On 10/15/2012 08:30 PM, Uwe Rathmann wrote: > On Mon, 15 Oct 2012 07:32:52 +0000, Knoll Lars wrote: > >> to make it really logical would probably also require to >> make 0 width pens invisible > > You are hitting the point: obviously setting a pen width of 0 enables a > special mode and doesn't really set a pen width: > > - disabling the hardware accelerated path ( on systems that have one ) > - making the pen cosmetic > - ... ? > > In opposite to Samuel I see this hardware accelerated mode as important - > even so important that it should stay the default mode. And when this > mode only works for cosmetic pens ( ? ) I strongly argue for not changing > the default setting.
The point about hardware acceleration is a bit moot at the moment since Qt 5 doesn't include an X11 paint engine. In any case, instead of having the "0" value magically be rendered differently from a 1-width pen it would probably be nicer to have that as a RenderHint if we should at some point re-introduce an X11 paint engine in the 5.x series. Alternatively, always use the fastest available line rendering for 1-width cosmetic pens. > IMO a clean API should offer all attributes explicitly - not hidden > behind an invalid width of a pen. So after reading Samuels explanations I > would see the proposal of Jens as an improvement for the API - but not > enough to clean up the API. Right :) > So when you really want to introduce an incompatibility like this now, > please clean up the "pen width 0" API completely. Indeed, the idea would be that 0-width pens get treated the same as pens with style Qt::NoPen. -- Samue _______________________________________________ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development