preferences - safe to merge (finally)

[Dirk Hohndel]

On Sat, Aug 27, 2016 at 12:36:25PM -0700, Dirk Hohndel wrote:
> 
> > While this is mostly finished and a *bit* tested, I didn't do a ton of
> > testing, so I'll send more patches over those, but this is a goos start, if
> > I break anything on any preferences - you'd been warned, not finished, but
> > mergeable.
> 
> It applies nicely. I did my usual rewriting of the commit messages and a
> couple of tiny whitespace fixes.
> 
> > I'd say this is 80% ready. I'v foud a few places where a setting is changed
> > the 'wrong' way (and by wrong way I mean that a setting is changed, the
> > setting should trigger code but some unrelated signal is send that in turn
> > will trigger the correct code. But the signal (or the class) shouldn't be
> > doing that)
> 
> I compiled it and started testing. First thing I notice, I get asked if I
> allow Subsurface to check for updates - and if I accept then it asks me
> again, every time I start Subsurface. Only by declining can I get rid of
> the dialog.
> 
> That's not what we want, is it? :-/
> 
> This is caused by your implementation losing sight of the fact that there
> are THREE states. The DontCheckForUpdates can be unset, true, or false.
> Your code turns this into a binary case. I added a commit to fix that.

Oh, and I noticed that you didn't make the corresponding changes for
Subsurface-mobile to keep compiling, so I added those as well.

/D