Comment 98 for bug 66566

(In reply to Florian Quèze [:florian] [:flo] from comment #90)
> (In reply to :Gijs Kruitbosch from comment #86)
> > Also, if we're going to change this on only some platforms, given the
> > contention, the most straightforward thing would be to have a pref that we
> > default to different values on Windows - then Windows users who don't care
> > for the IE behaviour can flip it, or vice versa.
>
> A pref would make sense, especially if we want to be able to test both
> behaviors on all platforms. I didn't suggest it mostly because of comment
> 62, but that comment was arguably more about exposing the pref in the UI
> than about having a pref in the implementation.

Yeah - I don't think we want UI, but if we're going to do different things per-OS we might as well implement-via-hidden-pref than hardcode the OS distinctions. There's plenty of precedent for that, e.g. backspace-for-history.back being a hidden pref (browser.backspace_action) that's different on Linux vs. elsewhere.