Block methods splitting

[charig]

See: SOM's related issues

In Mate the situation is slightly different because SInvokable's are dynamic objects.

Relevant code in Mate are the replace method of BlockNode:

BlockNode

[smarr]

Just out of curiosity, why are blocks dynamic objects? Do you want to be able to define fields on them? Or subclass them?

[charig]

Blocks are not dynamic objects, but yes methods. So, the reference from BlockNodes to their containing method is a reference to a dynamic object in Mate. I think that is the main diff with SOM which has a reference to an SInvokable. Does that make sense?

[smarr]

Ok, right, so but then the question about methods. Why did you make methods instances of DynamicObject? I am not intending to say that it's bad, because it probably fixes some limitations and non-uniformaties (but apparently not all if Blocks aren't DynamicObjects :) ). Just want to hear what reason you had for the change.

[charig]

I do not think it is a matter of limitations. Yes, uniformity. And sticking to ST nature where everything is an object... I moved everything subclassing SObject in SOM to dynamic objects. Then I started to do the same with things subclassing SAbstractObjects. I stopped with Blocks (Arrays are different but they may be included too) mainly because of time but also I was not sure what would be the impact on performance because Blocks heavily leverage type specialization at the truffle level. Now I know there should not be a performance impact because the same should be achieved using type guards, but who knows. 2018-01-08 11:28 GMT-03:00 Stefan Marr <notifications@github.com>:

…

Ok, right, so but then the question about methods. Why did you make methods instances of DynamicObject? I am not intending to say that it's bad, because it probably fixes some limitations and non-uniformaties (but apparently not all if Blocks aren't DynamicObjects :) ). Just want to hear what reason you had for the change. — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#47 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA5gGUomt2WUFeVTncAedk0zooPdiIbvks5tIiYDgaJpZM4RUo_4> .