Sam Steingold <sds(a)goems.com> writes:
I did not set `debug-ignored-errors', so it was nil.
what I was observing: (with debug-ignored-errors = nil; debug-on-error =
t):
Wierd indeed. Do you ever get a backtrace? i.e. with other errors? I
noticed that if XEmacs cannot find the xemacs-base package where
debug.el is then it just does nothing when a backtrace should happened.
>> that, but I was overruled. A sample setting for
debug-ignored-errors
what was the rationale?
Running with debug-on-error set to t permanently is an expert thing.
Experts can be expect to know how to set the variable themselves. For
normal users you want to be able to tell them
"If you get an error, set debug-on-error and mail me the backtrace".
You don't want to have "Except when the error is in
debug-ignored-errors then you'll have to set that for."
And yes it happens that end-of-buffer is a reall error in package.
What I see there is what I get for free in Emacs.
I guess I'll just remove the comments and re-dump.
That's what the GPL is for, right? :-)
More conventional is to set it in you site-start.el or even .emacs.
However as you said you can do what you like.
Signaling: (void-variable last-buffer-undo-list)
(eq last-buffer-undo-list nil)
)
(or (eq buffer-undo-list t) (eq last-buffer-undo-list nil) (not (or ... ...)) (or (eq
buffer-undo-list pending-undo-list) (eq ... pending-undo-list)))
This seems to be from the code in menubar-items.el
["Redo" redo
:included (fboundp 'redo)
:active (not (or (eq buffer-undo-list t)
(eq last-buffer-undo-list nil)
(not (or (eq last-buffer-undo-list buffer-undo-list)
(and (null (car-safe buffer-undo-list))
(eq last-buffer-undo-list
(cdr-safe buffer-undo-list)))))
(or (eq buffer-undo-list pending-undo-list)
(eq (cdr buffer-undo-list) pending-undo-list))))
:suffix (if (eq last-command 'redo) "More" "")]
However that code should never be run unless (fboundp 'redo) is
non-nil. Do you have something in your .emacs that defines a function
redo but does not load redo.el?
Jan