message-newline-and-reformat and supercite

message-newline-and-reformat and supercite

[Michael R Cook]

If I do M-RET (message-newline-and-reformat) in a Supercite context,
such as:

 +-----
 | FOO> hello_world
 +-----

(the underscore `_' represents point), the result is

 +-----
 | FOO> hello
 |
 |_
 |
 | world
 +-----

I was expecting

 +-----
 | FOO> hello
 |
 |_
 |
 | FOO>  world
 +-----

Is this a bug in message.el?
Or is my supercite installation misconfigured?

Michael.

Re: message-newline-and-reformat and supercite

[Lars Magne Ingebrigtsen]

Michael R Cook <mcook@cognex.com> writes:

> If I do M-RET (message-newline-and-reformat) in a Supercite context,
> such as:

That command doesn't know anything about Supercite, so it won't do
anything useful there...

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: message-newline-and-reformat and supercite

[Wes Hardaker]

>>>>> On 04 Jan 1998 10:29:46 +0100, Lars Magne Ingebrigtsen <larsi@gnus.org> said:

Cook> If I do M-RET (message-newline-and-reformat) in a Supercite
Cook> context, such as:

Lars> That command doesn't know anything about Supercite, so it won't
Lars> do anything useful there...

I think the point was to change the regex used to do the inserting of
">"s to something that would recognize text in front of the '>'?  IE,
replace message-yank-prefix with a regexp.  That way it would catch
">> " as well.  A bit more complex of a function, but worth it.

-- 
"Ninjas aren't dangerous.  They're more afraid of you than you are of them."

Re: message-newline-and-reformat and supercite

[Lars Magne Ingebrigtsen]

Wes Hardaker <wjhardaker@ucdavis.edu> writes:

> I think the point was to change the regex used to do the inserting of
> ">"s to something that would recognize text in front of the '>'?  IE,
> replace message-yank-prefix with a regexp.  That way it would catch
> ">> " as well.  A bit more complex of a function, but worth it.

Yes, but I think this should be up to Supercite, if possible.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: message-newline-and-reformat and supercite

[Kai Grossjohann]

>>>>> Wes Hardaker <wjhardaker@ucdavis.edu> writes:
  >> I think the point was to change the regex used to do the inserting of
  >> ">"s to something that would recognize text in front of the '>'?  IE,
  >> replace message-yank-prefix with a regexp.  That way it would catch
  >> ">> " as well.  A bit more complex of a function, but worth it.

>>>>> On 05 Jan 1998, Lars Magne Ingebrigtsen said:
  Lars> Yes, but I think this should be up to Supercite, if possible.

Well, Supercite has C-c C-p o already.  It is more like C-o than RET,
but it comes close...

Do we need a hook for doing an intelligent newline so that SC can
provide a function and message-mode can provide a function and the
user can decide which one to use?

kai
-- 
Kai Grossjohann, Informatik VI        grossjohann@ls6.cs.uni-dortmund.de
Uni Dortmund, D-44221 Dortmund        http://ls6-www.cs.uni-dortmund.de/
                                      Vox +49 231 755 5670, Fax -2405
I like both kinds of music.

Re: message-newline-and-reformat and supercite

[Lars Magne Ingebrigtsen]

Kai Grossjohann <grossjohann@charly.cs.uni-dortmund.de> writes:

> >>>>> On 05 Jan 1998, Lars Magne Ingebrigtsen said:
>   Lars> Yes, but I think this should be up to Supercite, if possible.
> 
> Well, Supercite has C-c C-p o already.  It is more like C-o than RET,
> but it comes close...
> 
> Do we need a hook for doing an intelligent newline so that SC can
> provide a function and message-mode can provide a function and the
> user can decide which one to use?

I think Supercite should just bind `M-RET' to its intelligent
newline. 

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: message-newline-and-reformat and supercite

[Lars Magne Ingebrigtsen]

Francisco Solsona <flsc@hp.fciencias.unam.mx> writes:

>         I don't use supercite either, though I found M-RET useful, so
> what about this modified version (based on Vladimir's):

I've taken bits of your and Vladimirs versions and rolled a slightly
different version.  It's in the CVS.

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen

Re: message-newline-and-reformat and supercite

[Francisco Solsona]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

> Vladimir Volovich <vvv@vvv.vsu.ru> writes:
> 
> >       (if (looking-at (sc-cite-regexp))
> 
> One potential problem with this is that it depends on Supercite.
> Supercite already offers C-c C-p o which I used to use a lot when I
> used Supercite.

        I don't use supercite either, though I found M-RET useful, so
what about this modified version (based on Vladimir's):

,--------------------
| (defun message-newline-and-reformat (&optional no-fill)
|   "Insert four newlines, and then reformat if inside quoted text. If
| called with a prefix argument it avoids filling the paragraph."
|   (interactive "P")
|   (let ((point (point))
|         (gnus-prefix "[]>»|:}+ \t]*") ;; gnus-cite-prefix-regexp is more than we need.
|         (gnus-supercite-thing "[-._a-zA-Z0-9]*[>]+[ \t]*")
|         quoted)
|     (save-excursion
|       (beginning-of-line)
|       (if (looking-at (concat gnus-prefix
|                               gnus-supercite-thing)) ;; (sc-cite-regexp)
|           (setq quoted (buffer-substring (match-beginning 0) (match-end 0)))))
|     (insert "\n\n\n\n")
|     (when (null no-fill) ;; No filling.
|       (delete-region (point) (re-search-forward "[ \t]*")))
|     (when quoted
|       (insert quoted))
|     (when (null no-fill) ;; I said: No filling!.. Thank you.
|       (setq fill-prefix quoted) ;; For me fill-paragraph sucks without this. :-(
|       (fill-paragraph nil))
|     (goto-char point)
|     (forward-line 2)))
`--------------------

which avoids using sc-cite-regexp (but mimics its functionality, of
course). It also tries to solve:

Laura Conrad <lconrad@world.std.com> writes:

[...]

>     Vladimir> 	(re-search-forward "[ \t]*")) Vladimir> (when quoted
>     Vladimir> 	Vladimir> (insert quoted)) Vladimir> (fill-paragraph
>     Vladimir> 	nil) Vladimir> (goto-char point) Vladimir>
>     Vladimir> 	(forward-line 2)))
> 
> 
> <end test 2>
> 
> This is a bad example, since you wouldn't really want the lisp code
> reformatted, but I frequently reply to correspondents who haven't
> filled their lines any better than the lisp code did.

        The new message-newline-and-reformat avoids filling the
paragraph if you pass it a prefix argument, like in: `C-u M-RET'.

works for me, so hth,
Francisco

P.S. Thank yuo Vladimir. :-)

-- 
Alcoholics Anonymous is when you get to drink under someone else's name.

Re: message-newline-and-reformat and supercite

[Vladimir Volovich]

"LC" == Laura Conrad writes:

 LC> However, if there are spaces at the ends of lines, the
 LC> attributions don't get removed:

i have to agree. the current implementation of reformatting should be
made a bit more robust... it relies on fill-paragraph, which works not
perfect in some situations.

	Best regards, -- Vladimir.

Re: message-newline-and-reformat and supercite

[Laura Conrad]

>>>>> "Vladimir" == Vladimir Volovich <vvv@vvv.vsu.ru> writes:

       CC> Do you have a workaround ?

    Vladimir> the current message-newline-and-reformat in gnus 0.97
    Vladimir> behaves like you said. i already posted the corrected
    Vladimir> function. here is it again:
...
    Vladimir> does it work for you, or is *the above* function behaving buggy?

It seems to work some of the time, for instance, on the above,
correctly filled text, I can do:

<begin test 1>
    Vladimir> the current message-newline-and-reformat in gnus 0.97
    Vladimir> behaves like you said. 

Here I have inserted a break, and the next paragraph displays correctly.

    Vladimir> i already posted the corrected function. here is it
    Vladimir> again:
<end test 1>

However, if there are spaces at the ends of lines, the attributions
don't get removed:

<begin test 2>


    Vladimir> (defun message-newline-and-reformat ()
    Vladimir>   "Insert four newlines, and then reformat if inside quoted text."
    Vladimir>   (interactive)
    Vladimir>   (let ((point (point))

Here's where I inserted the break.

    Vladimir>   
    Vladimir> 	quoted) Vladimir> (save-excursion Vladimir>
    Vladimir> 	(beginning-of-line) Vladimir> (if (looking-at
    Vladimir> 	(sc-cite-regexp)) (setq quoted (buffer-substring
    Vladimir> 	(match-beginning 0) (match-end 0))))) Vladimir>
    Vladimir> 	(insert "\n\n\n\n") Vladimir> (delete-region (point)
    Vladimir> 	(re-search-forward "[ \t]*")) Vladimir> (when quoted
    Vladimir> 	Vladimir> (insert quoted)) Vladimir> (fill-paragraph
    Vladimir> 	nil) Vladimir> (goto-char point) Vladimir>
    Vladimir> 	(forward-line 2)))


<end test 2>

This is a bad example, since you wouldn't really want the lisp code
reformatted, but I frequently reply to correspondents who haven't
filled their lines any better than the lisp code did.

-- 
Laura (mailto:lconrad@world.std.com , http://www.world.std.com/~lconrad/ )
(617) 661-8097	fax: (801) 365-6574 
233 Broadway, Cambridge, MA 02139

Re: message-newline-and-reformat and supercite

[Kai Großjohann]

Vladimir Volovich <vvv@vvv.vsu.ru> writes:

>       (if (looking-at (sc-cite-regexp))

One potential problem with this is that it depends on Supercite.
Supercite already offers C-c C-p o which I used to use a lot when I
used Supercite.

So, maybe we should lobby Barry to provide such a function, and maybe
we should lobby Barry and Lars to find some way for Message and
Supercite to better interoperate.  Maybe it is sufficient to introduce
a variable message-newline-function which is called when non-nil, and
Barry provides one which is Supercite aware?

What do you think?  (I don't use Supercite now and the standard
definition for M-RET does rather well for me.  I still need to do M-q
on the second paragraph, and whitespace isn't deleted, but I don't
perceive this as a real problem.)

kai
-- 
Life is hard and then you die.

Re: message-newline-and-reformat and supercite

[Vladimir Volovich]

"CC" == Christophe Cuq writes:

 VV> Note that the rest of the quote was micely reformatted, and
 VV> cursor was put at the convenient place. Doesn't this work for
 VV> you?

 CC> I have the same problem as François. When I use M-RET it does not
 CC> reformat the quoted material. The first line appears without the
 CC> quote mark.

 CC> Do you have a workaround ?

the current message-newline-and-reformat in gnus 0.97 behaves like you
said. i already posted the corrected function. here is it again:

(defun message-newline-and-reformat ()
  "Insert four newlines, and then reformat if inside quoted text."
  (interactive)
  (let ((point (point))
	quoted)
    (save-excursion
      (beginning-of-line)
      (if (looking-at (sc-cite-regexp))
	  (setq quoted (buffer-substring (match-beginning 0) (match-end 0)))))
    (insert "\n\n\n\n")
    (delete-region (point) (re-search-forward "[ \t]*"))
    (when quoted
      (insert quoted))
    (fill-paragraph nil)
    (goto-char point)
    (forward-line 2)))

does it work for you, or is *the above* function behaving buggy?

	Best regards, -- Vladimir.

Re: message-newline-and-reformat and supercite

[Christophe Cuq]

>> "DSG" == David S Goldberg <dsg@mitre.org> writes:

DSG> Did you use Vladimir's definition or the default?  Vladimir's works
DSG> great for me; this is something I'd wanted for quite some time.
DSG> I'd meant to send him a "thank you" in private email but since I
DSG> forgot, Thanks Vladimir.

I didn't see his message, and so I didn't apply his func and use the
default :(

Sorry for the noise

And thanks to Vladimir and you.

-- 
CHC

Re: message-newline-and-reformat and supercite

[David S. Goldberg]

> I have the same problem as François. When I use M-RET it does not
> reformat the quoted material. The first line appears without the
> quote mark.

Did you use Vladimir's definition or the default?  Vladimir's works
great for me; this is something I'd wanted for quite some time.  I'd
meant to send him a "thank you" in private email but since I forgot,
Thanks Vladimir.
-- 
Dave Goldberg
Post: The Mitre Corporation\MS B325\202 Burlington Rd.\Bedford, MA 01730
Phone: 781-271-3887
Email: dsg@mitre.org

Re: message-newline-and-reformat and supercite

[Christophe Cuq]

>> "VV" == Vladimir Volovich <vvv@vvv.vsu.ru> writes:

VV> Note that the rest of the quote was micely reformatted, and cursor
VV> was put at the convenient place. Doesn't this work for you?

I have the same problem as François. When I use M-RET it does not
reformat the quoted material. The first line appears without the quote
mark.

Do you have a workaround ?

Best Regards

-- 
CHC

Re: message-newline-and-reformat and supercite

[Vladimir Volovich]

"FP" == Francois Pinard writes:

 >> message-newline-and-reformat (M-RET) does not work with supercited
 >> messages, which is very inconvenient: i find this function VERY
 >> useful in general, so it would be great to make it SC-aware.

 FP> How do you use it?

well, i just put cursor to the place where i wish to break the quote,
and press M-RET (as i did just now ;). Then gnus automagically inserts
a prefix (` FP> ' in this case) behind the first line, and re-formats
the rest of the quoted paragraph (making it both look beautigul and
reserving a bit of space :), and puts the point in the convenient
place, so i can type an answer immediately without moving cursor.

 FP> I just tried it, and it seems be convenient for inserting
 FP> comments within a reply, maybe.  Yet, it leaves two white lines
 FP> before point, which is surely one too much.  Also, it does not
 FP> seem to properly discover the fill prefix for the following
 FP> paragraph, at least in my little tries.  The amount of extra work
 FP> needed to repair these things, even if tiny, outweighs the
 FP> benefits of using the command, at first glance.  But since you
 FP> underlined `VERY', above, I guess I may be missing something?
 FP> :-)

Well, did you re-defined the message-newline-and-reformat function
before trying? Let me give an example of how it works for me.

1) the quoted paragraph, as it looked after following up (prepared by
   supercite). Cursor is shown by `_' (at the first line, just after
   the question sign), and is at the place where i wished to insert a
   reply:

 FP> How do you use it?_ I just tried it, and it seems be convenient
 FP> for inserting comments within a reply, maybe.  Yet, it leaves two
 FP> white lines before point, which is surely one too much.  Also, it
 FP> does not seem to properly discover the fill prefix for the
 FP> following paragraph, at least in my little tries.  The amount of
 FP> extra work needed to repair these things, even if tiny, outweighs
 FP> the benefits of using the command, at first glance.  But since
 FP> you underlined `VERY', above, I guess I may be missing something?
 FP> :-)

2) when i press M-RET, the paragraph looks like:

 FP> How do you use it?

_

 FP> I just tried it, and it seems be convenient for inserting
 FP> comments within a reply, maybe.  Yet, it leaves two white lines
 FP> before point, which is surely one too much.  Also, it does not
 FP> seem to properly discover the fill prefix for the following
 FP> paragraph, at least in my little tries.  The amount of extra work
 FP> needed to repair these things, even if tiny, outweighs the
 FP> benefits of using the command, at first glance.  But since you
 FP> underlined `VERY', above, I guess I may be missing something?
 FP> :-)

Note that the rest of the quote was micely reformatted, and cursor was
put at the convenient place. Doesn't this work for you?

	Best regards, -- Vladimir.
-- 
	"What terrible way to die."
	"There are no good ways."
		-- Sulu and Kirk, "That Which Survives", stardate unknown

Re: message-newline-and-reformat and supercite

[François Pinard]

Vladimir Volovich <vvv@vvv.vsu.ru> writes:

> message-newline-and-reformat (M-RET) does not work with supercited
> messages, which is very inconvenient: i find this function VERY useful
> in general, so it would be great to make it SC-aware.

How do you use it?  I just tried it, and it seems be convenient for inserting
comments within a reply, maybe.  Yet, it leaves two white lines before point,
which is surely one too much.  Also, it does not seem to properly discover
the fill prefix for the following paragraph, at least in my little tries.
The amount of extra work needed to repair these things, even if tiny,
outweighs the benefits of using the command, at first glance.  But since
you underlined `VERY', above, I guess I may be missing something? :-)

-- 
François Pinard   http://www.iro.umontreal.ca/~pinard

message-newline-and-reformat and supercite

[Vladimir Volovich]

Lars,

message-newline-and-reformat (M-RET) does not work with supercited
messages, which is very inconvenient: i find this function VERY useful
in general, so it would be great to make it SC-aware.

i modified it as follows, and it "works for me" (c):

(defun message-newline-and-reformat ()
  "Insert four newlines, and then reformat if inside quoted text."
  (interactive)
  (let ((point (point))
	quoted)
    (save-excursion
      (beginning-of-line)
      (if (looking-at (sc-cite-regexp))
	  (setq quoted (buffer-substring (match-beginning 0) (match-end 0)))))
    (insert "\n\n\n\n")
    (delete-region (point) (re-search-forward "[ \t]*"))
    (when quoted
      (insert quoted))
    (fill-paragraph nil)
    (goto-char point)
    (forward-line 2)))

The line (delete-region (point) (re-search-forward "[ \t]*"))
is in general useful (not only for SC), as it removes extra spaces at
the beginning of the recited message, preserving good indentation.

Could you please merge something like the above into the mainstream
gnus?

	Best regards, -- Vladimir.