G> Martin Buchholz <martin(a)xemacs.org&gt; writes: G> This is probably (somehow) related to the redisplay fix for the G> widening crash (1998-09-11 in the ChangeLog). There were some other G> reports of this happening but nobody could ever reproduce it. I still G> hope that the fix is OK and we are just exposing different redisplay G> bugs this way... G> Are you using scroll-in-place? That seems to be pretty good at G> triggering problems in redisplay (which is a Good Thing). Yes, I am. G> Yeah, I know what you mean, but it would certainly be great for XEmacs G> if *you* spend a year improving/fixing redisplay :-) I really think we G> should do something about redisplay: either rewrite it or at least G> generate some helpful documentation. G> could you print the values of: w-> start_at_line_beg G> beginning_of_line_p(b,startp) G> in this frame? G> Currently I don't have any time at all to look into this (I got a G> deadline for some papers on Friday and still need to prepare the talk G> for next week :-) Maybe we can talk about it in Japan. I'll do the crash demo for you in Japan. I think I can reproduce it at will, albeit in a heavily Martinized xemacs. Hmmm. Ok, I can now reproduce this in a vanilla xemacs as follows: Eval: (progn (setq split-window-keep-point nil) (popper-install)) Then: M-x shell, C-x 1, C-x 2, button 1 in empty window. KABOOM! Martin