[webkit-reviews] review granted: [Bug 233358] [WebGPU] Implement GPU process's message receivers : [Attachment 444800] Patch

[bugzilla-daemon at webkit.org]

Dean Jackson <dino at apple.com> has granted Myles C. Maxfield
<mmaxfield at apple.com>'s request for review:
Bug 233358: [WebGPU] Implement GPU process's message receivers
https://bugs.webkit.org/show_bug.cgi?id=233358

Attachment 444800: Patch

https://bugs.webkit.org/attachment.cgi?id=444800&action=review




--- Comment #2 from Dean Jackson <dino at apple.com> ---
Comment on attachment 444800
  --> https://bugs.webkit.org/attachment.cgi?id=444800
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=444800&action=review

> Source/WebKit/GPUProcess/graphics/WebGPU/RemoteBuffer.cpp:52
> +    // FIXME: Implement this.

ASSERT_NOT_IMPLEMENTED()?

> Source/WebKit/GPUProcess/graphics/WebGPU/RemoteCommandEncoder.cpp:72
> +    auto computePassEncoder =
m_backing->beginComputePass(*convertedDescriptor);

What happens if convertedDescriptor still doesn't have a value at this point?

> Source/WebKit/GPUProcess/graphics/WebGPU/RemoteCommandEncoder.cpp:88
> +    auto convertedSource =
m_objectRegistry.convertBufferFromBacking(source);
> +    ASSERT(convertedSource);
> +    auto convertedDestination =
m_objectRegistry.convertBufferFromBacking(source);
> +    ASSERT(convertedDestination);
> +    m_backing->copyBufferToBuffer(*convertedSource, sourceOffset,
*convertedDestination, destinationOffset, size);

Should this (and the other similar code) actually test that the converted
buffers exist rather than just ASSERT?