On 07-Jan-2002 Lars Gullik Bjønnes wrote:
> Pointer * pointer;
> ...
> if (!pointer) {
>pointer = new Pointer;
> }
>
> will always have problems.
I don't get what you mean here (especially in the contents we are talking
about), but as I told you I don't argue that all class variables shoul
On 07-Jan-2002 Lars Gullik Bjønnes wrote:
> Class variables should be initialized anyway, especially pointers.
Well if this is the rule just commit your change it surely won't do anything
bad, the only thing we won't get debug notifies anymore and so we cannot be
pointed to problems in the code
On 05-Jan-2002 John Levon wrote:
> hmm, this looks like the_locking_inset is referring to a deleted inset. I don't
> get this code too much ...
Nope the_locking_inset is ok inset_y is not ok (it was not initialized), but
I explained this already in my former mail!
Jug
--
-._-._-._-._
On 06-Jan-2002 Lars Gullik Bjønnes wrote:
>| - Some (biggest) memory leaks
[snip]
>| createUndo(BufferView*,Undo::undo_kind,const Paragraph*,const Paragraph*)
>
> I really cannot see how and why this happens.
Well I can and we ALWAYS had this problem. It's easy when undoing/redoing
changes t
On Sat, Jan 05, 2002 at 12:04:07PM +0100, Michael Schmitt wrote:
> - Operation not known exactly:
I bet this is the same as your other bug where the delete empty para
mechanism comes in and removes a paragraph from under BufferView::removeAutoInsets()
(after all it's a crash, what else could it
