On 2010-11-25 at 21:31:12 [+0100], Stephan Aßmus <superstippi@xxxxxx> wrote: > Am Donnerstag, den 25.11.2010, 21:07 +0100 schrieb ingo_weinhold@xxxxxx: > > +void > > +SATDecorator::GetComponentColors(Region component, ComponentColors > > _colors) > > +{ > > + switch (component) { > > + case REGION_TAB: > > + if (fBordersHighlighted) { > > + _colors[COLOR_TAB_FRAME_LIGHT] = > > kHighlightFrameColors[0]; > > + _colors[COLOR_TAB_FRAME_DARK] = kHighlightFrameColors[5]; > > + } else { > > + _colors[COLOR_TAB_FRAME_LIGHT] = kFrameColors[0]; > > + _colors[COLOR_TAB_FRAME_DARK] = kFrameColors[3]; > > + } > > this is something that came to my mind from reading the initial patch: > The type name "Region" is perhaps misleading especially in the > app_server context. In the above method, you already call the parameter > "component". Another option is "part", which seems to be used in the Qt > API. Perhaps "DecoratorPart" (unless it's already "Decorator::Part" > anyway). What do you think? The reason why I called it "component" in this context instead of "region" is because the semantics is somewhat different. I intended the Region in the base class to identify the functional aspect, while in DefaultDecorator::GetComponentColors() I mean the concrete graphical components of the default decorator. Admittedly I was just too lazy to introduce a new enum for that purpose. Probably should do just that. Regarding the name of Decorator::Region, I doubt that there's much potential for confusing it with other regions in the app server, but I don't mind, if anyone renames it (after I'm done working in this area). CU, Ingo