Sauerbraten 2006-04-26 release!

by Aardappel_ on 04/26/2006 08:01, 106 messages, last message: 09/02/2006 04:19, 120816 views, last view: 05/03/2024 06:30

http://sourceforge.net/project/showfiles.php?group_id=102911

# tweaks & fixes to the capture gameplay mode
# added high resolution skyboxes, and some of the g_pack textures to the standard set
# fixed newent sending invalid attributes in coopedit mode
# models with a specularity of 0 now have a specialized shader ("nospecpvmodel") for faster drawing (mainly for vegetation)
# grenadelauncher damage increased to max 75
# multiplayer players & flag models will not be culled no matter how far away they are
# added waterfall effect for water material sides
# added pistol ammo (ent type "cartridges")
# added linearly-interpolated normals for lighting (controlled by "lerpangle", "lerpsubdiv", and "lerpsubdivsize" vars)
# added HW occlusion culling support (requires NV GeForce3-class hardware or better)
# height map mode. will hopefully allow for easier terrain editing.
# reduced overdraw dramatically
# reduced size of vertex data by using short instead of float
# removed ZYX order inconistency from math classes
# fixed lights placed outside the world not casting shadows
# fixed grenades being stopped by clip material and being stuck in walls
# made networking protocol use less bandwidth
# calclight now uses multi-sampling and adaptive sampling
# faster mapmodel ray intersects with spheretrees instead of bsp
# shorter fuse on grenades (2 seconds)
# fixed hudgun light computation related crash

   Board Index   

#50: backtrace

by Sat Apr 29 11:32:35 2006 on 04/29/2006 13:29

backtrace of crash if map nmp8_d is loaded:

#0 0x00000000 in ?? ()
#1 0x08066320 in newquery (owner=0x9118c98) at engine/octarender.cpp:1178
#2 0x08068d31 in rendermapmodels () at engine/octarender.cpp:1342
#3 0x08076565 in gl_drawframe (w=640, h=480, curfps=71.5291214) at engine/rendergl.cpp:495
#4 0x0805aa40 in main (argc=2, argv=0xbffad8b4) at engine/main.cpp:379

reply to this message

#51: Re: importcube crashes Sauer

by Aardappel_ on 04/29/2006 20:53, refers to #46

mitaman, if you could use default sauer textures that would be great. There are a lot of new ones since sauer you could look at.

We are interested in making your sp maps fast, and not crash :) can you please send the sauer versions of those maps to me and/or eihrul and we will do some performance analysis.

reply to this message

#52: Re: capture!

by Aardappel_ on 04/29/2006 20:59, refers to #47

thanks for the feedback on the capture mode. I am hoping that the troubles making teams is because its most people's first time playing sauer or capture, and that it will improve. But yeah, who knows we need a forced team setup. As long as people can still make their own team if they want to.

Maybe a "rpickup" command would be good (assigns people to teams), or otherwise one of those "join a team" menus upon connect.

reply to this message

#53: Re: importcube crashes Sauer

by MitaMAN in the cave on 04/29/2006 21:08, refers to #51

Sure, I'll send out mpsp4/mpsp5 Sauer versions to you Aard. The config files will also be included. If for some strange reason you need the "custom textures" for those maps, just copy the "mitaman folder" from cube.

I'm sure I can use many of the new textures from Sauer, but there are a few that are "dear to my heart" that i really enjoy using.

Thanks for looking into the performance issues - MitaMAN

reply to this message

#54: Re: ..

by absinth on 04/29/2006 23:19, refers to #37

>It would be kinda lame if we have to entirely disable occlusion for all ATI cards.

right. but since this looks like a common problem i fear that many people might conclude that sauerbraten sucks and not that their driver sucks and they have to do "/oqfrags 0" to fix it. perhaps this solution should be mentioned in a prominent position in the docs. do you have any estimates how much of a performance hit doing "/oqfrags 0" causes?

also is there any chance to extract an testcase for the hw-occlusion features as sauer needs them, so i can submit a bug-report?

perhaps we should look for a way to disable hw-occlusion for all ATI-cards on Mac OS X (until this issue is resolved)?

reply to this message

#55: Re: capture!

by Passa on 04/30/2006 01:30, refers to #52

aardappel: maybe give the master control of the teams?

absinth: are you sure its not just the macintosh port in general? i know some older macs came with nvidia cards.. if someone has one, could they test it out?

reply to this message

#56: ..

by makkE on 04/30/2006 01:40

I would welcome a "join team" screen at join.
Free team could still be available via vote, but 2 teams that you´re forced into is just more practical on a pub imho.

reply to this message

#57: Re: ..

by enigma_0Z on 04/30/2006 02:37, refers to #54

AAAnd in linux please?

reply to this message

#58: Re: ..

by Passa on 04/30/2006 03:46, refers to #56

the master having control would be nice too, so the master can make the teams 'even'. or that could be solved by having an auto team balance option, like in other games.

also, just a note on master, it would be real nice to have say a masterpass command,so we can enter a password and become master on our own server. im sick of having to ask my friends to disconnect and reconnect so i can be master.

reply to this message

#59: rANDOM crashes with ded.server

by Passa on 04/30/2006 08:46

ok i have a system with the following specs as my dedicated server
Pentium 3.4Ghz
1GB DDR2 RAM
Windows XP Professional SP2

now as its a family computer, i created a new user account, then from there i load the server, and if i leave it there for a while the sauerbraten server crashes with an error window, debug, dont send or send error report

i also run a cube server and that has no problems, could it have something to do with the server running after i have 'fast switched' users back to the main one?

reply to this message

#60: capture :D

by Passa on 04/30/2006 08:47, refers to #59

also, a capture mode bug, i played again today, and if the enemy is standing near the flag as well as you, then you kill him while still remaining within the radius of the flag, the flag stops switching teams, so you gotta leave its radius then re enter to continue capturing it

reply to this message

#61: ..

by CC_machine4 on 04/30/2006 21:35

hmm.. no sendmap yet..

reply to this message

#62: Re: ..

by Aardappel_ on 04/30/2006 23:38, refers to #54

people don't read docs, putting something in there more prominently won't help.

The only solution for now is to disable oq on those platform/card combinations where it doesn't work. So put an macosx if() or #ifdef in there that will force it off (in glinit).

reply to this message

#63: article

by rancor on 05/01/2006 21:08

The occlusion culling release made the front page of Phoronix (http://www.phoronix.com/ ) Unfortunately, it appears that they've taken all of their screenshots with a Nvidia driver with the texturing bug.

reply to this message

#64: Re: article

by Passa on 05/01/2006 23:03, refers to #63

I noticed that too.. how did they get the nvidia bug with the occlusion release? weird..

reply to this message

#65: Re: article

by kurtis84 on 05/02/2006 02:02, refers to #64

A shame actually. I wonder what video card they used to be having this problem? I didn't think anyone had this happening since the shader version came out.

I asked them what card they used to take those screenies, and warned them about the ATI problems, and the old nvidia cards running 8x.xx drivers that disable the shaders and cause those terrible looking texture streaks :/ ...hopefully they'll do an update.

reply to this message

#66: Re: article

by Aardappel_ on 05/02/2006 07:54, refers to #65

actually, with shaders the problem should be permanently fixed, as we bypass the opengl state that causes this bug. So the screenshots must be from an older release.

I wonder what the guy making the screenshots must have been thinking... the walls are supposed to look like that.

reply to this message

#67: Re: article

by Passa on 05/02/2006 08:42, refers to #66

What I do not understand is that he actually calls it the occlusion edition.

reply to this message

#68: mapmodel bug found

by MitaMAN in the cave on 05/03/2006 00:46

I found two bugs when placing mapmodels. (Or am I just doing something stupid?)

1. In a map originally created in Sauer, I place a mapmodel (from the editing menu), no problem. When I scroll through the various mapmodels Sauer crashes. I use a config that makkE showed me back in the CUBE days (and it worked great!) that allowed rotate, texture change, mapmodel, etc..

This is the config in my autoexec.cfg for editing models "in-game":

bind kp8 "entproperty 0 1"
bind kp9 "entproperty 0 -1"

bind kp5 "entproperty 1 1"
bind kp6 "entproperty 1 -1"

bind kp2 "entproperty 2 1"
bind kp3 "entproperty 2 -1"

bind kp_minus "entproperty 3 1"
bind kp_plus "entproperty 3 -1"

2. In a map converted from CUBE into Sauer I won't even see the mapmodel, just the "blue sparklies". I tried to save and reload the map, but I only see the blue sparklies.

Aard, I am using the newest sauer.exe you sent me (frame rate opptimized for cube import of mpsp4/mpsp5).

If I'm doing something dumb, let me know.

MitaMAN

reply to this message

#69: Re: mapmodel bug found

by eihrul on 05/03/2006 01:01, refers to #68

The reason it's crashing is because once the mapmodel index gets negative, it's now referencing an invalid model. I can make it not crash, but it's not going to cycle back to the highest mapmodel again.

reply to this message

   Board Index