work

Ruby 3.1 support branch

Fix opalopal <<~END support.

Opal-in-Opal workshop

* Separate Enumerator subclasses into separate files and autoload them.
* Introduce Enumerator::...

* Enumerable#tally to support an optional hash argument for accumulation
* Enumerable#each_{cons...

* Fixes #2169
* Fixes #2363

Avoid using String#slice! in the dedenter and expect a new string f...

Optimize truthiness/falsiness checks

They'll be patched in a following commit, this import is there just to
maintain a clean diff.

* Load corelib/main before corelib/error
* Correct an edge case where errors weren't readable at...

Second batch of generated code optimization

=== Files: benchmark-ips/bm_truthy.rb
bundle exec opal --dynamic-require ignore --missing-requir...

```
Comparison of the Asciidoctor (a real-life Opal application) compile and run:
...

This came from my cooperation with Elia on writing extensive benchmarks
for how to write the $tr...

The ! and != calls happen quite often. Let's also optimize them akin
to how we optimize == and =...

Are easier to read and still manageable with and editor supporting
multicaret changes.

Since we don't have a `case` node anymore, we can make an `if` node
generate code better.

This ...

This is very simple. I found a few places where we can make the
arguments handling boilerplate a...

Sometimes we don't care about the return value except of if it's
truthy or falsy.

Let's conside...

We used to generate helpers for $rb_plus, $rb_minus etc. for every top.
Now we generate them in ...

This may not be that important, but there was a `js_falsy` method
inside a compiler which isn't ...

Same with `===` -> `$eqeqeq`. This is so that we can safely cherry-pick case
node elimination. S...

This is a small patch, but as we free the temporary variables for the or
rewriter, we can also f...

This is a backport from an earlier pull request, but this time it makes
a lot of sense, because ...

Fix io puts

Copied and adapted some specs from rubyspec as Benchmark IPS was
messing up some output when pri...

Passing nil as the first argument was resulting in an obscure JS error.

Move aliases to the end of classes

We had a loose policy in the corelib to keep method definitions
alphabetized, that doesn't work ...

Spec updates chore