On the other hand, you'd think we would have seen this before,
that redisplay is entered very frequently. Also, the fact that it is
caught by an assert suggests that somebody thinks that the invariant
is satisfied by the code.
On the gripping hand, it's the first time I've seen it, and I've been
editing the offending file for 4.5 years, to say nothing of all the
other everyday work and VM I do...
I can reproduce it in a standard 21.4.21 (but with my extensive
.emacs, as it's font-dependent). Of course, a standard build doesn't
have error-checking turned on, so instead of a crash, one sees a
momentary glitch. Here's the sequence of events from the user
search to the end of the desired narrowing region, set mark.
Move (by C-p) to the beginning of the desired region.
C-x n n
The region narrows and displays as expected.
The end of the region is displayed as expected.
The start of the region is displayed centered in the window, and above
it are the preceding lines from the file, outside the narrowed region.
One can move about with keystrokes in the region, but changing it, or
even moving the mouse, causes the display to correct itself.
I can't get it in 21.5, because the font setup is different, and I
can't remember how to make 21.5 use the same fonts as I use in 21.4.
XEmacs-Beta mailing list