Yes, I defined it once, it's not difficult to implement. The protocol was
something like:

Symbol>>asUnarySelector
Symbol>>asBinarySelector
Symbol>>asKeywordSelector

Selector>>#suggestedSelectorFor: aString
Selector>>#precedence
Selector>>argumentsSize

I defined also compare protocol methods #unaryProtocol, #binaryProtocol,
#keywordProtocol.

And then I've tried to split global SymbolTable. My idea was to separate
symbols used for message sending, and those used for description (like when
you define a Symbol for a widget name).

But I should search it because it was time ago.

Cheers,

Hernán


2015-02-04 16:28 GMT-03:00 Sean P. DeNigris <[email protected]>:

> As I sit here writing this case statement, I'm thinking "Man, it sure would
> be nice to have:
>   Selector
>     UnarySelector
>     BinarySelector
>     KeywordSelector"
>
>
>
> -----
> Cheers,
> Sean
> --
> View this message in context:
> http://forum.world.st/Real-Selectors-tp4803732.html
> Sent from the Pharo Smalltalk Developers mailing list archive at
> Nabble.com.
>
>

Reply via email to