763904 - Detect when an add-on attempts to access cookies in private browsing mode

Status: NEW

(Firefox :: Private Browsing, enhancement)

Description

[David Teller [:Yoric] - still alive but not very active]

This is a splinter from bug 761950.

The idea is to provide facilities to help diagnose add-ons that, for whatever reason, attempt to access cookies while in private browsing mode. This can be useful both for add-on authors and for AMO testing.

Followup bugs will cover access to other sensitive information while in private mode.

Comment 1

[David Teller [:Yoric] - still alive but not very active]

The current idea is to post events to the nsIObserverService whenever an add-on attempts to access sensitive information while in private mode.

Comment 2

[(no longer active)]

So, two questions: why limit this to only accessing cookies?  Also, Ameya, does your approach differentiate between different method calls on an interface?  Because for example, reading cookies in PB mode is not a problem, writing them to disk is though.

Comment 3

[David Teller [:Yoric] - still alive but not very active]

(In reply to Ehsan Akhgari [:ehsan] from comment #2)
> So, two questions: why limit this to only accessing cookies?

Simply to start small and expand the scope only once we have reached firm ground.

Comment 4

[Ameya Bapat]

(In reply to Ehsan Akhgari [:ehsan] from comment #2)
> Also, Ameya,
> does your approach differentiate between different method calls on an
> interface?  Because for example, reading cookies in PB mode is not a
> problem, writing them to disk is though.

Approach is same for all interfaces. I also considered File Handling in it. 
http://pastebin.mozilla.org/1661227 has list of all considered interfaces. Need a discussion on which of them are really sensitive interfaces because I might have considered some unnecessary or I might be lacking several other.

Comment 5

[(no longer active)]

(In reply to Ameya from comment #4)
> (In reply to Ehsan Akhgari [:ehsan] from comment #2)
> > Also, Ameya,
> > does your approach differentiate between different method calls on an
> > interface?  Because for example, reading cookies in PB mode is not a
> > problem, writing them to disk is though.
> 
> Approach is same for all interfaces. I also considered File Handling in it. 
> http://pastebin.mozilla.org/1661227 has list of all considered interfaces.
> Need a discussion on which of them are really sensitive interfaces because I
> might have considered some unnecessary or I might be lacking several other.

Yes, but my question was specifically about the methods called on those interfaces.  For example, it's perfectly fine for an add-on to call getCookieString on the nsICookieService interface.

Comment 6

[Liz Henry (:lizzard) (relman/hg->git project)]

Hi Ameya, Are you still working on this bug? I see that it's been a while. Is there anything that might help you go further on this?  Let us know. Thanks!

Comment 7

[Nitin Garg [:musingmario]]

Hi Liz, David. I'm a beginner here. Did a Firefox build on Ubuntu couple of days ago and after that just been looking for mentored / good first bugs. I would like to take this up if someone can get me started on the basics of Private Browsing code.

Comment 8

[Josh Matthews [:jdm]]

I think you should look for a different bug to start with; this one is no longer well-defined since we introduced private windows instead of the global private browsing mode.

Comment 11

[random]

There is a problem connected with this bug: nsICookieService can't access cookies when 'Use custom settings for history' selected in Privacy settings (like this https://support.cdn.mozilla.net/media/uploads/gallery/images/2013-07-03-04-42-16-13bb88.png). 
And XMLHTTPRequest from chrome sent without cookies.

Comment 12

[Josh Matthews [:jdm]]

Bug 823941 is separate from this one.