Grammar extension: catching them all about textx HOT 2 CLOSED

Thank you for this interesting suggestion.

• From an end-user point of view it is a valid extension.
• From the implementation-of-textx point of view it should be carefully analyzed what kind of variations are possible and allowed and what are the consequences of such a change (e.g. is it also allowed to mix references and non-references: Rule: x=(A | [B]);?).
  • @pvmm Maybe the following simple alternative is also acceptable for you? Define a new rule FactionOrFloatOrInt: FRACTION | FLOAT | INTEGER; and use Number: value=FactionOrFloatOrInt;.
  • @igordejanovic Maybe such a transformation could be automated, if we include such an extension.

I think it is not a trivial change...

from textx.

@pvmm Thanks for the suggestion.

I agree with @goto40. It would be a non-trivial change. Furthermore, if we allow for arbitrary expressions in the RHS of the assignment it would be hard to deduce the meta-model. What would be the meta-class of the RHS for complex expressions? On the other hand, if allowed expressions are somehow constrained it would be difficult to explain those constraint to the end-users. Now it is very simple. You can use only rule references and plain matches (with repetition modifiers for repeatable assignments). I think the way it is now is the best trade-off between the expressive power and the implementation complexity.

I'm closing this issue as wontfix but feel free to reopen if you want to discuss further.

from textx.