Repository navigation
Region screenshare creates a managed session per frame at fractional scale #16953
Replies: 2 comments
|
This discussion has been automatically closed. |
0 replies
|
|
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Summary
On Hyprland 0.56.2, region screenshare at a fractional monitor scale can create and retain a new managed
CScreenshareSessionfor every captured frame. The managed-session cache compares the incoming logicalcaptureBoxwithm_captureBoxstored in physical pixels afterCScreenshareSession::init()scales and rounds it. The boxes therefore do not match at scale 1.25.Reproduction and evidence
efb50993780079460b0cbed1363e2166a2de1d9f), fractional scale 1.25, active region stream.CScreencopyClient::captureOutput()/Screenshare::CScreenshareManager::getManagedSession().CScreenshareSessionregion constructors per second; one sample observed 2,877 constructors in roughly 40 seconds, with no session destructors during the active stream.monitor->m_scaleandround()for cache comparison only, while passing the original logical box to session construction. In a centered 3440x1440 stream at scale 1.25, this produced one region session across 2,931 observed frames.Expected
Repeated frames for the same region and client reuse one managed session at fractional scale.
Question
Would a focused fix in
CScreenshareManager::getManagedSession()be appropriate upstream? The patch compares the cached physical box with a normalized incoming box only forSHARE_REGION; monitor/window/workspace matching is unchanged. I have a minimal patch available: CedrickMakhlouf@60edb8eThe equivalent change is not present in the inspected v0.57.2
ScreenshareManager.cpp.All reactions