Stephen J. Turnbull writes:
> I notice this problem in xemacs 21.4 patch 22 running in linux
> mode. It does not happen in xemacs 21.4 patch 8 (also linux, nw).
> I mark a region and then try some command like untabify or sort lines
> and get a complaint that the region is not active.
First, tell us, are these built by you from source, or are they
packaged from a Linux distribution?
This machine runs CentOS release 5.4
According to my notes there was no yum package available and I ended
up doing this:
gzip -dc xemacs-21.4.22.tar.gz | tar xf -
mv xemacs-mule-sumo.tar.gz /root/
tar xf /root/xemacs-mule-sumo.tar
tar xf /root/xemacs-all-packages.tar
I started a new terminal, ran xemacs -nw again and did not see the
same behavior. So I must have done something to put it into this
strange mode. But I can't figure out what.
The interesting thing I saw was that after marking the region I would
see the region highlighted, and then I'd type m-x and after typing
one more character, the s for sort lines or the u for untabify, the
highlighting would disappear. Is that a useful hint?
As it turns out, this is the first time I've seen this particular
problem, but there's another one that I've seen a lot and I'd be even
more interested in finding out how to fix. This one happens on many
different machines. Again I don't know how I get into this mode.
I'm in a dired buffer.
I type f to go to a buffer.
I see in the minibuffer "Failing I-search backward: f".
I type c-g and get a backtrace buffer:
Debugger entered--Lisp error: (wrong-type-argument integer-or-marker-p nil)
#<compiled-function nil "...(33)" [isearch-invalid-regexp isearch-opoint
As long as I leave that backtrace buffer I can then go back to the
dired and type f and get to the buffer. When I type q to get out of
the backtrace I'm back in the same state.
Also in that state I seem to be in the middle of a search, so that
when I type c-s it's as if I've already typed a bunch of text.
How do you mark the region?
I only know one way to do it, which
is c-space and then move.
This is in an xterm where I've run ssh to talk to another machine and
inside that ssh I've done xemacs -nw.
Given that simply setting and resetting zmacs-regions seems to help,
What should/could be different between the original state and the
result of setting and resetting zmacs-regions?
that's our first guess, but XEmacs should not surprise you this
If it's a side effect of something you did, that should either be
better documented or perhaps the side effect can be eliminated.
> I wonder what. And how I can tell.
If you notice an error that might be a bug, immediately doing C-h l
will give you your most recent hundred or so keystrokes as well as the
contents of the " *Message Log*" buffer. That will tell you what
you've done recently.
I've actually increased the number stored to 10000 in
most of my
xemacs's. That affects (recent-keys ) though help - l still only
XEmacs-Beta mailing list