PLEASE TEST: Nexuiz 2.0.1 pre-release

[dfkgmasdfnasodrg]

i use linux sdl and i now get that bug where when switching from fullscreen to window mode and back. the gamma is off again and i have to toggle hwgamma from menu.

also while testing i joined a random server and the scoreboard said next to my name time 12 ? isn't this supose to be the time i've spent on the server?

[GDI+]

Hello.

I’ve noticed a couple of slight issues (more niggles than anything ) with v2.0.1 although a number of them are minor and/or trivial and are probably best ignored. I also noted a couple of suggestions, for future versions, but I think several of these have already been mentioned before and are on the DarkPlaces TODO list.

Nexuiz binary: WGL
OS: Windows NT5.1 (XP) - Service Pack 1b
GFX: Nvidia GeForceFX 5950 Ultra – 256MB (ForceWare v91.31)
SFX: Creative SB Audigy 2 ZS Platinum Pro (v2.08.0002 Beta)

Bugs:
1) Crash when downloading map at 0% progress - RBI custom map server with tznex03. Re-launching the game and connecting again worked fine – no crash at the end of map download (only happened once, I played several severs prior to this and downloaded other maps in-game with no problem).

2) Last weapon switch (laser 2nd or 'weaplast' bind) infrequently forgets last weapon. When this occurs the 'weaplast' command displays invalid weapon error (I do use the new 'bestweapon 98765432' command, although I don't think it's the cause).

3) Dropping a weapon and then dieing before weapon switch completes causes another identical weapon to drop (v2.0.1 has faster weapon switch so this is unlikely now). One weapon has the default ammo count, i.e. Electro = 25, while the other weapon has the remaining ammo when dropped by the dead player up to the max. default for the weapon, i.e. Electro <= 25. This is potentially exploitable to gain more ammo for a weapon at the cost of a life, heavy price indeed.

To artificially test, try the following:

Code: Select all

alias "+test" "dropweapon"
alias "-test" "kill"
bind "ENTER" "+test"

...then press ENTER

4) Changing most r_* variables/commands for example, in the console, forces suicide (v2.0.1 only). This happened repeatedly the other day when I was constantly adjusting graphics-related variables. Restarting the client fixed the problem and has not occurred again despite attempts to reproduce the error (somewhat strange).

Minor Issues:
- Runningman 'plane mismatches dist' error on level load.

- Warfare invisible blocking volume on approaching left-side of lift (bottom-platform) (walk into the left corner wall, before the lift, at about 30 degrees facing towards the lift and it will slide past until it hits an invisible wall/notch, does this from other side of same corner - newbies & drunk people might suffer slight annoyance ).

Suggestions (future versions):
- Warfare lift button could be tiggered by weapon fire as well (enemy can shoot the button to stop a fleeing opponent from using the lift).

Engine-related:
- Max. particles hard-limit; useful for lower-end cards when rendering large amounts of blood/weapon fire, i.e. the Electro secondary detonation with large numbers of cells. When approaching user-specified limit, i.e. 500, use less particles for effects - dynamically adjust particle number/quality downwards on-the-fly.

- Dynamic lighting unusable on lower-end cards due to the per-pixel lighting introduced in v2.0 (appears to show bump-map type effect on surfaces when shells explode). Using ‘r_shadow_bumpscale_basetexture "0"’ & ‘r_shadow_bumpscale_bumpmap "0"’ shows v1.5 style, no bump-map effect, but no better FPS. Electro performance is by far the worst when detonating many secondary cells (by default I set ‘r_glsl_deluxemapping 0’ for better performance yet a similar effect still persists with dynamic lighting at a very high cost to performance).

- Render-to-texture support and motion blur effects.

- Better q3map support, i.e. shaders - texture bending and localised fog.

- Player-name indicator when acquiring target with crosshair.

Some of these errors might be specific to my machine/setup, although it does seem somewhat strange. Sorry if this is not the right place to post such suggestions you’ve probably heard many times before and if the list seems a little demanding it was not my intention.

If there’s any additional information or testing you need I’d be happy to help in anyway I can. BTW, thanks for letting us test the new pre-release version of Nexuiz.

EDIT 1:

also while testing i joined a random server and the scoreboard said next to my name time 12 ? isn't this supose to be the time i've spent on the server?

I've noticed this as well, happens both on-line and locally. Seems to be showing the time the map has been running on the server and is the same for everyone.

EDIT 2:

Minor Bugs:
- Hit sound sometimes plays when shooting corpses (presumably your weapon fire was in transit before the player died).
- F5 team selection menu available when playing demos (F1-F4 are correctly disabled).
- Screenshots are created locally on clients machine while viewing demos that were created with the 'sv_autoscreenshot' server option set.

Suggestions:
- Ability to switch players while viewing demos.
- Make decals appear correctly on curved surfaces and adjacent brushes.
- Prevent screensaver activating when watching/spectating for long periods (could send key-codes periodically to fool the screensaver user-idle timeout).

Sorry again, for the never-ending list of issues. Maybe I should just put a "Detrimental to your health" warning message at the beginning of my posts in future.

[divVerent]

Bugs:
1) Crash when downloading map at 0% progress - RBI custom map server with tznex03. Re-launching the game and connecting again worked fine – no crash at the end of map download (only happened once, I played several severs prior to this and downloaded other maps in-game with no problem).

Can you send me the Dr. Watson report? Start/Run, then call "drwtsn32", save the crash report and send it to polzer followed by the at sign followed by informatik.uni-frankfurt followed by a dot and "de".

2) Last weapon switch (laser 2nd or 'weaplast' bind) infrequently forgets last weapon. When this occurs the 'weaplast' command displays invalid weapon error (I do use the new 'bestweapon 98765432' command, although I don't think it's the cause).

Does it happen when you switch to the same weapon twice. Can you confirm that?

3) Dropping a weapon and then dieing before weapon switch completes causes another identical weapon to drop (v2.0.1 has faster weapon switch so this is unlikely now). One weapon has the default ammo count, i.e. Electro = 25, while the other weapon has the remaining ammo when dropped by the dead player up to the max. default for the weapon, i.e. Electro <= 25. This is potentially exploitable to gain more ammo for a weapon at the cost of a life, heavy price indeed.

Will look into that, although it doesn't sound like it's "generating" ammo.

4) Changing most r_* variables/commands for example, in the console, forces suicide (v2.0.1 only). This happened repeatedly the other day when I was constantly adjusting graphics-related variables. Restarting the client fixed the problem and has not occurred again despite attempts to reproduce the error (somewhat strange).

No idea, sounds like an engine problem.

Minor Issues:
- Runningman 'plane mismatches dist' error on level load.

Sounds quite obviously like a bug in the map editor, but shouldn't be serious.

- Warfare invisible blocking volume on approaching left-side of lift (bottom-platform) (walk into the left corner wall, before the lift, at about 30 degrees facing towards the lift and it will slide past until it hits an invisible wall/notch, does this from other side of same corner - newbies & drunk people might suffer slight annoyance ).

I also noticed that, but nobody wanted to hear it.

- Dynamic lighting unusable on lower-end cards due to the per-pixel lighting introduced in v2.0 (appears to show bump-map type effect on surfaces when shells explode). Using ‘r_shadow_bumpscale_basetexture "0"’ & ‘r_shadow_bumpscale_bumpmap "0"’ shows v1.5 style, no bump-map effect, but no better FPS. Electro performance is by far the worst when detonating many secondary cells (by default I set ‘r_glsl_deluxemapping 0’ for better performance yet a similar effect still persists with dynamic lighting at a very high cost to performance).

LordHavoc: PLEASE!

- Player-name indicator when acquiring target with crosshair.

Also an evil engine thing.

[Gut_Eater]

from Vociferous:

i use linux sdl and i now get that bug where when switching from fullscreen to window mode and back. the gamma is off again and i have to toggle hwgamma from menu.

I have noticed the same, but in version 2.0.0. It also appears if i am changing the resoluttion.

[morfar]

also while testing i joined a random server and the scoreboard said next to my name time 12 ? isn't this supose to be the time i've spent on the server?

It shows maptime instead of playtime. I hope LH will fix that soon. He knows about it

[SavageX]

gl_combine 0 should get rid of bumpmapping in the non-GLSL path. May have other side effects, though.

[GDI+]

Can you send me the Dr. Watson report? Start/Run, then call "drwtsn32", save the crash report and send it to polzer followed by the at sign followed by informatik.uni-frankfurt followed by a dot and "de".

I do recall the Dr. Watson error message that appeared; it had a debug button which then displayed additional data that didn't make much sense to me. I've tried executing "drwtsn32" however there doesn't appear to be a valid log available:

http://www.carrickj.plus.com/Nexuiz/DrWatson.png

I also tried the Event Viewer MMC snap-in as well but to no avail – it simply displayed that an error had occurred with nexuiz.exe and an associated time-stamp.

The error may have been caused by me first playing for extended periods on Qantourisc's "New Bot Test" server which is running modified server-side quakec with some non-standard settings such as game-speed. I do remember the initial drwtsn32 error debug mentioned "New Bot Test" or something similar. Sorry I can't be of more help, I'll try my best to reproduce the bug again for you.

Does it happen when you switch to the same weapon twice. Can you confirm that?

This happened multiple times when testing off-line on a local server. Haven’t noticed it on-line yet but then again they were v2.0 servers thus far.

I did switch between weapons and the laser very often. I used a combination of laser 2nd fire and ‘weaplast’ to return to last weapon. I also changed to a different weapon while currently using the laser and continued to use the combination of ‘weaplast’ and laser 2nd.

Just in case it might help I use the following config relating to my mouse:

Code: Select all

bind "MOUSE1" "+attack"
bind "MOUSE2" "+attack2"
bind "MOUSE3" "weaplast"
bind "MWHEELUP" "weapprev"
bind "MWHEELDOWN" "weapnext"
bind "MOUSE4" "bestweapon 98765432"
bind "MOUSE5" "impulse 1"
sensitivity "12"
scmenu_mouse_speed "2"

If you want my full config or any other information just let me know. I’ll see if I can replicate any of these problems, although they may just have been one-time anomalies caused by a combination of my machine, config and user error.

gl_combine 0 should get rid of bumpmapping in the non-GLSL path. May have other side effects, though.

Thanks for the tip and caution. It's certainly worth trying out.

[Urmel]

hi there, I doubt that it'll be helpful but I just have to post this (and I'm sure I can forget about the system specs):

Having setted up the forum pre a few minutes ago I first started a campaign and shortly after that I joined one of the RBI 2.0.1servers, playing against a bot. then the system crashed after about 10 minutes.

I'm used to my system crashing time after time in nexuiz, not a big deal. in the former versions however it was always a classical, immetiate crash. this time it was another style, no "crash", more like "being stopped in a forced way". like a marine fighter getting stopped when landing on an aircraft carrier. Strange new behaviour. never experienced a crash like that. anyway. fascinating.

ah, my system is very hot at the moment due to the high temperatures in my flat so this could be a reason...

[GDI+]

^^ Sounds not too dissimilar to the crash I experienced also on a RBI server. The game didn't crash to the desktop, it was still running full-screen but with no activity except for the Dr.Watson dialog waiting for my response. I also played some maps locally before playing online and then switching to the RBI extra-map server.

Curious coincidence?

My room and system are also very "warm", more like hot-as-hell actually.

[Urmel]

okay, now I'm quite sure it was a heat issue. system crashed shortly after posting that, playing the 2.0 version locally. but it crashed the old, immediate style then. different to GDI's crash my system freezes completely. nothing I can do than pushing the power button.

I took my hoover and cleaned my rig. Now it's showing up 2°C less than before, lets start the game and see