In a couple of worlds I have a built a long line of item sorters and in both worlds they have sometimes broken on chunk boundaries.  The symptoms are that the filter items leak from the filtering hopper into the storage chests.

I have attached a test world where I can reliably reproduce the problem.

• Place 3 stacks of yellow wool (exact amount doesn't matter) into the input storage chest.
• Check that it starts filtering into the chest for yellow wool, note that this is the one at X=15, on a chunk border, with red carpet in front.
• Go 4 chunks east to the chunk marked with a large square of blue carpet just beyond X=80.
• Watch the redstone lamp next to the input chest, when it goes off all items have left the chest/hoppers.
• Wait a little while (I don't know exactly how long - maybe a minute or two) and the items will keep filtering.
• All this time you will notice if you stand in the right place that the hopper is unlocked (redstone torch is off - a spyglass may help).
• If you start walking back towards the storage system while looking through a spyglass and watching coordinates the redstone torch should turn on exactly when you cross the 79/80 chunk boundary.  This is when the redstone for the filter wakes up again.
• Go back and look in the yellow wool chests, it is likely that the filter sticks have entered the chest (they may still be in the top chest).

What appears to be happening - and I don't know the exact mechanics involved - is that hoppers on the edge of a chunk are running even when the chunk is not fully loaded.  The redstone does not run, and therefore the hopper stays unlocked longer than it should.

It seems to me like the correct behaviour here would be that a hopper can push items into a hopper in an unloaded chunk (if that is needed for other reasons), but I feel like they shouldn't move between hoppers in an unloaded chunk where redstone doesn't run.

I have reproduced this in a local world on Windows 10 but have seen the same problem while playing in a realm on both Xbox and Windows.

If this hopper behaviour on chunk boundaries is intentional are there any possible workarounds so that my storage systems don't keep breaking?