It finally started happening when running QS with "-noglsl". Now once the weird visual glitches start appearing again, they show up every time I restart QS until I reboot my computer. Rebooting makes it go away for a while. This what it looks like, by the way.

Another weird thing I've noticed is that quite often when I've pressed down the run key (mapped to "w") for a while, I suddenly stop moving, and have to release and re-press the key. Don't know if it's related, but it never happened on my old OS.

The audio thing is most noticeable when using weapons with rapid fire (i.e. the nailgun or SNG). It's as if the firing sound doesn't loop smoothly, but rather stops briefly and then restarts.


PS: Sorry for spamming this thread with my problem. Hopefully it doesn't annoy anyone. 

Is there a way to create a dedicated server on Linux with QS? Preferably without the host acting as a client as well. I'm interested in spinning up something for co-op and ideally I'd like to be capable of hosting multiple mods like AD and Quoth. I see that FTE has a binary for servers but I'd first like to see if I could get something going without QuakeWorld physics i.e. ludicrously fast bunny hopping.

I know that QSS works well for co-op but I don't want to host from my computer and I also can't host without acting as a client myself. 

You can switch to Netquake physics in FTE with the console command sv_nqplayerphysics 1. 

I've switched to a new computer since posts #3601 and #3602, but I still get that weird stuttering audio thing that had never happened before.

The same thing happens on a third, much older computer too.

Each of these computers is running some version or derivative of Ubuntu 18.04 and has Intel integrated graphics, in case that's relevant.

You can actually already hear it the moment you start a new game in Quakespasm: the ambient sound that plays right at the beginning in the start map doesn't loop seamlessly as it should (and always used to), but instead clearly stops and then starts again. With louder sounds, e.g. during combat, it's much more noticeable. 

Tough to diagnose. I've got a few debugging things you can try:
- Try with a clean config. Make a copy of your quake directory, and delete everything except id1/pak0.pak and pak1.pak. You can launch in this gamedir with "quakespasm -basedir {directory here}"
- try the Linux binary from http://quakespasm.sourceforge.net/download.htm as well as the Ubuntu one you can install with apt. Does the issue happen on both?
- Type "condump" and upload the resulting file
- Do you get a reasonably good framerate? What does "host_maxfps 1000" then "scr_showfps 1" show? Make sure to reset "host_maxfps 72" afterwards. 

<quote>Does the issue happen on both?</quote>

This is with the Linux binary downloaded from http://quakespasm.sourceforge.net/download.htm (which is what I've also done in the past when things worked properly; I don't think I've ever installed Quakespasm with apt).

I'll definitely try the things you listed and then post the results. My apologies in advance if it takes a while. 

Does the issue happen on both?

^Obviously that's what I meant to do. How embarrassing. 

So I've added per triangle collision detection for entities (vs. bounding boxes) to my quakespasm project at home but I'm running into a peculiar collision detection issue.

My code works fine when the entity's baseline angles/origins remain 0, but not otherwise. I'm using the entities's v.origin and v.angles to construct an inverse transformation matrix. I perform my per-triangle traceline collision detection, find my hit point, and then transforming the hit point back into world space.

As stated, this works fine when the baseline is zero. If i need to do some more voodoo in my math to account for baseline angles/origins please let me know? 

That is odd. Assuming your collision detection is being done in server code, the baseline shouldn’t matter at all since that is merely used as part of the network protocol and client code. What types of collisions are you testing with this? Bullets against enemy models? 

Thank you for responding metlslime. I am testing bullets against enemy models. I have a modified traceline routine that iterates through all of the model's triangles looking for ray triangle intersection (terribly unoptimized, I know).

What you have helped me understand is that i don't need to worry about the baseline information. Is using the AngleVectors on the entity's v.angles call to extract the forward, right and up vectors, plus using the v.origin sufficient to build my world-to-model space matrix? Is there a better way?

The only time it does work correctly is when the model's angles are <0,0,0>, and i invert the y axis (right vector * -1) when projecting from world to model space. I don't know if this behavior pops any ideas into your head, but i thought id share.

Thanks for your help!

FYI: the collision data is built in model space during the the Mod_LoadAliasModel process. In pr_cmds.c I invoke my modified traceline routine to translate my ray origin & endpoint from world space into model space, perform my hit detection, and then transform it back from model space to world space. 

I'm really sorry it took me this long to respond.

- Try with a clean config.

With the binary downloaded from http://quakespasm.sourceforge.net/download.htm it happens with a clean config too.

- try the Linux binary from http://quakespasm.sourceforge.net/download.htm as well as the Ubuntu one you can install with apt. Does the issue happen on both?

Ok, I've installed Quakespasm via the Software Centre or whatever it's called nowadays (which I guess is the same as installing via the terminal with apt. It's called "quakespasm-beidl", though. Is that right, or did I accidentally install the wrong thing?). As far as I can tell, the problem doesn't happen with this version of Quakespasm, or at least it hasn't yet.

- Type "condump" and upload the resulting file

Should I do this for both versions of Quakespasm I now have on my computer? And where do I upload it?

- Do you get a reasonably good framerate? What does "host_maxfps 1000" then "scr_showfps 1" show?

Under both versions of Quakespasm, my framerate is initially just under 200 on the start map and then rises a little, while hovering between 200 and 300 on e1m1. Is that good? The version downloaded from http://quakespasm.sourceforge.net/download.htm seems to have a slightly higher framerate, though, for what it's worth. 

https://snapcraft.io/quakespasm-beidl 

It would be nice to be able to bind my own menu navigation keys in QS. My keyboard doesn't have regular arrow keys but can send the numpad direction keys. Is this feature already present and I just don't see it? It's not super annoying to have a second keyboard shoved in the back of my desk to do Quake menu nav, but it would be nice to use my aircraft carrier Model F for everything. 

Oooh, Model F, you say? o: Gotta love dem clicky keys! ;) I'd wager some hardcore typists would pay a hefty penny for one of those keyboards. :D 

I love my Model M keyboards. 

Condolences to anyone else in the same building as you when you're typing on that thing. 

where does one buy these? 

why does the Quakespasm still have that "skin is taller than 480 issue"? At number of other engines have took care of that, why can't Quakespasm? 

Version 0.93.2 of QuakeSpasm is released.
Changes since the previous version:
https://sf.net/p/quakespasm/news/2019/11/quakespasm-0932/
Downloads:
https://sourceforge.net/projects/quakespasm/files/
http://quakespasm.sourceforge.net/download.htm 

... please? 

what is the command to disable fullbright in the quakespasm engine? 

... in the console.

Posted in the vein of "give a man a fish" VS "teach a man to fish". 

got it. Thanks!