Comment 138 for bug 583797

(In reply to Amy Lee [:amylee] UX from comment #49)
> (In reply to :Gijs (Not available 3-19 Aug; he/him) from comment #48)
> > Second, users who are used to the current behavior with session restore
> > enabled (esp. on Windows, where mis-hitting ctrl-shift-q is also not a major
> > concern) will start getting warning dialogs when we implement the new
> > behavior. That seems disruptive. But we can't turn off the warning
> > everywhere because then they won't get a warning when closing a window,
> > which they might have relied upon. Any thoughts about how to address that?
>
> The default settings of Session Restore would be what we have currently (not
> having warnings).

That's not really what happens today. Today, there are warnings when closing windows, or using "close tabs to the right", or using "close other tabs" - but not when quitting, if and only if session restore is enabled.

Specifically, the preference you're suggesting we surface "under" Restore Previous Session already exists, and its default is true (ie warn when closing tabs). However, there are hardcoded checks for "does the user have session restore enabled?" in the code that avoid warning when quitting, irrespective of whether the "warn when closing tabs" checkbox is checked.

In your new design, we'd remove the hardcoded checks to ask session restore in the code that actually implements warning, and cosmetically we'd move the checkbox in about:preferences, and add some logic to tick/untick (and set/unset the pref) when the user toggles restoring their session. But that doesn't solve the conundrum here - the pref would be turned on by default, and we'd remove the checks, so then the user will start getting warnings.

We could have a one-time migration that turns off the "warn when closing tabs" pref for people who have session restore turned on, but that also seems like it'd potentially confuse users who would no longer see warnings when closing windows / multiple tabs (e.g. via "close tabs to the right") where they did see those before.

What would you prefer we do? Disable the warnings everywhere for session restore users, or accept that we'll now start warning session restore users who might not have expected this?

(In reply to Amy Lee [:amylee] UX from comment #49)
> I think we are getting very specific for an issue that I think affects a
> small population of users.

While I sympathise with the sentiment here, we already made a change in bug 1438499 (to show the same "closing multiple tabs" warning when quitting when session store is not enabled, removing a separate option to show an alternative dialog that people had to opt in to via about:config), and already people on Nightly are basically saying that that removed a "fail-safe":

(In reply to Kestrel from bug 1438499 comment #47)
> browser.showQuitWarning was a good fail-safe to avoid data loss but now that
> it has been removed a suitable replacement is needed, hopefully Bug 550559
> can be expedited to avoid a regression.

... because the "closing multiple tabs" warning on quit is enabled/disabled together with the same warning when closing windows/other/right-hand tabs, there's no way to force the quit warning to come up but not warn for closing windows / multiple tabs, and that's burning people (it seems).

(In reply to Amy Lee [:amylee] UX from comment #49)
> It starts to become overkill if we try to provide
> such specific checkbox options (ie. warn me when quitting, warn me when
> closing a window, etc.).

FWIW, I think the only requirement here is quitting vs. everything else, and I think an about:config pref would be fine. I don't think this is going to be a slippery slope. Quitting is by its nature *much* more destructive than closing a single window, and as noted, easier to do by accident (esp. on macOS/Linux) than all the other ways of closing multiple tabs.

If you're convinced that we don't need to provide an option here, that works, but I think we need to tread carefully and deliberately here.