ENH: Experimental option to fully disable solid infill against walls when ensure vertical thickness is turned off - alternative implementation

[igiannakas]

This PR achieves the same outcome as https://github.com/SoftFever/OrcaSlicer/pull/3235 only with a different implementation approach.

This one eliminates slightly less vertical infill however it is safer as it retains more of the top slanted surface infill to address issues like the below: https://github.com/SoftFever/OrcaSlicer/pull/3235#issuecomment-1868958227

@liftbag @ShryuKt @HakunMatat4 please test when you can.

Disabled:

First implementation: https://github.com/SoftFever/OrcaSlicer/pull/3235

This PR:

Heavily sloped model: Disabled

First implementation:

This PR Solid infill is generated to support the top surface

[ShryuKt]

Just based on the preview version I think this one is better. For information, I use this model for my tests, with a lot of round and sloppy surfaces. Yoshi #3285 👍 #3235 same layer, a lot more unneeded solid infill.

Thank you for taking the time to address this old issue.

[liftbag]

This PR is exactly what I expected. It works very well, it seems.

This is without further reduce solid infill on walls. Seven top layers set. Nine top layers are generated (including the supporting one).

This is with further reduce solid infill on walls enabled. Seven top layers set. Eight top layers are generated (including the supporting one).

And in any case the algorithm seems to work correctly where the inclinations are less accentuated.

I'd say the problem seems solved. Thank you so much.

[HakunMatat4]

@igiannakas , hey, I haven't had time to focus on this test since after the first run I shared with you on discord. I'll have a play tomorrow but it looks like other legends got you covered.

Thank you so much for looking into this 🙏

[liftbag]

And again, the algorithm also takes into account the top shell thickness. So if the object is printed with variable layer height, if the top shell thickness parameter is set, it is correctly applied when the number of top layers does not guarantee the desired thickness.

[igiannakas]

Excellent. so sounds like this PR is the way to go! I've marked the other one as "to-delete" but retaining it in case its ever needed.

[Eldenroot]

It seems that this is the better approach :)

Good job, great feedback

[igiannakas]

@SoftFever again I'm not too happy with this PR (my yardstick is whether I'd be happy to have it turned on all the time).

While it works great in retaining thickness with slanted surfaces, It creates issues with flat top surfaces which are way too common to ignore as an issue:

Layer 39

Layer 40:

Basically a one layer top surface...

That is a no-go for me for this PR. I'll revert to the original version here: https://github.com/SoftFever/OrcaSlicer/pull/3235

And look at tweaking the parameters a bit to tune out the issue with heavily sloped surfaces