On 01-Sep-00 at 08:49, Alexander Mai ([EMAIL PROTECTED]) wrote:
> On Fri, Sep 01, 2000 at 08:02:05AM -0400, Rick Scott wrote:
> > Analysis:
> >
> > During the operation in question, ml calls XmListReplaceItemsPos with a
> > position of 0. This replaces the entire list. I suspect that something is
> > going to need to be re-drawn. That is, if at least 1 item is showing :)
> > The last patch added in the need_refresh variable, line ~5468, which is
> > initialized to False, and in no case is ever changed to True. The missing
> > _XmListRedraw is causing my missing "D's". The other thing that I noticed
> > is in _XmListReplaceItemPos we are setting the name, width, height, and
> > length of the new item, but not the LastTimeDrawn. Setting this to False
> > at line ~5404 also seems to cure my problem.
> >
> > Peter, you've looked closer at this than I have recently, could you take
> > a look. My thinking is that _both_ places need to be fixed. The first
> > because need_refresh should be turning True in some cases, the second
> > because if it is a new item, how would it have been drawn last time??
> >
>
> Hmm, my guilty conscience forces me to interrupt here ...
>
> I had trouble applying Peter's patch and had to apply the patch.
> Now looking at the diff again (for unknown reasons I kept it ;-)
> it "might" be that I missed something, e.g. some lines
> which feature a
> need_refresh = True;
> ...
>
> So before bashing Peter for a broken patch someone should investigate
> what exactly happened here ...
I'm certainly not bashing him, or you. I've checked in things that have broken
things far worse than this did. After all, it's so subtle that it took me a few
weeks to convince myself there was a problem. Also it may be the correct thing
to do that has uncovered a problem elsewhere. I'm just not sure yet.
>
> --
> Alexander Mai
> [EMAIL PROTECTED]
>
>
