Christoph Daniel Schulze

Sure, we could replace the label-node spacing by margins, which would be configurable for all four sides independently. You don't see a bug here, do you? I'd consider this expected...

Ah, I see why you find it inconsistent. > If we were introduce `nodeLabels.margins`, `spacing.nodeLabel` would be obsolete, wouldn't it? Absolutely. I guess I'd simply remove it. The spacing options...

Another example is the node and label size calculation. We factored that code out of ELK Layered to be generally usable. Regardless of which layout algorithm calls the code, it...

We thought about this problem a bit. Here are the "results"... There are three ways to invoke automatic layout: the `DiagramLayoutEngine`, the `RecursiveGraphLayoutEngine`, and by invoking layout providers directly. The...

One of the most important aspects of this is code independent of particular layout algorithms, such as the node label placement and size calculation code, which can be called by...

> Eliminating either of the horizontal segments would cause segments of the edges to have a both-way flow of direction. That wouldn't have to be a problem; it's a hyperedge,...

We currently do not support such edges in ELK Layered. I'll classify this as an enhancement, but I have a hunch that implementing this would not be that easy. So...

@uruuru Usually, I would agree, but given that @BjBe82 may choose to ignore the issue for the time being, I wouldn't want to force him to exclude the edge from...

> I need to check with the enduser how he wants to deal with it. :) We'll be here waiting for the reply. 😉 > I was aiming in the...

Yeah, why not. That line should of course only be visible (or at least only enabled) if a diagram element is selected that is a parent.