Fast follow-up on v9.5.10.
v9.5.10 stopped the crash on every Linux launch but my defensive bytes-form slot identifier (b'1handle(bool)') was also rejected at runtime by PySide6 ("wrong argument values"), so Tee86's #144 sleep/resume hook silently failed even though the launcher worked.
Empirical bisect on PySide6 6.10.3: the binding's declared slot: bytes | bytearray | memoryview is wrong — only the str returned by PySide6.QtCore.SLOT('handle(bool)') is accepted at runtime. Switched to that and verified end-to-end: subscribe_power: PrepareForSleep listener active now logs against the live system bus.
This is the version that actually ships the #144 fix working as designed.
See CHANGELOG.md for the full story.