Hello.
This is similar but not the same as 11052.
This is reproducable on Gtk and Lucid so I think it is a general redisplay problem. It is harder to reproduce on those platforms, as they seem to redraw more often. It may also be a timing issue.
But I found a simpler way to reproduce the bug:
./emacs -Q
M-x scroll-bar-mode (to turn off scrollbars).
C-x 3
M-x linum-mode
Then double-click on a line to mark a word on that line. The redisplay engine will then redraw the whole line, but the black border between windows will be gone from that line and to the end of the window (why Emacs redisplays from the marked word to the end of the window I don't know, it does not seem efficient).
Screenshots attached, taken from a lucid build. If you look at the zoomed in version, you can see that if you extend the one-pixel black line down to the "4", it would actually draw over pixels belonging to the "4", thus indicating that this bug is due to the display engine redrawing line numbers over the black line.
This recepie also shows that there are cursor redrawing bugs in the X port, remnants show up in the line number column. But that is a separate issue :-).
Jan D.
3 sep 2012 kl. 12:38 skrev Robert Dallas Gray <mail@robertdallasgray.com>:
On 31 Aug 2012, at 10:05, Eli Zaretskii wrote:
From: Robert Dallas Gray <mail@robertdallasgray.com>
Date: Mon, 27 Aug 2012 09:37:14 +0100
Bearing in mind this affects Cocoa Emacs on OS X, in windowed mode:
cd /Applications/Emacs.app/Contents/MacOS
./Emacs -Q
M-x scroll-bar-mode (to turn scroll bars off)
(now visit a file with contents large enough to require scrolling)
C-x 3 (to split the window vertically)
M-x linum-mode
Now scroll the right-hand window (using the mouse wheel seems to reproduce the glitch more often). The black vertical bar between the two windows intermittently disappears or becomes 'broken'.
When the vertical line separating the two windows disappears or
becomes broken, does it get corrected if you wait for a while, or does
it stay corrupted forever, until the next scroll?
It stays corrupted until the next scroll (or M-x redraw-display, or a window resize).
FWIW, all I see on my MS-Windows machine is slow scrolling (due to
linum-mode), but no redisplay problems.
I wasn't able to reproduce it in Emacs 24.1 on Windows 7.
I have a screenshot of the bug occurring, but I'm not sure on the rules for posting images here.
Please do post the screenshots.
Attached.
<Screen Shot 2012-08-26 at 11.12.22.png>