Position of point in reply-with-original

Position of point in reply-with-original

[Antoine Levitt]

Hello,

This might be a very dumb question but I didn't find any posts or doc to
that effect. When I reply-with-original, the point is in weird
places. For instance,

*headers*
--text follows this line--
On XX, YY wrote
!> stuff

with point at !. Why isn't it either above or below by default, and why
isn't there a variable to change this behaviour? I'd expect this issue
to have been raised by now: how do people cope with this?

Antoine

Re: Position of point in reply-with-original

[Eric Abrahamsen]

On Thu, Feb 03 2011, Antoine Levitt wrote:

> Hello,
>
> This might be a very dumb question but I didn't find any posts or doc to
> that effect. When I reply-with-original, the point is in weird
> places. For instance,
>
> *headers*
> --text follows this line--
> On XX, YY wrote
> !> stuff
>
> with point at !. Why isn't it either above or below by default, and why
> isn't there a variable to change this behaviour? I'd expect this issue
> to have been raised by now: how do people cope with this?

I started a thread on this issue either here or in emacs.gnus.user
pretty recently, but now can't find it… Basically the variable
message-citation-line-function dictates what kind of citation line is
inserted, and where point ends up afterwards. All the built-in defaults
for that variable do what you've seen: place point after the citation
line. If you want open space *above* the citation line, and start point
at the top of the buffer, you'll have to write a custom function. Check
out section 3.6 of the info manual for Message (how do you all make
nifty clickable links to Info nodes, by the way?).

Briefly, the "interleaved" quotation style expects that you'd begin
composing your reply somewhere down in the midst of the quoted text.
"Top posting" style expects you to start typing a distinct reply at the
top of the message body. Gnus, having evolved for newsreading, assumes
you don't want to top-post.

Eric

Re: Position of point in reply-with-original

[Antoine Levitt]

03/02/11 17:24, Eric Abrahamsen
> On Thu, Feb 03 2011, Antoine Levitt wrote:
>
>> Hello,
>>
>> This might be a very dumb question but I didn't find any posts or doc to
>> that effect. When I reply-with-original, the point is in weird
>> places. For instance,
>>
>> *headers*
>> --text follows this line--
>> On XX, YY wrote
>> !> stuff
>>
>> with point at !. Why isn't it either above or below by default, and why
>> isn't there a variable to change this behaviour? I'd expect this issue
>> to have been raised by now: how do people cope with this?
>
> I started a thread on this issue either here or in emacs.gnus.user
> pretty recently, but now can't find it… Basically the variable
> message-citation-line-function dictates what kind of citation line is
> inserted, and where point ends up afterwards. All the built-in defaults
> for that variable do what you've seen: place point after the citation
> line. If you want open space *above* the citation line, and start point
> at the top of the buffer, you'll have to write a custom function. Check
> out section 3.6 of the info manual for Message (how do you all make
> nifty clickable links to Info nodes, by the way?).
>
> Briefly, the "interleaved" quotation style expects that you'd begin
> composing your reply somewhere down in the midst of the quoted text.
> "Top posting" style expects you to start typing a distinct reply at the
> top of the message body. Gnus, having evolved for newsreading, assumes
> you don't want to top-post.
>
> Eric

I see. Wikipedia has an amazingly long article about it at
http://en.wikipedia.org/wiki/Posting_style

There's message-cite-reply-above that does the trick, but is incredibly
condescending by

- insulting you for wanting the option in the first place in the
  docstring
- ignoring the setting in news posting
- asking you if you're really really sure when you post

It seems it's pretty standard use in mail, and almost all "modern"
clients (webmails, thunderbird, outlook and such) have this as
default. I'm not trying to start a flamewar as to which one is better,
but we should give the users the option.

A good long-term option would be to have a message-citation-style that
would include 'top-posting, 'interleaved-posting, 'bottom-posting, where
interleaved-posting would be what's currently in place, and top / bottom
would position point at the beginning/end of the buffer, creating the
necessary newlines. Users would then set that in gnus-posting-style, in
order to (for instance) top-post for mail and interleave for news. What
do people think?

Re: Position of point in reply-with-original

[Adam Sjøgren]

On Thu, 03 Feb 2011 18:10:49 +0100, Antoine wrote:

> There's message-cite-reply-above that does the trick, but is incredibly
> condescending by

> - insulting you for wanting the option in the first place in the
>   docstring
> - ignoring the setting in news posting
> - asking you if you're really really sure when you post

So, basically it is trying to tell you to not top-post, and helps you
not to do so in newsgroups where it is frowned upon, right?

> It seems it's pretty standard use in mail, and almost all "modern"
> clients (webmails, thunderbird, outlook and such) have this as
> default.

I just created an email-account in Thunderbird, and the default was
"[x] Automatically quote the original message when replying", Then,
"start my reply below the quote":

 * http://koldfront.dk/misc/thunderbird/reply-default.png

So it seems that Thunderbird (3.0.11) does not default to top posting.

Note the box right below, where Thunderbird recommends putting the
signature last if you choose to top-post (unlike Microsoft Outlook).

> I'm not trying to start a flamewar as to which one is better, but we
> should give the users the option.

Why?

[...]
> What do people think?

I don't understand why you wouldn't want people to jump through hoops to
do the - arguably - wrong thing? It is entirely possible to do so, but
why make it easier?


  Best regards,

    Adam

-- 
 "Här kommer rädslan, gamle vän                               Adam Sjøgren
  När alla fjärilar i magen vaknar upp                   asjo@koldfront.dk
  Viskar välkommen hem"

Re: Position of point in reply-with-original

[Steinar Bang]

>>>>> asjo@koldfront.dk (Adam Sjøgren):

> On Thu, 03 Feb 2011 18:10:49 +0100, Antoine wrote:

>> I'm not trying to start a flamewar as to which one is better, but we
>> should give the users the option.

> Why?

I don't always agree with Adam.  But here I do: why?

Re: Position of point in reply-with-original

[Antoine Levitt]

03/02/11 18:31, Adam Sjøgren
> On Thu, 03 Feb 2011 18:10:49 +0100, Antoine wrote:
>
>> There's message-cite-reply-above that does the trick, but is incredibly
>> condescending by
>
>> - insulting you for wanting the option in the first place in the
>>   docstring
>> - ignoring the setting in news posting
>> - asking you if you're really really sure when you post
>
> So, basically it is trying to tell you to not top-post, and helps you
> not to do so in newsgroups where it is frowned upon, right?

It doesn't seem to be frowned upon on all newsgroups, isn't in mail, and
presumably if the user has gone through all the trouble of setting the
variable, he knows what he's doing. The whole idea of having a variable,
and then making it useless by prompting you whenever you try to compose
a mail is just wrong.

>
>> It seems it's pretty standard use in mail, and almost all "modern"
>> clients (webmails, thunderbird, outlook and such) have this as
>> default.
>
> I just created an email-account in Thunderbird, and the default was
> "[x] Automatically quote the original message when replying", Then,
> "start my reply below the quote":
>
>  * http://koldfront.dk/misc/thunderbird/reply-default.png
>
> So it seems that Thunderbird (3.0.11) does not default to top posting.
>
> Note the box right below, where Thunderbird recommends putting the
> signature last if you choose to top-post (unlike Microsoft Outlook).

Sorry about that, it seems my example is wrong. All the webmails I've
seen do this by default, though.

>
>> I'm not trying to start a flamewar as to which one is better, but we
>> should give the users the option.
>
> Why?

I don't know, somehow I was under the impression that gnus was
customizable :)

>
> [...]
>> What do people think?
>
> I don't understand why you wouldn't want people to jump through hoops
> to do the - arguably - wrong thing? It is entirely possible to do so,
> but why make it easier?

Wow, I wasn't really expecting to get this kind of answer on the gnus
ML. I haven't given much thought as to the different posting styles, and
am not going to enter the debate. But I for one would like to post above
the cited text, at least in mail. In the current state of affairs, I
cannot do it without modifying gnus' source code.

Re: Position of point in reply-with-original

[Adam Sjøgren]

On Thu, 03 Feb 2011 22:52:15 +0100, Antoine wrote:

>>> I'm not trying to start a flamewar as to which one is better, but we
>>> should give the users the option.

>> Why?

> I don't know, somehow I was under the impression that gnus was
> customizable :)

Me too. You didn't quite answer the question, though?

>> I don't understand why you wouldn't want people to jump through hoops
>> to do the - arguably - wrong thing? It is entirely possible to do so,
>> but why make it easier?

> Wow, I wasn't really expecting to get this kind of answer on the gnus
> ML.

Sorry to disappoint you so. Apparantly the group isn't as homogeneous as
expected.

> I haven't given much thought as to the different posting styles, and
> am not going to enter the debate. But I for one would like to post
> above the cited text, at least in mail. In the current state of
> affairs, I cannot do it without modifying gnus' source code.

You're not going to enter the debate, but on the other hand you want it
to work your way. Sounds kind of self-contradictory to me.

Anyway, I think you could do exactly what you want by adding a hook
rather than modifying the source code of Gnus.

Perhaps message-setup-hook?


  Best regards,

    Adam

-- 
 "Pray, Mr. Babbage, if you put into the machine wrong        Adam Sjøgren
  figures, will the right answers come out?"             asjo@koldfront.dk

Re: Position of point in reply-with-original

[Antoine Levitt]

03/02/11 23:10, Adam Sjøgren
> On Thu, 03 Feb 2011 22:52:15 +0100, Antoine wrote:
>
>>>> I'm not trying to start a flamewar as to which one is better, but we
>>>> should give the users the option.
>
>>> Why?
>
>> I don't know, somehow I was under the impression that gnus was
>> customizable :)
>
> Me too. You didn't quite answer the question, though?

Please stop with the passive agression. In the end the decision to add
an option or not comes down to whether people will use it. I can't
magically poll all gnus users, but I don't think I'm the only one to
want it.

>
>>> I don't understand why you wouldn't want people to jump through hoops
>>> to do the - arguably - wrong thing? It is entirely possible to do so,
>>> but why make it easier?
>
>> Wow, I wasn't really expecting to get this kind of answer on the gnus
>> ML.
>
> Sorry to disappoint you so. Apparantly the group isn't as homogeneous as
> expected.

I wasn't expecting any consensus, but I was not expecting that kind of
frontal opposition.

>
>> I haven't given much thought as to the different posting styles, and
>> am not going to enter the debate. But I for one would like to post
>> above the cited text, at least in mail. In the current state of
>> affairs, I cannot do it without modifying gnus' source code.
>
> You're not going to enter the debate, but on the other hand you want it
> to work your way. Sounds kind of self-contradictory to me.

I'm not going to enter the debate of whether top posting or interleave
posting is better. I _am_ trying to discuss whether the option should be
added.

>
> Anyway, I think you could do exactly what you want by adding a hook
> rather than modifying the source code of Gnus.
>
> Perhaps message-setup-hook?

An option already exists, and is perfectly correct, except it's useless
because of limitations.

Therefore, I'm proposing this simple patch. It keeps the "it's evil" bit
in the docstring, it just doesn't make the variable impossible to use.

=== modified file 'lisp/gnus/message.el'                                       
--- lisp/gnus/message.el        2011-02-03 05:26:36 +0000                      
+++ lisp/gnus/message.el	2011-02-03 22:26:18 +0000
@@ -1082,6 +1082,17 @@
   :link '(custom-manual "(message)Insertion Variables")
   :group 'message-insertion)
 
+(defcustom message-cite-reply-above nil
+  "*If non-nil, start own text above the quote.
+
+Note: Top posting is bad netiquette.  Don't use it unless you
+really must.  You probably want to set variable only for specific
+groups, e.g. using `gnus-posting-styles':
+
+  (eval (set (make-local-variable 'message-cite-reply-above) t))"
+  :type 'boolean
+  :group 'message-insertion)
+
 (defcustom message-distribution-function nil
   "*Function called to return a Distribution header."
   :group 'message-news
@@ -3638,17 +3649,6 @@
       (while (re-search-forward citexp nil t)
 	(replace-match (if remove "" "\n"))))))
 
-(defvar message-cite-reply-above nil
-  "If non-nil, start own text above the quote.
-
-Note: Top posting is bad netiquette.  Don't use it unless you
-really must.  You probably want to set variable only for specific
-groups, e.g. using `gnus-posting-styles':
-
-  (eval (set (make-local-variable 'message-cite-reply-above) t))
-
-This variable has no effect in news postings.")
-
 (defun message-yank-original (&optional arg)
   "Insert the message being replied to, if any.
 Puts point before the text and mark after.
@@ -3665,17 +3665,11 @@
     (when (and message-reply-buffer
 	       message-cite-function)
       (when message-cite-reply-above
-	(if (and (not (message-news-p))
-		 (or (eq message-cite-reply-above 'is-evil)
-		     (y-or-n-p "\
-Top posting is bad netiquette.  Please don't top post unless you really must.
-Really top post? ")))
-	    (save-excursion
-	      (setq body-text
-		    (buffer-substring (message-goto-body)
-				      (point-max)))
-	      (delete-region (message-goto-body) (point-max)))
-	  (set (make-local-variable 'message-cite-reply-above) nil)))
+	(save-excursion
+	  (setq body-text
+		(buffer-substring (message-goto-body)
+				  (point-max)))
+	  (delete-region (message-goto-body) (point-max))))
       (if (bufferp message-reply-buffer)
 	  (delete-windows-on message-reply-buffer t))
       (push-mark (save-excursion

Re: Position of point in reply-with-original

[Ted Zlatanov]

On Thu, 03 Feb 2011 23:29:07 +0100 Antoine Levitt <antoine.levitt@gmail.com> wrote: 

AL> Therefore, I'm proposing this simple patch. It keeps the "it's evil" bit
AL> in the docstring, it just doesn't make the variable impossible to use.

I agree with the defvar -> defcustom change.

I think the top-posting setup should be a little more involved,
mirroring Outlook somewhat.  This is what I use (I got help from the
Gnus mailing list when I set it up):

;; outlooky
(defun sc-style-outlook ()
  (interactive)
  (setq
   message-cite-function 'message-cite-original
   message-citation-line-function 'message-insert-formatted-citation-line
   message-cite-reply-above 'is-evil
   message-yank-prefix ""
   message-yank-cited-prefix ""
   message-yank-empty-prefix ""
   message-citation-line-format
   "\n\n-----------------------\nOn %a, %b %d %Y, %N wrote:\n"))

I use this at work by necessity.

Ted

Re: Position of point in reply-with-original

[Lars Ingebrigtsen]

Ted Zlatanov <tzz@lifelogs.com> writes:

> I agree with the defvar -> defcustom change.
>
> I think the top-posting setup should be a little more involved,
> mirroring Outlook somewhat.  This is what I use (I got help from the
> Gnus mailing list when I set it up):
>
> ;; outlooky
> (defun sc-style-outlook ()
>   (interactive)
>   (setq
>    message-cite-function 'message-cite-original
>    message-citation-line-function 'message-insert-formatted-citation-line
>    message-cite-reply-above 'is-evil
>    message-yank-prefix ""
>    message-yank-cited-prefix ""
>    message-yank-empty-prefix ""
>    message-citation-line-format
>    "\n\n-----------------------\nOn %a, %b %d %Y, %N wrote:\n"))

Given that this much is needed to get a satisfactory top posting format,
I don't think making the defvar->defcustom change would help much.

Perhaps there should be a new meta-ish variable, say

(defcustom message-cite-style nil
  "The overall style to be used when yanking cited text.
Values are either `traditional' (cited text first),
`top-post' (cited text at the bottom), or nil (don't override the
individual message variables)."
  :version "24.1"
  :group 'message-various
  :type '(choice (const :tag "None" :value nil)
		 (const :tag "Traditional" :value traditional)
		 (const :tag "Top-post" :value top-post)))

This would bind the variables you mention above if set to `traditional'
or `top-post'.

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

Re: Position of point in reply-with-original

[Antoine Levitt]

04/02/11 09:45, Lars Ingebrigtsen
> Ted Zlatanov <tzz@lifelogs.com> writes:
>
>> I agree with the defvar -> defcustom change.
>>
>> I think the top-posting setup should be a little more involved,
>> mirroring Outlook somewhat.  This is what I use (I got help from the
>> Gnus mailing list when I set it up):
>>
>> ;; outlooky
>> (defun sc-style-outlook ()
>>   (interactive)
>>   (setq
>>    message-cite-function 'message-cite-original
>>    message-citation-line-function 'message-insert-formatted-citation-line
>>    message-cite-reply-above 'is-evil
>>    message-yank-prefix ""
>>    message-yank-cited-prefix ""
>>    message-yank-empty-prefix ""
>>    message-citation-line-format
>>    "\n\n-----------------------\nOn %a, %b %d %Y, %N wrote:\n"))
>
> Given that this much is needed to get a satisfactory top posting format,
> I don't think making the defvar->defcustom change would help much.
>
> Perhaps there should be a new meta-ish variable, say
>
> (defcustom message-cite-style nil
>   "The overall style to be used when yanking cited text.
> Values are either `traditional' (cited text first),
> `top-post' (cited text at the bottom), or nil (don't override the
> individual message variables)."
>   :version "24.1"
>   :group 'message-various
>   :type '(choice (const :tag "None" :value nil)
> 		 (const :tag "Traditional" :value traditional)
> 		 (const :tag "Top-post" :value top-post)))
>
> This would bind the variables you mention above if set to `traditional'
> or `top-post'.

I don't necessarily agree. My ideal setup would be just
message-cite-reply-above to t. The other variables are a matter of
preference, and shouldn't be tied to the position of point after a
reply.

Then there could be "presets" to mimic the behaviour of various other
clients. I'm not sure they should be as variables, that'd be confusing
as to the behaviour of setting both the preset and the individual
setting. What about functions that set the variables in a buffer-local
fashion and would be triggered by gnus-posting-styles?

Re: Position of point in reply-with-original

[Ted Zlatanov]

On Fri, 04 Feb 2011 00:45:53 -0800 Lars Ingebrigtsen <larsi@gnus.org> wrote: 

LI> Ted Zlatanov <tzz@lifelogs.com> writes:
>> I agree with the defvar -> defcustom change.
>> 
>> I think the top-posting setup should be a little more involved,
>> mirroring Outlook somewhat.  This is what I use (I got help from the
>> Gnus mailing list when I set it up):
>> 
>> ;; outlooky
>> (defun sc-style-outlook ()
>> (interactive)
>> (setq
>> message-cite-function 'message-cite-original
>> message-citation-line-function 'message-insert-formatted-citation-line
>> message-cite-reply-above 'is-evil
>> message-yank-prefix ""
>> message-yank-cited-prefix ""
>> message-yank-empty-prefix ""
>> message-citation-line-format
>> "\n\n-----------------------\nOn %a, %b %d %Y, %N wrote:\n"))

LI> Given that this much is needed to get a satisfactory top posting format,
LI> I don't think making the defvar->defcustom change would help much.

LI> Perhaps there should be a new meta-ish variable, say

LI> (defcustom message-cite-style nil
LI>   "The overall style to be used when yanking cited text.
LI> Values are either `traditional' (cited text first),
LI> `top-post' (cited text at the bottom), or nil (don't override the
LI> individual message variables)."
LI>   :version "24.1"
LI>   :group 'message-various
LI>   :type '(choice (const :tag "None" :value nil)
LI> 		 (const :tag "Traditional" :value traditional)
LI> 		 (const :tag "Top-post" :value top-post)))

LI> This would bind the variables you mention above if set to `traditional'
LI> or `top-post'.

I agree.  This would be useful for everyone using Emacs, not just Gnus
users.  Of course that means many will complain, but we'll just tell
them Gnus is now supported by Microsoft donations and being rewritten in
C#.  Which is, of course, all true.  Hail the Gnus Cabal.

On Fri, 04 Feb 2011 10:38:13 +0100 Antoine Levitt <antoine.levitt@gmail.com> wrote: 

AL> I don't necessarily agree. My ideal setup would be just
AL> message-cite-reply-above to t. The other variables are a matter of
AL> preference, and shouldn't be tied to the position of point after a
AL> reply.

We can make a choice 'top-post-traditional-quotes that reflects your
preference.  But I already said I'm in favor of changing the defvar to a
defcustom for `message-cite-reply-above', which is what you're asking
for.

AL> Then there could be "presets" to mimic the behaviour of various other
AL> clients. I'm not sure they should be as variables, that'd be confusing
AL> as to the behaviour of setting both the preset and the individual
AL> setting. What about functions that set the variables in a buffer-local
AL> fashion and would be triggered by gnus-posting-styles?

I think that's even more confusing, personally.

On Fri, 04 Feb 2011 10:21:43 +0100 pmlists@free.fr (Peter Münster) wrote: 

PM> And the remaining 10% (people like Ted), do top-posting at work, because
PM> they believe, that they need to conform to some "standard".

My concern is really about aesthetics and communication.  I hate my
messages to be mangled by Outlook and misinterpreted more than I hate
top-posting.

PM> There are many other examples, where 90% of the people are doing the
PM> "- arguably - wrong thing"
...
PM> So question the "standard"!
PM> Don't conform to the "standard" when you see, that's it's the "arguably
PM> wrong thing"! ;)

Personally, I'm with the Apathy Coalition.  "Not Our Problem" and "Join
Us.  Or Don't.  Whatever." are our mottos.

(yes, I have the T-shirt: http://shirt.woot.com/Derby/Entry.aspx?id=37083)

Ted

Re: Position of point in reply-with-original

[Lars Ingebrigtsen]

Ted Zlatanov <tzz@lifelogs.com> writes:

> I agree.  This would be useful for everyone using Emacs, not just Gnus
> users.  Of course that means many will complain, but we'll just tell
> them Gnus is now supported by Microsoft donations and being rewritten in
> C#.  Which is, of course, all true.  Hail the Gnus Cabal.

SHH!!!  IT"S A SECRET!!!1!

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

Re: Position of point in reply-with-original

[Steinar Bang]

>>>>> Lars Ingebrigtsen <larsi@gnus.org>:

> Ted Zlatanov <tzz@lifelogs.com> writes:
>> I agree.  This would be useful for everyone using Emacs, not just Gnus
>> users.  Of course that means many will complain, but we'll just tell
>> them Gnus is now supported by Microsoft donations and being rewritten in
>> C#.  Which is, of course, all true.  Hail the Gnus Cabal.

> SHH!!!  IT"S A SECRET!!!1!

http://dotlisp.sourceforge.net/dotlisp.htm

Re: Position of point in reply-with-original

[Peter Münster]

Ted Zlatanov <tzz@lifelogs.com> writes:

> My concern is really about aesthetics and communication.  I hate my
> messages to be mangled by Outlook and misinterpreted more than I hate
> top-posting.

That means, that Outlook renders interleaved postings quite differently
than the MUAs that I know. I'll take a look at this strange software at
the next opportunity...

So excuse me please: I thought, that the only reason for you to top-post
was "Because all the other people at work do it that way." ... ;)
(That's exactly the argument, that always bugs me.)

-- 
           Peter

Re: Position of point in reply-with-original

[Antoine Levitt]

[-- Attachment #1: Type: text/plain, Size: 2237 bytes --]

04/02/11 09:45, Lars Ingebrigtsen
>
> Given that this much is needed to get a satisfactory top posting format,
> I don't think making the defvar->defcustom change would help much.
>
> Perhaps there should be a new meta-ish variable, say
>
> (defcustom message-cite-style nil
>   "The overall style to be used when yanking cited text.
> Values are either `traditional' (cited text first),
> `top-post' (cited text at the bottom), or nil (don't override the
> individual message variables)."
>   :version "24.1"
>   :group 'message-various
>   :type '(choice (const :tag "None" :value nil)
> 		 (const :tag "Traditional" :value traditional)
> 		 (const :tag "Top-post" :value top-post)))
>
> This would bind the variables you mention above if set to `traditional'
> or `top-post'.

So I took a stab at this, and came up with the following patch.

So, what it does:

- The defvar -> defcustom change for message-cite-reply-above discussed
  before, with the prompt "are you sure you want to be evil?" removed,
  and allowing the choice in news posting

- message-cite-reply-above -> message-cite-reply-position, which can be
  traditional (default), above or below

- The new message-cite-style (apparently, you commited it in an
  unrelated commit, Lars:
  https://github.com/emacsmirror/emacs/commit/92a7a0fc1dcec08b923f267e53f1758507efd747)
  variable goes to the message-insertion group (that's where the other
  variables like this one are). Instead of taking symbol values, it
  takes either nil or a list of let-style (variable value) forms.

- This variables/values pairs are enforced by a eval-let in
  message-yank-original, binding the variables in the function. If
  there's a more clever approach, please tell me.

- message-cite-style presets for outlook, thunderbird and gmail
  (shamelessly ripped from Ted Zlatanov for the outlook one, roughly
  inferred for the other two) are available, allowing Gnus to
  impersonate other clients. Some of these are pretty ugly, though. The
  idea is that the user chooses between those styles or creates other
  ones, and then sets message-cite-style in gnus-posting-styles.

It Works For Me (TM), and looks generic and simple enough to suit the
needs of everybody. Here's the patch.


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: cite_style.diff --]
[-- Type: text/x-diff, Size: 7821 bytes --]

diff --git a/lisp/gnus/message.el b/lisp/gnus/message.el
index 58daf1b..78da5d7 100644
--- a/lisp/gnus/message.el
+++ b/lisp/gnus/message.el
@@ -129,17 +129,6 @@
   :group 'message-buffers
   :type '(choice function (const nil)))
 
-(defcustom message-cite-style nil
-  "The overall style to be used when yanking cited text.
-Values are either `traditional' (cited text first),
-`top-post' (cited text at the bottom), or nil (don't override the
-individual message variables)."
-  :version "24.1"
-  :group 'message-various
-  :type '(choice (const :tag "None" :value nil)
-		 (const :tag "Traditional" :value traditional)
-		 (const :tag "Top-post" :value top-post)))
-
 (defcustom message-fcc-handler-function 'message-output
   "*A function called to save outgoing articles.
 This function will be called with the name of the file to store the
@@ -1088,6 +1077,71 @@ needed."
   :link '(custom-manual "(message)Insertion Variables")
   :group 'message-insertion)
 
+(defcustom message-cite-reply-position 'traditional
+  "*Where the reply should be positioned.
+If `traditional', reply inline.
+If `above', reply above quoted text.
+If `below', reply below quoted text.
+
+Note: Many newsgroups frown upon nontraditional reply styles. You
+probably want to set this variable only for specific groups,
+e.g. using `gnus-posting-styles':
+
+  (eval (set (make-local-variable 'message-cite-reply-above) 'above))"
+  :type '(choice (const :tag "Reply inline" 'traditional)
+		 (const :tag "Reply above" 'above)
+		 (const :tag "Reply below" 'below))
+  :group 'message-insertion)
+
+(defcustom message-cite-style nil
+  "*The overall style to be used when yanking cited text.
+Value is either `nil' (no variable overrides) or a let-style list
+of pairs (VARIABLE VALUE) that will be bound in
+`message-yank-original' to do the quoting.
+
+Presets to impersonate popular mail agents are found in the
+message-cite-style-* variables.  This variable is intended for
+use in `gnus-posting-styles', such as:
+
+  ((posting-from-work-p) (eval (set (make-local-variable 'message-cite-style) message-cite-style-outlook)))"
+  :version "24.1"
+  :group 'message-insertion
+  :type '(choice (const :tag "Do not override variables" :value nil)
+		 (const :tag "MS Outlook" :value message-cite-style-outlook)
+		 (const :tag "Mozilla Thunderbird" :value message-cite-style-thunderbird)
+		 (const :tag "Gmail" :value message-cite-style-gmail)
+		 (variable :tag "User-specified")))
+
+(defconst message-cite-style-outlook
+  '((message-cite-function  'message-cite-original)
+    (message-citation-line-function  'message-insert-formatted-citation-line)
+    (message-cite-reply-position 'above)
+    (message-yank-prefix  "")
+    (message-yank-cited-prefix  "")
+    (message-yank-empty-prefix  "")
+    (message-citation-line-format  "\n\n-----------------------\nOn %a, %b %d %Y, %N wrote:\n"))
+  "Message citation style used by MS Outlook. Use with message-cite-style.")
+
+(defconst message-cite-style-thunderbird
+  '((message-cite-function  'message-cite-original)
+    (message-citation-line-function  'message-insert-formatted-citation-line)
+    (message-cite-reply-position 'above)
+    (message-yank-prefix  "> ")
+    (message-yank-cited-prefix  ">")
+    (message-yank-empty-prefix  ">")
+    (message-citation-line-format "On %D %R %p, %N wrote:"))
+  "Message citation style used by Mozilla Thunderbird. Use with message-cite-style.")
+
+(defconst message-cite-style-gmail
+  '((message-cite-function  'message-cite-original)
+    (message-citation-line-function  'message-insert-formatted-citation-line)
+    (message-cite-reply-position 'above)
+    (message-yank-prefix  "    ")
+    (message-yank-cited-prefix  "    ")
+    (message-yank-empty-prefix  "    ")
+    (message-citation-line-format "On %e %B %Y %R, %f wrote:\n"))
+  "Message citation style used by Gmail. Use with message-cite-style.")
+
 (defcustom message-distribution-function nil
   "*Function called to return a Distribution header."
   :group 'message-news
@@ -3645,17 +3699,6 @@ To use this automatically, you may add this function to
       (while (re-search-forward citexp nil t)
 	(replace-match (if remove "" "\n"))))))
 
-(defvar message-cite-reply-above nil
-  "If non-nil, start own text above the quote.
-
-Note: Top posting is bad netiquette.  Don't use it unless you
-really must.  You probably want to set variable only for specific
-groups, e.g. using `gnus-posting-styles':
-
-  (eval (set (make-local-variable 'message-cite-reply-above) t))
-
-This variable has no effect in news postings.")
-
 (defun message-yank-original (&optional arg)
   "Insert the message being replied to, if any.
 Puts point before the text and mark after.
@@ -3669,49 +3712,49 @@ prefix, and don't delete any headers."
   (interactive "P")
   (let ((modified (buffer-modified-p))
 	body-text)
-    (when (and message-reply-buffer
-	       message-cite-function)
-      (when message-cite-reply-above
-	(if (and (not (message-news-p))
-		 (or (eq message-cite-reply-above 'is-evil)
-		     (y-or-n-p "\
-Top posting is bad netiquette.  Please don't top post unless you really must.
-Really top post? ")))
+    ;; eval the let forms contained in message-cite-style
+    (eval
+     `(let ,message-cite-style
+	(when (and message-reply-buffer
+		   message-cite-function)
+	  (when (equal message-cite-reply-position 'above)
 	    (save-excursion
 	      (setq body-text
 		    (buffer-substring (message-goto-body)
 				      (point-max)))
-	      (delete-region (message-goto-body) (point-max)))
-	  (set (make-local-variable 'message-cite-reply-above) nil)))
-      (if (bufferp message-reply-buffer)
-	  (delete-windows-on message-reply-buffer t))
-      (push-mark (save-excursion
-		   (cond
-		    ((bufferp message-reply-buffer)
-		     (insert-buffer-substring message-reply-buffer))
-		    ((and (consp message-reply-buffer)
-			  (functionp (car message-reply-buffer)))
-		     (apply (car message-reply-buffer)
-			    (cdr message-reply-buffer))))
-		   (unless (bolp)
-		     (insert ?\n))
-		   (point)))
-      (unless arg
-	(funcall message-cite-function)
-	(unless (eq (char-before (mark t)) ?\n)
-	  (let ((pt (point)))
-	    (goto-char (mark t))
-	    (insert-before-markers ?\n)
-	    (goto-char pt))))
-      (when message-cite-reply-above
-	(message-goto-body)
-	(insert body-text)
-	(insert (if (bolp) "\n" "\n\n"))
-	(message-goto-body))
-      ;; Add a `message-setup-very-last-hook' here?
-      ;; Add `gnus-article-highlight-citation' here?
-      (unless modified
-	(setq message-checksum (message-checksum))))))
+	      (delete-region (message-goto-body) (point-max))))
+	  (if (bufferp message-reply-buffer)
+	      (delete-windows-on message-reply-buffer t))
+	  (push-mark (save-excursion
+		       (cond
+			((bufferp message-reply-buffer)
+			 (insert-buffer-substring message-reply-buffer))
+			((and (consp message-reply-buffer)
+			      (functionp (car message-reply-buffer)))
+			 (apply (car message-reply-buffer)
+				(cdr message-reply-buffer))))
+		       (unless (bolp)
+			 (insert ?\n))
+		       (point)))
+	  (unless arg
+	    (funcall message-cite-function)
+	    (unless (eq (char-before (mark t)) ?\n)
+	      (let ((pt (point)))
+		(goto-char (mark t))
+		(insert-before-markers ?\n)
+		(goto-char pt))))
+	  (case message-cite-reply-position
+	    ('above
+	     (message-goto-body)
+	     (insert body-text)
+	     (insert (if (bolp) "\n" "\n\n"))
+	     (message-goto-body))
+	    ('below
+	     (message-goto-signature)))
+	  ;; Add a `message-setup-very-last-hook' here?
+	  ;; Add `gnus-article-highlight-citation' here?
+	  (unless modified
+	    (setq message-checksum (message-checksum))))))))
 
 (defun message-yank-buffer (buffer)
   "Insert BUFFER into the current buffer and quote it."

Re: Position of point in reply-with-original

[Lars Ingebrigtsen]

Antoine Levitt <antoine.levitt@gmail.com> writes:

> So I took a stab at this, and came up with the following patch.

Looks good to me.  Do you have FSF copyright assignment papers on file
for Emacs and/or Gnus?

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

Re: Position of point in reply-with-original

[Antoine Levitt]

20/02/11 02:51, Lars Ingebrigtsen
> Antoine Levitt <antoine.levitt@gmail.com> writes:
>
>> So I took a stab at this, and came up with the following patch.
>
> Looks good to me.  Do you have FSF copyright assignment papers on file
> for Emacs and/or Gnus?

Not yet, but paperwork's on its way for emacs. I'll let you know when
it's done.

Re: Position of point in reply-with-original

[Lars Ingebrigtsen]

Antoine Levitt <antoine.levitt@gmail.com> writes:

> Not yet, but paperwork's on its way for emacs. I'll let you know when
> it's done.

Great!

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

Re: Position of point in reply-with-original

[Antoine Levitt]

21/02/11 10:35, Lars Ingebrigtsen
> Antoine Levitt <antoine.levitt@gmail.com> writes:
>
>> Not yet, but paperwork's on its way for emacs. I'll let you know when
>> it's done.
>
> Great!

It's done, you can merge the patch if you want.

Antoine

Re: Position of point in reply-with-original

[Lars Magne Ingebrigtsen]

Antoine Levitt <antoine.levitt@gmail.com> writes:

> It's done, you can merge the patch if you want.

Great; applied.

Could you also send a patch for message.texi to document this?

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

Re: Position of point in reply-with-original

[Antoine Levitt]

05/03/11 11:10, Lars Magne Ingebrigtsen
> Antoine Levitt <antoine.levitt@gmail.com> writes:
>
>> It's done, you can merge the patch if you want.
>
> Great; applied.
>
> Could you also send a patch for message.texi to document this?

Hmm, it seems my reply never made it to the list, not sure why. Here it
is:



I'm not sure what the convention is for the docs, so I opted for a
slightly higher-level description of what the variable does, rather than
the more technical info in C-h v.

diff --git a/doc/misc/message.texi b/doc/misc/message.texi
index b286399..9914235 100644
--- a/doc/misc/message.texi
+++ b/doc/misc/message.texi
@@ -1930,6 +1930,26 @@ posting a prepared news message.
 @section Insertion Variables

 @table @code
+
+@item message-cite-style
+@vindex message-cite-style
+The overall style to be used when replying to messages. This controls
+things like where the reply should be put relative to the original,
+how the citation is formatted, where the signature goes, etc.
+
+Value is either @code{nil} (no variable overrides) or a let-style list
+of pairs @code{(VARIABLE VALUE)} to override default values.
+
+See @code{gnus-posting-styles} to set this variable for specific
+groups. Presets to impersonate popular mail agents are available in the
+@code{message-cite-style-*} variables.
+
+@item message-cite-reply-position
+@vindex message-cite-reply-position
+Where the reply should be positioned. Available styles are
+@code{traditional} to reply inline, @code{above} for top-posting, and
+@code{below} for bottom-posting
+
 @item message-ignored-cited-headers
 @vindex message-ignored-cited-headers
 All headers that match this regexp will be removed from yanked

Re: Position of point in reply-with-original

[Lars Magne Ingebrigtsen]

Antoine Levitt <antoine.levitt@gmail.com> writes:

> I'm not sure what the convention is for the docs, so I opted for a
> slightly higher-level description of what the variable does, rather than
> the more technical info in C-h v.

Thanks; applied.

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

Re: Position of point in reply-with-original

[Reiner Steib]

On Thu, Feb 03 2011, Antoine Levitt wrote:

> There's message-cite-reply-above that does the trick, but is incredibly
> condescending by
>
> - insulting you for wanting the option in the first place in the
>   docstring

I think the doc string (I wrote it, as well as the code) is neither
condescending nor insulting:

,----[ <f1> v message-cite-reply-above RET ]
| message-cite-reply-above is a variable defined in `message.el'.
| Its value is nil
| 
| Documentation:
| If non-nil, start own text above the quote.
| 
| Note: Top posting is bad netiquette.  Don't use it unless you
| really must.  You probably want to set variable only for specific
| groups, e.g. using `gnus-posting-styles':
| 
|   (eval (set (make-local-variable 'message-cite-reply-above) t))
| 
| This variable has no effect in news postings.
`----

> - ignoring the setting in news posting

That's a feature.

> - asking you if you're really really sure when you post
>
> It seems it's pretty standard use in mail, 

Yes unfortunately, like HTML, missing line breaks, stupid disclaimers,
"Sent from my whatever...", signatures without a separator, ...

> and almost all "modern" clients (webmails, thunderbird, outlook and
> such) have this as default.

If you want , use it.

> I'm not trying to start a flamewar as to which one is better, but we
> should give the users the option.

You have the option, as you already discovered.

On Thu, Feb 03 2011, Antoine Levitt wrote:
> But I for one would like to post above the cited text, at least in
> mail. In the current state of affairs, I cannot do it without
> modifying gnus' source code.

Not true.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: Position of point in reply-with-original

[Antoine Levitt]

03/02/11 23:25, Reiner Steib
> On Thu, Feb 03 2011, Antoine Levitt wrote:
>
>> There's message-cite-reply-above that does the trick, but is incredibly
>> condescending by
>>
>> - insulting you for wanting the option in the first place in the
>>   docstring
>
> I think the doc string (I wrote it, as well as the code) is neither
> condescending nor insulting:
>
> ,----[ <f1> v message-cite-reply-above RET ]
> | message-cite-reply-above is a variable defined in `message.el'.
> | Its value is nil
> | 
> | Documentation:
> | If non-nil, start own text above the quote.
> | 
> | Note: Top posting is bad netiquette.  Don't use it unless you
> | really must.  You probably want to set variable only for specific
> | groups, e.g. using `gnus-posting-styles':
> | 
> |   (eval (set (make-local-variable 'message-cite-reply-above) t))
> | 
> | This variable has no effect in news postings.
> `----
>
>> - ignoring the setting in news posting
>
> That's a feature.

Why? Again, if the user sets the variable despite the warnings, there's
a good chance he knows what he's doing.

>
>> - asking you if you're really really sure when you post
>>
>> It seems it's pretty standard use in mail, 
>
> Yes unfortunately, like HTML, missing line breaks, stupid disclaimers,
> "Sent from my whatever...", signatures without a separator, ...
>
>> and almost all "modern" clients (webmails, thunderbird, outlook and
>> such) have this as default.
>
> If you want , use it.
>
>> I'm not trying to start a flamewar as to which one is better, but we
>> should give the users the option.
>
> You have the option, as you already discovered.
>
> On Thu, Feb 03 2011, Antoine Levitt wrote:
>> But I for one would like to post above the cited text, at least in
>> mail. In the current state of affairs, I cannot do it without
>> modifying gnus' source code.
>
> Not true.
>
> Bye, Reiner.

I _can't_ use it, because I want to be able to use it in news posting
and not answer a yes-no question whenever I send a mail. See the
proposed patch in the other thread.

Re: Position of point in reply-with-original

[Steinar Bang]

>>>>> Antoine Levitt <antoine.levitt@gmail.com>:

> Why? Again, if the user sets the variable despite the warnings,
> there's a good chance he knows what he's doing.

Impossible.

Nobody who would willingly pick top posting could possibly know what
he's doing.

Re: Position of point in reply-with-original

[Peter Münster]

Steinar Bang <sb@dod.no> writes:

>>>>>> Antoine Levitt <antoine.levitt@gmail.com>:
>
>> Why? Again, if the user sets the variable despite the warnings,
>> there's a good chance he knows what he's doing.
>
> Impossible.
>
> Nobody who would willingly pick top posting could possibly know what
> he's doing.

I agree: 90% of the top-posters never question what they are doing, they
just begin to write, where the MUA places the point. (5 lines of text,
500 lines of old messages attached, with headers and signatures...)

And the remaining 10% (people like Ted), do top-posting at work, because
they believe, that they need to conform to some "standard".

At my work, there are 99,9% top-posters, I'm the only one, who respects
the order in which people normally read text (from top to bottom). And
sometimes, people tell me, that my messages are more readable, and ask
me how I do it and how they could do it using Lotus Notes.

There are many other examples, where 90% of the people are doing the
"- arguably - wrong thing" (citing Adam):

- Typographic layout of a thesis
90% use the "standard" Word layout with very long lines on A4-paper
At my university institute, I was the only one using a professional
typesetting system with professional layout: a member of the jury said,
that it was for her like reading a book.

- Document format
99,9% of my colleagues use the "standard" proprietary Word-format to
save their documents (a lot of trouble, when sending to users with
different version of Word...). I use the ConTeXt typesetting system,
with version control, conditional compilation (one version for the
client, one for us), and so on..., and people keep asking me how I do
the things that I do...

Just 2 examples from my personal experience, but there are many others.

So question the "standard"!
Don't conform to the "standard" when you see, that's it's the "arguably
wrong thing"! ;)

-- 
           Peter

Re: Position of point in reply-with-original

[Ted Zlatanov]

On Thu, 03 Feb 2011 23:25:29 +0100 Reiner Steib <reinersteib+gmane@imap.cc> wrote: 

RS> On Thu, Feb 03 2011, Antoine Levitt wrote:

>> - ignoring the setting in news posting

RS> That's a feature.

Well, considering GMane, some people may want to top-post in news
servers.  So the feature is less applicable now than when GMane didn't
exist.

>> - asking you if you're really really sure when you post
>> 
>> It seems it's pretty standard use in mail, 

RS> Yes unfortunately, like HTML, missing line breaks, stupid disclaimers,
RS> "Sent from my whatever...", signatures without a separator, ...

This is like complaining that "hopefully" is misused too often.  Yeah,
but this is about language and linguists will tell you that what's
considered "bad" usage today may become tomorrow's standard.  We may not
like it but I think it should be easier for users to enable top-posting,
including revising the docstring to be less hostile to the idea and
mentioning it more in the manual.

Ted

Re: Position of point in reply-with-original

[Lars Ingebrigtsen]

Ted Zlatanov <tzz@lifelogs.com> writes:

> Well, considering GMane, some people may want to top-post in news
> servers.  So the feature is less applicable now than when GMane didn't
> exist.

That's true.

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