Lua 5.2 released, is there a plan to use it?

[Inny]

Rather than start a new thread, I wanted to revive the existing one, for the same stated reason, "Is there a plan to use it?"

Obviously the answer is yes, but I'm curious to if there's been any increased or compelling desire for it, now that Lua is officially at 5.2.1 as of a few weeks ago. Incidentally, my estimate that it might be in 2013 was clearly overshooting, as Ubuntu 12.04 has the Lua 5.2.0-2 available (though 5.1 is still the default). I thought it'd take a lot longer to make its way up from debian-sid.

[Robin]

Not sure how much work it would be, but my guess would be: version 0.10.0 at least (and there's no telling when that might be). I think that because you would probably need to start planning the port from the previous major release, which would be 0.9.0. (Technically that is a minor release, but isn't it different for pre-1.0? ANYWAY---)

So yeah, whatever. I know nothing. But you knew that already.

[Inny]

Yeah, I figured that bartbes and co would be the best to answer that. Or if people who have been doing their own builds and testing would like to chime in with what the pain points have been, or in any incompatibilities have been discovered.

Bear in mind, I totally understand the lifecycle of programming languages, and it would be totally rational and understandable to be a couple years behind the latest version (hell, in the python world, most people are over a decade behind python3). My curiosity is if the love devs are rational and understandable, or if they like to ride and bleed on the bleeding edge.

[bartbes]

I think robin's about right, of course there's no way of telling when 0.9.0 will come out, but it's probably not going to be in there. That said, the only reason we're lagging behind is because we want it commonly available just about everywhere. To be fair, it hasn't really been a point yet, and I think that with relatively little effort we could make it run on both.. we'll see, I'll need to discuss this with the other devs.

[_ex_]

I'm using Lua 5.2 with Killa, my Lua dialect and I was supporting Lua 5.1, 5.2 and Killa on the same code base without much trouble, I have tested a bit the physics part and think that Lua sockets could give you trouble. Lua 5.2 has a break keyword without limitations, goto, and the source code is a lot more easy to read and modify than 5.1

[Inny]

To put this in the right framing, Lua 5.2 deprecates more than it adds, and may have some significant Garbage Collection impacts. I wouldn't want to see the upgrade happen because "inny wanted the _ENV feature" but rather because there was some real tangible benefit and everyone was ready for it. There's still a lot of code where I find the module keyword kick it off, and all of that would break.

[kikito]

When the time comes to work towards 5.2, I would also consider using Luajit instead of the stock Lua. Luajit 5.2.0 is in beta10 right now, so I would expect it to be stable when the LÖVE ~0.10.0 arrives. So we get more features but also "fasta machines".

[Inny]

Adopting luajit5.1 would probably be a much higher value goal than lua5.2

[dreadkillz]

I would love to see an official adoption of LuaJIT 5.1 with Love. Free speed increase can't hurt.

[Boolsheet]

Don't confuse Lua's versions with the versions of LuaJIT.

LÖVE can already be built wtih LuaJIT 2.0.0. The reason you won't see official builds is because Lua 5.1 has proven to be very stable for the last 6 years and LuaJIT 2.0.0 still has to come out of beta. Bugs in a JITed language can be weird, it's better the official builds wait until everything gets smoothed out.