This shouldn't happen on a well UV-mapped, manifold mesh -- well, unless someone manages to click exactly on one of the triangle corners. Better to get this fixed now then to run into floating point shenanigans later.
part of CURA-12543
Refactoring that part to up top caused the problem I think -- getSelectedObject(0) over getAllSelectedObjects()[0] is clearly the better call in this case anyway.
part of CURA-12543
The most important thing to make it work is actually notifying the scene that something has changed -- the rest are just refactorings and (hopefully) optimizations.
part of CURA-12543
Also typing. The way it now works is way too slow though, and it doesn't add 'inbetween' the moude-move-positions yet. Also several other things of course.
part of CURA-12543
Should be able to paint pixels now if the tools is active, and the model loaded is with UV-coords (that rules out our current impl. of 3MF at the moment -- use OBJ instead), and you position the model outside of the build-plate so the paint-shadr that is temporarily replacing the 'disabled' one is showing.
Will need a lot of extra features and optimizations still!
part of CURA-12543
Currently replacing the 'disabled' batch until we can get it to switch out on command (when we have the painting stage/tool/... pluging up and running.
part of CURA-12543
CURA-12101
The printer parses the machine_start_gcode to allow selecting the filaments mapping at start time, without it the user has to set the filaments in fixed order. This is probably a security to ensure the proper filament is loaded at start.
CURA-12346
For some dark reason, importing the ExtruderStack class in the ThreeMFWriter somehow corrupts the loading of the extruders stacks. As it is imported only for typing, do the import only for type-checking.
- Fix: Need to fill the package's API version into the (template) URL, instead of using latest.
- Fix: Loop was closing over the package ID, which caused the recieving function to always only get the last handled package ID instead of the one it needed to handle.
part of CURA-12156
This change adds a setting for "minimum Z height" for the "nail down the string" moves.
Update PurgeLinesAndUnload.py
Bug fix. The prime blob code didn't account for the retraction at the end. This adds an 'unretract' to compensate.
It's now in a state where it can actually upload ... something (that should work). Not tested the restore yet.
I did run into trouble with the max concurrent requests, which I had to up to [more than 4, now on 8] to get it to work -- I'm not sure if I'm just working around a bug here, or if that's expected behaviour.
part of CURA-12156
