[KDE/Mac] Preparing the QSP patch for KDE CI

[René J.V. Bertin]

On Friday December 19 2014 05:36:24 Jeremy Whiting wrote:

> I think my strategy is to point to use cases where this feature is
> required. If they aren't open to these kinds of changes I'll ask how we
> would go about making these automated tests work without this feature. I
> think showing a use case is enough to convince them that this feature would
> be useful.

I've already tried that approach for the text-based heuristic guessmatism they use to determine whether an action being added to a menu should get a special role on OS X, just so t^Hh^He^H a settings menu ends up under <Application Menu>/Preferences, an about action under that same menu's About, and idem for Quit. That latter is about the only action that seems to work reliably, but there too it depends on how such an action is called (by the developer who could call it "Go with the wind" instead, but also in the language being used).
I've given ample evidence of KDE applications where these guesses lead to unexpected and inappropriate behaviour, but until now that's not had any success in driving interest into at least a mechanism to disable the feature for the most problematic actions/menuroles.

> > Only this morning I realised that Thiago was quite clear about
> > that such patch for our CI isn’t something he’d like to see
> > promoted. He considers testing apps in their default locations
> > as the only way to go.

Yes, I think they are *very* concerned about accepting anything that will endanger the needs of their paying customers, some of whom apparently consider it important that their apps can end up in the Mac Store without having to worry about incompatible Qt features.

I wonder what his (and Apple's...) attitude would be towards build options. After all, anything that we patch in could also be included as an optional but official build configuration. That would at least have the advantage of our adaptations being "officially included" and visible to anyone planning to modify something.

R.