On Fri Nov 11 14:46:02 2022 +0000, Rémi Bernon wrote:
It wasn't so much a matter of performance but rather to allow waiting on samples concurrently. `GetNextSample` will call the allocation callbacks, which may block. I'm not completely sure anymore in which case it was needed, maybe it was some pool size issue. I'll try removing it for now.
Yeah, if we need to do this to prevent a deadlock, that sounds to me like something we need to solve on the sync reader side.