Logging this based on a conversation with Bill. A major threat to the stability of a new Firefox version based on Photon could be old profile data containing bad prefs, known bad extensions / combinations or any number of other issues. I'd like to propose that on first run we check for which version of Firefox the current profile was last used. a) If the profile hasn't been used by t least the most recent version, we should run the "reset my Firefox" process and import old data into a fresh profile, automatically. b) If the profile has been used by the most recent version but was not created recently, we should offer and strongly encourage the "reset" process.

(In reply to Jeff Griffiths (:canuckistani) (:⚡︎) from comment #0) > a) If the profile hasn't been used by t least the most recent version, we > should run the "reset my Firefox" process and import old data into a fresh > profile, automatically. ...keeping around the old profile so people can revert. I'm not sure if that already happens with profile reset.

(In reply to Bill McCloskey (:billm) from comment #1) > (In reply to Jeff Griffiths (:canuckistani) (:⚡︎) from comment #0) > > a) If the profile hasn't been used by t least the most recent version, we > > should run the "reset my Firefox" process and import old data into a fresh > > profile, automatically. > > ...keeping around the old profile so people can revert. I'm not sure if that > already happens with profile reset. One thing you have to consider is - what would most people do in this case? I think it is something like: 1. install / upgrade to the new Firefox, check it out, decide to revert. 2. uninstall new Firefox, somehow get an installer for previous Firefox, install that. The user would expect everything to be unchanged once they're using the older Firefox. I think under the hood what we should do is just create a new v57+ profile that new Firefox will use and import data from the default / Firefox 56 profile. If the user starts up Firefox 56, they'll just get the old profile as-is.

Can we have a meeting about this? It seems clear from the first few comments that the discussion here would benefit from more detail about how Firefox Refresh works, what the pitfalls are/aren't, and what our options are here. We should make a decision, and it's been on the front-end team's radar, but bugzilla is a horrible place to flesh this out.

(In reply to :Gijs from comment #3) > Can we have a meeting about this? It seems clear from the first few comments > that the discussion here would benefit from more detail about how Firefox > Refresh works, what the pitfalls are/aren't, and what our options are here. > We should make a decision, and it's been on the front-end team's radar, but > bugzilla is a horrible place to flesh this out. Sure, no problem. I don't think this is the most pressing issue to discuss though, I mainly logged this bug so it wouldn't be forgotten after a discussion with Bill.

Just FYI, this is something I hoped to address in Bug 1268708 and that we could start working on long before Photon. Please keep me in the loop.

FWIW, we will be WebExtensions-only by the time Photon ships. Since removing add-ons is the thing in refresh that (to my knowledge) has the highest impact, we might get that effect for »free« already.

For the record in bug 1336576, we are planning on limiting what add-ons are available on release. We need to know the plans for this new profile. For example, would the new profile import over all extensions or just WebExtensions? Would it tell the users if it did not copy over some extensions?

(In reply to Andy McKay [:andym] from comment #7) > For the record in bug 1336576, we are planning on limiting what add-ons are > available on release. We need to know the plans for this new profile. For > example, would the new profile import over all extensions or just > WebExtensions? Would it tell the users if it did not copy over some > extensions? This bug is about what happens when we ship Photon, which will also be the version in which only web extensions are allowed. I imagine we would only copy web extensions, but I also think Photon's first run should communicate to the users which extensions they previously had installed that are no longer available, and I understand that work is under way on the addons team to provide alternate suggestions?

The main focus has been telling users during 55 and 56 about alternative extensions they could use. What would happen post 57 would have slightly different story, but could use a similar mechanism, if it could find out what the extensions removed are.

Is this still part of the plan for 57?

(In reply to Andy McKay [:andym] from comment #10) > Is this still part of the plan for 57? Yes, generally. I think we need to reach a decision on what to do specifically about extension migration. Maybe a separate bug, with Kev as the default owner.

Remove the whiteboard flag beacause now focusing on bug 1369255 as the bug 1351616 comment 6 & comment 7.

Remove the dependency per the comment 12.

Moving to p3 because no activity for at least 1 year(s). See https://github.com/mozilla/bug-handling/blob/master/policy/triage-bugzilla.md#how-do-you-triage for more information