Symbol's function definition is void: <

Symbol's function definition is void: <

[Lennart Borgman]

I have got a strange bug report for nXhtml. The report says that < is
not defined. Please see

    http://www.emacswiki.org/emacs/NxhtmlMode#toc13

What functions could make < void? fset?

Re: Symbol's function definition is void: <

[Davis Herring]

> I have got a strange bug report for nXhtml. The report says that < is
> not defined. Please see
>
>     http://www.emacswiki.org/emacs/NxhtmlMode#toc13
>
> What functions could make < void? fset?

fmakunbound is it AFAIK (I replied there too).

Davis

-- 
This product is sold by volume, not by mass.  If it appears too dense or
too sparse, it is because mass-energy conversion has occurred during
shipping.

Re: Symbol's function definition is void: <

[Lennart Borgman]

On Thu, Jun 25, 2009 at 5:13 PM, Davis Herring<herring@lanl.gov> wrote:
>> I have got a strange bug report for nXhtml. The report says that < is
>> not defined. Please see
>>
>>     http://www.emacswiki.org/emacs/NxhtmlMode#toc13
>>
>> What functions could make < void? fset?
>
> fmakunbound is it AFAIK (I replied there too).

Thanks Davis. So something like this would help

  (defadvice fmakunbound (before
                          test-ad-makunbound
                          activate
                          compile
                          )
    (message "fmakunbound %s, backtrace\n%s"
             (ad-get-arg 0) (with-output-to-string (backtrace))))

Re: Symbol's function definition is void: <

[Stefan Monnier]

> I have got a strange bug report for nXhtml. The report says that < is
> not defined. Please see

>     http://www.emacswiki.org/emacs/NxhtmlMode#toc13

> What functions could make < void? fset?

It might also just be an uninterned symbol, so it has the same name as
the "strictly greater" but doesn't have any function associated with it.


        Stefan

Re: Symbol's function definition is void: <

[Lennart Borgman]

On Thu, Jun 25, 2009 at 10:59 PM, Stefan
Monnier<monnier@iro.umontreal.ca> wrote:
>> I have got a strange bug report for nXhtml. The report says that < is
>> not defined. Please see
>
>>     http://www.emacswiki.org/emacs/NxhtmlMode#toc13
>
>> What functions could make < void? fset?
>
> It might also just be an uninterned symbol, so it has the same name as
> the "strictly greater" but doesn't have any function associated with it.


Thanks. What do we need to do to catch that possibility? I suggested a
defadvice on fmakunbound (which the user reporting the problem now is
using). Maybe defadvice on make-symbol and unintern could catch this
possibility? Some more function to watch?

Re: Symbol's function definition is void: <

[Thien-Thi Nguyen]

() Lennart Borgman <lennart.borgman@gmail.com>
() Fri, 26 Jun 2009 01:11:25 +0200

   Maybe defadvice on make-symbol and unintern could catch this
   possibility? Some more function to watch?

How about enabling `debug-on-error' then adding to `post-command-hook':

  (lambda () 
    (unless (boundp '<) 
      (error "unbound <")))

thi

Re: Symbol's function definition is void: <

[Andreas Roehler]

Lennart Borgman wrote:
> I have got a strange bug report for nXhtml. The report says that < is
> not defined. Please see
> 
>     http://www.emacswiki.org/emacs/NxhtmlMode#toc13
> 
> What functions could make < void? fset?
> 
> 
> 

Encountering that error from time to time, reason here is regulary:

a crashed beginning-of-buffer command --wrongly typed-- left a "<", which was saved
 when closing Emacs.

If I start Emacs again, that error is signaled.

So sorry, should I'm be the only stupid one doing such things.

But, earnestly, would expect just this reason for the OP.

Cheers

Andreas

Re: Symbol's function definition is void: <

[Alan Mackenzie]

On Fri, Jun 26, 2009 at 01:10:01PM +0200, Thien-Thi Nguyen wrote:
> () Lennart Borgman <lennart.borgman@gmail.com>
> () Fri, 26 Jun 2009 01:11:25 +0200

>    Maybe defadvice on make-symbol and unintern could catch this
>    possibility? Some more function to watch?

> How about enabling `debug-on-error' then adding to `post-command-hook':

>   (lambda () 
>     (unless (boundp '<) 
>       (error "unbound <")))

fboundp  :-)

But thanks for this tip.  It will surely come in handy.

> thi

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: Symbol's function definition is void: <

[Lennart Borgman]

On Fri, Jun 26, 2009 at 1:10 PM, Thien-Thi Nguyen<ttn@gnuvola.org> wrote:
> () Lennart Borgman <lennart.borgman@gmail.com>
> () Fri, 26 Jun 2009 01:11:25 +0200
>
>   Maybe defadvice on make-symbol and unintern could catch this
>   possibility? Some more function to watch?
>
> How about enabling `debug-on-error' then adding to `post-command-hook':
>
>  (lambda ()
>    (unless (boundp '<)
>      (error "unbound <")))

Yes, that is a good idea if combined with the backtrace from the
defadviced functions.

Re: Symbol's function definition is void: <

[Lennart Borgman]

On Fri, Jun 26, 2009 at 2:10 PM, Andreas
Roehler<andreas.roehler@online.de> wrote:
> Lennart Borgman wrote:
>> I have got a strange bug report for nXhtml. The report says that < is
>> not defined. Please see
>>
>>     http://www.emacswiki.org/emacs/NxhtmlMode#toc13
>>
>> What functions could make < void? fset?
>>
>>
>>
>
> Encountering that error from time to time, reason here is regulary:
>
> a crashed beginning-of-buffer command --wrongly typed-- left a "<", which was saved
>  when closing Emacs.
>
> If I start Emacs again, that error is signaled.
>
> So sorry, should I'm be the only stupid one doing such things.
>
> But, earnestly, would expect just this reason for the OP.


Andreas, could you please explain what you did? I do not understand.

Re: Symbol's function definition is void: <

[Andreas Roehler]

Lennart Borgman wrote:
> On Fri, Jun 26, 2009 at 2:10 PM, Andreas
> Roehler<andreas.roehler@online.de> wrote:
>> Lennart Borgman wrote:
>>> I have got a strange bug report for nXhtml. The report says that < is
>>> not defined. Please see
>>>
>>>     http://www.emacswiki.org/emacs/NxhtmlMode#toc13
>>>
>>> What functions could make < void? fset?
>>>
>>>
>>>
>> Encountering that error from time to time, reason here is regulary:
>>
>> a crashed beginning-of-buffer command --wrongly typed-- left a "<", which was saved
>>  when closing Emacs.
>>
>> If I start Emacs again, that error is signaled.
>>
>> So sorry, should I'm be the only stupid one doing such things.
>>
>> But, earnestly, would expect just this reason for the OP.
> 
> 
> Andreas, could you please explain what you did? I do not understand.
> 
> 
> 


Hi Lennart,

seems I mistook something.

When starting Emacs, occassionaly I get that warning:

,----
| Warning (initialization): An error occurred while loading `MY-PATH/.emacs':
|
| Symbol's value as variable is void: <
`----

Thats caused by a "<" wrongly inserted into some
lisp-file from failed executiong of keys
"\M-<". (beginning-of-buffer).

But seems not your problem, as it is no function mentioned.

Sorry for the noise.

Andreas

Re: Symbol's function definition is void: <

[Davis Herring]

>> I have got a strange bug report for nXhtml. The report says that < is
>> not defined. Please see
>
>>     http://www.emacswiki.org/emacs/NxhtmlMode#toc13
>
>> What functions could make < void? fset?
>
> It might also just be an uninterned symbol, so it has the same name as
> the "strictly greater" but doesn't have any function associated with it.

Then it would have to be called with `funcall' or `eval', which seems
unlikely.

I suppose it could also be the case that the "real" '< got uninterned, and
then any Lisp read after that would generate a new '< that had no
function.  (Any Lisp already read would still use the old symbol, of
course, even through `eval' and such.)

Lennart, you can use functions like these to test for these cases:

;; Non-nil if sym is in the obarray OB (or `obarray').
(defun internp (sym &optional ob) (if sym (intern-soft sym ob) t))
;; Non-nil if sym is the '< that was current when this function was read.
(defun <p (sym) (eq sym '<))
;; Non-nil if sym is the current '<.
(defun is< (sym) (eq sym (intern "<")))
;; Non-nil if '< when called is not the same as when read.
(defun <-changed () (not (is< '<)))

Davis

-- 
This product is sold by volume, not by mass.  If it appears too dense or
too sparse, it is because mass-energy conversion has occurred during
shipping.