> PS Some bright person might want to code up some RB find templates to > look for methods > where if you use the closure VM it will possibly give you different > behaviour than with the > current VM.
Wouldn't it also require to recompile all methods to get closures? Have a look at the senders of #fixTemps. I have 37 senders in the latest Pharo image. These are all places where people had to work around the shortcomings of the Squeak blocks. With block closures enabled these blocks should behave the same without fixing their temps. For more examples the Seaside code base has a vast collection of places where it has to work around missing closures. For portability reasons this is done using #fixCallbackTemps. > Unless of course everyone feels comfortable in turning block closures > on, and assuming nothing > will change? It could be that some code breaks, because closures behave subtly different. However if people depend on such code, then it is ambiguous and buggy in the first place. I think that closures should be included into mainstream VMs and images as fast as possible. It is a **major** annoyance that Squeak has no full block closures even in 2009. Lukas -- Lukas Renggli http://www.lukas-renggli.ch _______________________________________________ Pharo-project mailing list [email protected] http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
