Mc 1.7.10
OpenComputers-MC1.7.10-1.8.2+94f6405-universal

I ran a program that kept repeating os.sleep(0.05) and noticed that at some times it sleeps for 0.05 seconds and at other times the sleep time fluctuates between 0.05 and 0.1 seconds. This happens systematically and predictably, and is related to the current value of computer.uptime(). For example, if you run such a program immediately after starting the computer, then fluctuations in sleep time begin at 8 seconds and end at about 30 seconds, and then start again at 128 seconds. If you assign 0.049 or 0.025 as sleep time, then the fluctuations almost disappear. If you replace os.sleep() with computer.pullSignal(), the results are the same. It seems that the fluctuations are caused by the inability to accurately represent most decimal fractions in the binary system and the resulting errors in arithmetic operations. Because of this, sometimes when a wake-up should occur, the deadline is slightly higher than computer.uptime() and the sleep is extended by 1 game tick. Event.timer() can theoretically also sometimes call a function a little later than it should. And also, if I understand the implementation of event.timer() correctly, it can drift if a number that does not have an exact binary representation is chosen as the interval, although a significant shift requires a large number of cycles and a large computer.uptime() value.

Is this really a bug? Slight deviations are inevitable for timers, right?

Could you please test the proposed fix? Build should be available here in a few minutes: https://github.com/MightyPirates/OpenComputers/actions/runs/5401692819

I tested the os.sleep() and computer.pullSignal() functions. It looks like the issue is still not resolved.

-0.001 should theoretically be enough, but I think -0.025 would be more reliable.
Although the fix most likely does not work for some other reason.