Non-real actualBounds has been set. Object Shape(None)#12794, actualBounds: Rect(NaN,NaN,45,30)


I have this weird problem where I’m getting this error for a link between two nodes that are slightly overlapping:

Uncaught Error: Non-real actualBounds has been set. Object Shape(None)#12794, actualBounds: Rect(NaN,NaN,45,30)
at s (gojs)
at it.l.arrange (gojs)
at dn.arrange (gojs)
at (gojs)
at c.y.arrange (gojs)
at u.n.dw (gojs)
at wf (gojs)
at sp (gojs)
at u.n.hd (gojs)
at gojs

I only tried as high as goJS 2.0.18… didn’t test with goJS 2.1.

what I found so far is that the values of the waypoints sent to the link do matter…

BAD -> that’s the initial set of waypoints that is causing the crash
GOOD -> similar waypoints that work fine
FIXED -> whatever goJS sets for the BAD link if I sent an empty array of points as the initial bindings

as far as I can tell the link template is pretty standard… just a line and an arrow…

any ideas where to look next?.. or how to fix this?

What is going on when the error occurs?
At that time are there any nodes that do not have real actualBounds? Or any ports that do not have a real actualBounds?
Is there any way for us to reproduce the problem?

this is on diagram load… the diagram doesn’t load at all.

if I just don’t set links all the nodes load up fine… if I only remove that one bad link… everything loads fine…

no idea how to reproduce it… :|
I obviously can’t send you my entire thing…

seems that also if I move one of the nodes 1 pixel to the left or to the right it loads up fine… :|
so it’s just those exact values that make this happen.

I will try to make a separate project to simulate this.

What is the link data object for that one bad link?

OK, but we’re going to need more information to be able to figure out what’s wrong.

well… I failed to isolate the problem in a separate project… but… good news!.. seems to be working fine in 2.1.0… :)


^^ that’s the issue…

still have it in 2.1.0 unfortunately… :(

That fiddle doesn’t do or show anything. How can I reproduce the problem?

the error will be in the browser’s console once you run it…

Ah, sorry about that – I should have looked there too. OK, thanks – we can reproduce the problem, so I expect we should be able to track down the bug.

great… let me know how that goes :)

This has been fixed and will be out with the next releases (both in the 2.0.x branch and the 2.1.x branch)

Thanks for reporting.