Client freezes desktop when closed

[Pardok]

Ok, the Gentoo mirrors updated to the most recent Nvidia kernel set.....I downloaded that and everything works fine now concerning closing the client.

However, I'm stuck on how to play ATITD with sound while listening to music at the same time.

Currently, I'm using ALSA 0.9 with the OSS emulation. I am able to play 2 or more audio files at once, so my mixer is set up ok. However, with ATITD, this isn't possible. If I'm listening to music when the client starts up, the sound doesn't initialize.

I have tagged my start-up script with the -alsa tag. So, essentially, I'm running ./elaunch -alsa.

My assumption is that the sound would then be piped through the Alsa server, mixed, and spit out.....like with all my other audio.

Only, this doesn't happen.

I tried using the -oss tag....and managed to get sound out of the client. But it then locked all other output. I'd try playing music (via Amarok with GStreamer with alsasink) and would only get a frozen media player or a crash.

Any Linux experts know how to remedy this problem? I used to just do "artsdsp" when running the client.....but I found Arts to be fairly mediocre.

[Pluribus]

Using ALSA with OSS emulation, emulates the blocking as well... the sound library only supports ALSA 0.9 as well so that doesnt help much...

ESD works well for both, but unless you run Gnome, you probably dont run ESD...

Chalk it up to crappy sound driver support in linux.

[Pardok]

Bah.....so nothing to be done for it yet?

Well, I was never a big fan of cicadias anyway. :P

I'll just run it with the -nosound

Thanks Pluribus.

[Corwitep]

As Pluribus said install esd. Run elaunch with -esd and use the appropriate output plugin in xmms or whatever you are using.

Works like a charm for me. I never managed to get ALSA with non-blocking to work properly.

[Pardok]

Well, I'm not too fond of Esd. My music sounds 20x better coming through ALSA (don't know why though).

I've managed to configure ALSA to use multiple channels and mixers. Thus, I can be watching a Flash video with sound and an MP3 at the same time. Just configured my .asoundrc file to use dmix as the default.

For some reason, though, ATITD HATES this.....don't know why.

Anyway, I don't need the sounds anyway. :D

[Yoru]

I'm having the same problem mentioned in this thread's title; when I close the client conventionally, my computer is rendered completely unresponsive. No screen updates, no mouse input accepted, no keyboard input accepted; in fact, the capslock and scroll-lock LEDs on my keyboard blink repeatedly.

System: 1.4Ghz Athlon-Tbird, 1024MB RAM, GeForce 3 (64MB) graphics card running Gentoo Linux (2.4.26-r9 gentoo-sources kernel) with the latest nvidia-kernel/nvidia-glx libraries in Portage (1.0.5336-r4) and ALSA (alsa-driver-1.0.5a/alsa-oss 1.0.5/alsa-lib 1.0.5-r3) for sound output.

I've found a stopgap solution in the meantime: xkill terminates the app with no problems, so I just use the logoff option and xkill it from the login screen. :)

Addendum: Ok, it seems that xkill is no longer sufficient if I actually enter the game, even if I drop to the login screen before xkill'ing. Pardok, what version of the Nvidia driver kernel are you using? Maybe the newer version re-breaks whatever it is your version fixed.

[Yoru]

Seems things have taken a turn for the worse. My client now periodically freezes during gameplay as well; it seems to be related to being around an anvil with a piece in progress, as I've only ever frozen in-game while near one.

Upgraded my drivers to 1.0.6111 in an effort to resolve the issue, to no avail. Client still freezes the entire computer periodically, or when being closed either conventionally (quit menu option) or via being killed.

Interestingly, if I do manage to make it through a session without freezing, I can successfully close the program with xkill if, and only if, I wait for the CPU usage to drop from 100% to about 85% at the login screen. This usually happens about 3-6 seconds after returning to the login screen. Xkilling from in-game or prior to that drop in CPU usage still causes catastrophic freezing.

[Pluribus]

Have you checked for any log file messages etc? The egenesis.bug file is the useful one for the client. There isnt anything oddball about your setup. (My machine is a Athon XP 2000+ w/1GB RAM and a GF3 Ti500. running 61.06 drivers on Fedora Core 2) Couple of things to try and isolate your problem...

1) try running it with the -nosound switch to see if it is the audio subsystem. (ATITD doeesnt lose much at this stage without audio :-) ) If it doesnt lock play it until you can definitively say that was it... (My hope that it IS there since that code is getting a rewrite shortly)

2) Make sure your cards heatsinks and fans are CLEAN... Blow out the dust... ATITD taxes video cards harder than any game on linux that I have seen. (Nvidia cards will stop rendering 3D when they overheat and will deadlock the system in those situations) While this may not be the issue with shutting down, it very well could be for the issues while it running.

3)Try -noext (I doubt it will do anything, but it disables all non-vital OpenGL extentions used by the client)

As for the audio blocking... OSS and by extension OSS emulation stink horribly... They only support a single application having the sound device open at any one time. Alas, the audio driver in the client uses OSS whenever it is found and the ALSA support in it has not been upgraded to ALSA 1.0 yet. (Yet another reason the audio subsystem is slated for a rewrite across the board)

[Yoru]

Today's results...

Ran elaunch with just -nosound, froze after about 30 minutes.
Opened case, dusted it out, washed the air filters, ran with -nosound, froze after 45 minutes.
Ran elaunch with -nosound -noext, no problems for the 4 hours I had it open. Froze when using the normal quit option.
I'll try running with just -noext tonight and see if there's any issues.

When run with just -nosound, I noticed these sprinkled around in my egenesis.bug:
Rendering mode ripplegen: Line 582: ARBFP10 not supported by this driver.
Rendering mode sprite_marble_N0: Line 1456: ARBFP10 not supported by this driver.
Rendering mode sprite_marble_nsc_N0: Line 1471: ARBFP10 not supported by this driver.
(and several others with the same label, ARBFP10)

When run with -nosound -noext, new stuff appeared (I've truncated, there were several identical ones):
Rendering mode dynclouds: Line 259: nvrc2 not supported by this driver.
Rendering mode grass: Line 458: nvrc not supported by this driver.
Rendering mode ripplegen: Line 582: ARBFP10 not supported by this driver.

Also, at the tail end of the -nosound -noext log, presumably where the system hung, three lines of unprintable characters have been saved (represented in less as '^@'). Strange.

[Pluribus]

No nvrc2???? WTF? My GF3 supports it.... can you send me pluribus@pluribus.org a copy of your egenesis.bug file.... /me thinks something is afoot at the circle K....