[0.15.37] Splitter Throughput Strangeness

[MrGrim]

I've spent the last hour trying to see if this has been reported, but I can't find anything. I'm seeing a weird issue where one belt backs up on the input of a standard 4x4 balancer despite no output belts being blocked.

You can observe the behavior here:
https://imgur.com/LewXrJy

The input belts are being controlled by a circuit system that restricts the throughput of all belts to the slowest running belt. A screenshot of the wider area is here:
https://imgur.com/Dt1G9ti

I can see no reason why that one input belt gets backed up when there's no backup on the output side. You can find a save file with this in action here:
https://www.dropbox.com/s/cylfv11qlr56y ... e.zip?dl=0

I'm at a loss to explain this. I can only assume this behavior is a bug. I'm guessing the unique way that this balancer is being fed exposes it in a way that isn't normally seen.

Thanks

[Klonan]

Well, this behaviour might just be a quirk or the update order, or some internal logic

Belt logic is massively changed in 0.16, and this might not even still happen

So I will move this to 'Not a bug', but if you find the same thing is happening after we release 0.16, please make a new bug report