In [1] ferjm suggested we check the DOM to determine if our code was preffed off. However, Arthur asked in [2] that we get the settings directly from JS, without going through the DOM. We should update the other gaia code to use this instead, it's much cleaner. [1] https://bugzilla.mozilla.org/show_bug.cgi?id=964390 [2] https://github.com/mozilla-b2g/gaia/pull/15692#discussion_r10418248

Jared, do you want to just back this code out? How does the risk/effort of: 1) moving the pref code, turning the pref on, waiting a week or three, removing the moved pref code compare to: 2) removing the existing pref code; patching it back in if we need to You can probably guess that I vote for 2) :-)

That's an awesome idea, Sam :-)

per discussion in bug 967783, maybe we want to promote this to blocker?

Updating title, as it turns out FTU, not System, looked for the pref.

If I split this change across the two apps, then if (say) the FTU commit removes the pref, the Settings marionette test will fail. If I remove the pref in the Settings commit, then the FTU marionette test will fail. Simplifying by marking this FTU bug as a dupe of 993794, and we'll get two reviews for one still small commit.

I was incorrect in c3, actually bug 996476 contains the discussion of problems with OTA updates and this pref.