Join GitHub today
GitHub is home to over 50 million developers working together to host and review code, manage projects, and build software together.
Sign upSend epoch timestamp with the frame #106
Conversation
| @@ -447,6 +448,7 @@ impl DeviceAPI<Surface> for OpenXrDevice { | |||
| transform, | |||
| inputs: vec![right_input_frame, left_input_frame], | |||
| events, | |||
| time_ns | |||
This comment has been minimized.
This comment has been minimized.
jdm
Dec 11, 2019
Member
Shouldn't this come from https://docs.rs/openxr/0.10.0/openxr/struct.FrameState.html#structfield.predicted_display_time instead?
This comment has been minimized.
This comment has been minimized.
Manishearth
Dec 11, 2019
Author
Member
That's the predicted display time, not the time for the actual frame data.
That said the frame data may itself be predicted. I'm not sure. There are WPT tests ensuring that Performance.now() is greater than the frame timestamp.
This comment has been minimized.
This comment has been minimized.
Manishearth
Dec 11, 2019
Author
Member
Basically, I'm not sure what to do here and need spec clarification
|
LGTM. |
| @@ -26,6 +26,9 @@ pub struct Frame { | |||
|
|
|||
| /// Events that occur with the frame. | |||
| pub events: Vec<FrameUpdateEvent>, | |||
|
|
|||
| /// Value of time::precise_time_ns() when frame was obtained | |||
| pub time_ns: u64, | |||
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
Manishearth
Dec 11, 2019
Author
Member
Instant can only be compared with other Instants, and we don't use that in servo for Performance.now()
|
@bors-servo r=asajeffrey |
|
|
Send epoch timestamp with the frame Fixes #36 r? @asajeffrey
|
|
Manishearth commentedDec 11, 2019
Fixes #36
r? @asajeffrey