This reminds me that the decor plugin really needs to be a priority in terms of getting an acceptance test framework up and running - the amount of regressions we get here is *stupidly* high because the behaviour is so tightly coupled with the rest of the geometry-changing plugins.
Should work as expected now.
This reminds me that the decor plugin really needs to be a priority in terms of getting an acceptance test framework up and running - the amount of regressions we get here is *stupidly* high because the behaviour is so tightly coupled with the rest of the geometry-changing plugins.