[Cuis-dev] Behavior hierarchy type error fixes

[Juan Vuletich]

On 6/29/2020 2:28 PM, Hernan Wilkinson via Cuis-dev wrote:
> Thank you for taking the time to look at it. I was not sure about 
> sending these changes because they are more cosmetic/formal but one 
> fix (the one of #allRegularInstVarNames. I wrote a test for it)
>
> I agree with your changes, although I have comments on two of them:
> 1) The reason I made #allClassVarNames abstract in Behavior is because 
> superclass can be nil and the type checker shows the implementation 
> you left as warning. Moving it to Metaclass (where superclass can not 
> nil) solves the problem. The final result is the same.

Oh, yes. That makes sense!

> 2) I see you removed #allSharedPools from Metaclass, is that correct? 
> I mean, currently it is possible to send that message to a Metaclass.

Yes, but (as you said) metaclasses don't define sharedPools. So, if 
anyone is calling that method, they are most likely confused. I think it 
is best to remove it, and if some day we learn something else about 
this, then do whatever is best.

>
> Cheers!
> Hernan.

Thanks,

-- 
Juan Vuletich
www.cuis-smalltalk.org
https://github.com/Cuis-Smalltalk/Cuis-Smalltalk-Dev
https://github.com/jvuletich
https://www.linkedin.com/in/juan-vuletich-75611b3
@JuanVuletich