Steps to reproduce: 1.Load any page in FF 12. Ex: google.com 2.Inspect page and enter 3D view(Right click -> Inpect Element -> 3D View) 3.Positionate the page upside down facing backwards. 4.Move the page up and down using right click(or up & down buttons) Expected results: Page moves correctly. Actual results: Page moves in the opposite direction.
the arcball doesn't reorient itself based on your current orientation. This is currently working as-designed, though I can understand the confusion.
Let's turn this bug into something more general. From an IRC talk: "the controls we have manipulate the orientation of the visualization, the controls I want put me in a spaceship flying around the visualization". An enhanced arcball and/or an alternate control system would be nice.
Created attachment 602642 [details] [diff] [review] v1 Recycled from experiments in bug 731652. Panning works like it should, both via keyboard/mouse. Added Q and E as roll controls. You can't really roll with the mouse. Tests fail.
not bad. It's an extra range of motion that could be helpful for positioning. The roll is fixed about the center origin again though so at extreme ends of the page, the effect is much more pronounced. i.e., less precise.
As discussed on IRC, actually making this work properly would require a significant refactoring effort, to distinguish between a "model" and a "camera" (basically they're respective matrix transformations) in Tilt. Since this doesn't directly concern this bug, filed bug 734038.
Created attachment 611416 [details] [diff] [review] v2 Played with this a bit during the weekend.
New component triage. Filter on HORSE MASKS.
Filter on TILT QUODLIBET.
The 3D View has been removed in Firefox 47. Closing all bugs inside that component as the component will be moved to Firefox Graveyard. Filter on #BYETHREEDVIEW.