lua-users home
lua-l archive

[Date Prev][Date Next][Thread Prev][Thread Next] [Date Index] [Thread Index]


Makes sense. Didn't think about it this way. Thanks for the new perspective!

On Tue, Mar 8, 2016 at 1:59 AM, Karel Tuma <kat@lua.cz> wrote:
> Excerpts from Mateusz Czaplinski's message of 2016-03-08 01:08:49 +0100:
>> Those being external unofficial tools/projects has a whole lot of
>> disadvantages vs. if it was an official part of the language. That's
>> why I'm interested in them being part of the core, and that's
>> specifically what I'm asking about. And also the existence of those
>> projects (esp. Typed Lua, given its PUC roots) is why I'm even having
>> any hopes for that and dare ask at all.
>>
>> Also, as to the "non-goal" support voices: I explicitly mentioned an
>> *optional* type system, which I'd imagine could hopefully be made
>> perfectly backwards-compatible with the current approach? (as seems to
>> be the case with e.g. the Typed Lua project, no?) So I genuinely don't
>> really understand what's a problem here one would feel a need to "hope
>> against"?
>
> Given that typing on top of dynamic languages are usually used only for safety,
> not performance there is little to no benefit adding it to core Lua when we are
> talking about it's core niche - embeddability. Full blown type system would complicate
> compiler a great deal, if we're talking about something proper and not just
> mere syntactic sugar - doing type checks at parse time.
>
> Lua offers very little in terms of code safety for a reason - it's much simpler to
> implement it that way.
>
> Such a project is simply better served as a separate recompiler targeting Lua.
> It could easily surpass current Lua codebase in line count if you're wishing
> for something robust enough to allow for large scale application development.
>
> But you don't see people doing dedicated VMs for TypeScript either.
> These tools are more akin to linter, and linter should be a separate tool.
>