key in the multi-material painting gizmo.
Fixed also a similar issue in support and seam gizmos where an object was displayed with the wrong color after pressing the ESC key.
multi_material_segmentation_by_painting is now returning only the painted region. Regions with default colors that aren't painted by multi-material gizmo aren't returned.
This commit also fixed the following issues:
1) After loading a 3MF with painted triangles using the MMU painting gizmo, the painted triangles might not be displayed correctly in the MMU painting gizmo.
2) The MMU segmentation was unnecessarily executed for all layers and not just for the painted layers.
3) Object's base color wasn't changed when the assigned extruder for that object was changed while the MMU paint gizmo was opened.
4) Changing the base color of an object was only possible by removing all painted triangles.
The cracks were caused by using glPolygonOffset to resolve Z-fighting. All triangles of the object were firstly rendered entirely with a base color. And then paint triangles were drawn over these already rendered triangles.
Based on information in https://stackoverflow.com/a/46115028, structured bindings in C++17 don't name variables, and because of that, it is impossible to capture these variables in lambda functions.
The modification is based on a preallocation space in the Extra field using a phony custom block in the Local file header. This space is eventually overwritten by the ZIP64 extension if the file exceeds 4GiB.
Also were fixed differences from PKZip specification in the streaming interface.
Occasionally, some input polygons contained self-intersections that caused problems with Voronoi diagrams and consequently with the extraction of colored segments by function extract_colored_segments.
Also, occasionally input polygons contained several points very close together (distance between points is 1 or so). Such close points sometimes caused that the Voronoi diagram has self-intersecting edges around these vertices. This consequently leads to issues with the extraction of colored segments by function extract_colored_segments.
MMU segmentation no longer works directly on lslices, instead of it works on custom merged regions. So lslices in PrintObject are no longer overwritten because of MMU segmentation.
All regions are scaled by SCALED_EPSILON before merging and shrunk back by SCALED_EPSILON after merging. That fixed issues with multi-volume objects when very close regions weren't merged.
Also, small expolygons and holes are filtered out that fixed missing segmentation at the boundary of two volumes in the case of multi-volume objects.
Serialization and deserialization in TriangleSelector were extended to support up to 16 materials (3 unused states left for possible later extension). These changes also affect the encoding of data from custom supports and seams, but it is backward compatible with the previous encoding. And for custom supports and seams, it is produced exactly the same data encoding as before.
When was set use_external_mp_once to true then after first calling the avoid crossing perimeters this flag was reset which cases that on the second call of the avoid crossing perimeters the travel move didn't process as external.
When an island contains exactly one infill ExtrusionEntity, this extrusion was oriented to maximize travel distance from the end of the previous extrusion of perimeters instead of minimizing it.
When the avoid crossing perimeters was enabled, and the wipe was enabled, there were unnecessary travels in the opposite direction than the wipe travel.
The solution adds a status flag to GUI_App::init_openg and ensures that GUI_App::post_init isn't called before initialization in GUI_App::init_openg is done.
The strategy for the avoid crossing perimeters algorithm has been redesigned. But external travels (travel between objects or supports) have not been solved yet. For these travels is used a direct path between two points.
Much of the code has been reworked, which leads to significant speedup compared to the previous implementation.
Also, several potential bugs have been fixed.
Calling std::move on itself causes that the first polygon is empty, which results in disabling this feature on Linux.
This was fixed before, but I accidentally reverted it when AvoidCrossingPerimeters was moved to separate file.