@User: It has something to do with the way that I am reading boundary layers; there are several different ways that SkyWater defines boundaries, and it may be that there is more than one of those in the cell. Magic should probably stop issuing a message as long as any boundary it reads is the same as the one it already knows about.

But this isn't a cell boundary but a shape boundary. For example, the bitlines on met2.

It cannot (?) do that on metal2. It comes from the "boundary" input operator, which is trying to find a layer in the GDS that corresponds to a cell boundary. The layers used to make boundaries are 235:4, 81:4, and 81:3. If there are two of those layers in the cell, you get the message.

This is a boundary record in the GDS file format, not a cell boundary.

That should not trigger that error message. Do you have a smallish GDS that produces that error message?

Let me stare at this more. Maybe it's something else.

No, only the three above; ideally I should have only one type for a boundary, but SkyWater uses those several different types interchangeably (and apparently in combination as well). I think the easiest thing to do is to stop issuing a message unless the boundary area is different. Or else just ignore it.

That fixes that!