Null implementations

[John J. Camilleri]

I have certain abstract categories & functions which I don't want to implement in some concrete syntaxes.

I could just leave them out and get some compiler warnings, but is there such a thing as a null implementation in GF? What I mean is some way of creating a linearisation placeholder which satisfies the compiler, but in reality produces nothing. I could construct such things manually for each type, but I wonder if there is some oper which already does this?

Is this acceptable as a design pattern, or is having null implementations somehow discouraged in GF?

[peter ljunglöf]

How about this:

lin f = variants {};

/Peter

[Krasimir Angelov]

you can do something like:

lin f = variants {}

This will not generate any code in the compiled grammar and you will
avoid the warnings.

2012/5/22 John J. Camilleri <jo...@johnjcamilleri.com>:

[John J. Camilleri]

Yes thanks that works.

I was trying to use Predef.Error, which is the same type but apparently is treated specially by the compiler.

What is the difference between

lin f = variants {} ;

and

lin f = {} ;

?

[John J. Camilleri]

Never mind, got it now. Thanks.