Fwd: Re: problem in porting from smalltalk

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Fwd: Re: problem in porting from smalltalk

Nicolas Cellier-3

> There is no point in putting expressions into a block that are
> executed exactly once. This also violates the SmallLint rule "Block
> immediately evaluated".
>
> Lukas

Thanks Lukas, one point.

So this work around is not a good idea.
Have to revert to a <pragma> thing...
3.9 image added the framework for such things i think.
Are there teams working on it and previews of code ?

Nicolas


Reply | Threaded
Open this post in threaded view
|

Fwd: Re: problem in porting from smalltalk

Bryce Kampjes
nicolas cellier writes:
 > Thanks Lukas, one point.
 >
 > So this work around is not a good idea.
 > Have to revert to a <pragma> thing...
 > 3.9 image added the framework for such things i think.
 > Are there teams working on it and previews of code ?

Nicolas,
You could just disable that optimisation in your image. It
shouldn't take more than a day including learning your way around the
bytecode compiler. I doubt that optimisation provides a measurable
gain in most situations.

Removing it yourself is simple and it'll let you implement your
project cleanly. If it's too slow for your use then think about
optimising again. Squeak is much faster than Ruby or Python which
are fast enough for many practical uses.

The clean solution is to dynamically inline message sends so the
simple unoptimised version of and: is as fast as the optimised
version. A system with dynamicly inlined messages would be much
faster for other things as well.

Bryce