[13.20] Weird behaviour with cross-chunk inserting

[XKnight]

Testing build: furnaces +500% speed, input fast inserters with stone (activated in the same moment), output fast inserters.
Signals from output inserters is compared each tick and if it is not equal lamps will be switched on.
Initial configuration: 1 iron ore is thrown in each furnace, 24 stone will be smelted after iron ore.
Current behaviour #1:

Video

During stone smelting lamps were switched off.

And current behaviour #2:

Video

During stone smelting lamps were switched on.

And what is the difference between first and second experiment?

Difference

Expected behaviour: chunk existence should be invisible for players because this is an implementation details, so behaviour should be the same in both experiments.

Although this weird behaviour might be hidden if viewtopic.php?f=7&t=32444 bug is resolved, but I still will be able to reproduce this bug using inserters.

[Rseding91]

This is just how it works. Chunks are updated in specific orders and the order changes when things are built in the chunk, go inactive in the chunk, and wake up later.

[XKnight]

This is just how it works. Chunks are updated in specific orders and the order changes when things are built in the chunk, go inactive in the chunk, and wake up later.

Wow, so many weird behaviours are explained in a single sentence...
Some time ago I noticed unexpected 1-tick decompression with a long straight yellow belt exactly near the chunk edge. Unfortunately, I didn't manage to reproduce it, but now I understand it.