message-send-comfirm saved me again, do implement

message-send-comfirm saved me again, do implement

[jidanni]

Ha ha ha, I told you C-c C-c was too dangerous.
Well, there I was, hitting
  C-c C-f C-c runs the command message-goto-cc
and guess which keypress I pressed too lightly?
And guess who almost didn't get on the CC list as the incomplete
message almost flew out the door? Obama? No, actually I forgot who,
all I know is that
 (defun jidanni-message-confirm ()
   "Confirm that we really want to send the message."
   (interactive)(or (y-or-n-p "Send?")(keyboard-quit)))
 (add-hook 'message-send-hook 'jidanni-message-confirm)
saved the day, and should be made into an option so users can do
(setq message-send-comfirm t) if they prefer not to every so often
"spray" others with half-baked messages that aren't fully comple

Re: message-send-comfirm saved me again, do implement

[Ted Zlatanov]

On Sat, 30 Aug 2008 13:03:28 +0800 jidanni@jidanni.org wrote: 

j> Ha ha ha, I told you C-c C-c was too dangerous.
j> Well, there I was, hitting
j>   C-c C-f C-c runs the command message-goto-cc
j> and guess which keypress I pressed too lightly?
j> And guess who almost didn't get on the CC list as the incomplete
j> message almost flew out the door? Obama? No, actually I forgot who,
j> all I know is that
j>  (defun jidanni-message-confirm ()
j>    "Confirm that we really want to send the message."
j>    (interactive)(or (y-or-n-p "Send?")(keyboard-quit)))
j>  (add-hook 'message-send-hook 'jidanni-message-confirm)
j> saved the day, and should be made into an option so users can do
j> (setq message-send-comfirm t) if they prefer not to every so often
j> "spray" others with half-baked messages that aren't fully comple

In my experience this will only make the user learn `C-c C-c y' as the
sequence.

Ted

Re: message-send-comfirm saved me again, do implement

[Tassilo Horn]

Ted Zlatanov <tzz@lifelogs.com> writes:

Hi Ted,

> j> Ha ha ha, I told you C-c C-c was too dangerous.

[...]

> In my experience this will only make the user learn `C-c C-c y' as the
> sequence.

But nobody will hit `C-c C-c y' accidentally when he wanted to type `C-c
C-f C-c'.

Bye,
Tassilo
-- 
When Richard Stallman goes to the loo, he core dumps.

Re: message-send-comfirm saved me again, do implement

[Eric Schulte]

Ted Zlatanov <tzz@lifelogs.com> writes:

> On Sat, 30 Aug 2008 13:03:28 +0800 jidanni@jidanni.org wrote: 
>
> j> Ha ha ha, I told you C-c C-c was too dangerous.

I agree, as someone who is just switching from VM and spent many minutes
yesterday looking for this exact option, I have to say, it saved me many
times, and I'm adding the confirmation function to my gnus.el.

> In my experience this will only make the user learn `C-c C-c y' as the
> sequence.

`C-c C-c y' adds another key, and the right hand to the equation, I've
had many moments of relief while hovering over the `y' key.

-- Eric

Re: message-send-comfirm saved me again, do implement

[Ted Zlatanov]

On Sat, 30 Aug 2008 08:08:00 -0700 Eric Schulte <schulte.eric@gmail.com> wrote: 

ES> `C-c C-c y' adds another key, and the right hand to the equation, I've
ES> had many moments of relief while hovering over the `y' key.

I still think that's just conditioning you to hit the extra key, but if
it makes you happy :)

On Sat, 30 Aug 2008 12:44:10 +0200 Tassilo Horn <tassilo@member.fsf.org> wrote: 

TH> But nobody will hit `C-c C-c y' accidentally when he wanted to type `C-c
TH> C-f C-c'.

Good point, I didn't think of it.

I added the boolean `message-confirm-send'.  It uses `keyboard-quit'
which may not be the right approach in `message-send'.  Opinions
welcome.  It works for me.

Ted

Re: message-send-comfirm saved me again, do implement

[Miles Bader]

Ted Zlatanov <tzz@lifelogs.com> writes:
> TH> But nobody will hit `C-c C-c y' accidentally when he wanted to type `C-c
> TH> C-f C-c'.
>
> Good point, I didn't think of it.

Though the right solution to _that_ particular little nit would probably
be to change the binding "C-c C-f C-c" to somethign else.

-Miles

-- 
Helpmate, n. A wife, or bitter half.

Re: message-send-comfirm saved me again, do implement

[Tassilo Horn]

Miles Bader <miles.bader@necel.com> writes:

Hi Miles,

>> TH> But nobody will hit `C-c C-c y' accidentally when he wanted to
>> TH> type `C-c C-f C-c'.
>>
>> Good point, I didn't think of it.
>
> Though the right solution to _that_ particular little nit would
> probably be to change the binding "C-c C-f C-c" to somethign else.

IMO the (optional) additional y-or-n-p is better, because the command
mnemonics stay consistent.  `C-c C-f X' goes to or inserts a header,
where X is

     C-t ==> To
     C-c ==> Cc
     C-b ==> Bcc
     C-s ==> Subject
     ...

Just my 2 cents,
Tassilo

Re: message-send-comfirm saved me again, do implement

[Drew Raines]

jidanni@jidanni.org wrote:

>  (defun jidanni-message-confirm ()
>    "Confirm that we really want to send the message."
>    (interactive)(or (y-or-n-p "Send?")(keyboard-quit)))

I don't really like the modal nature of being prompted.  I just throw
in a prefix when I'm intentionally sending.  For example:

  (defun aar/delay-send (prefix)
    "C-c C-c is too easy to hit."
    (interactive "p")
    (if (= 1 prefix)
        (message "Prefix with C-u to send...")
      (message-send-and-exit)))
  (define-key message-mode-map "\C-c\C-c" 'aar/delay-send)

-Drew

Re: message-send-comfirm saved me again, do implement

[Reiner Steib]

On Wed, Sep 03 2008, Ted Zlatanov wrote:

> On Sat, 30 Aug 2008 12:44:10 +0200 Tassilo Horn <tassilo@member.fsf.org> wrote: 
>
> TH> But nobody will hit `C-c C-c y' accidentally when he wanted to type `C-c
> TH> C-f C-c'.
>
> Good point, I didn't think of it.
[...]
> I added the boolean `message-confirm-send'.

IMHO, we don't need such an option.  Users who have problems with
accidentally `C-c C-c' should simply setup a different key binding in
their init file.

+(defcustom message-confirm-send nil
+  "Non-nil means when sending a message ask for y/n confirmation."

How about this?

   "When non-nil, ask for confirmation when sending a message.

+  :group 'message-sending
+  :group 'message-mail
+  :link '(custom-manual "(message)Sending Variables")

I can't find it documented there.

+  :type 'boolean)

Please add an appropriate :version tag.

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

Re: message-send-comfirm saved me again, do implement

[Ted Zlatanov]

On Tue, 23 Sep 2008 20:20:12 +0200 Reiner Steib <reinersteib+gmane@imap.cc> wrote: 

RS> +(defcustom message-confirm-send nil
RS> +  "Non-nil means when sending a message ask for y/n confirmation."

RS> How about this?

RS>    "When non-nil, ask for confirmation when sending a message.

RS> +  :group 'message-sending
RS> +  :group 'message-mail
RS> +  :link '(custom-manual "(message)Sending Variables")

RS> I can't find it documented there.

RS> +  :type 'boolean)

RS> Please add an appropriate :version tag.

Fixed both version and manual link.  Thanks for catching those.

Ted

Re: message-send-comfirm saved me again, do implement

[Reiner Steib]

On Wed, Sep 24 2008, Ted Zlatanov wrote:

> On Tue, 23 Sep 2008 Reiner Steib <reinersteib+gmane@imap.cc> wrote: 
> RS> +(defcustom message-confirm-send nil
> RS> +  "Non-nil means when sending a message ask for y/n confirmation."
>
> RS> How about this?
>
> RS>    "When non-nil, ask for confirmation when sending a message.

I still think we should avoid this "y/n" formulation (also in
message.texi).

> Fixed both version and manual link.  

Thanks.

You also checked in an unrelated change: Francis Litterio's patch.

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

Re: message-send-comfirm saved me again, do implement

[Ted Zlatanov]

On Wed, 24 Sep 2008 22:40:41 +0200 Reiner Steib <reinersteib+gmane@imap.cc> wrote: 

RS> On Wed, Sep 24 2008, Ted Zlatanov wrote:
>> On Tue, 23 Sep 2008 Reiner Steib <reinersteib+gmane@imap.cc> wrote: 
RS> +(defcustom message-confirm-send nil
RS> +  "Non-nil means when sending a message ask for y/n confirmation."
>> 
RS> How about this?
>> 
RS> "When non-nil, ask for confirmation when sending a message.

RS> I still think we should avoid this "y/n" formulation (also in
RS> message.texi).

OK, fixed everywhere.

RS> You also checked in an unrelated change: Francis Litterio's patch.

I was testing with it for a while and missed it.  I added a ChangeLog
entry but I can't disassociate the old commit.  I hope it's OK as it
stands now.

Ted