bug#10304: 24.0.92: display bug

[Lars Magne Ingebrigtsen <larsi@gnus.org>]

Eli Zaretskii <eliz@gnu.org> writes:

> Is this reproducible, e.g., with that blog post?

Nope.  After refreshing the display once, it doesn't happen again.

I'd say that it happens about, er, once a thousand times that Emacs
displays a message, and it's totally unreproducible.

I *think* what happens is always that this only happens when I scroll
the buffer down and it reaches the end.  After the end of the buffer,
the bottom of the window should be totally blanked out, but sometimes
(if there's an image in the part that should have been blanked out after
scrolling), it doesn't happen.

So my feeling is that there may perhaps be an optimisation relating to
clearing the window after end-of-buffer.  But that may be totally wrong,
of course.  :-)

> I've seen in the past redisplay problem with images, but they would
> disappear in a fresh session of Emacs, even if I made exactly the same
> gestures as the ones which reproduce the problem in a long-running
> session.  Do you have the same experience?

Yes.

> When you do have Emacs built with -DGLYPH_DEBUG=1, and this happens,
> invoke "M-x trace-redisplay RET", and tell what it prints to stderr
> when you do whatever is needed to reproduce the problem.

I'm now running with GLYPH_DEBUG and STDERR to a file, so I should be
able to get a trace the next time...

-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/