Re: 23.0.50; Display bug with vline.el and viper

[Lennart Borgman (gmail)]

Vinicius Jose Latorre wrote:
> > >     Try to insert:
> > >
> > >     (redisplay t)
> > >
> > >     At the end of vline-show function.
> > >
> > > If this is something magic, perhaps we need to document in the Lisp
> > > manual the need for this.  Can someone propose a patch to the manual?
> >
> > Maybe it is magic, but then the magic is not strong enough for this 
> > display bug. But maybe something like that function should be called 
> > internally for certain cases? Here it seems like wrapped lines are 
> > involved.
>
> I've just tested vline.el with the fix above, and it works.
>
> Probably Lennart inserted (redisplay t) inside the last while.

Strange. I just tested again with the latest vline.el from EmacsWiki and 
inserted (redisplay t) as the last thing in vline-show. Same bug.

When I tested I had (window-height) => 4, ie 3 lines are shown, and 
(window-height) => 46. I had vline.el in the buffer. Cursor at left side 
(is it column 0 or 1?;-). Just hold down arrow-down.

It does not get past the row starting with

 ;; Author: Taiki SUGAWARA <address@hidden>

> The problem is the following:
>
> 1. there are long lines (lines which go beyond window width) and one 
> long line is at end of window (one part of the line is in the current 
> window and the rest of line is in the next window);
>
> 2. when a scroll occurs (via next-line, for example), the vline-show is 
> called during window scroll process;
>
> 3. vline-show changes the display while the display is being changed by 
> window scroll, this gives some kind of conflict that (redisplay t) seems 
> to fix.
>
> Maybe the explanation is not so precise, but I hope it helps.