To evaluate this, I would like to understand how the Scala 3 spec came to differ from the Scala 2 spec on this point. Who did that when, and why? Was it done on purpose, or could be accidental fallout from some sort of refactoring?

To evaluate this, I would like to understand how the Scala 3 spec came to differ from the Scala 2 spec on this point. Who did that when, and why? Was it done on purpose, or could be accidental fallout from some sort of refactoring?

The Scala-2 syntax reads

Generator ::=  Pattern1 ‘<-’ Expr {[semi] Guard | semi Pattern1 ‘=’ Expr}

I believe the intention was that the optional semicolon before a Guard should be dropped, and that caused the refactoring in Scala 3. But now I see it's not so simple, since semis are also inserted by newlines. So I think it is better to go back to Scala 2's syntax here.

@unkarjedy Can you make the change here, and also in docs/docs/reference/syntax.md, and also in the Parser? We do keep the syntax in several places. Their meaning is as follows:

• internals/syntax.md Syntax with internal trees, and experimental additions supported by the Parser
• reference/syntax.md Official language syntax
• Parser.scala Syntax serves as documentation for parsing methods.

The three syntaxes should be kept in sync.

updated

Merged the changes

As I understand there is an option to "Squash and Merge" directly in Github, so do I need to manually squash the commits?

@bishabosha What's the status of this PR? Can we merge it?