Ensure that user gets updates during long passes The patcher can potentially generate log messages faster than the main thread can log them, causing frames that are noticeably long with no updates. This ensures that yields still happen then. Verified that this does not meaningfully affect performance. Previous tests suggest that the time wasted by waiting until the next frame is relatively small. (commit: 30eeb6f)
Ensure time between each check of the log queue This prevents the queue from being locked too often, slowing down the patching thread (commit: dc1a167)