'New Tree' Refactors #7
Conversation
meshfix_maximum_extrusion_area_deviation is already in millimeters. using get<coord_t> divides the value by 1000, instead it should be get<size_t> or get<int>.
Simplify: fix scaling of maximum area deviation setting
Contributes to CURA-9879
Revert before merging to main Contributes to CURA-9879
Revert before merging to main Contributes to CURA-9879
Contributes to CURA-9879
Mac is being Mac again, also the range-v3 library is used consistently throughout the rest of the code Contributes to CURA-9879
Put code in our code style -- including moving comments around, for better readability so definitely not all that is said there is mine ;-) part of CURA-9879
Put code in our code style -- including moving comments around, for better readability so definitely not all that is said there is mine ;-) -- Includes workaround for Mac compiler bug stumbled into in last commit. part of CURA-9879
Pass sby reference was erronously introduced by refactoring. part of CURA-9879
…ish. Manually cherry picked for CURA-9879 (git gave a bit of a mess when git cherry-pick was run...)
Will not do that much, but standard in the rest of the code. part of CURA-9879
Added a note as well. I have no idea why the (in context) equivalent (even when the static_cast<size_t> is replaced by a simple cast to int) statement, using iota, doesn't work, but this does. It may just expose another bug, but best not to take the risk right now. part of CURA-9879
It (moving the area-calculation to outside of the loop) _shoulnd't_ have caused issues (indicating that there's probably some deadlock possibility in the code as-is), but it _did_, so undo this part (for now). part of CURA-9879
|
This would also undo "take disallowed areas into account when printing tree support". Is this intentional or should i cherry pick this commit? |
src/TreeSupport.cpp
Outdated
| // smooth upwards | ||
| for (LayerIndex layer_idx = 0; layer_idx < LayerIndex(layer_tree_polygons.size()) - 1; layer_idx++) | ||
| // Smooth upward. | ||
| for (const auto layer_idx : ranges::views::iota(0UL, layer_tree_polygons.size() - 1UL)) |
There was a problem hiding this comment.
I do not think it is very relevant as i don't think that this code can even be reached in the relevant edge-case, but theoretically if layer_tree_polygons.size() == 0 then this would underflow, while it would not underflow in my version as LayerIndex is a signed type.
|
Why is PropertyAreas and PropertyAreasUnordered in TreeSupportElements.h and not inside TreeSupport.h? |
|
Hi @ThomasRahm . Sorry it took a bit longer to get back.
I should also really mention that our team has decided that 5.3 is already very 'full' of features, and we've got bad experiences with putting everything in at once. As such, we'd like to take some more time on the New Tree Support as well. So it'll be in 5.4. (We try to release every two to three months.) In the mean-time, the alpha is still out, so people can continue with that for a while. |
Propose to keep the code bases closer together, so it will be easier to merge (new) fixes.