inlet timestamp post processing could lead to timeout_error if using custom TimeUpdateInterval

[Chadwick Boulay]

Prompted by a conversation with a user in Slack (whose issue turned out to be firewall blocking UDP), I discovered that clock offset post-processing uses a default timeout of 5.0 seconds. At least I think that's what's going on here:

liblsl/src/stream_inlet_impl.h

Line 44 in 5198054

 postprocessor_([this]() { return time_receiver_.time_correction(5); }, 

If a user uses lsl_api.cfg and sets TimeUpdateInterval to >= 5.0 then there's a good chance that on the first pull the inlet will timeout while waiting for the clock offset.

liblsl/src/time_receiver.cpp

Lines 54 to 63 in f048715

 if (!timeoffset_available()) { 
 	// start thread if not yet running 
 	if (!time_thread_.joinable()) time_thread_ = std::thread(&time_receiver::time_thread, this); 
 	// wait until the timeoffset becomes available (or we time out) 
 	if (timeout >= FOREVER) 
 		timeoffset_upd_.wait(lock, timeoffset_available); 
 	else if (!timeoffset_upd_.wait_for( 
 				 lock, std::chrono::duration<double>(timeout), timeoffset_available)) 
 		throw timeout_error("The time_correction() operation timed out."); 
 }