XEmacs does abort, but only in X. I agree that X server crashing is not an XEmacs problem, but the abort may be something to be looked-at.
In the past year, I've used 21.5.18-23, and have regularly done this <cr> removal from files, but only have seen this recently. Hmm. "Recently" for me means I used GCC 4 to compile XEmacs, so that, too, may be something to consider. (The libraries and the rest of the system is also compiled with GCC 4, with Mandrake 2006, come to think of it.)
Clay Haapala, Symantec Corp. Roseville, MN 651-746-7232
From: Aidan Kehoe [mailto:kehoea＠parhasard.net]
Sent: Thursday, November 03, 2005 4:38 PM
To: Clay Haapala
Cc: XEmacs Beta
Subject: Re: [Bug: 21.5-b23] 2.5.23 core with replace-regexp in X
Ar an triú lá de mí na Samhain, scríobh Clay Haapala:
> [...] To reproduce this:
> 1) Must be in X (the bug does not manifest in a console 'xemacs -nw'
> session, and I haven't tried it under Windows)
> 2) Meta-x replace-regexp<ret>^^Q^M
> And that's about as far as I get. Replace all carriage-returns with
> nothing is my intent. Either the control-Q or the control-M does it.
> I know the athena=3d config option does not matter. The --with-gtk
I've seen this in most beta versions I've used in the last couple of years,
without GTK. Since this is an X server bug--nothing XEmacs draws to the
screen should cause the X server to crash--here are the details of mine;