Proposal for new modifiers/shortcuts system

[Dmitry Kazakov]

> This would be a major change and basically turns the whole tool system
> upside down. I would prefer the changes to be much less invasive and I think
> that is possible.
>

Well, there are not much changes in flake. I just added new classes there.
And new tools can be written with them. Old ones can work as usual. All the
"invasive" changes will happen in KisTool and lower. Take into account that
these classes will be rewritten anyway while encapsulating their actions for
the scheduler. So these shortcut changes will be a kind of side-work =)

I think the two level with the tool at global level and strategies at local
> level should stay.


Yep, that is already done in flake. But noone uses it. We have 8
interactional tools only (5 in flake itself, 1 in each of karbon, tables,
treeshape).


> Tools need to have control about the strategy that is activate e.g. look at
> the default tool, the next strategy depends on the content and state of the
> tool.


The state machine of the switching is a good point actually. I need to think
it over. It should be generalized as well.



> That's why it's not possible to move that to a higher level as that manager
> can only decide based on shortcuts and gestures.
>

For most of the tools we can assume that we can do panning and rotation when
the user does not hold any mouse button pressed. This is the idea of
KoToolBaseStrategyWrapper.
For all the interactional tools there is no such problem, because you can
know what it is doing by checking currentStrategy() pointer. And that is the
idea of KoInteractionToolStrategyWrapper.


> One global manager should be enough for that. Additionally tools could
> block global switches while working.
>

How are you going to communicate between this global manager and all the
tools? And how will it process semi-global shortcuts, like the ones those
should be activated for KisToolPaint only? There should be more than two
layers of shortcuts.

-- 
Dmitry Kazakov
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.kde.org/pipermail/kimageshop/attachments/20110613/94cb0807/attachment.htm