Open theosib opened 3 weeks ago
BTW, I can do up to 25. If I increase it to 26, it crashes. This attachment has it at 25. So all you have to do to get the crash is increase the count to 26. knotplate2.FCStd.zip
Roy, assigned to you so you see this. Feel free to unassigned if in error
There are a number of issues.
The most important one is that setting "Fuse" to true for a Link array actually does not make sense, you can't fuse such an array. But the task panel does allow checking both the Link and Fuse checkboxes. What is worse is that apparently the code does not ignore the "Fuse" setting for Link arrays. Both issues need to be addressed and would solve (avoid) the reported crash.
If a fused array is the target then the arrayed shape should be simplified. Currently the top and bottom faces are not planar. which results in unnecessarily long calculation times and unwanted faces. You must then recreate the array and uncheck the Link checkbox in the task panel. Of course the resultant array must have overlapping objects for the "Fuse" property to actually have an effect.
Looking at the model I don't think a fused array is intended, so for now the advice is to set "Fuse" to false for this array.
I think this highlights one of the controversies over making fuse the default. I didn't set it to true. It just came that way, and I didn't even notice it.
I suspect the faces are not planar as a side-effect of the way the Curves workbench works. This was the only way I could find to warp a sketch around the edge of a disc, but I can't set a cone's height to zero (possibly another bug).
As for the Link setting, I'm not sure which option that refers to since for me "Link Transform" is already set to false.
Finally, thank you for the analysis since unchecking "fuse" does mitigate the crash.
Incidentally, this inability to fuse does cause problems down the line. Notice how it's impossible to fuse this array with anything else. That's okay since what I really want is probably a compound so that I can color the pattern differently in the slicer if I want. But this might be a bug to look into.
This, BTW, is another of numerous cases where exporting as STEP produces a wrong result, while exporting as STL works fine.
I didn't set it to true. It just came that way, and I didn't even notice it.
That is not correct. You must have accidentally checked the Fuse checkbox at some point.
Incidentally, this inability to fuse does cause problems down the line.
There is a mix up here. The Fuse property of Draft arrays only pertains to the fusion of objects within the array. The problem here seems to be related to this particular model. See the attached file (remove the zip extension): knotplate3-roy.FCStd.zip.
This, BTW, is another of numerous cases where exporting as STEP produces a wrong result.
I have not checked that. Please create a separate issue, thanks.
I did not explicitly enable fuse when creating the array. However, if there's a global setting, I think I may have changed it from the default based on advice to work around a problem with an earlier version of FreeCAD.
It looks like you intersected the extrusion with a cube in order to flatten it. That's a fantastic solution. And it solves some problems I encountered in the slicer. Thank you!
I can't find a global setting for this. But by default, whenever I create a Draft Array, Fuse and Link Array are both set by default.
But by default
There are default settings that are used if these options have never been set. Fuse is not checked by default, you must have done this yourself. You can verify this by renaming your CFG file and restarting FreeCAD.
These 'modes' are not exposed in the Preferences. You can use Std_DlgParameter to inspect or change them. The names of the parameters are Draft_array_Link
and Draft_array_fuse
.
Is there an existing issue for this?
Problem description
Steps:
This results in a crash. The crash report from MacOS is attached.
knotplate.FCStd.zip knot crash.txt
Full version info
Subproject(s) affected?
None
Anything else?
No response
Code of Conduct