Skip to content

Commit

Permalink
Still seeing asserts in this function, weakening remaining to debug-o…
Browse files Browse the repository at this point in the history
…nly.. but these really shouldn't trigger.
  • Loading branch information
hrydgard committed Nov 30, 2017
1 parent d884eed commit b6f69b7
Showing 1 changed file with 2 additions and 2 deletions.
4 changes: 2 additions & 2 deletions ext/native/thin3d/VulkanQueueRunner.cpp
Expand Up @@ -594,8 +594,8 @@ void VulkanQueueRunner::PerformBindFramebufferAsRenderTarget(const VKRStep &step
w = vulkan_->GetBackbufferWidth();
h = vulkan_->GetBackbufferHeight();
renderPass = GetBackbufferRenderPass();
assert(step.render.color == VKRRenderPassAction::CLEAR || step.render.color == VKRRenderPassAction::DONT_CARE);
assert(step.render.depthStencil == VKRRenderPassAction::CLEAR || step.render.depthStencil == VKRRenderPassAction::DONT_CARE);
_dbg_assert_msg_(G3D, step.render.color == VKRRenderPassAction::CLEAR || step.render.color == VKRRenderPassAction::DONT_CARE, "Backbuffer pass should only be CLEAR or DONT_CARE");
_dbg_assert_msg_(G3D, step.render.depthStencil == VKRRenderPassAction::CLEAR || step.render.depthStencil == VKRRenderPassAction::DONT_CARE, "Backbuffer pass should only be CLEAR or DONT_CARE");
Uint8x4ToFloat4(clearVal[0].color.float32, step.render.clearColor);
numClearVals = 2; // We don't bother with a depth buffer here.
clearVal[1].depthStencil.depth = 0.0f;
Expand Down

2 comments on commit b6f69b7

@unknownbrackets
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This happens, at least, during debugging. I'd really hate for it to have to clear, and on ALL backends even, every time you step. It's very annoying.

Not sure what's causing it elsewhere, but RebindFramebuffer() has a KEEP from fbo_unbind() lineage, and so does Vulkan's BlitFramebuffer (with buffered rendering off.)

-[Unknown]

@hrydgard
Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Oh yeah, hm. I should probably just kill this assert, I just wanted to make sure we don't try to KEEP anything from the backbuffer from one frame to the next (since that's invalid) but for cases like stepping, it's indeed necessary.

Please sign in to comment.