Proposal: Mixins? about jsonata HOT 6 CLOSED

It isn't entirely clear to me what the semantics of what you propose would be. Is there some reason you can't simply bind this function and use it in your expressions directly?

from jsonata.

In my understanding a mixin is supposed to replace the standard code without altering the code itself. The purpose is to have custom extensions to standard code without having to worry every time the standard code gets upgraded.

If I changed something inside jsonata, then I would have to modify every time I upgrade. This concept is supposed to avoid that propblem. So a simple function can't do it since I want to change standard code.

Hope this is clear.
Martin

Example use in jsonata:

    function evaluateName(expr, input, environment) {
        // lookup the 'name' item in the input
        var result;
        var mixins = environment.lookup('mixins');
        
        if(mixins && mixins.evaluateName) return mixins.evaluateName(expr, input, environment);

If mixin exist, it is called instead of the default implementation.

usage:

        expression.evaluate(input,{},(rej,res)=>{
          if(rej) reject(rej);
          resolve(res);
        },mixins);

from jsonata.

I think I understand what you mean here. I had some similar use cases which is what brought me here. I think the best approach here would be to actually implement some kind of API for developers to override the various tree walking functions. I've initiated a discussion with @mattbaileyuk and @andrew-coleman about this to get their feedback. If they are not opposed to the idea, I could write a prototype of what I have in mind.

from jsonata.

Great! Wonderful idea! A central approach would be fantastic.

I also have a minimal approach that works for my limited needs. In fact, that is the one you see above. Might be useful as an inspiration.

Many thanks for your effort!

from jsonata.

@Allam76 Glad you like the idea. No need to thank me for my effort since I haven't actually made any effort yet. That's going to depend on the feedback I get from @mattbaileyuk and @andrew-coleman. 😃

from jsonata.

Closing this issue. Sounds like @xogeny's implementation might address this which will hopefully be the basis for the next major version of JSONata.

from jsonata.