Issue with Multisamples and Depth Texture Causing OpenGL Errors

Scripting Issues
1

Hi everyone,

I’ve encountered a strange issue when enabling multisampling and using a depth texture with the FilterManager in Panda3D. Below is a detailed description of the problem and a minimal reproducible example.

Problem Description

When I set Multisamples greater than 0 in FrameBufferProperties:

fbprobs = FrameBufferProperties()
fbprobs.setMultisamples(2)

and then use a depth texture:

colortex = Texture()
depthtex = Texture()
manager = FilterManager(base.win, base.cam)
quad = manager.renderSceneInto(colortex=colortex, depthtex=depthtex, fbprops=fbprobs)

I receive the following errors repeatedly:

Known pipe types:
CocoaGraphicsPipe
(all display modules loaded.)
2025-01-08 00:40:33.918 Python[61495:5469081] +[IMKClient subclass]: chose IMKClient_Modern
2025-01-08 00:40:33.918 Python[61495:5469081] +[IMKInputSession subclass]: chose IMKInputSession_Modern
:display:gsg:glgsg(error): GL error 0x502 : invalid operation
:display:gsg:glgsg(error): An OpenGL error has occurred. Set gl-check-errors #t in your PRC file to display more information.

When enabling gl-check-errors in the PRC file, I get:

:display:gsg:glgsg(error): at 2074 of panda/src/glstuff/glGraphicsBuffer_src.cxx : invalid operation
...
:display(error): Deactivating CocoaGraphicsStateGuardian.

This issue does not occur in the following scenarios:

  1. When I omit the depth texture:
quad = manager.renderSceneInto(colortex=colortex, fbprops=fbprobs)
  1. When I set Multisamples to 0:
fbprobs.setMultisamples(0)
quad = manager.renderSceneInto(colortex=colortex, depthtex=depthtex, fbprops=fbprobs)

Additional Information

I’m using Panda3D version 1.10.15. Here is my system configuration:

• Python version: 3.12.6

• Machine: arm64

• Graphics Driver Vendor: Apple

• Graphics Renderer: Apple M1

• Graphics Driver Version: 4.1 Metal - 89.3

• Supports Basic Shaders: False

The specific line in panda/src/glstuff/glGraphicsBuffer_src.cxx (line 2074 for 1.10.15) triggering the error is:

report_my_gl_errors();

Minimal Reproducible Example

Here is a small script to reproduce the issue:

from direct.filter.FilterManager import FilterManager
from direct.showbase.ShowBase import ShowBase
from panda3d.core import (
    AntialiasAttrib,
    FrameBufferProperties,
    Texture,
    loadPrcFileData,
)

# Uncomment this to enable detailed OpenGL error reporting
# loadPrcFileData("", "gl-check-errors #t")

base = ShowBase()
base.render.setAntialias(AntialiasAttrib.MAuto)

fbprobs = FrameBufferProperties()
fbprobs.setMultisamples(2)
colortex = Texture()
depthtex = Texture()
manager = FilterManager(base.win, base.cam)
quad = manager.renderSceneInto(colortex=colortex, depthtex=depthtex, fbprops=fbprobs)

model = base.loader.loadModel("box")
model.setY(2)
model.reparentTo(base.render)

base.run()

Questions

  1. Is this a known issue with multisampling and depth textures on Apple M1 devices or with Panda3D 1.10.15?

  2. Could this be a driver-related limitation or something misconfigured in the FilterManager setup?

  3. Are there any workarounds to enable both multisampling and depth textures without causing this error?

Any help or pointers would be greatly appreciated!

Thanks in advance,

Miklesz