I'm getting an odd error when I edit my custom flex axle in a certain way. Attached is a test case file.
Step to reproduce:
- Open test.ldr
- Enter nested mode
- Click on either control point
- Hit Ctrl-R to enter rotation
- Enter an arbitrary number into the rotation field
- Hit accept and notice odd output
[
attachment=3589]
(2019-04-29, 20:28)Orion Pobursky Wrote: [ -> ]I'm getting an odd error when I edit my custom flex axle in a certain way.
That's because the caps are scaled, LDCad assumes all matrices are non scaled and non mirrored.
This so position and rotation can be uniformly manipulated during editing.
The solution is to move the caps into a donor and checking the caps using them inline option.
That way you can also add snapping info to the caps.
I attached the donor and template I just made for 1.6c.
Place ldcFlexAxleCap.dat in %appdata%\LDCad\donors
(2019-04-29, 20:45)Roland Melkert Wrote: [ -> ]That's because the caps are scaled, LDCad assumes all matrices are non scaled and non mirrored.
This so position and rotation can be uniformly manipulated during editing.
The solution is to move the caps into a donor and checking the caps using them inline option.
That way you can also add snapping info to the caps.
I attached the donor and template I just made for 1.6c.
Place ldcFlexAxleCap.dat in %appdata%\LDCad\donors
Ok. The odd thing is this doesn't happen when rotating with the edit pin or rotating a group of parts with the primary part not the control point.
Is this due to that those flex-hoses use a stud3a-prim, instead of a cyli?
Is there today a good reason to keep the design of those flex-hoses?
They all seam to have a very unconventional design, producing a lot of triangles to render.
The cross axle in the longest hose, 32235, is made using the axlehol8 primitive 75 times.
(2019-04-29, 21:02)Magnus Forsberg Wrote: [ -> ]Is this due to that those flex-hoses use a stud3a-prim, instead of a cyli?
Is there today a good reason to keep the design of those flex-hoses?
They all seam to have a very unconventional design, producing a lot of triangles to render.
The cross axle in the longest hose, 32235, is made using the axlehol8 primitive 75 times.
If we get some sort of flex object standard in place then we can switch over to using that instead of how we do it now.
(2019-04-29, 20:57)Orion Pobursky Wrote: [ -> ]The odd thing is this doesn't happen when rotating with the edit pin or rotating a group of parts with the primary part not the control point.
I'll have to dive in the source to explain that in detail but the main difference is one applying a new absolute matrix while the pin applies a relative one.
The matrix math involved is open for many improvements as seen here
(2019-04-29, 21:02)Magnus Forsberg Wrote: [ -> ]Is there today a good reason to keep the design of those flex-hoses?
They all seam to have a very unconventional design, producing a lot of triangles to render.
The cross axle in the longest hose, 32235, is made using the axlehol8 primitive 75 times.
In a template small segments help the quality of the bend, but I see no reason to use 75 segments in the static official library part version.