Josh Goebel

Yes, it seems to me like if the patch was simple and non-invasive (if it would somehow redefine existing macros, as mentioned above) - and made Wren measurably faster on...

Would a PR for this be welcome?

There has been a lot of discussion on this and proper handling of UTF-8 but if we say "we just do what C does" as a first-step I don't think...

> The world is Unicode, now, _Even if this is conceded..._ I'd be in favor of **continuous improvement vs postponed perfection**. If we agree upcase and downcase are good things...

> Unfortunately, migrating ASCII to Unicode in this case is a breaking change. I'm really hoping someone will clear up what the official policy is on breaking changes. I feel...

It would be _both an added feature and a breaking change_ for anyone depending on the initial implementations "only alters ASCII character" behavior...

> Technically true, but how many people would use .lower and hope that it doesn't lowercase UTF-8? _With breaking changes one should be technical._ I'm of course in favor of...

> that the vm can hook to call an external lib i18n lib Is there something special about string processing that warrants a new hook? (perhaps it is just that...

I think a counter inside `void expression(Compiler* compiler)` would help here (though it would likely _slightly_ slow down parsing expressions), but I'm not sure what the correct behavior is when...

I'd have no idea how to pick a limit. I just started with 1000 locally and that solved all the crashes. > It manages to use up all 32 GB...