Porting Voxlap

[Sonarpulse]

It's coming along nicely!

(screen-shot by mach1ne -- proof that this will eventually compile on other peoples computers.)
Original Post:
So awhile ago I was working on porting voxlap. I almost finished, but then I got busy, and figured with the GL version working under linux it would be released on all three platforms anyways.

Well, now I am increasingly doubtfull of that. And even if it was, it wouldn't have the performance of the original which is one of the reasons I liked the game in the first place. So now I want to start working on this again. I am probably a week of solid work away from finishing my initial port. The next step is to use this community's in depth knowledge of server and protocol to recreat AoS on top.

Yes in a short run an open-source game will be greifers delight, but I beleive open-source this communities best bet at staying alive. And the only truely robust anti-cheat is server side anyways.

[CommieBuffalo]

You do know about Project Iceball, right?

[Sonarpulse]

Haha yes. I learned about it right after I made this thread. GreaseMonky wants to keep IceBall GPL, so I am thinking of contacting ken to see if we can use it under the GPL. IceBall has already made alot of progress on the rendering front, but using this for rendering would ensure 0.75's level of performance for the first release.

[rakiru]

I just heard about this yesterday and was delighted. Great work so far, and good to hear you're so close to finishing it.

As for the AoS protocol, the only complete source of information on it is in the pyspades/pysnip source, but I am (slowly) working on rectifying this on my wiki.

[TB_]

Haha yes. I learned about it right after I made this thread. GreaseMonky wants to keep IceBall GPL, so I am thinking of contacting ken to see if we can use it under the GPL. IceBall has already made alot of progress on the rendering front, but using this for rendering would ensure 0.75's level of performance for the first release.

Sorry if I got this wrong, but are you saying that contacting ken silverman and releasing it under GPL will make rendering faster?

[Triplefox]

Implied is that we would use the Voxlap code and not the half-finished rendering system GM made. Which I'm sort of ambivalent about, Voxlap is exactly what we're used to, so it would help give the game the same feel, but it may also put a chill on further rendering improvements, since it would take time to know it well enough to do useful maintenance(vs. learning as we go and gradually accumulating the features for a new renderer). I'm not a rendering junkie but I get the impression that everyone else is, so maybe it won't be a problem.

[Des]

This sounds pretty good. keep up the work and maybe you can help with the iceball devs sometime later

[HoboHob]

Haha yes. I learned about it right after I made this thread. GreaseMonky wants to keep IceBall GPL, so I am thinking of contacting ken to see if we can use it under the GPL. IceBall has already made alot of progress on the rendering front, but using this for rendering would ensure 0.75's level of performance for the first release.

Iceball does not use Voxlap. It uses a custom rendering engine made by GreaseMonkey.

[Articsledder]

I was just thinking about this the other day, being able to mix TC, server stuff and network optimizations into .26. Will definitely be cool to play around with. This is a huge step towards creating a BnS version of AoS, I'm interested o see how this and Iceball turn out.

[Sonarpulse]

Iceball does not use Voxlap. It uses a custom rendering engine made by GreaseMonkey.

But based on how development goes there, it maybe wise to use this at least temporarily. I'll admit my bias, after putting 3/4s of a year's effort into my port, only to see the openGL the version render this useless. Now is probably my last chance to see it used. Conversely, while Voxlap his it's problems (we would have to reimplement multirse support for example), it could be that the optimizations that went in to Voxlap would be very hard to reproduce. Looking at Iceball's TODO list, it seems most of the remaining problems are graphical.

Grease Monkey's main concern as expressed in his PM to me was licensing, not engineering, so maybe this will be useful.

I was just thinking about this the other day, being able to mix TC, server stuff and network optimizations into .26. Will definitely be cool to play around with. This is a huge step towards creating a BnS version of AoS, I'm interested o see how this and Iceball turn out.

What is the big deal about 0.2x anyways? I am actually curious. Anyways, I recommend that regardless of whether this is reused, you make your project a fork of or mod for Iceball. That project has made enough progress where it no longer makes sense to try to reverse engineer old AoS Binaries. If you could figure out why the rendered voxels aren't getting copied to the screen when gcc is used to compile this program though, that would be much appreciated.

[Articsledder]

.26 was generally seen as the peak of the game, community wise, however the game itself needed improvements. With each update however, we got some good things (network optimizations and votekicks) and some bad things (Autoclimb, SMG, etc)

[HoboHob]

-snip-

I think it would definitely be worth my time to make a voxlap port of Iceball.

It probably wouldn't take to long provided I can get both easily to compile in MinGW.

[Sonarpulse]

Thanks for the support

the GPL allows you to do whatever with the code for your own personally copy. So you or I can try to combine voxlap and iceball locally, and if I am sucessfull in getting my port GPL-able, we can release it.

[HoboHob]

Thanks for the support

the GPL allows you to do whatever with the code for your own personally copy. So you or I can try to combine voxlap and iceball locally, and if I am sucessfull in getting my port GPL-able, we can release it.

Can I get a link to the MinGW port of voxlap?

[Sonarpulse]

Sure! https://github.com/Ericson2314/Voxlap
I will admit though I probably jumped the gun, because my port is a critical bug away from completion

A) Port is not finished, things compile and run without crashing, but so far, only some text is actually coppied to the screen. I know the routines to actually rendered the voxells are being called but they don't appear in the window when gcc/mingw is used. If you have a copy of MSVC I would also compile with that for reference. Whatever is causing this issue is simple, but I have no idea how to track it down.

B) On Linux, it will grab all your imput and you CANNOT alt-tab out. the one solutino I know is to use ctrl-alt-f# to switch from a text terminal and kill it from there.

C) if you are running 64-bit, you must compile a 32-bit binary because it still contains some 32-bit inline asm (removing it though would be super handy, as that should help track down the LAST GREAT BUG, as i think of it, too.