pkarashchenko commented on pull request #5424: URL: https://github.com/apache/incubator-nuttx/pull/5424#issuecomment-1030901325
> I'm not pushing back; I am thinking about what @patacongo wrote (also at [here](https://cwiki.apache.org/confluence/display/NUTTX/Short+Time+Delays)) about the API contract being to wait at least the given time and not knowing whether the (in that case delay, but in other cases non-delay usage) is starting in phase with the tick timer or not. I also see what you're saying about the timing being off by (in this case) 10%, which also seems wrong. I'm not sure what to suggest here. All I know if what I've always done: since software execution is sequential in nature and (with compiled code + multiple threads + CPU caches + other factors) it is not possible to predict what its timing will be, so my go-to solution is a hardware timer when I care about timing and software timers when all I want is to ensure that some time passes without caring too much about accuracy. I appreciate that this is not a good solution for everyone. I really believe that this change does not contradict with anything that is written in https://cwiki.apache.org/confluence/display/NUTTX/Short+Time+Delays -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
