[tor-bugs] #17446 [Tor Browser]: Canvas image extraction prompt logic

[Tor Bug Tracker & Wiki]

#17446: Canvas image extraction prompt logic
---------------------------------+--------------------------------
     Reporter:  arthuredelstein  |      Owner:  tbb-team
         Type:  defect           |     Status:  new
     Priority:  Medium           |  Milestone:
    Component:  Tor Browser      |    Version:
     Severity:  Normal           |   Keywords:  tbb-fingerprinting
Actual Points:                   |  Parent ID:
       Points:                   |    Sponsor:
---------------------------------+--------------------------------
 For our canvas image extraction patch (#6253), I think the following
 describes the logic in
 `bool IsImageExtractionAllowed(nsIDocument *aDocument, JSContext *aCx)`
 in [https://gitweb.torproject.org/tor-
 browser.git/diff/dom/canvas/CanvasUtils.cpp?h=tor-
 browser-38.3.0esr-5.5-2&id=65a1a301e584d28f84879625f0dd0e8cd708e5a4
 CanvasUtils.cpp]:

 1. Get the first party URI of the aDocument
 2. Check Permission Manager to see if that first party URI has permission
 to access canvas data; return "true" if so.
 3. Otherwise, check if aDocument is "third party" (meaning, presumably, an
 iframe or similar)
 4. If aDocument is not "third party", then show a prompt allowing user to
 give permission (to the first party) to access canvas data.
 5. If the user gives permission ("always"), then add the first-party URI
 to the Permissions database

 Is there a reason we are preventing third parties from requesting
 permission on behalf of the first party?

 My feeling is we should either (a) allow third parties to request
 permission to extract canvas data, but assign that permission to the first
 party, or (b) prevent third parties from extracting canvas data at all. I
 might be confused about this, though.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/17446>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online