Jan Vroonhof <vroonhof(a)math.ethz.ch&gt; writes: Very interesting. "xwininfo -tree" informs me... xwininfo: Window id: 0x4c00029 "emacs: *Summary comp.emacs.xemacs*" Root window id: 0x26 (the root window) (has no name) Parent window id: 0x1808752 (has no name) 1 child: 0x4c0002a (has no name): () 1267x976+0+0 +5+21 10 children: 0x4c002db (has no name): () 1244x15+4+933 +9+954 0x4c002a2 (has no name): () 15x866+1248+67 +1253+88 0x4c002b1 (has no name): () 15x206+1248+67 +1253+88 0x4c002d9 (has no name): () 481x15+4+258 +9+279 0x4c0015f (has no name): () 15x191+485+67 +490+88 0x4c0012b (has no name): () 15x661+485+287 +490+308 0x4c0014b (has no name): () 481x15+4+258 +9+279 0x4c002b3 (has no name): () 1244x15+4+258 +9+279 0x4c0002c (has no name): () 1267x950+0+26 +5+47 0x4c0002b (has no name): () 1265x24+0+0 +5+21 And gdb reveals... (gdb) p/x x_event->xany.window $49 = 0x4c002d9 ...being the 4th child down in the list. BTW... (gdb) p x_event->xkey.keycode $50 = 46 ...which is an `l', the 2nd of the keystrokes I'd hit when XEmacs tripped over the SEGV. Hm, given the window name cited by xwininfo, I guess my recollection about Gnus having cleaned itself out to just the *Group* buffer was incorrect -- it hadn't gotten quite that far yet. If you need to see how my Gnus-involved XEmacs usually looks (if it matters to understanding what all the child windows are), take a peek at http://www.cs.cmu.edu/~karl/gnus/pterodactyl/sample.gif. --karl