Andre Poenitz wrote:
Woo! Things move on bloody quickly!
I find it frankly amazing that the LCursor class shrinks to almost
nothing. I even feel I understand how it works which is just
gob-smacking!
Index: cursor.h
+ /// mainly used as stack, but wee need random access
Swap wee for we.
Index: lyxfunc.C
- FuncRequest cmd(view(), LFUN_INSET_EDIT, "left");
- par.inset()->dispatch(cmd);
+ par.inset()->edit(view(), true);
I don't really see why you are introducing a new function here. Why
can't this be handled by the dispatch machinery?
Index: insets/insetbase.C
+ /// cursor enters
+ virtual void edit(BufferView * bv, bool left);
Shouldn't we really use the concept of 'front' and 'back' here,
rather than 'left' and 'right'? I'm thinking the current notation
sits uncomfortably with the RtL stuff.
also, why not introduce a new
enum EnterInset {FRONT, BACK};
It will be more readable to have
cursor.pos())->edit(bv(), FRONT;
don't you think?
Index: text3.C
+ inset->edit(bv, true);
+ bv->cursor().push(inset, inset->getText(0));
Very nice!
Index: insets/insetbibitem.C
- case LFUN_INSET_EDIT:
+ case LFUN_MOUSE_PRESS:
PRESS or RELEASE? Honestly, I don't know. I do see, however, that the
graphics inset acts on MOUSE_RELEASE.
Index: insets/insetcite.C
- case LFUN_INSET_EDIT:
- InsetCommandMailer("citation", *this).showDialog(cmd.view());
+
+ case LFUN_MOUSE_PRESS:
+ InsetCommandMailer("bibitem", *this).showDialog(cmd.view());
I don't think I want to open the "bibitem" dialog here ;-)
--
Angus
