Fix tree support fake roof

[ThomasRahm]

Description

Fixes Tree Support fake roof not working and having no changes in it's pattern with increasing branch density.

Type of change

• Bug fix (non-breaking change which fixes an issue)

How Has This Been Tested?

Before (fake roof not working):

After (fake roof working again):

Project File:
TestOverhang.zip

Test Configuration:

• Operating System: Windows

Checklist:

• My code follows the style guidelines of this project as described in UltiMaker Meta
• I have read the Contribution guide
• I have commented my code, particularly in hard-to-understand areas
• I have uploaded any files required to test this change

[casperlamboo]

@ThomasRahm, thanks for the fix looks good/clean! Had one teeny tiny small remark. Any idea why the regression occurred? Why did this only become an issue since the 5.5/5.6 release?

[ThomasRahm]

Based on what I could see 7c02094 was the commit where the regression was introduced. I do not know which versions included it.

[casperlamboo]

Just tested the proposed solution, and all seems to work! Going to merge. Have one idea/suggestion though (not to add as part of this PR, but something to think about). One downside of tree support is that it introduces a lot of island, and thereby a lot of retracted (or possibly unretracted) travels which is bad for flow control. I was investigating the layer below the support roof, and I noticed that the edges almost describe the straight line edges from the layer above. Why not connect the edges like I've drawn very crudely in the image below?

if there would be $n$ horizontal edges, and $m$ vertical edges this would reduce the number of travels from (approximately) $n \cdot m$ to (approximately) $n + m$.

[rburema]

That commit was in order to deal with fractional layer heights throughout the code (in this case tree-support). It seems I didn't pay enough attention to the differences between the pieces of code I consolidated into 1 there.

(Thanks Casper, doing this ticket saves me some work. We should probably also couple this back to internal ticket CURA-11368.)