Bugs

by Al_Capone on 06/06/2002 11:00, 92 messages, last message: 10/31/2002 08:08, 15531 views, last view: 05/05/2024 15:57

Anyone here tired of loosing because of some stupid bug in cube ?

   Board Index   

#74: Bug ???

by =Riot777= on 09/21/2002 22:36

This was a false alarm people, Eihrul has right. I use for mistake an Sdl mixer v 1.2.0 the game works only on 1.2.1 and above I thik. Sorry folks again for false alarm. But you have an another CUBE community Member ;)

BIG thanx to SirLiveAlot wich is a great man.

Regards Riot777

reply to this message

#75: np

by SirLiveAlot. on 09/22/2002 02:21

no problem

reply to this message

#76: Voodoo3 GLX Linux alpha tex problem

by Corvus on 10/18/2002 21:34

Well thats it, under glx Xserver i have textures like the fonts , water surface and such where you should be able to see through, i have solid white boxes where there should be text letters and solid white polygons where there should be water surface. Not only in cube but in several other games, too. this is definitely a driver problem. nevertheless some games (like Quake 1,2,3) do it right with a "voodoo-workaround".

does anyone know what to change where to make thosose alpha textures work with Voodoo GLX driver, so i could implemet a workaround for a cube Voodoo patch? Would be helpful since i'm planning to do some 3D projects myself and fear to run into the same prob.


(the Voodoo cards are unfortunately the last cards which a fully supported and not proprietary and buggy Linux driver. (at least it has far less bugs that the Nvidia shit or ATI crab) and i wont buy a new graphic card until thers a nother "nearly perfect" driver for it. what use is a card which could draw 256bit Textures and 16 quadrillion polygons a second if thers no driver to support even simple 2d acceleration like its necessary to view fullscreen videos in more than 10fps? dont believe it? look at that ATI driver, the card could even do mpeg decompression, if the driver would :-(
)

reply to this message

#77: Re: Voodoo3 GLX Linux alpha tex problem

by Shockie on 10/19/2002 06:02, refers to #76

there is a great work about i know, but it might be expensive =D

buy a new card =)

geforce 2's are under AU$100 these days, that ~US$50

reply to this message

#78: Re: Voodoo3 GLX Linux alpha tex problem

by Corvus on 10/19/2002 07:01, refers to #77

yeah thanx, i know that, do you know the NVidia GLX driver for this card?

it replaces the whole DRI part of the Xserver with its buggy proprietary and pre-compiled for 80386 or whatever trash.

This way compromising all attempts of running dual-screen configurations with more than 1 graphic-card, (xinerama), borking the GL libs to be no longer software rendering capable (important if u want to render a GL szene not for gameplay but for work), crashing the system randomly by borking the kernel with closed code nobody really knows what it does, ...

i can say, if that nvidia driver would be an option for me i had bought an nvidia card long ago. the speed of the Voodoo3 is far behind todays cards.

Curse Nvidia for killing 3Dfx !

if i am wrong since nvidia meenwhile completly redesigned their driver, please tell me, but it would surprise me, if so.

reply to this message

#79: Re: Voodoo3 GLX Linux alpha tex problem

by Shockie on 10/19/2002 07:18, refers to #78

just go buy a new card and be done with =)

reply to this message

#80: skybox bug

by spentron@PP on 10/19/2002 19:46

I'm seeing a triangle in the corner or skyboxes when fog distance is increased beyond 160. Affacts all skyboxes and maps. Increasing fog distance actually reduces wqd and improves framerate on my machine on my large map areas, although I saw no change on the startup map. This default fog setting also seems a bit closer than the previous Cube version. It also seems slightly slower at times but that can be fixed by increasing fog distance, but then I get the triangle. System is 366 Celeron, vanilla TNT. I can supply a screenshot and we should see how widespread the problem is if there's a problem recreating this bug.

Other than that, thanks and congrats to everyone responsible for the current release. The hudguns create a much more professional and slick impression, beyond anything I expected. The new physics are great, and I can now upgrade my machine without worrrying about creating maps unplayable on lower spec machines.

reply to this message

#81: con't

by spentron@PP on 10/19/2002 19:50

that was corner OF skyboxes. Visible when looking up or down ~45 degrees and 45 degrees from the grid. With a high fov I can get an untriangled skybox corner on the screen if I look at a different angle. The triangle schrinks and dissappears 2 mm from the top or bottom of the screen when looking up and down at fov=80.

reply to this message

#82: skybox bug -- hello??

by spentron@PP on 10/24/2002 19:20

Bump or something --
anyone else tried this? Bug or not? :

"I'm seeing a triangle in the corner of skyboxes when fog distance
"is increased beyond 160. Affacts all skyboxes and maps.
"Visible when looking up or down ~45 degrees
"and 45 degrees from the grid. I can get an untriangled
"skybox corner on the screen if I look at a different angle.

reply to this message

#83: sounds like the farplane is set incorrectly somehow...

by Aardappel on 10/25/2002 11:00

which would be a bug. it however doesn't happen on my machine and I have tried a variety fog distances. I have no idea why different drivers/card may affect this; I could fix it by increasing the farplane arbitrarily but that is not very desirable.

reply to this message

#84: Re: sounds like the farplane is set incorrectly somehow...

by spentron@PP on 10/25/2002 17:55

"it however doesn't happen on my machine" Ok. Was starting to expect that.

However I will still deal this possibly happening through using map.cfg's with high fog distance areas not being used where there is skybox (I'm using a fog 400 for an underwater area -- far better than the old set limit).

Another thing that's almost certainly drivers is the desktop resolution doesn't get reset now on my machine. It was fine on the previous version though except it would leave a blank taskbar icon that went away when clicked on, which it doesn't do now.

reply to this message

#85: Re 2: sounds like the farplane is set incorrectly somehow...

by spentron@home on 10/28/2002 01:18

There's still an issue of the closer default fog and some slowdown.

Finally did a straight compare of the Cube versions and rather shocked. On Shadowland start map, I'm getting wqd 5 times what I get with the old Cube version. With fog set to 1000, it drops to only slightly higher than the old version. A wqd of over 5000 in little more than a gray box is absurd. The new version requires fog about 250 to actually look the same, also, otherwise it is foggier.

Other maps show some difference but not near as large. I am quite as a loss to understand why this would be, that my little start map has a higher wqd than any point in Infidel, a big SP map, for example. I'm seeing an increase of about 4000 everywhere in my map, also.

It also seems odd that the default fog setting is exactly the maximum before I get this "triangle bug" ... is it possible to set that farplane based on the fog setting, maybe? Like I said, I get better performance at high fog settings.

I was also noticing some other vision distance reduction in the new version. As Cube already does a good job with big spaces, I would consider this an aspect to build on, rather than allow to reduce.

I realize this is a new version with some definite improvements, but should avoid 1 step forward, one step back ... while it should run fine on good hardware, something has definitely changed.

reply to this message

#86: Re: Re 2: sounds like the farplane is set incorrectly someho

by Aardappel on 10/28/2002 09:56, refers to #85

ok, I looked at it, and the wqd is indeed a bug in cube, it doesn't cull things beyond the foglimit at all anymore, even if occluded.

This bug however only shows up in your map because it has such large areas which are not properly occluded. Rise above your map in edit mode, and note (for example in the start area) how you can look through to areas only separated by a corner. if you make these corners a proper wall, you will stop seeing into the other rooms and your wqd will go back down.

As to farplane, it is being set depending on the fog, though I guess on your card gives the incorrect results. I will test both problems for the next release.

reply to this message

#87: Re: Re 2: sounds like the farplane is set incorrectly someho

by spentron@PP on 10/28/2002 22:13

Cool, explained. The contradiction between the readme on fog and actual results was one confirmation of something not right (I don't think it's possible for a machine to be "fog limited" without being "geometry limited" as it's essentially the same thing)... Shorter fog differences shouldn't always improve performance, but they should sometime. To a lesser extent, other maps are affected by the fog differences. I'll check the rest out.

reply to this message

#88: "Triangle" screenshot

by spentron@PP on 10/29/2002 18:35

I've put up a screenshot of the "triangle" temporarily: http://spentron.web1000.com/cube/misc.html ... it's funny looking ;) .

Aard, I can't see anything in Shadowland that would cause the problem nor do I see much not occluded with the raise-above-the-level trick. The start area has full solid walls (they fall away with non-solid command). The first play area does have some walls that only come together at an edge, but changing that only bought down the wqd by 5 or 10 or so (from 9000, ouch). Any possibility it's just too big (total stuff, not just the size of the level "blank") or even that maybe I cut away an original perimeter wall or that it has an extreme height differential? I didn't try putting a second wall entirely around the start area, that's the only other thing I can think of to try.

reply to this message

#89: wqd mystery: solved

by spentron@home on 10/30/2002 07:54

Water.

I could see some small light blue areas when above the map below. Moving towards them below the map, they are huge, several times the largest water area. Went back to the totally dry, sealed start room and cranked up watersubdiv. Bingo. wqd down to ~normal.
Explains why the wqd increase seemed additive, about a 4400 increase over a large section of the map.

I don't pretend to understand much more than that. Anyway, a quick fix, for now. 6/24 version doesn't show this behavior.

reply to this message

#90: Re: wqd mystery: solved

by Aardappel on 10/30/2002 10:54, refers to #89

the size of the water is the square which contains all water areas visible... as I explained you can see far out int the level from the start room which causes this. Look for the diagonal staircasy stuff in the distance. If you fix this (use solid walls in those places, not mere corners) the water problem will also go away.

reply to this message

#91: Re: wqd mystery: solved

by pushplay on 10/30/2002 23:33, refers to #90

Water's a little bit of a problem because if you slope from dry land to beneath the water level just right, then the water won't be rendered. But you can still hear yourself splash around in it, and the screen does the usual blue thing when you submerge.

reply to this message

#92: Re: wqd mystery: solved

by spentron@home on 10/31/2002 06:13

"Look for the diagonal staircasy stuff in the distance" [when above the room in edit mode]

The only stuff like that in the start map (with "START") is *inside* the room. I do see some dark areas right outside the room, through a solid room, which oddly go away if I light them (in both versions on that). The walls are solid and have proper corners, and I do know exactly what you mean. I did even try adding another 8 cube thick layer of wall completely around the room, no change in the water effect. There also may be water being rendered and then covered over so I can't even see it, because watersubdiv affects wqd even when no water is in the fov.

Anyway, I feel like I'm getting negative harping on this, but I also don't think you're going to convince me there isn't something funny going on. I will find ways to do cool things with the engine either way ;) .

Pushplay, I know about that (#91), in fact I mentioned it in the hieghtfield/type/tag discussion a couple weeks back. I even considered doing something gimmicky with the effect. I just played around with it, and it does seem like the new version is more predictable in that respect (I can't get a water surface partially over a pool anymore, strangely floating in "midair", although the water surface disappears when the token nonsolid at the bottom of a heightmapped pool goes outside the fov).

reply to this message

#93: Re: wqd mystery: solved

by Shockie on 10/31/2002 08:08, refers to #91

oooha i would like to see this!

reply to this message

   Board Index