Re: What should the bbox of a path without a d attribute be?

The contribution of invalid shapes bounding box in parents is an oversight
in Firefox that I think should be corrected. Which would lead to this...

'path' boundingClientRect: 18,18,100,100 bbox: 10,10,110,110 (inner bbox:
0,0,0,0)
'rect' boundingClientRect: 18,18,100,100 bbox: 10,10,110,110 (inner bbox:
0,0,0,0)
'circle' boundingClientRect: 18,18,100,100 bbox: 10,10,110,110 (inner bbox:
0,0,0,0)
'ellipse' boundingClientRect: 18,18,100,100 bbox: 10,10,110,110 (inner
bbox: 0,0,0,0)
'polygon' boundingClientRect: 18,18,100,100 bbox: 10,0,110,110 (inner bbox:
0,0,0,0)
'polyline' boundingClientRect: 18,18,100,100 bbox: 10,0,110,110 (inner
bbox: 0,0,0,0)
'line' boundingClientRect: 18,18,100,100 bbox: 0,0,110,110 (inner bbox:
0,0,0,0)
'line' boundingClientRect: 18,18,100,0 bbox: 0,0,110,10 (inner bbox:
0,0,0,0)

Note that it says nothing in the specification about lines being invalid if
they don't have attributes set, whereas an ellipse with a zero radius, or a
path without any d contents is explicitly not rendered for instance. I
think polygon and polyline without attributes or with empty attributes
should have text which says not rendered to be consistent with path
elements.

Robert