Hard to say without seeing the code, but it looks like there's some condition that makes it possible for layoutColorbar to complete without newpos ever being defined -- that's the initial warning, which then leads to the error on line 125 of f_colorbar. Presumably it's a subtle bug in the logic which hasn't manifested until now.
With the Editor, you can highlight a variable and see where it shows up in that file. Maybe you can use that to trace the creation of newpos and see if there's anything that could prevent it. You could also simply try putting a breakpoint in layoutColorbar and stepping through (when running an example you know is failing).
Best Answer