Fork Irrlicht

[ThomasMonroe314]

Ok, so before you go all crazy saying, "NOOOOOOO" and "WHAAAAAA?" and "not gonna happen." Caleb and I were talking about where Minetest was headed, where it should go, what needs work, etc. We started talking about the possible replacement of Irrlicht and how much work would go into that, then we had a crazy, but possibly ground-breaking idea(and no its not a shovel ;) ). What if, instead of replacing Irrlicht, we make a customized version of it and use that for MT instead of trying to replace it. It has already been proven to work, just take a look at SuperTuxKart. Irrlicht had what they wanted, just they tweaked it a bit to fit their needs. By having a custom rendering engine just for MT, you guys could create the tools necessary for adding, say shadows. Improve what Irrlicht doesn't have instead of limiting yourself to its own boundaries

[ThomasMonroe314]

Not only would it allow for the above possibilities, but it would also be isolated from MT, so it would not affect the structure of the code at all.

[Calinou]

I don't think the gameplay and general user experience will be improved by spending dozens (if not hundreds) of hours working on an engine fork. Not to mention that nobody other than us would use our fork, making it not very useful to the general public.

[rubenwardy]

Waste of time for not much gain :-1:

It would be better to look into Antarctica or Build a World's fork (which they released as FOSS)

[rubenwardy]

5365

[paramat]

:-1: for reasons above.

[ThomasMonroe314]

@rubenwardy OK, that would be a little of what I'm talking about. @Calinou the game's "play" might not be improved much, but it would open up a lot of possibilities for "eye-candy" type stuff, like shadows and shader packs(poor example, ik). I understand that at this point in time you guys are all busy with 0.5.0 coming out, so something like this may not sound like it would be worth it. But think about after 0.5 comes out, most of the bugs have been fixed,the question is what comes next? more minute features? or is MT going to improve in a way that is fairly noticeable by the players? Sorry for being blunt, I'm not very good at tip-toeing around controversial subjects.

[ghost]

:-1: it will eat much time for maintaining. So, it's easier to switch to another 3D engine. /me crazy saying "NOOOOOOO", "WHAAAAAA?" :rofl:

[nerzhul]

No need to fork irrlicht, we just need either use it more, and enhance client intelligence as now it's just a stupid client, and CSM is here for that, or change the engine, but it doesn't change that the client is a stupid box. Another point, client rendered and event handlers should be on separated threads

[ThomasMonroe314]

@nerzhul ok, good point.

[rubenwardy]

Shapes and shader packs is already possible with irrlicht

[paramat]

There already is plenty of potential for eye-candy, we just lack people working on it.

I understand that at this point in time you guys are all busy with 0.5.0 coming out, so something like this may not sound like it would be worth it.

Limited core dev time will continue after 0.5.0 is released, 0.5.0 isn't making much difference to how busy we are.

But think about after 0.5 comes out, most of the bugs have been fixed

Bugs are not falling in number, we have 248 bug issues.

the question is what comes next? more minute features? or is MT going to improve in a way that is fairly noticeable by the players?

That's insulting, we are always making significant improvemwnts and adding significant new features.

[numberZero]

@paramat http://irc.minetest.net/minetest-dev/2017-12-28#i_5182234

But instead of forking IrrLicht ourselves, we can use some existing fork like https://github.com/buildaworldnet/IrrlichtBAW, as @rubenwardy just pointed out.