Will pay for cone light implementation

cj45 · July 22, 2011, 12:46pm

well, atleast you won’t need to rework 1 month and 3 weeks!

Addez · July 22, 2011, 12:59pm

Yea, but it’s gonna be hell of a problem trying to work with this now.

I will probably take a break and then start all over later in the autum.

normen · July 22, 2011, 1:19pm

Get some application for continuous backup. My laptop backs itself up to a disk connected to my TV-MacMini as soon as I am in my local WiFi and then every hour.

Empire_Phoenix · July 22, 2011, 1:20pm

Well actually one week of work can be easily redone especially since you already know what you have to do, since you already did it once. This shouldn’t need that long actually. My current project I’m now working on since over 2 years could probably redone in like 3 month.

normen · July 22, 2011, 1:27pm

Yeah, probably you could write the whole game in two days now given what you’ve learned along the way (and clean up the code a lot probably ;))

Addez · July 22, 2011, 1:30pm

Thanks for the incouragement guys

I’m actually a very stubborn guy, I’ll probebly start fixing it tomorrow again…

tehflah · July 22, 2011, 1:35pm

normen said:
Yeah, probably you could write the whole game in two days now given what you've learned along the way (and clean up the code a lot probably ;))

heh. I hear that. I think I rewrote my little project about 3 times and each time it was progressively cleaner and easier to work with. You just gotta remember those moments of working on the older copy when you went "Man, I wish I'd done this like that," and then actually DO THAT. :p

You'll be alright. Probably for the better as well.
~FlaH

Addez · July 22, 2011, 2:57pm

Yea, problem is that I was actually scripting fairly clean this time.

It’s not my first big game, I’v made a fairly big 2d game too, so I learned cleanliness from there.

Ohwell, Maybe the only way is to start all over. Reuse the good import files, delete everything else…

starcom · July 23, 2011, 12:44am

Hi!

I have the same error since the last svn checkout:

svn rev: 7904 does not work

svn rev: 7788 worked

I tested on 2 different Hardware machines (one with linux ubuntu and one with XP)

After comment out the lights, it works, but i can not see shaded objects.

*************************

SCHWERWIEGEND: Uncaught exception thrown in Thread[LWJGL Renderer Thread,6,main]

com.jme3.renderer.RendererException: Shader link failure, shader:Shader[language=GLSL100, numSources=2, numUniforms=17, shaderSources=[ShaderSource[name=Common/MatDefs/Light/Lighting.vert, defines, type=Vertex], ShaderSource[name=Common/MatDefs/Light/Lighting.frag, defines, type=Fragment]]] info:Fragment info

0(86) : error C5041: cannot locate suitable resource to bind parameter "spotVec"
***********************

Thanks in advance.

erlend_sh · July 23, 2011, 6:14am

Why not set yourself up with an online repository for backups? GoogleCode (SVN/Mercurial) or GitHub are both excellent services if you’re okay with making your code available in the public domain. If not, the recently recommended private hosting service Unfuddle (SVN/Git) or BitBucket should be what you need, as they both seem to offer private hosting for free.

nehon · July 23, 2011, 8:43am

ok according to this thread http://www.gamedev.net/topic/447733-glsl-display-and-compilation-problems/page__forum_id__12

this can be due to too much varying used. I guess the limit depends on the hardware, because i can’t get it to fail whatever i try on my computer.

@starcom what is your hardware? i mean your graphic card?

Also, could you try to add this line in your simpleInit and report back the result?

[java]

System.out.println("nb max varying "+GL11.glGetInteger(ARBVertexShader.GL_MAX_VARYING_FLOATS_ARB));

[/java]

the result is 128 for me

starcom · July 23, 2011, 2:07pm

Hi!

I tried:

[java]System.out.println("nb max varying "+org.lwjgl.openg.GL11.glGetInteger(org.lwjgl.openg.ARBVertexShader.GL_MAX_VARYING_FLOATS_ARB));[/java]

The result is:
Ubuntu: 32
XP: 32
And now i tried it on my Laptop, the result is 64, and on my Laptop this svn rev 7904 works!
Ubuntu PC: Nvidia GeForce 7025
MS XP PC: Nvidia GeForce 6600 GT (2 Cards on Mainboard)
And my Laptop uses any ATI Graphics card.
Is there any way, to get this version working on my PCs?

tralala · July 23, 2011, 3:03pm

i tried spotlight.

set direction has no effect.
the spotlight always looks towards the camera. = the character has the spotlight in his back.

i dont know if its a problem of spotlight, or of “Common/MatDefs/Light/Lighting.j3md”. PointLight has the exact same problem. Maybe its a problem of shiness in Common/MatDefs/Light/Lighting.j3md.

Anyway this means that jme will never have normal lights.

normen · July 23, 2011, 3:19pm

Rofl, dude, it works for me and for most everyone else I guess. You sure you didn’t do anything wrong? Otherwise its a parallel universe kind of thing maybe

Momoko_Fan · July 23, 2011, 6:47pm

tralala said:
i tried spotlight.
1) set direction has no effect.
2) the spotlight always looks towards the camera. = the character has the spotlight in his back.

i dont know if its a problem of spotlight, or of "Common/MatDefs/Light/Lighting.j3md". PointLight has the exact same problem. Maybe its a problem of shiness in Common/MatDefs/Light/Lighting.j3md.
Anyway this means that jme will never have normal lights.

Do the spot light tests work for you?

nehon · July 24, 2011, 12:45am

tralala said:
Anyway this means that jme will never have normal lights.

uh ...what do you call normal lights?

also @Kirill this 32 varying limitation is gonna be a pain...but we ca try to reduce the number of varyings I guess

Momoko_Fan · July 24, 2011, 3:57am

I reduced the number of varyings by 2. We can go further at the expense of performance loss (shouldn’t be too bad though)

EDIT: It seems we are at our limit at this point at exactly 32 varyings. If you enable reflection or separate texcoord lightmapping then you go above the limit.

nehon · July 24, 2011, 9:19am

Nice,

I thought of some things too, but I think it will be slower like : remove the spotVec varrying, and use the g_LightDirection uniform in the frag shader, but we would have to transform it in view space and decode the cos values stored in the w attribute on a pixel basis.

Also, we could transform g_LightDirection uniform in viewSpace on the java side, before sending it to the shader, this would avoid a matrix operation every pixel.

The gain would be 4 float varying, so we would be at 28. but still…reflection and separate tex coord make it 34…

tralala · July 24, 2011, 12:44pm

Do the spot light tests work for you?

I just tested them and they work fine, i will copy paste the code and put it in my project to see if the problem is fixed.

miki100 · July 26, 2011, 6:51am

Were these lighting shaders actually fixed? I still see this:

[java]com.jme3.renderer.RendererException: Shader link failure, shader:Shader[language=GLSL100, numSources=2, numUniforms=18, shaderSources=[ShaderSource[name=Common/MatDefs/Light/Lighting.vert, defines, type=Vertex], ShaderSource[name=Common/MatDefs/Light/Lighting.frag, defines, type=Fragment]]] info:Fragment info

0(88) : error C5041: cannot locate suitable resource to bind parameter "spotVec"[/java]
Trying to enable tri-planar texturing in TerrainTestAdvanced gives this:
[java]com.jme3.renderer.RendererException: Shader link failure, shader:Shader[language=GLSL100, numSources=2, numUniforms=26, shaderSources=[ShaderSource[name=Common/MatDefs/Terrain/TerrainLighting.vert, defines, type=Vertex], ShaderSource[name=Common/MatDefs/Terrain/TerrainLighting.frag, defines, type=Fragment]]] info:Fragment info

0(162) : error C5041: cannot locate suitable resource to bind parameter "wVertex"
0(163) : error C5041: cannot locate suitable resource to bind parameter "wNormal"[/java]

As to :
[java]System.out.println("nb max varying "+org.lwjgl.opengl.GL11.glGetInteger(ARBVertexShader.GL_MAX_VARYING_FLOATS_ARB));[/java]
My result is 32. I'm using latest nightly. My gfx is nVidia Geforce 7600GS. OS: openSUSE Tumblewed.