939269 - Build broken with (17) SecurityError: The operation is insecure.

Status: RESOLVED FIXED

(Firefox OS Graveyard :: Gaia::Browser, defect)

Description

[Gareth Aye [:gaye] (back from PTO)]

1) Settings "before each" hook:

JavaScriptError: (17) SecurityError: The operation is insecure.

Remote Stack:

@app://browser.gaiamobile.org/js/browser_db.js, line 177

at Error.MarionetteError (/home/travis/build/mozilla-b2g/gaia/node_modules/marionette-client/lib/marionette/error.js:67:13)

at Object.Client._handleCallback (/home/travis/build/mozilla-b2g/gaia/node_modules/marionette-client/lib/marionette/client.js:474:19)

at /home/travis/build/mozilla-b2g/gaia/node_modules/marionette-client/lib/marionette/client.js:508:21

at TcpSync.send (/home/travis/build/mozilla-b2g/gaia/node_modules/marionette-client/lib/marionette/drivers/tcp-sync.js:94:10)

at Object.send (/home/travis/build/mozilla-b2g/gaia/node_modules/marionette-client/lib/marionette/client.js:455:36)

at Object.Client._sendCommand (/home/travis/build/mozilla-b2g/gaia/node_modules/marionette-client/lib/marionette/client.js:501:19)

at Object._executeScript (/home/travis/build/mozilla-b2g/gaia/node_modules/marionette-client/lib/marionette/client.js:1448:19)

at Object.executeAsyncScript (/home/travis/build/mozilla-b2g/gaia/node_modules/marionette-client/lib/marionette/client.js:1229:19)

at Object.Client.waitForSync (/home/travis/build/mozilla-b2g/gaia/node_modules/marionette-client/lib/marionette/client.js:674:14)

at Object.Client.waitFor (/home/travis/build/mozilla-b2g/gaia/node_modules/marionette-client/lib/marionette/client.js:648:60)

at Object.MarionetteHelper.waitForElement (/home/travis/build/mozilla-b2g/gaia/node_modules/marionette-helper/index.js:129:12)

at Context.<anonymous> (/home/travis/build/mozilla-b2g/gaia/apps/browser/test/marionette/settings_test.js:13:19)

at Hook.Runnable.run (/home/travis/build/mozilla-b2g/gaia/node_modules/mocha/lib/runnable.js:211:32)

at next (/home/travis/build/mozilla-b2g/gaia/node_modules/mocha/lib/runner.js:246:10)

at /home/travis/build/mozilla-b2g/gaia/node_modules/mocha/lib/runner.js:253:7

at done (/home/travis/build/mozilla-b2g/gaia/node_modules/mocha/lib/runnable.js:185:5)

at /home/travis/build/mozilla-b2g/gaia/node_modules/mocha/lib/runnable.js:197:9

at Object.executeHook (/home/travis/build/mozilla-b2g/gaia/node_modules/marionette-client/lib/marionette/client.js:370:18)

at process._tickCallback (node.js:415:13)

Comment 1

[Gareth Aye [:gaye] (back from PTO)]

Needinfo'ing the browser guys...

Comment 2

[Michael Henretty [:mikehenrty][:mhenretty]]

Just throwing in my two cents here: this could easily be a gecko thing. Looking at travis logs, here's a build that worked:

http://ftp.mozilla.org/pub/mozilla.org/b2g/tinderbox-builds/mozilla-central-linux64_gecko/1384516920/

Here's one that didn't work:

http://ftp.mozilla.org/pub/mozilla.org/b2g/tinderbox-builds/mozilla-central-linux64_gecko/1384530185/


Commits:
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=b2fab608772f&tochange=60a117ce0fab


Looking at the commits, this one looks like it pertains to network service: http://hg.mozilla.org/mozilla-central/rev/f49134e2a4dc, but that is just a guess.

Comment 3

[Michael Henretty [:mikehenrty][:mhenretty]]

Actually, this one looks more likely the culprit:
http://hg.mozilla.org/mozilla-central/rev/e6a9c98d62d0

Comment 4

[Andreas Gal :gal]

Can you confirm that that patch regressed us and if so, please back it out? Thanks.

Comment 5

[:gerard-majax]

I'm on this, I have the bug reproduced locally and I will start searching for the culprit.

Comment 6

[:gerard-majax]

I can confirm that reverting the commit pointed by :mhenretty in comment #3 fixes the issue: I can now run the test suite for the Browser app.

Comment 7

[Ed Morley [:emorley]]

(In reply to Michael Henretty [:mhenretty] from comment #3)
> Actually, this one looks more likely the culprit:
> http://hg.mozilla.org/mozilla-central/rev/e6a9c98d62d0

Backed out:
remote:   https://hg.mozilla.org/mozilla-central/rev/beddd6d4bcdf

Comment 9

[Anthony Ricaud (:rik)]

Travis is green, this was the good culprit. Thanks :mhenretty for finding the patch and :gerard-majax for confirming!

Comment 10

[Martijn Wargers (dead)]

Isn't this external code?

Comment 11

[Martijn Wargers (dead)]

This line needs to have added mozAnon: true
https://github.com/mozilla-b2g/gaia/blob/abff7bf12d5033201405d12d3e8d8bcd6f1dfdff/apps/browser/js/browser_db.js#L177
So instead of this:
      var xhr = new XMLHttpRequest({mozSystem: true});
It needs to be:
      var xhr = new XMLHttpRequest({mozAnon: true, mozSystem: true});

And these instances also need this treatment, I guess:
https://github.com/mozilla-b2g/gaia/search?p=2&q=mozSystem&ref=cmdform

Comment 12

[Julien Wajsberg [:julienw]]

This issue came back.

Comment 13

[Julien Wajsberg [:julienw]]

Closing again.

It's here but it's another issue.

Comment 14

[Martijn Wargers (dead)]

I don't understand. The issue is still here, then it's not fixed, is it? But then you say it's another issue? What other issue is it then? Are you saying that bug 927196 didn't cause it?

Comment 15

[Julien Wajsberg [:julienw]]

Sorry, I was confusing.

It happens that our Travis setup is currently fetching a build that doesn't have the backout. I filed bug 939787 to track this.

We had a small moment where Travis got a build with the backout and it actually fixed the issue we were seeing here.

Comment 16

[Martijn Wargers (dead)]

Thanks for the explanation, sorry about the problems I created.

Comment 17

[Andrew McCreight [:mccr8]]

It would probably be good to file a new bug to get Gaia to update and fix these issues.  Would that work with or without your patch, Martijn?

Comment 18

[Martijn Wargers (dead)]

I already put a patch in bug 927196.

Comment 19

[Anthony Ricaud (:rik)]

I can't see bug 927196 so I don't understand why this change is needed. But I'm concerned about breaking 3rd party apps that use mozSystem already.

Comment 20

[Martijn Wargers (dead)]

(In reply to Anthony Ricaud (:rik) from comment #19)
> I can't see bug 927196 so I don't understand why this change is needed. But
> I'm concerned about breaking 3rd party apps that use mozSystem already.

Yes, that would be a problem. They would have to update their apps.