dmn_clown
Posts a lot
Cakes 1
Posts: 1324
|
 |
« Reply #50 on: July 14, 2007, 07:11:57 PM » |
|
Changing sound quality crashes the engine when using OpenAL: http://bugs.icculus.org/show_bug.cgi?id=3261Bug in OpenAL |
|
|
|
« Last Edit: July 22, 2007, 04:02:57 PM by dmn_clown »
|
Logged
|
|
|
|
amdfanatyk
Nub
Cakes 0
Posts: 19
|
|
« Reply #51 on: July 15, 2007, 02:34:09 AM » |
|
No, it means that your setup has issues with one map and ends gracefully with a signal 11 error, you are using a kernel with a brand new memory scheduler that has bugs and is far from stable. Have you tried the game on an earlier kernel or the 2.6.22 kernel?
Do U know what U talk about? "memory scheduler ... is far from stable" - or just want make me laugh loud. Sorry but it's 2.6.21.5 and it cannot contain UNSTABLE solutions. |
|
|
|
|
Logged
|
Occupation:
programmer
Fluent in:
Polish, Assembler, C/C++, Java, bash/awk
|
|
|
Fitacus
Lesser Nub
Cakes -3
Posts: 142
|
|
« Reply #52 on: July 15, 2007, 03:32:20 AM » |
|
Old bug also exists in 0.7:
- Game stucks very often, when new players join the game
|
|
|
|
|
Logged
|
"The things you own, they end up owning you."
|
|
|
amdfanatyk
Nub
Cakes 0
Posts: 19
|
|
« Reply #53 on: July 15, 2007, 09:09:06 AM » |
|
Currently (svn version newer than 0.7.0) maps that always cause crash are OA_DM2 and OA_DM5. I found another annoying issue. Looks like every map has bot_minplayers variable set to fixed value so I cannot add as many bots as I want even in skirmish mode. Only OA acts in this way, both Q3A and UT2k4 allow any number of bots that is limited only by available slots' number.
|
|
|
|
|
Logged
|
Occupation:
programmer
Fluent in:
Polish, Assembler, C/C++, Java, bash/awk
|
|
|
|
GrosBedo
Member
Cakes 20
Posts: 710
|
|
« Reply #54 on: July 15, 2007, 11:17:41 AM » |
|
@GrosBedo, have you tried deleting your q3config.cfg file completely and starting over from scratch?
Off course man, in fact I didnt put my old configured q3config.cfg, and when I encountered those bugs I tried cvar_restart AND delete the q3config (but cvar_restart do the same). It would be cool if the dev team could do a patch for the bugs they can correct before next release, else it would really lower down the quality of the game the time being. |
|
|
|
|
Logged
|
|
|
|
dmn_clown
Posts a lot
Cakes 1
Posts: 1324
|
|
« Reply #55 on: July 15, 2007, 01:30:37 PM » |
|
Do U know what U talk about? "memory scheduler ... is far from stable" - or just want make me laugh loud. Sorry but it's 2.6.21.5 and it cannot contain UNSTABLE solutions.
I guess you told me. All of the people that have played dm2 and dm5 without issue are obviously wrong and it has absolutely nothing to do with your setup. Thank's. |
|
|
|
|
Logged
|
|
|
|
dmn_clown
Posts a lot
Cakes 1
Posts: 1324
|
|
« Reply #56 on: July 15, 2007, 01:48:30 PM » |
|
Off course man, in fact I didnt put my old configured q3config.cfg, and when I encountered those bugs I tried cvar_restart AND delete the q3config (but cvar_restart do the same).
It would be cool if the dev team could do a patch for the bugs they can correct before next release, else it would really lower down the quality of the game the time being.
The only time I've ever seen the camera lag other than at the very start of a match when the vm is still compiling, is when I set my maxfps too high (in a network game) reducing the maxfps solved any camera lag that I've experienced, though it kills your trickjumping. |
|
|
|
|
Logged
|
|
|
|
amdfanatyk
Nub
Cakes 0
Posts: 19
|
|
« Reply #57 on: July 15, 2007, 02:18:52 PM » |
|
Do U know what U talk about? "memory scheduler ... is far from stable" - or just want make me laugh loud. Sorry but it's 2.6.21.5 and it cannot contain UNSTABLE solutions.
I guess you told me. All of the people that have played dm2 and dm5 without issue are obviously wrong and it has absolutely nothing to do with your setup. Thank's. I play Q3A, UT2k4 and Painkiller via Cedega and imagine - IT WORKS, there are no stupid crashes! So stop abusing kernel developers cause they know better what they put into stable branch. These maps cause crash same as fan did and it's because they are crappy made not because my kernel, gpu, cpu or because I live in Poland and use Neostrada. |
|
|
|
|
Logged
|
Occupation:
programmer
Fluent in:
Polish, Assembler, C/C++, Java, bash/awk
|
|
|
anyone
Half-Nub
Cakes 2
Posts: 86
I may weep openly
|
|
« Reply #58 on: July 15, 2007, 02:40:43 PM » |
|
Hey chill ^^
Let's see that bug as an exception, incompatible lines or a misplaced dot somewhere around the calculation of the next alignment of three mystical planets before the back of Cthulhu. Of course there are bugs around (correct me if I'm wrong, but Open Arena is still Beta) but it's nice to tell the dev about them.
As far as I know, OpenArena is based on the source released by Id for Q3A. Unfortunately, some part of this code has been chopped because of legal issues. That's why OA doesn't behave exactly like Q3A.
|
|
|
|
|
Logged
|
I'll wear my badge, a vinyl sticker with big block letters adherent to my chest, tells to all your friends: "I am a visitor here, I am not permanent"
(The postal service "district sleeps alone tonight")
Promoted Mapper of the Century by some CPMA mid/high experienced player (luvyoalot, man)
|
|
|
dmn_clown
Posts a lot
Cakes 1
Posts: 1324
|
|
« Reply #59 on: July 15, 2007, 03:03:59 PM » |
|
I play Q3A, UT2k4 and Painkiller via Cedega and imagine - IT WORKS, there are no stupid crashes! So stop abusing kernel developers cause they know better what they put into stable branch. These maps cause crash same as fan did and it's because they are crappy made not because my kernel, gpu, cpu or because I live in Poland and use Neostrada.
I offered suggestions on getting the game working on your setup, you didn't try them and became combative instead. OpenArena is based on the ioquake3 engine NOT Q3A. Also if you were talking about the old kernel dev. model I would agree with you, with the 2.6 branch stability was pushed off on the distro's so the kernel dev's could focus on new features/performance. That is not an attack on the kernel dev's that is their development model. I could be extremely impolite here, much like Linus, and start calling you F*ing this and F*ing that for not bothering to even try an easy suggestion on your end to debug what exactly is wrong on your setup that no one else is experiencing, instead I'll just end this conversation which is going nowhere because you refuse to try anything. Have a nice day  |
|
|
|
|
Logged
|
|
|
|
kit89
Member
Cakes 6
Posts: 636
Shoot him..
|
|
« Reply #60 on: July 15, 2007, 03:17:42 PM » |
|
I play Q3A, UT2k4 and Painkiller via Cedega You do know that Q3A & UT2004 have native Linux versions right? Which if you use the native binaries would improve performance for these games. Also about the dev of the kernel, didnt a kernel developer say that the 2.6v was unstable and more bugs where appearing than fixes? Please also bear in mind that without actually seeing your system with our own eyes, its abit of guess work to find the actual solution to your problem. Unless you are willing to help find a solution and try suggestion brought forward, there is very little we can do. Also I've tested OA_DM2 with 6 bots, bloom on, & using com_hunkmegs 256. The map ran flawlessly in skirmish mode. Using the official 0.7.0 release. Also what are your game settings? Do you have stencil shadows on? Textures set to high? High Geometric detail etc? AA & AF set high? I once cranked all the Doom 3 settings to high(Even tho my system was only suitable for low/meduim settings), which caused an identical error after I played for a while, bringing the settings back down to normal removed this problem. |
|
|
|
« Last Edit: July 15, 2007, 03:40:24 PM by kit89 »
|
Logged
|
|
|
|
amdfanatyk
Nub
Cakes 0
Posts: 19
|
|
« Reply #61 on: July 16, 2007, 03:15:04 AM » |
|
I play Q3A, UT2k4 and Painkiller via Cedega You do know that Q3A & UT2004 have native Linux versions right? Which if you use the native binaries would improve performance for these games. I play Painkiller via Cedega. Also about the dev of the kernel, didnt a kernel developer say that the 2.6v was unstable and more bugs where appearing than fixes?
It's so hard for some people to understand difference between containing bugs and being unstable. I'm very sorry because of it. Try to show me software without bugs. Every version that have even number in the middle is stable, so if You want try sth unstable, try 2.5. Please also bear in mind that without actually seeing your system with our own eyes, its abit of guess work to find the actual solution to your problem. Unless you are willing to help find a solution and try suggestion brought forward, there is very little we can do.
Also I've tested OA_DM2 with 6 bots, bloom on, & using com_hunkmegs 256. The map ran flawlessly in skirmish mode. Using the official 0.7.0 release.
UT2k4 and PainEngine are more demanding for whole system than ioquake3 engine which is in fact based on Q3A engine 1.32 which is 8 years old. So it's not fault of my system. Also what are your game settings? Do you have stencil shadows on? Textures set to high? High Geometric detail etc? AA & AF set high?
I once cranked all the Doom 3 settings to high(Even tho my system was only suitable for low/meduim settings), which caused an identical error after I played for a while, bringing the settings back down to normal removed this problem.
Predefined settings called "Very High Quality" which includes 1024x768 resolution, highest details, trilinear filtering etc. I've attached my conf file. com_hunkMegs is set to 64 but as I said increasing to 512 gives no result. |
|
|
|
|
Logged
|
Occupation:
programmer
Fluent in:
Polish, Assembler, C/C++, Java, bash/awk
|
|
|
|
GrosBedo
Member
Cakes 20
Posts: 710
|
|
« Reply #62 on: July 16, 2007, 08:45:12 AM » |
|
The only time I've ever seen the camera lag other than at the very start of a match when the vm is still compiling, is when I set my maxfps too high (in a network game) reducing the maxfps solved any camera lag that I've experienced, though it kills your trickjumping.
Ok I'll try that, and I'll try to set higher my com_hunkmegs and every other memory parameters (it's at 128 Mo at the moment). Some other infos about this problem: it actually happens only on new maps, particulary when there is a big area to gith (and weirdly the bug doesn't happen on these maps when I look in a little room or a wall). The older maps that were already included in 0.6 don't produce any bugs. /EDIT: Okay I tested and maxfps does nothing at all but reduce my ability to trickjump :p But com_hunkmegs seems to reduce the occurence of this bug, it's still here, but now it's playable. I noticed that launching many rockets with projectiletrail activated produces this bug, but removing this parameter reduce the bug. I will do some other tests, but it seems that the number of polygons or particules are the source of this bug (and my gpu isnt old at all). /EDIT2: Okay, I've put com_hunkmegs from 256 to 384 and nothing happened. But with com_maxfps set to a number higher than 80, you then encounter this weird camera bug, particularly in the Hydronex (or whatever the name) CTF level. And it happens on old maps too, so it's a bug from the modifications of the engine (I didn't put r_bloom). Ill submit new infos as I find them. /EDIT3: I've just tested locally on Hydronex map, and no bug happened, even with very very high FPS and with maximum bots I could add. Could it be linked to packet management as I suggested before ? Note that I dont have any network problem, no router, no firewall, and a ping between 15 and up. |
|
|
|
« Last Edit: July 16, 2007, 10:02:27 AM by GrosBedo »
|
Logged
|
|
|
|
kit89
Member
Cakes 6
Posts: 636
Shoot him..
|
|
« Reply #63 on: July 16, 2007, 12:14:14 PM » |
|
I've attached my conf file. I've tested your config, and I played the entire map without a crash. The only thing I can think of is you may have some library conflicting with OA. |
|
|
|
|
Logged
|
|
|
|
amdfanatyk
Nub
Cakes 0
Posts: 19
|
|
« Reply #64 on: July 16, 2007, 01:35:42 PM » |
|
I've attached my conf file. I've tested your config, and I played the entire map without a crash. The only thing I can think of is you may have some library conflicting with OA. Which map? and tell sth about Ur setup. ldd ioquake3.i386
linux-gate.so.1 => (0xb7fab000)
/lib/libsafe.so.2 (0xb7fa4000)
libSDL-1.2.so.0 => /lib/libSDL-1.2.so.0 (0xb7ed5000)
libpthread.so.0 => /lib/libpthread.so.0 (0xb7ec2000)
libopenal.so.0 => /usr/lib/libopenal.so.0 (0xb7e89000)
libvorbisfile.so.3 => /lib/libvorbisfile.so.3 (0xb7e82000)
libvorbis.so.0 => /lib/libvorbis.so.0 (0xb7e5a000)
libogg.so.0 => /lib/libogg.so.0 (0xb7e55000)
libdl.so.2 => /lib/libdl.so.2 (0xb7e51000)
libm.so.6 => /lib/libm.so.6 (0xb7e2b000)
libc.so.6 => /lib/libc.so.6 (0xb7d04000)
libX11.so.6 => /lib/libX11.so.6 (0xb7be8000)
libXext.so.6 => /lib/libXext.so.6 (0xb7bd7000)
/lib/ld-linux.so.2 (0xb7fac000)
libXau.so.6 => /lib/libXau.so.6 (0xb7bd3000)
libXdmcp.so.6 => /lib/libXdmcp.so.6 (0xb7bcd000)
|
|
|
|
|
Logged
|
Occupation:
programmer
Fluent in:
Polish, Assembler, C/C++, Java, bash/awk
|
|
|
kit89
Member
Cakes 6
Posts: 636
Shoot him..
|
|
« Reply #65 on: July 16, 2007, 02:04:35 PM » |
|
Tested OA_DM2.
Spec:
AMD Sempron 1.5Ghz
1GB DDR 400MHz
GeForce 6 6600 256MB
Kubuntu Feisty Fawn
Running Open Arena 0.7.0
|
|
|
|
|
Logged
|
|
|
|
amdfanatyk
Nub
Cakes 0
Posts: 19
|
|
« Reply #66 on: July 16, 2007, 04:16:34 PM » |
|
Because of quite large size of executable file (1.6 MiB) I think it has been compiled with -O3 - some funny person chose that flag as most people know it almost always breaks everything. -O2 -fomit-frame-pointer -ffast-math -funroll-loops -falign-loops=2 -falign-jumps=2 -falign-functions=2 -fstrength-reduce -fstack-protector -pipe -march=athlon-xp
Also q3a executable file is smaller than 1 MiB so it probably uses -O2 too. I've tested with fun, oa_dm2, oa_dm5 and everything works properly. Just build it with -O2 and everything will be ok. I think there will be no need to limit bots' number. Probably most problems occur just because of too strong optimization. |
|
|
|
|
Logged
|
Occupation:
programmer
Fluent in:
Polish, Assembler, C/C++, Java, bash/awk
|
|
|
amdfanatyk
Nub
Cakes 0
Posts: 19
|
|
« Reply #67 on: July 16, 2007, 05:18:53 PM » |
|
Finally I have removed more shitty flags. Now game runs faster and executable file is even smaller (1 MiB). -march=athlon-xp -O2 -pipe -fstack-protector -Wall -fno-strict-aliasing -Wimplicit -Wstrict-prototypes
|
|
|
|
|
Logged
|
Occupation:
programmer
Fluent in:
Polish, Assembler, C/C++, Java, bash/awk
|
|
|
anyone
Half-Nub
Cakes 2
Posts: 86
I may weep openly
|
|
« Reply #68 on: July 16, 2007, 06:25:27 PM » |
|
Small request: could it be possible to have randomly chosen footsteps instead of cycle through the four samples ?
|
|
|
|
|
Logged
|
I'll wear my badge, a vinyl sticker with big block letters adherent to my chest, tells to all your friends: "I am a visitor here, I am not permanent"
(The postal service "district sleeps alone tonight")
Promoted Mapper of the Century by some CPMA mid/high experienced player (luvyoalot, man)
|
|
|
dmn_clown
Posts a lot
Cakes 1
Posts: 1324
|
|
« Reply #69 on: July 16, 2007, 09:38:17 PM » |
|
/EDIT3: I've just tested locally on Hydronex map, and no bug happened, even with very very high FPS and with maximum bots I could add. Could it be linked to packet management as I suggested before ? Note that I dont have any network problem, no router, no firewall, and a ping between 15 and up.
It's possible, try adjusting your network configuration in the cfg file to match your connection. http://www.techspot.com/tweaks/quake3/quake-9.shtml |
|
|
|
|
Logged
|
|
|
|
|
GrosBedo
Member
Cakes 20
Posts: 710
|
|
« Reply #70 on: July 17, 2007, 05:56:17 AM » |
|
Thank you very much for this link it's very useful, but I found where the problem exactly was. It was effectively a problem of network, through on my side. I thought I disabled firewall (ZoneAlarm) completely but no, there was some residue. Now I've completely turned it off and it works very smoothly, it's completely playable (but I still miss Void4  (( ). Sorry for all these troubles, but now if someone has some similar problems with FPS it's because of network configuration for sure. Now for another bug that is for sure in the game, and which is new, is a bug disabling people from connecting to the game. It seems this bug comes when the players are playing (not spec) in a map, and the server set as nextmap the same map as previously. The players then don't connect and keeps in the score window, and their status is marked as "connecting...". Solution: command /reconnect or go to Spectator. |
|
|
|
|
Logged
|
|
|
|
anyone
Half-Nub
Cakes 2
Posts: 86
I may weep openly
|
|
« Reply #71 on: July 17, 2007, 01:20:26 PM » |
|
(Don't know if it has been made on purpose -ie bunnyjumping- but) when you press and hold the "jump" key while in air, the player will jump once more when he/she reach the floor.
|
|
|
|
|
Logged
|
I'll wear my badge, a vinyl sticker with big block letters adherent to my chest, tells to all your friends: "I am a visitor here, I am not permanent"
(The postal service "district sleeps alone tonight")
Promoted Mapper of the Century by some CPMA mid/high experienced player (luvyoalot, man)
|
|
|
dmn_clown
Posts a lot
Cakes 1
Posts: 1324
|
|
« Reply #72 on: July 18, 2007, 11:04:08 AM » |
|
Many people would be upset if that were "fixed" consider it a feature |
|
|
|
|
Logged
|
|
|
|
amdfanatyk
Nub
Cakes 0
Posts: 19
|
|
« Reply #73 on: July 18, 2007, 12:21:19 PM » |
|
I'm not sure if it was intended but name of bot should be rather "Widow" than "Widowe".
|
|
|
|
|
Logged
|
Occupation:
programmer
Fluent in:
Polish, Assembler, C/C++, Java, bash/awk
|
|
|
|
fromhell
|
|
« Reply #74 on: July 18, 2007, 02:16:08 PM » |
|
I asked lazureus about that spelling too, the e is intentional
|
|
|
|
|
Logged
|
asking when OA3 will be done won't get OA3 done. Progress of OA3 currently occurs behind closed doors alone I do not provide technical support either.new code development on github |
|
|
|
