Logic error in ImGui_ImplVulkan_SwapBuffers() function.

[NostraMagister]

Version/Branch of Dear ImGui:

Version 1.90.9 Docking branch

Back-ends:

imgui_impl_vulkan.cpp

Compiler, OS:

All

Full config/build information:

All

Details:

Function: ImGui_ImplVulkan_SwapBuffers()

Partial Code:

    err = vkQueuePresentKHR(v->Queue, &info);
    if (err == VK_ERROR_OUT_OF_DATE_KHR || err == VK_SUBOPTIMAL_KHR)
    {
        vd->SwapChainNeedRebuild = true;
        return;
    }
    check_vk_result(err);

    wd->FrameIndex = (wd->FrameIndex + 1) % wd->ImageCount;             // This is for the next vkWaitForFences()
    wd->SemaphoreIndex = (wd->SemaphoreIndex + 1) % wd->SemaphoreCount; // Now we can use the next set of semaphores
	

Comment:
The return value VK_SUBOPTIMAL_KHR must/should, IMO, be treated like VK_SUCCESS after the SwapChainNeedRebuild
flag has been set. The Vulkan specs qualify it as success which means the vkQueuePresentKHR() has properly
executed and the Frame and Semaphore indices must be increased. In the code above with VK_SUBOPTIMAL_KHR the indices are not increased.

Info Source: [https://registry.khronos.org/vulkan/specs/1.3-extensions/man/html/vkQueuePresentKHR.html]

Screenshots/Video:

No response

Minimal, Complete and Verifiable Example code:

[ocornut]

Hello,

Thanks for the report. I generally have a principle issue where I don't want to "blind fix" issues based on theory or documentation only, I would like to confirm that this can run correctly when it happens.

At this point I'm not even sure why we don't increase the two values even on the VK_ERROR_OUT_OF_DATE_KHR path?

Pushed e212511, this amends 085cff2 (#3881)

[NostraMagister]

The answer is partially in a new ticket with two suggestions i just posted. The frame-index shouldn't be increased at all IMO, as i explain. There are IMO quite some other issues in the vulkan backend code. I'll report them when I get there and then you can decide whether you do something with them or not.