clearCoatMap : reason: sampler named "clearCoatMap" not found

[ccworld1000]

Describe the bug
clearCoatMap : reason: sampler named "clearCoatMap" not found

To Reproduce
Steps to reproduce the behavior:

1. Import a new glb model into the project hello-gltf in Maos 12 to replace the original model

Expected behavior
A clear and concise description of what you expected to happen.
crash : reason: sampler named "clearCoatMap" not found

Screenshots
If applicable, add screenshots to help explain your problem.

Logs
If applicable, copy logs from your console here. Please do not
use screenshots of logs, copy them as text.
2022-09-28 11:18:30.982812+0800 hello-gltf[4591:105410] Metal GPU Frame Capture Enabled
2022-09-28 11:18:30.983199+0800 hello-gltf[4591:105410] Metal API Validation Enabled
FEngine (64 bits) created at 0x7fe5b180c400 (threading is enabled)
FEngine resolved backend: Metal
Selected physical device 'Apple iOS simulator GPU'
FEngine feature level: 1
Panic
in getSamplerInfo:97
reason: sampler named "clearCoatMap" not found
CoreSimulator 842.8 - Device: iPhone 13 (D1A3AAD9-E8DE-4E05-9889-8D12FB891708) - Runtime: iOS 16.0 (20A5303f) - DeviceType: iPhone 13

Desktop (please complete the following information):

• OS: [iOS]
• GPU: [AMD]
• Backend: [Metal]

Smartphone (please complete the following information):

• Device: [MacOS12/ iPhone]
• OS: [iOS16]
• Xcode: [14]
• filament: [filament-1.27.1]

Additional context
Add any other context about the problem here.
I found that this problem is very similar, but the original library file seems to be missing. I don't know how to deal with it in the new version
#4143

[prideout]

@bejado this bug is probably my fault but I don't have an iOS device. I wonder what's going on. Do you recommend using the iOS simulator to repro?

@ccworld1000 does this happen if you import any model, or does it only happen with models that have clearcoat?

[bejado]

Let me try to reproduce on the simulator

[bejado]

I'm unable to reproduce; I can successfully load a glb file with clear coat in the iOS simulator. I can try on device, but I doubt that would change anything.

[bejado]

My only idea is that you might be trying to apply clear coat to a model that has some other special feature (e.g., sheen). I don't think we support this. @ccworld1000 are you able to share your model with us?

[ccworld1000]

The simulator can support running @prideout @bejado

[ccworld1000]

The model is also very simple, which is a cube.
CCCubeTest123.glb.txt

rename "CCCubeTest123.glb.txt" -> "CCCubeTest123.glb"

@prideout @bejado

[ccworld1000]

CC # file *
CCCubeTest123.glb: glTF binary model, version 2, length 2104 bytes
CCCubeTest123.glb.txt: glTF binary model, version 2, length 2104 bytes
CC #

It doesn't seem to matter if you don't rename

[ccworld1000]

model Screen Shot

[bejado]

I'll try on device tomorrow.

[ccworld1000]

The simulation should also run

[daemyung]

@ccworld1000 Will you try #6122?

[romainguy]

Looks like the issue is due to KHR_materials_ior, which causes us to pick the clear coat base spec. The fix proposed by @daemyung will work once my comments are taken into account.

[romainguy]

See issue #6121 instead

[ccworld1000]

I use the binary library on tag instead of compiling it myself.
It is estimated that the test can only be performed on my side after the new library is updated.
The associated model is already above.

The model is also very simple, which is a cube. CCCubeTest123.glb.txt

rename "CCCubeTest123.glb.txt" -> "CCCubeTest123.glb"

@prideout @bejado

@daemyung

[daemyung]

@ccworld1000 I have tested CCCubeTest123.glb and it's fixed by #6122.

[ccworld1000]

I just used the binary library of the new tag 1.27.2 for testing
Not sure if the problem is completely solved?
1.gltf_viewer may be fixed
2. It seems that the iOS simulator still has problems

@daemyung

[ccworld1000]

call stack text file

call_stack.txt

[ccworld1000]

* thread #1, queue = 'com.apple.main-thread', stop reason = signal SIGABRT
    frame #0: 0x000000010f94100e libsystem_kernel.dylib`__pthread_kill + 10
    frame #1: 0x000000010f68d1ff libsystem_pthread.dylib`pthread_kill + 263
    frame #2: 0x000000010f576fef libsystem_c.dylib`abort + 130
    frame #3: 0x000000010ba14a02 hello-gltf`utils::TPanic<utils::PreconditionPanic>::panic(char const*, char const*, int, char const*, ...) + 242
    frame #4: 0x000000010ba0ca51 hello-gltf`filament::SamplerInterfaceBlock::getSamplerInfo(std::__1::basic_string_view<char, std::__1::char_traits<char> >) const + 97
    frame #5: 0x000000010b9bb761

[ccworld1000]

in SamplerInterfaceBlock

[ccworld1000]

The top screenshot is still valid

[romainguy]

The fix is not that in release. You'll have to wait a week or two.