When is a closure a real one? [was: Real closures]

Marcus Denker denker at iam.unibe.ch
Mon Oct 9 11:38:39 UTC 2006


On 09.10.2006, at 12:54, Klaus D. Witzel wrote:

> Thank you master for making this crystal clear!
>
> I think that distiguishing block locals and also scope of locals  
> could have been solved by the old (current) compiler+decompiler,

Yes, it just needs to do a real scope analysis and understand that  
variables can be different even is the name is different.
The new compiler does this, even when in compatibility mode (thus  
generating old style non-closure code).

The problem is that lexical scoping was kind of new when Smalltalk  
was invented... so from today's perspective
it is easy to critize.

> but for the rest I don't think so.
>

Indeed.

       Marcus



More information about the Squeak-dev mailing list