HI, Eric

You are right, I remove the usage of advice now, and use the method you (nearly) gave, only 1 little change:
I found the code:
----
(defun t1 () (message "abc"))
;;(symbol-function 't1)

(let ((hold #'t1))
  (defun t1 () (message "def"))
  (setq t1 hold))
;;(symbol-function 't1)
----

did recover the t1 function after it executed, so I use this way:
----
(defun t1 () (message "abc"))
;;(symbol-function 't1)

(let ((hold (symbol-function 't1)))
  (defun t1 () (message "def"))
  (fset 't1 hold))

;;(symbol-function 't1)
----

And the new patch is attached.

BTW: I received a PDF assignment form from FSF, but the developer name and the target program on it were wrong (It's for another person who contributes to GCC, I think), so I reply that mail for a new PDF assignment form, I 'll tell you after these things done.

Thanks.


On Sat, Apr 12, 2014 at 3:18 AM, Eric Schulte <schulte.eric@gmail.com> wrote:
Hmmm,

Not to be overly nitpicky here, but I see two issues.

1. You should use unwind-protect, to ensure that (ad-unadvise #'message)
   is run even if @body throws an error, and

2. This will remove any advise which the user has placed on #'message.

How about something shaped like the following.

    (defmacro with-weird-message (&rest body)
      `(let ((hold #'message)
             current-message)
         (unwind-protect
             (progn
               (defun message (&rest args)
                 (setq current-message (apply #'format args)))
               ,@body
               current-message)
           (setq message hold))))

Best,

P.S. I know this is a lot of process for a small patch, but from this
     point forward once you have the FSF assignment you can much more
     easily contribute to ob-scheme and org in general

KDr2 <killy.draw@gmail.com> writes:

> Hi, Eric
>
> I'm sorry for that I used `flet' in the patch, It's a easy way to let
> function `current-message' work in batch mode, so I used it even I saw that
> emacs says `flet' is obsolete, I'm sorry for that.
>
> And I made a new patch(attachment) using `defadvice' for `message' to
> capture the message in batch mode, after the message being captured, the
> advice function is removed. Is this way OK?
>
> And I also sent a request email to assign@gnu.org, and now waiting the
> reply.
>
> Thanks.
>
>
> On Fri, Apr 11, 2014 at 10:45 AM, Eric Schulte <schulte.eric@gmail.com>wrote:
>
>> We can no longer use `flet' in the Org-mode code base, please re-work
>> this patch w/o flet.
>>
>> Also, I don't see your name in the list of contributors, and (I believe)
>> this patch is too large to apply w/o FSF assignment.  See the following
>> page on how to contribute to Org-mode.
>>
>>   http://orgmode.org/worg/org-contribute.html
>>
>> KDr2 <killy.draw@gmail.com> writes:
>>
>> > The bug:
>> > write file ~/scheme-test.org with the content below:
>> > -------8<--------------
>> > #+BEGIN_SRC scheme :exports results :results output raw
>> >   (display "Hello Scheme in OrgMode")
>> > #+END_SRC
>> > -------8<--------------
>> >
>> > and run:
>> >
>> > emacs --batch --eval='(load "~/.emacs.d/init.el")' ~/scheme-test.org -f
>> > org-html-export-to-html
>> >
>> > you will find the bug:
>> >
>> > `org-babel-scheme-execute-with-geiser' uses `current-message' to get the
>> > results of scheme code blocks, but `current-message' always returns nil
>> in
>> > batch mode, and this patch fixes this.
>> >
>> > --
>>
>> --
>> Eric Schulte
>> https://cs.unm.edu/~eschulte
>> PGP: 0x614CA05D
>>
>
>
>
> --
> --
>
> KDr2, http://kdr2.com
>
> From fe5549f3f48acf9b51aeb3706eb8dd3d76ab18c1 Mon Sep 17 00:00:00 2001
> From: KDr2 <killy.draw@gmail.com>
> Date: Fri, 11 Apr 2014 12:56:24 +0800
> Subject: [PATCH] lisp/ob-scheme.el: Fix scheme code blocks execution error in
>  batch mode
>
> * lisp/ob-scheme.el (org-babel-scheme-capture-current-message, org-babel-scheme-execute-with-geiser): Capture scheme code results via current-message both in interactive mode and noninteractive mode.
>
> `org-babel-scheme-execute-with-geiser' uses `current-message' to get the results of scheme code blocks, but `current-message' always returns nil in batch mode, and this patch fixes this.
>
> Modified from a patch proposal by KDr2(killy.draw@gmail.com)
> ---
>  lisp/ob-scheme.el | 20 +++++++++++++++++---
>  1 file changed, 17 insertions(+), 3 deletions(-)
>
> diff --git a/lisp/ob-scheme.el b/lisp/ob-scheme.el
> index b7117e9..6b82c6e 100644
> --- a/lisp/ob-scheme.el
> +++ b/lisp/ob-scheme.el
> @@ -118,6 +118,19 @@ org-babel-scheme-execute-with-geiser will use a temporary session."
>              (name))))
>      result))
>
> +(defmacro org-babel-scheme-capture-current-message (&rest body)
> +  "Capture current message in both interactive and noninteractive mode"
> +  `(if noninteractive
> +       (let ((current-message nil))
> +         (defadvice message (after capture-current-message activate)
> +           (setq current-message ad-return-value))
> +         ,@body
> +         (ad-unadvise #'message)
> +         current-message)
> +     (progn
> +       ,@body
> +       (current-message))))
> +
>  (defun org-babel-scheme-execute-with-geiser (code output impl repl)
>    "Execute code in specified REPL. If the REPL doesn't exist, create it
>  using the given scheme implementation.
> @@ -142,10 +155,11 @@ is true; otherwise returns the last value."
>                            (current-buffer)))))
>       (setq geiser-repl--repl repl-buffer)
>       (setq geiser-impl--implementation nil)
> -     (geiser-eval-region (point-min) (point-max))
> +     (setq result (org-babel-scheme-capture-current-message
> +                   (geiser-eval-region (point-min) (point-max))))
>       (setq result
> -           (if (equal (substring (current-message) 0 3) "=> ")
> -               (replace-regexp-in-string "^=> " "" (current-message))
> +           (if (and (stringp result) (equal (substring result 0 3) "=> "))
> +               (replace-regexp-in-string "^=> " "" result)
>               "\"An error occurred.\""))
>       (when (not repl)
>         (save-current-buffer (set-buffer repl-buffer)

--
Eric Schulte
https://cs.unm.edu/~eschulte
PGP: 0x614CA05D



--
-- 

KDr2, http://kdr2.com