Andre Poenitz <[EMAIL PROTECTED]> writes:
| On Wed, Aug 28, 2002 at 08:38:39AM +0200, Lars Gullik Bjønnes wrote:
| > Not too bad. But I feel that the switch/cases are a bit too big and
| > really should be kept in separate functions.
|
| I said I'll move them down to LyXText (which will make them slimmer).
| If they are still big than, I'll put them into seperate functions.
| Certainly today...
|
| > Also, one thing I'd really like to see (not necessarily in this
| > patch), is to split the mouse lfuns ... not that is not what I want...
| >
| > I want it to be possible to do a
| >
| > \bind mouse-button-1 place-cursor
| >
| > kindo of binding. This means that the mouse buttons must be seen as a
| > kind of keyboard-key. (keysym if you want)
|
| Sort of LFUN re-mapping. Should be possible. Actually I like the idea
| because it makes things more flexible...
|
| > This also implies that there should not be _any_ kind of special LFUNS
| > for mouse handling, but that you use key-bindings to achieve that
| > instead. Then we should not need any kind of special casing in the
| > code on "which button was pressed" or any explicit mouse button
| > handling at all.
|
| I'd approach this a bit differently. Assume for a while we have
| inset-local bindings and that lfun re-map feature.
|
| The workarea could still generate mouse FuncRequest. After all, the
| workarea is completely ignorant of text, formulas etc. It only knows it got
| clicked and has to do something with it. So it passes the mouse lfun to
| BufferView::dispatch(), and as BufferView is ignorant of this, too, it gets
| passed in the default branch to LyXText::dispatch().
I do not really agree, and feel that this would be more cleanly
supported by "context"-bindings.
But of course if we spread the dispatch out to thin, that will be a
real hassle.
What I really want is for all dispatch to be really ignorant about
mouse buttons. (and the rest of the code as well...)
--
Lgb
