[Qt5-feedback] more specific mouse questions.

[Rick Stockton]

Thanks for advising me that Qt5 will probably use Wayland! My immediate 
curiosity, however, is mostly about our public API for mouse/pen type 
"pointers" (I.e., old-style "core pointer" devices which contact the 
Screen(s) and Windows at one point, defined by 2D coordinates.)

Ignoring X11's various bits of "lateness", Wayland and X11 both use 
evdev drivers, so the information we can acquire is almost identical. My 
REAL interest is in the higher level design and API which we will offer 
to users- And right now, I've got two questions:

(1) Already mentioned, but I didn't explain my interest: We will have 
the capability, underneath, of providing raw events to ALL Windows and 
their objects. I've seen many 'newby' Qt Users asking "why didn't I get 
that Button Click?" -- and of course, it's because the button was 
"sucked in" by a child object, higher on the stack, by "accident". An 
API which allowed a user to request event/signal notification all the 
time, even if another Window has grabbed the device, might help such 
programmers get their work done faster. And with fewer lines of code. 
(And actually, there a quite a few EXPERTS who want this sort of thing too.)

(2) What about unifying the "WHEEL" API with the "BUTTON" API? They're 
really all just buttons, except for the compression of repeat events to 
create a "scroll speed" adjustment. (In the long term, that should 
probably be pushed down into Wayland, or even the Driver, eliminating 
LOTS of throw-away interrupts.) In "compatibility" code, the two sets of 
functions would share the new code, but the "wheel" functions would 
execute under a mask.

Thanks again!