Fix #4195: Preserve transforms when switching PBR materials

[Aqil-Ahmad]

Description

Fixes texture transforms being reset when switching from Blinn-Phong to PBR materials and between PBR. Previously, custom scale, offset, and rotation settings would be lost, making it tedious to switch between PBR materials.

Related Issues

Issue Link: #4195

---

Checklist

Please ensure the following before requesting review:

• I have provided a clear title and detailed description for this pull request.
• If useful, I have included media such as screenshots and video to show off my changes.
• The PR is linked to a relevant issue with sufficient context.
• I have tested the changes locally and verified they work as intended.
• All new and existing tests pass.
• Code follows the project's style guidelines.
• Documentation has been updated if needed.
• Any dependent changes have been merged and published in downstream modules
• I have reviewed the contributing guidelines.

---

Additional Notes

scndLife_vid.mp4

[akleshchev]

pre-commit fails due to spaces.
Build fails due to outdated branch, I suggest updating to a newer develop (and updating your local build variables then wiping build folder)

I might be missing something, but the variant from previous PR looked better, sans lack of convertTextureTransformToPBR.

[Aqil-Ahmad]

pre-commit fails due to spaces. Build fails due to outdated branch, I suggest updating to a newer develop (and updating your local build variables then wiping build folder)

I might be missing something, but the variant from previous PR looked better, sans lack of convertTextureTransformToPBR.

In the previous PR, applying PBR had an issue: texture values were only preserved in the preview. When clicking OK, the values reset to defaults (though changing them after applying PBR did preserve them). In addition, PBR-to-PBR conversions didn’t preserve values and subsequent PBRs still picked up texture values.

[akleshchev]

When clicking OK, the values reset to defaults

I suspect that's the issue with queueApply, basically doing 'queueApply(hit_obj, hit_face, NEW_asset_id, preserved_override)' works fine, but doing 'queueApply(hit_obj, hit_face, OLD_asset_id, preserved_override)' would ignore preserved_override and drop the overrides.

If that's indeed the cause, the solution is to compare current material id and if it didn't change then not send anything at 'OK', nothing changed after all.

[Aqil-Ahmad]

If that's indeed the cause, the solution is to compare current material id and if it didn't change then not send anything at 'OK', nothing changed after all.

You were right: the OK issue came from calling queueApply with the same material ID, which cleared overrides.
I tried to preserve values in the PBR→PBR case but couldn’t get it to work; I suspect a clear-then-override ordering, even though the local logic looks correct.

[akleshchev]

I tried to preserve values in the PBR→PBR case but couldn’t get it to work;

I think it did work in your closed PR with a 'drag&drop material over previous material' case.

[bennettgoble]

@Aqil-Ahmad, @akleshchev has done some investigation and determined that we are blocked from completing this feature because of a server-side issue. I'd like to propose the following:

1. We pay you out for the reward for your work so far: your efforts are appreciated and this problem is on our end
2. We close this PR and the related issue until the issue is fixed on our server

[akleshchev]

P.S. I will create a server ticket to fix it.

I was wrong and it was a viewer issue, fixed. Please update from develop.

I tested code from this PR (removed everythign from llpanelface.cpp and from llgltfmateriallist.cpp sans convertTextureTransformToPBR ) and the only issue appears to be a missed existing_override check (convertTextureTransformToPBR fires in case of a pbr-pbr transition), otherwise everything works as expected.

[akleshchev]

Why are you removing this part?

A 'send' happening here is a questianable practice, but removal doesn't seem relevant to transform preservation. Otherwise changes seem solid if a bit too many repeats of the same code.

[akleshchev]

Did some additional testing and everything seems to be in order, will squash tommorow then remove the 'material list not sending' part since that seems to be unintended result from a merge conflict.