Re^6: Perl 5 Optimizing Compiler

[Corion]

IMO, not every rant needs to be published, especially when it does not contribute positively to the issue at hand.

On the other hand, it doesn't help much to claim that certain things cannot be discussed on p5p without any explanation as to why when it's not obvious that previous discussion has taken place.

Still, I do prefer that silence over rants and flamewars that do nothing to improve the situation.

[rurban]

I have now 4 big p5p performance rants stored under ~/Documents/.
But I'm working on an official non-rant design draft https://gist.github.com/3740313 for the most common optimizing possibilities. Problem is that people are already implementing parts of it without knowing the internal problems behind. So e.g. a function signature implementation makes no sense without lexicals const's ("is ro") being worked out before. Not talking about "is bind".

Nobody knows about perl6 and perl5i. Nobody is talking to them.

class/methods need to define the immutable/final semantics first. And I fear the decisions will be wrong. At least Stevan is a clever person not rushing things.

[chromatic]

Nobody knows about perl6 and perl5i. Nobody is talking to them.

Ahem.

[rurban]

Yes.

It is very apparent reading the p5p threads on these matters. Esp. the named arguments (again) threads, but also general comments.

Damians perl5i proposals are even offline and only readable via archive.org. Small parts, isolated and destroyed, appear under schwerns project.