cast a float result back to uint32_t #1241
Conversation
|
CI gfxreconstruct build queued with queue ID 33175. |
|
CI gfxreconstruct build # 3150 running. |
|
CI gfxreconstruct build # 3150 failed. |
bradgrantham-lunarg
force-pushed
the
brad-cast-scale-result
branch
from
9a0e188
to
436853c
Compare
|
CI gfxreconstruct build queued with queue ID 33188. |
|
CI gfxreconstruct build # 3151 running. |
|
CI gfxreconstruct build # 3151 passed. |
|
CI gfxreconstruct build queued with queue ID 33276. |
|
CI gfxreconstruct build # 3153 running. |
|
CI gfxreconstruct build # 3153 passed. |
| @@ -2147,12 +2147,12 @@ void VulkanReplayConsumerBase::WriteScreenshots(const Decoded_VkPresentInfoKHR* | |||
| // If both copy_scale and copy_width are provided, use copy_scale. | |||
| const uint32_t screenshot_width = | |||
| options_.screenshot_scale | |||
| ? (options_.screenshot_scale * swapchain_info->width) | |||
| ? static_cast<uint32_t>(options_.screenshot_scale * swapchain_info->width) | |||
There was a problem hiding this comment.
constexpr uint32_t truncate_to_uint32(const float f)
{
GFXRECON_ASSERT_DEBUG(f >= 0.0f);
return static_cast<uint32_t>(f);
}somewhere in a util header to name what we are doing here and add the sanity precondition check on the param?
There was a problem hiding this comment.
I read static_cast<uint32> on float as truncating so for my taste there's not much reason to have a function that has that name. (To me it would be like a function called "copy_string" that calls "strcpy".)
This PR is just to fix warnings in /W3 build in Visual C++, but -
It does seem like a good idea to check the range of the scale but that should probably be validated in input. Want to add that in a PR?
* cast a float result back to uint32_t * fix some strict warnings
No description provided.