Need more developers

[Matzon]

Bug fixing has been a bit slow lately. This has been because of multiple issues such as the lack of ... "interesting" bugs and vacation to name but a few. However another issue is also we're simply too few that has the needed qualifications to work on both native code and java code, coupled with the requirement of having access to - and understanding - of multiple platforms.

We have, so to speak, been running in maintenance mode for a while (except for spasis excellent GL updates). Elias has been doing a lot of the mac and linux fixes for a while, but he is scaling down his commitment due to other work.

LWJGL works pretty well - so there isn't a lot to "do". However there are some issues that still need to be fixed and to that end I would like to call for more developers to help out.

I dont have a list of specific issues to look at, but I know that there are some input issues (utf8, keyboard, exotic characters), AWTGLCanvas issues on mac (I think), same with setParent, appletloader speed?, Maven support, updated buildscript etc...

I am not asking people to commit full time to the project or anything - but submitting patches or reproducible tests will go a long way in getting things fixed quickly. That said, someone to maintain the platforms - especially mac, would be nice.

[elias4444]

I can help with testing on the Mac side of things. I'm not a very good C/C++ programmer yet though, so I wouldn't trust myself to help build or develop the jni code.

Do we need to be worried about the future of LWJGL here? Or is this pretty much "par for the course?"

[Matzon]

nothing to be worried about at all. lwjgl is in a state where it doesn't need that much work anymore - its just some smaller issues that someone needs to dig into.

[princec]

"Necessity is the mother of invention" and all that. I think that people needing things to work properly for them are likely to fix problems. (Hence me doing bugger all for several years - everything works fine for me).

Cas

[kappa]

Theres been a lot of interest in lwjgl applets lately, especially with recent success of games like minecraft. The deployment method is just so cool and cuts out a large barrier (of download exe, install, run).

appletloader speed?

just thought I'd point out this bug is fixed now

As for AWTGLCanvas, it should really just die. Display.setParent() pretty much does everything it does (but better, faster and more stable) the only advantage it has is it allows multiple canvas's.

Biggest change that is needed to make lwjgl applets really awesome, is support for a native Display on mac, currently Display uses AWT which causes a lot of the issues. This is by no means an easy task and probably requires a lot of work. JOGL2 has this type of native window on mac (maybe we could borrow some of the code). Having this will allow some really cool stuff (and avoid some AWT bugs) like being able to run applets in a seperate vm (like GetDown but embedded). This will allow all sorts of cool stuff like setting memory requirements, catch native crashes without blowing up browser, very stable and be compatible all the way down to Java 1.4.

Only real showstopper bug with Display.setParent() atm is mouse grab does not work on windows, other than that i think it works pretty well. I had a go at trying to see if i could fix this but just know enough about how mouse grab works on windows to be able to fix it.

[broumbroum]

Well I think you must be aware that not everyone would use LWJGL openGL impl., as JOGL is still a good alternative. That's merely "killing" this project to suppress such "gateways" to programming in Java, especially with talking about erasing AWTGLCanvas which the far most important reason for I've switched to LWJGL from JOGL...

Never mind if you want all of the Java games over LWJGL but it's worth keeping all so-called gateways to programming OpenGL. I mean OpenGL was still very obscure for me, since I noticed that a Java application could render games as well as a Mario would on SNES or Half-life on PC-CDROM.

Then Java has many advantages to other languages such as multi-inheritance that is worth keeping "front of the scene", as a Canvas would extend to GameCanvas (as suggested in many books), AWTGLCanvas is worth to this approach, too. 

[kappa]

just curious, why are you using AWTGLCanvas instead of Display.setParent() ?

[broumbroum]

I've got my own impl. of Canvas which allows bypassing some configuration such as fullscreen switch and other AWT events.

[Villane]

I'm creating a patch for the Eclipse update site and adding a new plug-in to use in plain java projects, but I could also help with Maven.

What do you want to do with Maven? Move the build to Maven entirely or just provide a maven repo? (your own repo? central?)

[Matzon]

imo, we dont want to use maven for building, but rather to provide a repo.