|
I haven't tracked down the hang yet, but I at the moment, I suspect it's because we're not calling close on the pdb service. I need to go double-check just how System/exit interacts with daemon and non-daemon threads.
As a "cleaner" alternative, I've also tried using tk's request-shutdown, and that seems promising, but it causes new exceptions when close fails on unexpected nils. We might want to use tk's shutdown-on-error. It'll produce an unwanted backtrace, but might be the only way to shutdown "normally" for now with a non-zero exit status.
|