On Mon, Feb 3, 2014 at 1:06 AM, Lars Ingebrigtsen <larsi@gnus.org> wrote:

> "Lennart Borgman" <lennart.borgman@gmail.com> writes:
>
> > On Tue, Oct 14, 2008 at 7:57 PM, Lennart Borgman (gmail)
> > <lennart.borgman@gmail.com> wrote:
> >> I have found a bug where point jumps to a new postion instead of a new
> >> line beeing scrolled into the selected window. I am unable to narrow it
> >> down, but I can reproduce it (but I am not sure about what makes it
> >> happens). Just to get some thoughts I write down what I have seen so far
> >> here.
> >>
> >> The scenario is this:
> >> - Point is on the last line in the window.
> >> - I press "o" in viper. This opens a line below the current line and
> >> puts the point on this line.
> >>
> >> What I expect to happen is that this new line is scrolled into the
> >> window. Sometimes this happens. Sometimes instead point jumps up, maybe
> >> 10 lines (I did not count them at all) and the window is not scrolled so
> >> the new line is not visible.
> >>
> >> There are some other ingredients too:
> >> - I believe that nxml-mode (or a derivative) must be the major mode.
> >> - If I remove nxml-after-change from after-change-functions the bug
> >> disappears.
> >> - If I try to use edebug it also disappears.
> >>
> >> Maybe those ingredients also are required, I am not sure since I can't
> >> easily reproduce the bug yet:
> >> - visual-line-mode.
> >>
> >> Does anyone have any idea of how to find out what the problem is? In
> >> nxml-after-change there is a whole bunch of "save-*" macros. I commented
> >> out them all, but the bug still appears. But where is the scrolling
> done?
> >
> > It would help if someone could tell me where the scrolling should have
> > taken place. To begin with:
> >
> > - Where in command_loop_1?
> > - And then of course a little bit more exact ...
>
> Is this problem still present in Emacs 24?
>
>
Unfortunately I have no idea. I am still using my old patched version of
Emacs. And I do not have time to upgrade since that would require either
patching again or getting the patches inside Emacs. Both alternatives takes
more time than I can afford now.