3d graphics: unintentional mutation

[williamstein]

This is a significant bug in the 3d graphics that Theron Hitchman found. It's easy to replicate on the command line, and doesn't involve actually rendering anything. Basically, in some cases, if g and h are 3d scenes, then doing "k = g+h" actually mutates g itself. This ends up breaking multi-step constructions of scenes (e.g., [1]).

sage: g = point3d([0,-2,-2]); g += point3d([2,-2,-2])
sage: print len(g.all)
2
sage: h = g + arrow([0,-2,-2], [2,-2,-2])   # this should *NOT* change g
sage: print len(g.all)   # it changed g
3

[1] https://cloud.sagemath.com/projects/d115ee6d-f378-4f70-b400-0f2aba16f10c/files/ProjectionOntoAPlane.sagews

Component: graphics

Author: Frédéric Chapoton

Branch/Commit: 7b1682a

Reviewer: Karl-Dieter Crisman

Issue created by migration from https://trac.sagemath.org/ticket/17258

[fchapoton]

Commit: 7b1682a

[fchapoton]

Branch: u/chapoton/17258

[fchapoton]

New commits:

7b1682a

trac #17258 add method of Graphic3dGroup should copy its argument

[fchapoton]

Author: Frédéric Chapoton

[kcrisman]

comment:2

Wow, interesting bug, and relatively subtle since people usually just add plots willy-nilly. I'll try to confirm this works right in a few minutes, but the code seems fine.

[kcrisman]

comment:3

Great, this is very helpful. I don't see any problems with this.

[kcrisman]

Reviewer: Karl-Dieter Crisman

[vbraun]

Changed branch from u/chapoton/17258 to 7b1682a