Just following up on my own report --
I rebuilt xemacs with error-checking enabled (the default for beta
releases) and the problem disappeared, just as reported by others.
Eric Benson wrote:
I'm getting this error on a variety of commands in 21.5.14, built
on
Mac OS X with X11 support. This has been reported recently by a few
other people on other systems:
--------------------------------
http://list-archive.xemacs.org/xemacs-beta/200305/msg00354.html
From: "Schmitt, Christian (ext.)" <Christian.Schmitt(a)dresdner-bank.com>
Symbol's value as variable is void: region
Hi,
I (finally) came around last night and built a new binary on Windows
for the latest XEmacs code from CVS.
I built two versions:
One with error checking and debug enabled and one without (see below).
The version with debug and error checking enabled does not show
the "Symbol's value as variable is void: region" error but the one
without debugging enabled does.
Here's a stacktrace that happens when I try to select a piece
of text using SHIFT-left arrow:
# bind (standard-output stack-trace-on-signal debug-on-signal
stack-trace-on-error debug-on-error)
# (unwind-protect ...)
# bind (buffer valid start end n)
zmacs-make-extent-for-region((#<marker at 424 in *scratch* 0xe69e54> .
#<marker at 424 in *scratch* 0xe69e3c>))
zmacs-activate-region()
# bind (buffer activate-region nomsg location)
push-mark(nil nil t)
--------------------------------
http://list-archive.xemacs.org/xemacs-nt/200305/msg00031.html
From: Rick Rankin <rick_rankin(a)yahoo.com>
(Add'l info) Build problem and problem with region in 21.5 under Cygwin.
Here's some additional info. I set debug-on-error to true and got the
following
backtrace when I tried to set the mark:
Signaling: (void-variable region)
zmacs-make-extent-for-region((#<marker at 1 in .emacs 0x1028cd44> .
#<marker at 1 in .emacs 0x1028cd2c>))
zmacs-activate-region()
push-mark(nil nil t)
set-mark-command(nil)
call-interactively(set-mark-command)
--------------------------------
http://list-archive.xemacs.org/xemacs-nt/200305/msg00034.html
From: Markus Alt <altmark(a)de.ibm.com>
I might also have corrupt sources, but I doubt it (as a build for Linux
from the same source tree works fine). In any case I have the very same
error with 21.5.13 on AIX 5.1. Here's my backtrace (similar to Rick's):
Signaling: (void-variable region)
zmacs-make-extent-for-region((#<marker at 387 in init.el 0x2039640c> .
#<marker at 387 in init.el 0x20396424>))
zmacs-activate-region()
push-mark(nil nil t)
set-mark-command(nil)
--------------------------------
I also tried setting debug-on-error to t and got the same backtrace as
others.
Can I assume this was already fixed on those other platforms, but the
fix did not make it to Darwin? Can someone fill me in on the fix, so I
can apply it to the Darwin version as well?
By the way, this was built on Mac OS X 10.2.6, with X11 from
www.osxgnu.org. I configured with --error-checking=none. The error
happens when running in X11 or TTY mode.