Sauerbraten release 2006-06-11 out!

by Aardappel_ on 06/12/2006 07:21, 322 messages, last message: 07/31/2007 16:33, 375994 views, last view: 11/01/2024 14:36

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

As usual, feedback SPECIFIC TO THIS RELEASE in this thread! thanks!

* scripting language: every command/block/exp can now return string values (no more use of variable "s", much cleaner code)
* scripting language: reduced allocations a great deal, should execute faster now
* scripting language: can now write X = Y instead of alias X Y for more readable code
* If sauer crashes on a non-programmer machine on Windows, it will now output a stacktrace (both in a popup and the console window) which can be copy-pasted and sent to the developers (do this before you press "OK" to close the program, from the console windows menu).
* added provisional "sendmap"/"getmap" commands for coopedit mode
* explicit sky polies are now outlined in editmode (toggled by "showsky" var)
* now shows names of entities floating in edit mode (in green, closest = red)
* any non-multiplayer games can now be paused (variable "paused", default key F1)
* fixed rockets exploding mid-air if they were aimed at a moving target
* copy/paste now work in coop-edit
* moved all mapmodels into the menus
* new SP respawning system! read "SP Respawning" in game.html
* console language macros now allow code generation (see config.html)
* added customizable file completion (via "complete" command)
* aliases and vars set in map cfgs are now sandboxed (don't persist)
* world shaders can now work with any number of textures (see "texture" command).
* support for specularity maps and glow maps on md2/md3 models (see md2.txt/md3_sauer_howto.txt)
* new "editbind" works like "bind", only the key is only available in edit mode. cleaned up keymap/default cfg files (no more bindings in keymap)
* added triggerable mapmodels (see editref.html for more info)
* removed the z-offset attribute from mapmodel info and ents and moved mapmodel ent skin attribute to mapmodel info (see editref.html)
* added support for full screen pixel shader effects (see setfullscreenshader, and the "option" menu)
* removed obsolete editheight command
* added powerful new heightmap mode! use getheightmap command [default H key] when there is a selection box to activate (see editref.html for more info)
* added selectbrush modifier [default J key] that switches between brushes defined in data/brush.cfg
* added copybrush [default K key] to use current heightmap as brush. see data/brush.cfg for more info.
* editable brushes for heightmap mode. see data/brush.cfg for more info.
* capture gameplay mode fixes
* ip address lookups now use the concurrent resolver
* added redo command [default I key]
* made importcube remip while importing to save memory
* removed some lighting/physics bottlenecks in SP mode
* removed 256 textures per map limit (is now 64k)
* fixed incorrect mapmodel shadows on Windows
* fixed invincibility bug

   Board Index   

#130: Re: OSX resolver bug

by eihrul on 06/20/2006 17:34, refers to #129

You are using the old resolver implementation. That's what you added the #ifdef for.

reply to this message

#131: HOM

by jengelh2 on 06/20/2006 17:39

I get Hall of mirrors everywhere. Not limited to this release, I have tried all past version, same problem. Is this normal? Any ideas on how to resolve this?

I have an NVIDIA GF2 MX400 (chip NV11 - yes, one of those which sets /nvidia_texgen_bug in Sauerbraten), running it under Linux 2.6.17 with NVIDIA 1.0-8756, xorg 6.9.0, SDL 1.2.9.

reply to this message

#132: Re: OSX resolver bug

by guest57 on 06/20/2006 20:44, refers to #130

Something must be different - but I can't see it either. Deadlock on the mutex occurs:

Thread 1: resolverclear(): in the SDL_SemTryWait

Thread6: resolverloop(): in the first SDL_LockMutex

- aka rpointon

reply to this message

#133: Re: Multiplayer

by bloo on 06/20/2006 21:38, refers to #120

Definitely not alone. Somewhere in this thread are my problems with gamma as well.

I also have the 'slideshow' problem on networked games.

reply to this message

#134: Re: OSX resolver bug

by eihrul on 06/20/2006 21:57, refers to #132

I implemented a workaround in CVS, but it may leave around stale threads depending on how nice the resolver is on MacOS about timing out.

OTOH, it gets rid of that pthread kluge that was there earlier that never really worked.

reply to this message

#135: Re: OSX resolver bug

by guest57 on 06/20/2006 22:52, refers to #134

Waw... sorry but no. The thread exits from the while() immediately because rt->killtime==0 the first time, or it rt->killtime==killtime other times...

reply to this message

#136: Re: OSX resolver bug

by eihrul on 06/21/2006 02:45, refers to #135

Try again.

reply to this message

#137: Re: HOM

by kurtis84 on 06/21/2006 03:47, refers to #131

I would be willing to bet that an mx400 is not enough to run sauerbraten...at all. Theres plenty of warnings about htis all over this forum.

After reading all of these bugs, etc, I cannot help but notice how many of them pertain to video cards/chipsets that are simply too damn old. So much discussion about how to overcome this issue. In the long run, this is just going to get worse. More GPU-intensive features are going into maps right now.

I know, at first glance Sauerbraten doesn't look any different than glquake1, as far as rendering...but it CAN do a whole lot more, and the future maps will show this. Of course, this means using things like the -f switch will ruin maps that use the shaders for geometry textures.

You say you don't need shaders?...well the engine DOES. Everything is being rendered via the shader system. By disabling the shaders, you're undoing all the hard work the programmers have done for the last 3 releases.

I hope at some point people stop asking these guys to work forward, and backward at the same time.

reply to this message

#138: Re: HOM

by Passa on 06/21/2006 04:52, refers to #137

Problem is, people think that if their PC can run Cube, it can run Sauerbaten (An example was some guy complaining the minimum system requirements I made were too high). What they don't realise is that Sauerbraten is now an exremerly modern engine.

Newer maps and new shader features along with normal maps is making my 7800GT lag, Sauerbraten now rivals modern engines.

reply to this message

#139: Error 0xc0000005 [0x4]

by karel on 06/21/2006 09:13

Hi,

I tried to run Sauerbraten on WinXP (notebook IBM T40). During startup Error 0xc0000005 [0x4] appears.

Any hints how to solve this?

thanks a lot,
Karel

reply to this message

#140: Re: Error 0xc0000005 [0x4]

by Passa on 06/21/2006 09:46, refers to #139

Laptop is too old no doubt, I have a T50 and it BARELY gets 7 FPS on metl3. Around 25 on aard3 though.

reply to this message

#141: Why is Cube2 Engine called "Sauerbraten"

by mdkuser_____ on 06/21/2006 12:43

Did anybody knows why the cube2 engine is called "Sauerbraten"? It's a funny name for a FPS game engine.

reply to this message

#142: Re: Why is Cube2 Engine called "Sauerbraten"

by Passa on 06/21/2006 13:26, refers to #141

Apparently Aardappel was eating it at the time he was thinking of a name. Its German for sour roast.

reply to this message

#143: Re: Why is Cube2 Engine called "Sauerbraten"

by mdkuser_______ on 06/21/2006 14:41, refers to #142

Oh I see, that's why the name of the cube2 engine cames from. But still one question: Is the cube2 GAME also called "Sauerbraten" or is this only the name of the ENGINE and the game has to be called cube2? The RPG also seems to have a german name, "Eisenstern" means iron star.

reply to this message

#144: Re: Error 0xc0000005 [0x4]

by TomG on 06/21/2006 14:48, refers to #140

I also get this- only for this new build however- it worked fine in the last release.
For me it appears to be when the map is loading- I get the first frame up, and then it bombs out. Stangely, it lets me load my own maps.
Here is the log:

init: sdl
init: enet
init: video: mode
init: video: misc
init: console
init: gl
Using GL_ARB_vertex_buffer_object extension.
WARNING: Using floating point vertexes. (use \"/floatvtx 0\" to disable)
WARNING: No shader support! Using fixed function fallback. (no fancy visuals for
you)
Using GL_ARB_occlusion_query extension.
WARNING: Non-power-of-two textures not supported!
init: world
game mode is ffa/default
init: sound
init: cfg
init: localconnect
init: mainloop
read map packages\\base\\metl2.ogz (0.8 seconds)
Elsinore by metlslime
game mode is ffa/default
Sauerbraten Win32 Exception: 0xc0000005 [0x4]

reply to this message

#145: Re: Why is Cube2 Engine called "Sauerbraten"

by Greywhind on 06/21/2006 17:26, refers to #143

Yes, as I understand it the game and the engine are called "Sauerbraten."

"Cube2" just makes it easier for those who don't already know to relate Sauer to Cube.

reply to this message

#146: Re: OSX resolver bug

by guest57 on 06/21/2006 19:48, refers to #136

Sorry still broken...
Main Thread: resolverclear() SDL_SemTryWait
Resolver Thread: terminating

Because SDL threads on OSX use pthread_cancel, you will either have to let the thread wait until the lookup completes, or lose most of the resolver and rewrite it with native pthreads (pthread_kill and signal handler) - which seems a bit silly just to suport one platform... sigh...

Meanwhile I'll have a go at tweaking (in a crossplatform way) and will email you if I have any luck.

- aka rpointon

reply to this message

#147: Re: OSX resolver bug

by guest57 on 06/21/2006 20:34, refers to #147

No offense intended - just an offer to help... See email...

reply to this message

#148: Re: OSX resolver bug

by absinth on 06/21/2006 21:55, refers to #147

if you find a working solution please tell me so i can respin the release...thx

reply to this message

#149: Re: OSX resolver bug

by eihrul on 06/21/2006 23:17, refers to #147

I incorporated your changes, with a few changes here and there. Thanks. Please check CVS to make sure it works now.

reply to this message

   Board Index