debug-on-call

debug-on-call

[Dan Jacobson]

I wanted to have emacs stop if it ever called a function, but no:
debug-on-entry: Function base64-encode-region is a primitive

I don't need to debug inside that function. All I wanted was to see a
backtrace so I can tell the layers of stuff calling that function.

So provide a debug-on-call.

And, inside the Emacs Info manual, all we can find about debugging is
about debugging non emacs lisp.  If we want to learn about the emacs
debugger then we have to know to leave the emacs manual and search in
a neighboring manual not mentioned via any searching in the emacs
manual.

BTW
debug-on-entry-1: gnus-uu-post-news not user-defined Lisp function
same deal. not too many things i can debug these days. i know. no more
debugging for me. i will post my problems instead.
-- 
http://jidanni.org/ Taiwan(04)25854780

Re: debug-on-call

[Kevin Rodgers]

Dan Jacobson wrote:

> I wanted to have emacs stop if it ever called a function, but no:
> debug-on-entry: Function base64-encode-region is a primitive
> 
> I don't need to debug inside that function. All I wanted was to see a
> backtrace so I can tell the layers of stuff calling that function.

(defadvice base64-encode-region (before debug activate)
   "Debug (when called from an Emacs Lisp function)."
   (debug))

-- 
<a href="mailto:&lt;kevin.rodgers&#64;ihs.com&gt;">Kevin Rodgers</a>

Re: debug-on-call

[Dan Jacobson]

>>>>> "K" == Kevin Rodgers <kevin.rodgers@ihs.com> writes:

K> Dan Jacobson wrote:
>> I wanted to have emacs stop if it ever called a function, but no:
>> debug-on-entry: Function base64-encode-region is a primitive
>> 
>> I don't need to debug inside that function. All I wanted was to see a
>> backtrace so I can tell the layers of stuff calling that function.

K> (defadvice base64-encode-region (before debug activate)
K>    "Debug (when called from an Emacs Lisp function)."
K>    (debug))

OK, maybe this could be the guts of the new debug-on-call function, or
at least documented as what to do when debug-on-entry isn't usable...
Of course an antidote should also be provided to turn it off.

P.S. Kev, your References:
<mailman.1280.1044050572.21513.bug-gnu-emacs@gnu.org> does not give my
original post the fame and glory of its original Message-ID causing
its thread to be broken 'in many newsreaders'.  Did you try reading
this group via NNTP of gmane.org for maximal comfort?
-- 
http://jidanni.org/ Taiwan(04)25854780

Re: debug-on-call

[Richard Stallman]

    > I wanted to have emacs stop if it ever called a function, but no:
    > debug-on-entry: Function base64-encode-region is a primitive

It worked for me when I tried it.

Re: debug-on-call

[Alexander Pohoyda]

>     > I wanted to have emacs stop if it ever called a function, but no:
>     > debug-on-entry: Function base64-encode-region is a primitive
> It worked for me when I tried it.

Worked for me as well.


-- 
Alexander Pohoyda
<alexander.pohoyda@gmx.net>

Re: debug-on-call

[Kevin Rodgers]

[-- Attachment #1: Type: text/plain, Size: 1883 bytes --]

Dan Jacobson wrote:

>>>>>>"K" == Kevin Rodgers <kevin.rodgers@ihs.com> writes:
>>>>>>
> 
> K> Dan Jacobson wrote:
> 
>>>I wanted to have emacs stop if it ever called a function, but no:
>>>debug-on-entry: Function base64-encode-region is a primitive
>>>
>>>I don't need to debug inside that function. All I wanted was to see a
>>>backtrace so I can tell the layers of stuff calling that function.
>>>
> 
> K> (defadvice base64-encode-region (before debug activate)
> K>    "Debug (when called from an Emacs Lisp function)."
> K>    (debug))
> 
> OK, maybe this could be the guts of the new debug-on-call function, or
> at least documented as what to do when debug-on-entry isn't usable...


Please find attached a tentative user (i.e. non-integrated) implementation.

> Of course an antidote should also be provided to turn it off.


Yep, but I haven't gotten that far yet.


> P.S. Kev, your References:
> <mailman.1280.1044050572.21513.bug-gnu-emacs@gnu.org> does not give my
> original post the fame and glory of its original Message-ID causing
> its thread to be broken 'in many newsreaders'.


That's odd, google finds it under that Message-ID.  I used

	http://www.google.com/advanced_group_search?hl=en

to find

	From: Dan Jacobson <jidanni@dman.ddts.net>
	Newsgroups: gnu.emacs.bug
	Subject: debug-on-call
	Date: 01 Feb 2003 03:03:22 +0800
	Lines: 23
	Sender: news <news@main.gmane.org>
	Approved: bug-gnu-emacs@gnu.org
	Message-ID: <mailman.1280.1044050572.21513.bug-gnu-emacs@gnu.org>
	Reply-To: Dan Jacobson <jidanni@dman.ddts.net>
	NNTP-Posting-Host: monty-python.gnu.org

> Did you try reading
> this group via NNTP of gmane.org for maximal comfort?
 
I primarily read gnu.emacs.bug and only check gmane.emacs.bug (both via

NNTPSERVER=news.cis.dfn.de) to make sure I don't miss any articles.


-- 
<a href="mailto:&lt;kevin.rodgers&#64;ihs.com&gt;">Kevin Rodgers</a>

[-- Attachment #2: debug-primitive-on-entry.el --]
[-- Type: text/plain, Size: 777 bytes --]

(defadvice debug-on-entry (around advise-primitive activate)
  "If an error is signalled because FUNCTION is a primitive, query the user
whether to debug Lisp calls via `defadvice'."
  (condition-case condition-data
      ad-do-it
    (error (let ((error-message (format "Function %s is a primitive"
					(ad-get-arg 0))))
	     (if (and (equal (car (cdr condition-data)) error-message)
		      (interactive-p)
		      (yes-or-no-p
		       (concat error-message
			       "; debug Lisp calls via defadvice? ")))
		 (eval `(defadvice ,(ad-get-arg 0) (before debug activate)
			  "Debug (when called from an Emacs Lisp function)."
			  (debug)))
	       (signal 'error (cdr condition-data)))))))

;; (defadvice cancel-debug-on-entry (around advise-primitive activate)
;;   ...)

[-- Attachment #3: Type: text/plain, Size: 148 bytes --]

_______________________________________________
Bug-gnu-emacs mailing list
Bug-gnu-emacs@gnu.org
http://mail.gnu.org/mailman/listinfo/bug-gnu-emacs