I know I’ve raised this issue before, and I know Pascal disagrees with me on this, but Distribution does not work the way it should (based on Rhino Docs or every other 2d or 3d app out there).
In any 2 or 3 dimensional space where points {x, y, z} are at {x1, yn, zm}, {x2, yn, zm}, {x3, yn, zm}, {x4, yn, zm} … {x10, yn, zm} - where n and m are constant values or not - it shouldn’t matter, then Distribute in the x-axis should only adjust the vector locations of the x coordinates between {x1, yn, zn} and {x10, yn, zn) leaving the array of y and z coordinates as they are
If we’re distributing objects and we select a set of points as the reference point for each object being distributed, (just like move, rotate, move etc) then the other points in the object retain their relative coordinates to the ref points even as the intermediate ref points are repositioned.
It would appear:
a) we don’t respect (in my example of distributing {x}) the {y, z} coordinates
b) we don’t adjust the z-axis when looking at a {z, x} or {z, y} viewports.
I see a crash reported that appears to be related to this as well. If you have steps to reproduce that crash, please let us know (I realize sometimes it’s difficult to reproduce the steps).
It’s completely non-deterministic. However I would say it looks like a memory leak as, when I first open the model everything is fine, but as I begin to manipulate control points, the frequency of crashes increases. What argues against (perhaps) the memory leak is that when I reopen the model (after sending you guys the umpteenth crash report) and do exactly the same operation, it will likely crash immediately. If I go an do some other work on the model them come back to the operation that caused the last crash it may or may not crash again…
That is a good guess. Normally, with a memory leak, we’d likely see some warnings in the kernel triage in the crash reports. The weird thing is we don’t here. Another oddity is that I see a whole bunch of different call stacks in different crashes, but - as you observe - it’s probably just crashing a different “points” during the Distribute operation, which seems to match your description.
Regardless, I can’t seem to reproduce this here and the information in the backtraces is not great.
