[tor-bugs] #14429 [Tor Browser]: Automated rounding of content window dimensions

[Tor Bug Tracker & Wiki]

#14429: Automated rounding of content window dimensions
-------------------------+-------------------------------------------------
     Reporter:           |      Owner:  arthuredelstein
  arthuredelstein        |     Status:  needs_revision
         Type:  defect   |  Milestone:
     Priority:  normal   |    Version:
    Component:  Tor      |   Keywords:  tbb-fingerprinting-resolution, tbb-
  Browser                |  torbutton
   Resolution:           |  Parent ID:
Actual Points:           |
       Points:           |
-------------------------+-------------------------------------------------

Comment (by arthuredelstein):

 Replying to [comment:99 bernie.allen]:
 > Replying to [comment:98 arthuredelstein]:
 > > What is the zoom level?
 >
 > 119%
 >
 > > The reason for zooming is that, if we don't zoom at all, then there
 will be a large grey margin at the bottom. Zooming out in this case be
 nicer, but it's impossible to get precise multiples of 100 when we zoom
 out.
 >
 > That makes sense.
 >
 > The zoom ratio is potentially higher for lower display resolutions
 because 100 pixels is a greater proportion of the total height.  In my
 case I think I am just short of the next 100 pixel cutoff when maximized,
 which causes a large zoom.  So I reconfigured the Tor Browser UI to give
 more vertical space, and now maximizing only causes a 101% zoom, which
 looks great.

 What kind of reconfiguring did you do?

 > 1024x600 is probably an unusually low resolution, and reconfiguring the
 UI is a fine solution to me.  Two other possible solutions though:
 > 1. Zoom out instead of in when you are close to the next multiple.  For
 instance, if you are at 499 pixels you only need the tiniest zoom out to
 make it to 500.  But you need a large zoom in to remove a 99-pixel border.
 I think the zoom out should be small enough that it does not cause an
 accessibility problem.
 > 2. Make the quantization interval smaller for smaller sizes (e.g., 50
 pixels), so the maximum zoom is less.  But I understand there is already
 concern that 100 pixels adds too much entropy.
 >
 > Edit: Nevermind on 1, I see you already said that won't work.

 Well, it might work sometimes. Basically if you zoom out, it's a little
 unpredictable whether an exact multiple of 100px can be obtained. (Zooming
 in always works.) We could use code that under, certain circumstances,
 invisibly tries zooming out and falls back to zooming in in the case of
 failure. The code gets more and more complex, though, so I'm hesitant to
 it. I'll need to think about it more.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/14429#comment:100>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online