Double (mirrored) ephemeris lines in perspective/orthographic projections #811
Assignees
Labels
bug
Something likely wrong in the code
importance: medium
A bit annoying, minor miscalculation, but no crash
state: confirmed
A developer can reproduce the issue
Milestone
Comments
alex-w
added
bug
|
Same problem for planetary trails! |
guillaumechereau
added a commit
that referenced
this issue
Oct 23, 2019
This is an attempt to fix the bug #811. Line clipping cannot be done using only two dimension coordinates, because of cases where one point is in front of us, and the other one is behind us. In fact we should actually be using 4 dimensions to properly handle all the clipping cases, but this is not supported by the projection class. Not tested much.
|
I just pushed a fix attempt in a branch: fix-ephemeris-lines-perspective. |
alex-w
added a commit
that referenced
this issue
Oct 23, 2019
This is an attempt to fix the bug #811. Line clipping cannot be done using only two dimension coordinates, because of cases where one point is in front of us, and the other one is behind us. In fact we should actually be using 4 dimensions to properly handle all the clipping cases, but this is not supported by the projection class. Not tested much.
|
I've tested it and apparently it solve the bug, because I don't see rendering issue now. |
|
Oops... drawing in orthographic projection still have problem |
|
I push a fix for the orthographic projection. This is not a perfect solution: it might still shows some errors at the border, but I think it should be enough for the moment. |
guillaumechereau
added a commit
that referenced
this issue
Dec 18, 2019
This should fix bug #811. The problem is that with the ortho projection the points that are 'behind' the sphere should be hidden using culling, but at the moment we don't really support that. Here I add a clipping test at the projection, even though this is not optimal (two clipped points could still be linked by a partially visible line). Hopefully this is enough to fix the bug. I only activate this with the ortho projection.
gzotti
pushed a commit
that referenced
this issue
Dec 23, 2019
This is an attempt to fix the bug #811. Line clipping cannot be done using only two dimension coordinates, because of cases where one point is in front of us, and the other one is behind us. In fact we should actually be using 4 dimensions to properly handle all the clipping cases, but this is not supported by the projection class. Not tested much.
gzotti
pushed a commit
that referenced
this issue
Dec 23, 2019
This should fix bug #811. The problem is that with the ortho projection the points that are 'behind' the sphere should be hidden using culling, but at the moment we don't really support that. Here I add a clipping test at the projection, even though this is not optimal (two clipped points could still be linked by a partially visible line). Hopefully this is enough to fix the bug. I only activate this with the ortho projection.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Expected Behaviour
Ephemeris lines display as normal, even when using perspective and orthographic projections.
Actual Behaviour
Lines show up in doubles, or "mirrored". That is, a mirrored version of the line is displayed in the opposite direction of the real line. It's not attached to the sky and floats around when panning. See this screenshot:
The mirrored line (note that it has neither markers nor lables) is located between Jupiter and Venus. This screenshot uses perspective projection, and the same result is seen when using orthographic projection. The rest of them are fine.
Steps to reproduce
System
The text was updated successfully, but these errors were encountered: