Re: [TCLCORE] TIP 474: "Treat the mouse wheel events in a uniform way" again

[Harald O. <har...@el...>]

Am 26.06.2018 um 10:10 schrieb Arjen Markus:
> Hi everyone,
> 
>> -----Original Message-----
>> From: Harald Oehlmann [mailto:har...@el...]
>> Sent: Monday, June 25, 2018 4:31 PM
>> To: tcl...@li...
>> Subject: Re: [TCLCORE] TIP 474: "Treat the mouse wheel events in a uniform
>> way" again
>>
>> Am 25.06.2018 um 16:24 schrieb Brad Lanam:
>>> On Mon, Jun 25, 2018 at 6:29 AM, Arjen Markus
>>> <Arj...@de... <mailto:Arj...@de...>> wrote:
>>>     As the various widgets cope with this variety, I wonder how many
>>>     packages would be affected if I try to make %D behave uniformly.
>>>     Would it be a better idea to introduce a new binding code, like,
>>>     say, %r (for rotation – angle (a, A), and d are already taken)? Its
>>>     values would be -1 or +1. %D could still contain the current
>>>     platform-dependent value.____
>>> A new binding code would be better.   I could simplify my code.
>>
>> +1
>>
> So we have consensus there. It would reduce the incompatibility as the original bindings can still be used and new bindings can be introduced to eventually replace them.
> 
> As for high-resolution mouse wheels: the new binding code can introduce floating-point numbers. Do we have examples of such mouse wheels? (I am not at all well versed in such matters and seldom use a mouse wheel ;)).

We have input devices, which do the following scroll cases:
Directions:
D1) explicite x-scroll
D2) explicite y-scroll
D3) just scroll, the direction is given by the context
D4) 1+2 together at the same time

Scroll distance:
M1) Turn angle
M2) Pixels
M3) Click count
M4) Flat Distance

Examples:
E1) Mouse with 4 scroll buttons: D1+D2,M3
E2) Windows Mousewheel: D3,M1
E3) Screen gesture: D4,M4
E4) Mouse pad gesture: D4,M4
E5) Brads magic MAC-mouse (supposed): D4,M2

I would appreciate a new binding which supports all directions and uses
screen pixels as unit.

IMHO the aim is to support screen gestures. Most other input shemes
should (in a way) emulate that.
The only exception is the mouse wheel, where the direction depends on
the context. This should be transmitted.

Harald