Re: [ANN] Lua 5.2.0 (beta-rc3) now available

[Xavier Wang <weasley.wx@...>]

在 2011-6-28 晚上9:22，"Roberto Ierusalimschy" <roberto@inf.puc-rio.br>写道：
>
> > I wish with the new RC announcements would include a summary of what
> > changed since the last RC. The diff doesn't help much as it doesn't
> > give a lot of context and is full of minor formatting changes.
>
> The relevant changes were:
>
> - restrictions on labels is back to one-use-per-block. It is useful to
> be able to isolate a piece of code (e.g., a state machine) inside a
> pair do-end and be sure that its labels do not interfere with external
> labels. For those that think that homonymous labels are confusing (with
> good reason), simply do not use homonymous labels. It is easy to avoid
> and easy to spot.
>
> - ctype now tests whether the system is ASCII (#if 'A' == 65). For
> ASCII systems, it uses internal tables (as in 5.2 alpha). For other
> systems (namely EBCDIC), it uses the standard ctype, without changing
> the locale.  (So, yes, on EBCDIC systems variables are locale
> dependent...) However, for some reason the version we shipped was wrong
> (it still has the setlocales), so wait for rc4.
>
> -- Roberto
>
Hi Roberto,
Is there any plans to introduce on-stack-tuple (yes, I'm meaning sth like foo... that imply multiple value in lua stack) and lazy evaluation _expression_ (maybe with macro support, sth like macro(...) ... end, macro is just like function, except its argument is not evaluate utill they occur in the body of macro)?
Is it so interesting that I need explain the details of these proposal?
Thank you :)