fix Firefox color rendering by respecting surface texture format

[Ayush2k02]

Replace direct copy_texture_to_texture() with a shader-based blitter that handles format conversion between Rgba8Unorm (Chrome) and Bgra8Unorm (Firefox)

Screen.Recording.2026-03-18.at.5.48.13.AM.mov

Fixes #3880

[gemini-code-assist]

Summary of Changes

Hello, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request resolves a color rendering discrepancy observed in Firefox by implementing a more robust texture handling mechanism. It introduces a shader-based blitter to facilitate texture copying and on-the-fly format conversion, ensuring that the output correctly matches the surface's native format. This change enhances cross-browser compatibility and visual fidelity, particularly for users on Firefox.

Highlights

• Firefox Color Rendering Fix: Addressed an issue where Firefox displayed incorrect colors by ensuring the rendering surface respects its preferred texture format, which can differ from Chrome's.
• Shader-based Blitter Implementation: Replaced the direct copy_texture_to_texture() operation with a new shader-based blitter. This blitter is capable of handling texture format conversions, specifically between Rgba8Unorm (common in Chrome) and Bgra8Unorm (preferred by Firefox).
• Dynamic Surface Format Configuration: Modified the surface configuration to dynamically determine and use the surface's preferred texture format, improving compatibility across different browsers and WebGPU implementations.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request addresses an issue with color rendering on Firefox by correctly handling different surface texture formats. The change replaces a direct texture copy with a more flexible blitter that can perform format conversions. The approach of querying the surface capabilities for a suitable format is correct. I've identified a minor code duplication issue that could be addressed to improve maintainability. Overall, this is a good fix.

[gemini-code-assist]

This logic to determine the surface format is also present in editor/src/node_graph_executor/runtime.rs. To avoid this duplication and improve maintainability, consider storing the determined surface_format in the Surface struct and reusing it.

You could add a format field to the Surface struct:

pub struct Surface {
	pub inner: wgpu::Surface<'static>,
	pub target_texture: Mutex<Option<TargetTexture>>,
	pub blitter: TextureBlitter,
	pub format: wgpu::TextureFormat,
}

Then, you can populate it in create_surface_inner and use surface.surface.format in runtime.rs to avoid recalculating it.

[cubic-dev-ai]

1 issue found across 2 files

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="editor/src/node_graph_executor/runtime.rs">

<violation number="1" location="editor/src/node_graph_executor/runtime.rs:352">
P1: Surface format selection ignores wgpu’s preferred format ordering by forcing an sRGB match first.</violation>
</file>

_{Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.}

[TrueDoctor]

This change will degrade performance. (I recently changed it from a blitter to the direct texture copy) Can you give more details for when this workaround is needed?

[Ayush2k02]

Oh, this is only needed for Firefox based browsers , while importing an image in Firefox, the colours are inverted
@TrueDoctor

[Ayush2k02]

can you tell me which exact test would you use here to benchmark the performance here, that way I might be able to come up with a better solution

[TrueDoctor]

can you tell me which exact test would you use here to benchmark the performance here, that way I might be able to come up with a better solution

I previously used the wgpu inspector extension to debug this. The issue with using a blitter is, that it creates bind groups which we need to wait for javascript to garbage collect (and it is also just less efficient). So it would be could, if we could avoid it.
So this is only an issue on Firefox + MacOS? Because for me, it works fine in Firefox + Linux.
I'm basically wondering if we could check if the formats are different and only in that case use the blitter?