You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As the doc comment explains, timestamp_period is used when the wgpu-profiler. Should the timestamp_period member therefore be put behind the wgpu-profiler flag, like so:
pubstructRendererOptions{/// The format of the texture used for surfaces with this renderer/device/// If None, the renderer cannot be used with surfacespubsurface_format:Option<TextureFormat>,/// The timestamp period from [`wgpu::Queue::get_timestamp_period`]/// Used when the wgpu-profiler feature is enabled#[cfg(feature = "wgpu-profiler")]pubtimestamp_period:f32,}
Or is it not being behind the feature flag intentional?
The text was updated successfully, but these errors were encountered:
The intention there was to not require code changes on the user side because wgpu-profiler was enabled.
For example, if a library depends on Vello , and then a crate depends on that library and Vello, then enables profiling in the top-level crate would break the library, if it managed its own renderers.
Perhaps we should instead accept a reference to the queue in the options? It's a bit messy any way around
The
RendererOptions
struct has an additionaltimestamp_period
member due to #304:https://github.com/linebender/vello/blob/6d57093cc275589843766b8c898793f8147ed323/src/lib.rs#LL81C4-L81C4
As the doc comment explains,
timestamp_period
is used when thewgpu-profiler
. Should thetimestamp_period
member therefore be put behind thewgpu-profiler
flag, like so:Or is it not being behind the feature flag intentional?
The text was updated successfully, but these errors were encountered: