Doug Wright

Yes, 3.x please as it's easier to merge up than backport. X axis is width, Y is length, Z is depth. Thinking about it a bit now, the dimension comparison...

@neodisco for php-laff, the calculated width of the fake container is the width of the widest item (ignoring any rotation), and the length is the length of the longest item...

Hello @giozecc What you're seeing is currently "by design" - the algorithm works by building "layers". Once each layer is built, the next layer on top is built assuming a...

Hi @neodisco. Thanks for the report. Several things to pick through here: 0 isn't supposed to be a supported value, by using it I think you've accidentally trigged a 2D...

> 1. Are there any guidelines for how to best prepare the data before submission to BoxPacker? For example, I am now assuming that all dimensions on both box and...

Hello Choosing the shape of a hypothetical box isn't something that's currently possible, the hypothetical calculation is currently just an "add-on" to the physical box calculation. I'll keep this open...

> it works fine when we split this box into 2 and trying to pack 3 and 8 Sounds like you might have found an edge case where the algorithm...

Hi @kolevich There is no perfect box-packing algorithm in existence, so like all such algorithms BoxPacker is essentially a set of heuristics which due to their very nature will sometimes...

I've had a look at this, sadly haven't found a revised heuristic yet that works yet without regressing more cases than it improves

Hello @hudaibrhm Have you see https://boxpacker.io/en/stable/weight-distribution.html? Try turning that behaviour off and you may get the results you're expecting 🤞