@utri007
Ultimate Stunts - This game lacks polish, it's not worth to port in in its current state.
Unknown Horizons - Already on my list, see the first post.
PlaneShift - No dice, uses Cg shaders.
Glest - I'm pretty sure this one has been already ported by someone else.

I'd rather not add any more FPS' to my list :P

@SinanSam460
A RadeonHD Warp3D driver wouldn't help much, especially on Sam machines. A more modern graphics stack (Gallium3D) would be necessary. The current OpenGL implementation translates between two high level APIs, which always comes with a performance penalty. There's no support for vertex buffer objects either, which could drastically reduce the amount of data that has to be pumped through the system bus for every frame.

Using smaller textures won't give you better speed, it just reduces the VRAM footprint. It's far from stable, the aforementioned problems are still there. Unfortunately I don't know how to get any debug output from Warp3D, so the visual glitches (menu, track map) will stay there for the time being.

@samo79
No such problem here. Just to make sure, could you try running the game with 8x8 textures and multitexturing enabled?

edit: I uploaded TORCS too, it has a slightly lower polygon count.

Kas1e has mentioned 7KAA, which is an open source continuation of the RTS called Seven Kingdoms. Other than that I don't really know many open source RTS games which haven't been ported already.

For 128MB VRAM 512x512 textures are the best you can do without the hardware compositioning enabled. Other than that you must use 256x256 textures. Without texture compression, the textures of track alone will consume most of your VRAM.
Unfortunately the OS freezes if all the VRAM is consumed, this is not my fault.

I uploaded a test version to my website. Known problems:
1. Broken menu backgrounds. Since they are upscaled into 2048*1024 textures, this is probably caused by the W3D driver claiming to support 2048*n textures, but failing to create them. I had the same issue with RTTR, and others have confirmed it too.
2. If you don't quit after a race, the game locks up.
3. Broken track map. It's hard to tell what goes wrong without debug output from W3D.

If you have any issues beside this, feel free to report them.

@328gts
They went alright, I passed all of them except one. Differential geometry is so much fun... not!
Back to Speed Dreams, I have an initial version up and running. Don't expect miracles, this game is very demanding. Super Tux Kart should give you and idea of what to expect performance wise. I'll do a few test-runs with the profiler enabled to see if there's anything I can do to make it faster. I'll also add AmigaInput support.

I finished my last exam this Friday, so I'm back in action! I plan to crank out a few ports next week. Speed Dreams will be definitely one of them.

The latest version of iksemel comes with TLS support:
http://code.google.com/p/iksemel/source/browse/#git%2Fsrc
It could be worth giving a shot, if they didn't break the API.

@magnetic
I did read it, and being the scatterbrain I am I misread one word, for which I apologize.

@magnetic
You are basing this on the premise that people who port toolkits an linux stuff would otherwise develop native software.

WxWidgets is not just a toolkit, it has various support libraries just like Qt. This is the same old native application vs foreign toolkit debate.
Why wait for a toolkit? Who should I encourage to add Odamex support for their native server browser? I know where you are coming from, but I was looking forward to this port, so it was a letdown for me.

@kas1e
I'm with you on this. There were no real updates for more than a year, and then he comes back with a somewhat pretentious post... We'll see if there will be any progress.

@Severin
So it should just insert the string where the cursor is? Also it can't really be called a bug, because it's most likely the intended behavior. Just because something is counter-intuitive, it won't become a bug.

@ScottCabit
You get DSIs, because IPC_FindProc is called with NULL name.

Then it compares NULL to the task's name, and you get a DSI. I could add some sanity checks, but it would only treat the symptom. If you can pack up your whole dopus5 install for me, I'll see what I can do later.

PS: Next time you could post your logs as code, or attachment.

No log, no help

It looks doable, but it's written in C++, so the old compiler could be a problem if it uses C++11 features.

A small update to Odamex:
The icons have been converted to OS4 native format. I also removed the libsmpeg.so dependency due to popular demand.

@Hans
RTTR doesn't use GLUT at all, only SDL. The problem is with MiniGL + the fullscreen switching indeed, but I have no means of fixing that from the game side, I can only disable it.