pelado

There seems to be a problem with material with jitter gl objects. In the patch below, when auto material is toggled the material settings act correctly for the nurbs and gridshape objects. However, if you turn on automatic for the sketch object and then toggle auto_material again, the color settings override those of material, only shininess, mat_emission and specular continue to function. I can get back the correct settings by forcing render to rebuild the scene (I guess that's what happens?) for example, by toggling the fsaa attribute.

Mac OS X 10.4.8, powerbook G4 1.5GHz, Jitter 1.6.3b2, max 4.6.2

There seems to be a problem with material with jitter gl objects.  In the patch below, when auto material is toggled the material settings act correctly for the nurbs and gridshape objects.  However, if you turn on automatic for the sketch object and then toggle auto_material again, the color settings override those of material, only shininess,  mat_emission and specular continue to function.  I can get back the correct settings by forcing render to rebuild the scene (I guess that's what happens?) for example, by toggling the fsaa attribute.

Mac OS X 10.4.8, powerbook G4 1.5GHz, Jitter 1.6.3b2, max 4.6.2 

bug-report-material-issues-with-ji-gl-objects

On Feb 16, 2007, at 12:16 PM, pelado wrote:

> There seems to be a problem with material with jitter gl objects. 

> In the patch below, when auto material is toggled the material 

> settings act correctly for the nurbs and gridshape objects. 

> However, if you turn on automatic for the sketch object and then 

> toggle auto_material again, the color settings override those of 

> material, only shininess, mat_emission and specular continue to 

> function. I can get back the correct settings by forcing render to 

> rebuild the scene (I guess that's what happens?) for example, by 

jit.gl.model uses displaylists internally which expoit these 

properties at the time of displaylist build. Not all of the ob3d 

attributes are set up to force a rebuild of the displaylists for 

jit.gl.model. You should be able to force a rebuild by sending the 

> There seems to be a problem with material with jitter gl objects.   
> In the patch below, when auto material is toggled the material  
> settings act correctly for the nurbs and gridshape objects.   
> However, if you turn on automatic for the sketch object and then  
> toggle auto_material again, the color settings override those of  
> material, only shininess,  mat_emission and specular continue to  
> function.  I can get back the correct settings by forcing render to  
> rebuild the scene (I guess that's what happens?) for example, by  
> toggling the fsaa attribute.

jit.gl.model uses displaylists internally which expoit these  
properties at the time of displaylist build. Not all of the ob3d  
attributes are set up to force a rebuild of the displaylists for  
jit.gl.model. You should be able to force a rebuild by sending the  
message "rebuild_geometry".

Sorry Joshua but I don't understand your reply. The issue I highlighted was not with jit.gl.model but with gridshape and nurbs. I suspect it happens with other objects too but haven't tested them properly. The issue seems to be that if I set up a shape in gridshape with auto_material turned off and material values explicitly specified, then these are overidden should I enable a sketch object in the patch. I don't think that this is correct behaviour.

Sorry Joshua but I don't understand your reply.  The issue I highlighted was not with jit.gl.model but with gridshape and nurbs. I suspect it happens with other objects too but haven't tested them properly.  The issue seems to be that if I set up a shape in gridshape with auto_material turned off and material values explicitly specified, then these are overidden should I enable a sketch object in the patch.  I don't think that this is correct behaviour. 

On Feb 16, 2007, at 3:13 PM, pelado wrote:

> Sorry Joshua but I don't understand your reply. The issue I 

> highlighted was not with jit.gl.model but with gridshape and nurbs. 

> I suspect it happens with other objects too but haven't tested them 

> properly. The issue seems to be that if I set up a shape in 

> gridshape with auto_material turned off and material values 

> explicitly specified, then these are overidden should I enable a 

> sketch object in the patch. I don't think that this is correct 

Sorry, only quickly glanced at the message and totally misread it. 

The issue is that there is some OpenGL lighting state which is 

"leaking" into the other objects. We can attempt to fix this for the 

actual 1.6.3 release, but for the meantime you can protect against 

such state bleed through by typing "; jitter glattribpush 1" in a 

What this does when turned on is for each object, Jitter pushes and 

pops the OpenGL state stack to prevent this sort of state "leakage". 

However if you have lots of instances of jit.gl.* objects you may 

discover it impedes performance (which is why this was changed to 

> Sorry Joshua but I don't understand your reply.  The issue I  
> highlighted was not with jit.gl.model but with gridshape and nurbs.  
> I suspect it happens with other objects too but haven't tested them  
> properly.  The issue seems to be that if I set up a shape in  
> gridshape with auto_material turned off and material values  
> explicitly specified, then these are overidden should I enable a  
> sketch object in the patch.  I don't think that this is correct  
> behaviour.

Sorry, only quickly glanced at the message and totally misread it.  
The issue is that there is some OpenGL lighting state which is  
"leaking" into the other objects. We can attempt to fix this for the  
actual 1.6.3 release, but for the meantime you can protect against  
such state bleed through by typing "; jitter glattribpush 1" in a  
message box and clicking it.

What this does when turned on is for each object, Jitter pushes and  
pops the OpenGL state stack to prevent this sort of state "leakage".  
However if you have lots of instances of jit.gl.* objects you may  
discover it impedes performance (which is why this was changed to  
default off in Jitter 1.5 and later).

OK the message to Jitter "glattribpush 1" works in the example patch that I posted. However, in a larger patch I now get these error messages repeatedly in the max window:

ob3d_draw_begin push attr: GL Error: Stack overflow

ob3d_draw_end popattrib: GL Error: Stack underflow

It's not a big problem for me - just took a while to work out what was going wrong, but thought you should know that the problem existed.

OK the message to Jitter "glattribpush 1" works in the example patch that I posted.  However, in a larger patch I now get these error messages repeatedly in the max window:

ob3d_draw_begin push attr: GL Error: Stack overflow
ob3d_draw_end popattrib: GL Error: Stack underflow

bug report: material issues with ji.gl objects