stoxx
commented
3 weeks ago I'm thinking it needs to know about each and every line of bridge infill, the directionality and range of freedom in placement of support.
That might have been too much to ask. Perhaps it is enough for Lightning to know that it needs to support all around any intrusion that is too far away from outer walls. Or perhaps just do what it now does for ceilings, but with a wide buffer (long enough for the bridge to snap there).
Description of the bug
When there is an intrusion into object from the top, infill is generally required in order to support the initial bridge that is the first layer of solid infill. This bridge layer knows to extend to next infill for end support, but sometimes one is not available and the bridge gives up.
See screenshot from attached project.
I didn't look into more details but the issue https://github.com/prusa3d/PrusaSlicer/issues/10222 seems plausible that the related change/fix might be the reason bridge is not extended infinitely far to reach something to support it.
My opinion is that correct fix for issues with lightning infill not being good bridge support in some cases, is not to weaken the bridger, but improve the infill. In this case, lighning should be aware of the specific areas that it needs to cover fully - the ends of the bridge or something close enough outside it in the correct direction. I'm thinking it needs to know about each and every line of bridge infill, the directionality and range of freedom in placement of support. That way every piece of bridge would get both ends supported near enough, no matter how sparse the lightning is hoping to be.
Project file & How to reproduce
bugreport.zip
Checklist of files included above
Version of PrusaSlicer
2.7.4-win64
Operating system
Windows 11 23H3 22631.3447
Printer model
Ender 3 S1 and other profiles