Question about using named arguments with named references about abstracttext HOT 3 OPEN

If this was only about Z56/negate, then I could probably make the parser recognize the function call in the implementation relatively easily. It doesn’t make very much sense to call an eneyj function with zero arguments, so if the parser doesn’t see any ZabcKi or Ki arguments, it can probably just guess that any other ZxyzKi key in the JSON (other than basic keys like Z1Ki or Z7Ki) is actually an argument. That works for the Z56/negate implementation, where all the other keys are indeed arguments (Z31K1...Z31K3); but that won’t be enough to support this strange “call different functions with different arguments” case, hence the question whether that needs to be supported at all.

from abstracttext.

No, I think you're asking the right question. To the best of my memory, this is not an intentional pattern. I have somewhere in my Todo list "go and clean up the if functions", because ... there was a bit of experimenting with different implementations of the if function. In short, no, this is just a mistake. A validator should have caught that.

Thanks for reporting! I'll keep this open until it is fixed.

from abstracttext.

Alright, thanks! Then I’ll just make the GraalEneyj parser look a bit harder for arguments for now :)

from abstracttext.