Filing this bug to address this comment https://bugzilla.mozilla.org/show_bug.cgi?id=1289001#c39 which suggests backing out bug 1279568 to address content crashes in Beta50. Also see https://bugzilla.mozilla.org/show_bug.cgi?id=1279568#c46

I would like the backout or conditional pref change mentioned here https://bugzilla.mozilla.org/show_bug.cgi?id=1289001#c44 to land on moz-beta branch before I gtb 50.0b7 tomorrow noon PST. If this happens tonight PST or Taipei/Europe time, please work with :tomcat to get this landed on Nightly, and :sylvestre/:gchang to get Aurora/Beta patch uplift approved and then

Hi Jonathan, Do you have time to do this bug today (Taipei's Thursday)? You can send review request to baku or ehsan (whoever can get to it sooner). We need to do one of two things: 1) Add a pref, privacy.usercontext.about_newtab_segregation.enabled. Take your patches in bug 1279568 that caused newtab to use usercontextid=5 and make them conditional on that pref. When you set the value of the pref, make it false for everything but Nightly. In the future, we can consolidate privacy.usercontext.about_newtab_segregation and privacy.usercontext.enabled into one pref, but for now I think we need two. This solution would need to land on Nightly, Aurora, and Beta. 2) Back out your patches in bug 1279568 from Beta only. I prefer number 1, but if you run into problems, then do number 2. Once the code is done and reviewed, tell tomat, sylvestre, and/or gchang where the patches should land (all channels, or just beta). Thanks!

I'm on it.

Hi Baku, would you review this patch, please?

Pushed by cbook@mozilla.com: https://hg.mozilla.org/mozilla-central/rev/9a047a39cb2f Add pref for bug 1279568 to address content crashes in Beta50. r=baku, a=tomcat

https://hg.mozilla.org/mozilla-central/rev/9a047a39cb2f7e4f97a8b5c0bccc3096efd0aa05 landed this directly on central to get nightly coverage

Comment on attachment 8800540 [details] [diff] [review] Add pref for bug 1279568 to address content crashes in Beta50. [Triage Comment] adds pref to fix content crashes in beta50, needed for b7

clearing ni for sledru/gchang

https://hg.mozilla.org/releases/mozilla-aurora/rev/cd067bd545083034fe8c1a9c5f1c39cf4fd257ea https://hg.mozilla.org/releases/mozilla-beta/rev/4e6e75e9e5ed all done :)

and backed out for test failures in beta and aurora like https://treeherder.mozilla.org/logviewer.html#?job_id=3829008&repo=mozilla-aurora

I should've turn on the new pref in that test.

Hi Mark, would you please review this patch? We had to turn off the loading thumbnails in a separate container thing to avoid crashes, but it breaks the no_cookie_stored test.

nightly: https://treeherder.mozilla.org/#/jobs?repo=try&revision=f3c4c48cadf0bfac6e8328a90edaaf9b57384152 aurora: https://treeherder.mozilla.org/#/jobs?repo=try&revision=c980ffdafd3ade86aa6a19e13ce35f9e1fdc8fdc beta: https://treeherder.mozilla.org/#/jobs?repo=try&revision=f3c4c48cadf0bfac6e8328a90edaaf9b57384152

Comment on attachment 8800646 [details] [diff] [review] Fix the test (also applies to aurora and beta) Review of attachment 8800646 [details] [diff] [review]: ----------------------------------------------------------------- stealing the review!

Thanks for the rescue, Ehsan! Because I just realized what Mark's time zone is and he probably won't see this before tomorrrow noon PST. Add r=ehsan to commit message.

Comment on attachment 8800699 [details] [diff] [review] Fix the test (also applies to aurora and beta) Approval Request Comment [Feature/regressing bug #]: [User impact if declined]: Blocking the uplift of the other patch which solves a crash. [Describe test coverage new/current, TreeHerder]: It's a test itself. [Risks and why]: None. It's just a test. [String/UUID change made/needed]: None.

Comment on attachment 8800699 [details] [diff] [review] Fix the test (also applies to aurora and beta) Fixing a test for a patch needed in beta

Pushed by ryanvm@gmail.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/559fa765dbf1 Turn on the new pref in browser_thumbnails_bg_no_cookies_stored.js. r=ehsan

https://hg.mozilla.org/releases/mozilla-aurora/rev/a9a6e6eda3cb https://hg.mozilla.org/releases/mozilla-aurora/rev/39de9bdccf75

https://hg.mozilla.org/releases/mozilla-beta/rev/b4314d810229 https://hg.mozilla.org/releases/mozilla-beta/rev/2a05cc73a28a

Pushed by kwierso@gmail.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/829c16e72fb7 Spell the preference name correctly so it works a=me

https://hg.mozilla.org/releases/mozilla-aurora/rev/1f5f0d39b41a https://hg.mozilla.org/releases/mozilla-beta/rev/a072309812bd

Pushed by kwierso@gmail.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/b10e48ea5d20 Make super-triple-extra sure we're using the right preference name a=me CLOSED TREE

https://hg.mozilla.org/releases/mozilla-aurora/rev/3a7f992a6cd5 https://hg.mozilla.org/releases/mozilla-beta/rev/6efc0964ec62 If tests are still broken after this, I'm backing out.

I embarrassed myself. I should've notice the spelling error earlier. Thank you, Ryan and Wes.

Tests on aurora and beta are still failing like https://treeherder.mozilla.org/logviewer.html#?job_id=3846062&repo=mozilla-aurora even after getting the preference spelled correctly.

NI Tanvi and Baku in case they can get to this sooner. We really need to understand if the test failure is a code bug/new regression due to this change or a test related problem. I would like to push 50.0b7 with this conditional pref tomm am PST. I'd appreciate a prompt reply.

Backed out all four beta csets in https://hg.mozilla.org/releases/mozilla-beta/rev/688d16cc31ada89128bcf1ee0fd81391757f46ac (without looking first, sorry, but, let's go with one push which has actually been tested and will actually work, rather than a stream of wild guesses).

FWIW, this looks to me like a test-only issue - bug 1279568 landed some additional tests to demonstrate the "old way" of doing background captures would cause cookies to be stored via XHR requests or scripts referencing document.cookie, and I believe it is these tests which are failing (although the taskcluster stack traces are useless so I can't be 100% sure, but I strongly suspect that is the case)

oops - bugzilla is dumb

And backed out of aurora in https://hg.mozilla.org/releases/mozilla-aurora/rev/47f70e2af768c0a48ac6c938a1cf89f1e9d1138c

Comment on attachment 8800942 [details] [diff] [review] Backout changes in Bug 1279568. Hi Mark, would you review this patch that backout all the changes in bug 1279568?

Rebased it upon previous backouts. r=markh

Comment on attachment 8800965 [details] [diff] [review] Backout changes in Bug 1279568. Requesting to uplift to beta, and preferably NOT to nightly and aurora. We'd like to keep the newtab segregation in aurora and nightly. We're trying to fix the crash in the right way in bug 1289001. This backout patch is just a work-around for beta. Approval Request Comment [Feature/regressing bug #]: [User impact if declined]: this patch solves a crash in beta. [Describe test coverage new/current, TreeHerder]: toolkit/components/thumbnails/test [Risks and why]: Minimum. No new code added. Only backing out previous bugs. [String/UUID change made/needed]: None

Comment on attachment 8800965 [details] [diff] [review] Backout changes in Bug 1279568. OK, let's try it again. Should be in 50 beta 8

https://hg.mozilla.org/releases/mozilla-beta/rev/4d0656400d100f93edcf3031cb8ba453c68c092c fingers crossed!

https://hg.mozilla.org/mozilla-central/rev/559fa765dbf1 https://hg.mozilla.org/mozilla-central/rev/829c16e72fb7 https://hg.mozilla.org/mozilla-central/rev/b10e48ea5d20

What's the status of this bug then? Should we call 51/52 wontfix in favor of the work happening in bug 1289001? This is turning into a mess to track :(

This bug should only land in Beta 50. In Firefox 51 and 52, we should land the pref instead of the full backout; that work will happen in bug https://bugzilla.mozilla.org/show_bug.cgi?id=1310112.

(In reply to Carsten Book [:Tomcat] from comment #42) > https://hg.mozilla.org/mozilla-central/rev/559fa765dbf1 > https://hg.mozilla.org/mozilla-central/rev/829c16e72fb7 > https://hg.mozilla.org/mozilla-central/rev/b10e48ea5d20 Can we back this out of central?

Here's a backout of those three commits from m-c: https://hg.mozilla.org/mozilla-central/rev/8ca6edde188aa0bf9ac707b5e1e7b95792e413ff

I was explicitly requested to land attachment 8800699 [details] [diff] [review] on trunk as well yesterday, but OK.

Follow-up: Bug 1310112