1063186 - aWindow is null in e10s builds

Status: RESOLVED WORKSFORME

(Firefox :: Private Browsing, defect)

Description

[(no longer active)]

STR: Open a new window.

aWindow is null PrivateBrowsingUtils.jsm:24

  privacyContextFromWindow: function pbu_privacyContextFromWindow(aWindow) {
    return aWindow.QueryInterface(Ci.nsIInterfaceRequestor)
                  .getInterface(Ci.nsIWebNavigation)
                  .QueryInterface(Ci.nsILoadContext);
  },

Comment 1

[Brad Lassey [:blassey] (use needinfo?)]

Ehsan, is there any effect of this? or just console spew

Comment 2

[:Felipe Gomes (needinfo for replies!)]

And could you get a call stack? I thought most usage of PrivateBrowsingUtils.isWindowPrivate should be passing a chrome win as a param

Comment 3

[(no longer active)]

(In reply to Brad Lassey [:blassey] (use needinfo?) from comment #1)
> Ehsan, is there any effect of this? or just console spew

Hard to say without a callstack.

(In reply to :Felipe Gomes from comment #2)
> And could you get a call stack? I thought most usage of
> PrivateBrowsingUtils.isWindowPrivate should be passing a chrome win as a
> param

I would be happy to.  How do I do that?

Comment 4

[:Felipe Gomes (needinfo for replies!)]

You could use the browser debugger to set a conditional breakpoint there [1], or if you can reproduce the error with your own build you can dump((new Error).stack);

[1]
Set devtools.debugger.remote-enabled and devtools.chrome.enabled to true
Go to Tools > Web Developer > Browser Toolbox

Comment 5

[:Felipe Gomes (needinfo for replies!)]

FWIW I can't reproduce it here (testing with a clean profile), so that's why I can't get the stack myself

Comment 6

[Brad Lassey [:blassey] (use needinfo?)]

NI to Ehsan to get stacks

Comment 7

[(no longer active)]

Attachment: Callstack

Sorry that this is a screenshot, it seems impossible to copy a callstack from the debugger.

Comment 8

[:Felipe Gomes (needinfo for replies!)]

That's add-on SDK code. Zombie, care to take a look?

This may be breaking some add-ons, so i'm marking this as blocking e10s-addons

Comment 9

[Tomislav Jovanovic :zombie]

we have some private-browsing refactoring in flight (landed to github, to be uplifted to m-c), and with bug 1062024 close to landing (to turn on addon interposition for jetpack addons), it's probably a good idea to delay testing this until next week after both of these land in Nightly..

Comment 10

[Brad Lassey [:blassey] (use needinfo?)]

(In reply to Tomislav Jovanovic [:zombie] from comment #9)
> we have some private-browsing refactoring in flight (landed to github, to be
> uplifted to m-c), and with bug 1062024 close to landing (to turn on addon
> interposition for jetpack addons), it's probably a good idea to delay
> testing this until next week after both of these land in Nightly..

What's your ETA for that?

Comment 11

[Tomislav Jovanovic :zombie]

jetpack uplifts from github ~weekly, and with all the parties in bug 1062024 happy with proposed patches, i expect end of this week or beginning of next.

Comment 12

[Tomislav Jovanovic :zombie]

hey Ehsan, is this still an issue? 

lots of code changed since then, so could you please test again, or at least post STR so that i could test it?

thanks.

Comment 13

[(no longer active)]

The STR is in comment 0.  I do not get this any more, but I'm not sure if that's because the bug was fixed or if the call was moved out of the code path for opening a new window.

Comment 14

[Tomislav Jovanovic :zombie]

(In reply to :Ehsan Akhgari (not reading bugmail, needinfo? me!) from comment #13)
> The STR is in comment 0.  I do not get this any more, but I'm not sure if
> that's because the bug was fixed or if the call was moved out of the code
> path for opening a new window.

well, isn't that the same thing? that method shouldn't be called with anything but a chrome window, and now it isn't..

i'm gonna mark this WORKSFORME, if it happens again, feel free to re-open.