Which language version should lezer grammar target?

While reading through the lezer grammar for PHP, I noticed an oddity that turned out to be a PHP feature I didn’t know about.

At some point in time, it was possible to use continue and break statements with expressions in PHP (e.g. continue $someVar;).
This is no longer allowed since PHP 7.1: Online PHP editor | output for okXui
And since PHP 7.1-4 and 8.0 are no longer supported, I thought the grammar here and here could be simplified to use an integer instead of an expression.

I didn’t report this as a “bug” yet because it isn’t clear to me which language version the grammar should target. I figured it might be desired to support the widest grammar possible and have these sort of things in a different layer, such as inspections.

I generally try to cover as much syntax as possible, so that users don’t have to worry about a mismatch between the parser and the language they are writing. If versions allow the same input to mean something significantly different, I’d add a dialect for that.

Thanks, that pretty much answers my question.

May I suggest perhaps mentioning that somewhere?
For one thing, it would be a good marketing move to claiming that CM supports PHP 8 syntax. It would also be useful for users to know that they might have to watch out for deprecated PHP syntaxes/features. (*PHP => could be any other language, really)