I've posted this before, but the post must have been lost in the Great Coding Help Culling of 2015.
Blockland stores time in a 32-bit signed integer. The maximum positive value this data type can store is 2,147,483,647. Now, last time I made a very broad, sweeping statement that all of the ways you could get the simtime used this value at their core. I don't actually know if this is true, but it would amaze me if it were not. For one thing, it would mean that $Sim::Time would rapidly build up inaccuracies and deviate wildly from the correct value.
Assuming, however, that it is true, and every possible way you can get the simtime uses the same 32-bit integer at their core, then all of them will eventually overflow at the same time as the game passes 2147483647 milliseconds of uptime. With some basic arithmetic, we know that this occurs after:
24 days, 20 hours, 31 minutes, 23 seconds, and 647 milliseconds.
After this period, any scripts or code that use getSimTime() or $Sim::Time and aren't designed to be Ragnarok-proofed against simtime suddenly jumping to -2147483648 milliseconds will probably experience glitches. These can range from momentary bizarreness at the discontinuity, or extremely long-lived glitches that last until the simtime reaches a value of 0 milliseconds again, after 24 days, 20 hours, 31 minutes, 23 seconds, and 648 milliseconds. In theory, at this point, all of the problems would stop occurring for another 2147483647 milliseconds.
In theory, anyway. I don't think anyone's ever hosted a server for 50 continuous days without restart. If anyone wants to try, go for it; I'd be interested to see if the problems actually did stop or if they continued indefinitely.
