I think this PR is too premature. There is no decision made neither in #6083 (which is originally an unrelated issue) nor anywhere else. And making a decision requires more discussion about how the compiler, compiler tools and standard library should be developed and integrated. This also touches #5215 and #4613. Until there is a plan for all of this, I don't think we need any PR.

@straight-shoota sorry I don't quite understand what you mean. The change is only about the Playground, nothing else.
People thoughts on this are quite clear, 19 👍 to move out vs 3👎 to do nothing.

This doesn't stop to have the same discussion again here of course, but I'm afraid having general discussions won't lead to any concrete actions.

Please let's talk on what the PR is about here – move out the Playground, and its dozens of related issues. One thing at a time...

@j8r I agree with @straight-shoota. Whatever the actions needed to be taken, IMHO they should be stewarded by someone from @crystal-lang core team, not outside contributors and for sure the repo shouldn't be outside of @crystal-lang org.

BTW some core members has already expressed opinions:
@asterite, @bcardiff and RX14

The creation of https://github.com/crystal/playground has been asked

We know what most people think on this, the decision is up to @crystal-lang members now.

BTW some core members has already expressed opinions:
@asterite, @bcardiff and RX14

Opinions are not actions.

The creation of https://github.com/crystal/playground has been asked

Yet not answered (yet), hence 404 on your link.

We know what most people think on this, the decision is up to @crystal-lang members now.

Amen.

The playground is intrinsically linked to the compiler internals. Putting it inside a different repository will just make refactoring the compiler harder. This is a bad idea.

The playground is what makes all those if flag openssl needed. The playground can catch up to compiler changes. It really doesn't belong in the compiler.

I really appreciate that the playground is extracted in a new repo while keeping the history.

But, this PR can't be accepted until:

1. the distribution scripts are updated to deliver the same experience ($ crystal play)
2. a ci in playground is setup to use nightly releases of the compiler to avoid issues on release
   or define brew formulas / linux packages to distribute the playground, which would lead probably to have a compiler, playground and whole bundle packages.

Also, as stated here:

I would like to see this done only and as long as the compiler is extended to let the program be instrumented in a better way as it is today. Otherwise changes to the language or the the ToSVisitor are too sensible to break the playground.

Another cons of splitting is that the bin wrapper of the compiler will need to be duplicated for the playground probably, and actually the playground will have a fully featured compiler, but is not the compiler...

I don't think we are going to invest in the required tasks mention earlier so unless those want to be tackled by someone, this leave things stalled for this PR for the moment.

I acknowledge the will to split it from the compiler, but I don't think is going to happen soon.

It could happen in the far future, not anytime soon – closing.