[PATCH] Autoload `message-y-or-n-p' macro in gnus.el at compile time.

[PATCH] Autoload `message-y-or-n-p' macro in gnus.el at compile time.

[Steve Youngs]

[-- Attachment #1: Type: text/plain, Size: 1807 bytes --]

Over at XEmacs.org, they've got a couple of awk scripts to help them
find mis-compiled macros and stuff.  It turned up this
one... `message-y-or-n-p', which is defined in message.el isn't
getting picked up in gnus.el (where it's used in `gnu-read-group').

Anyway, I was just about to commit this when I had a thought (yep,
occasionally I do have em).  Should this patch go into gnus.el, or
into dgnushack.el?

Gnus patch:
ChangeLog files diff command: cvs -q diff -U 0
Files affected:               lisp/ChangeLog
Source files diff command:    cvs -q diff -uN
Files affected:               lisp/gnus.el

Index: lisp/ChangeLog
===================================================================
RCS file: /usr/local/cvsroot/gnus/lisp/ChangeLog,v
retrieving revision 7.86
diff -u -U0 -r7.86 ChangeLog
--- lisp/ChangeLog	14 Jan 2004 14:32:21 -0000	7.86
+++ lisp/ChangeLog	15 Jan 2004 09:30:19 -0000
@@ -0,0 +1,4 @@
+2004-01-15  Steve Youngs  <sryoungs@bigpond.net.au>
+
+	* gnus.el: Autoload `message-y-or-n-p'.
+
Index: lisp/gnus.el
===================================================================
RCS file: /usr/local/cvsroot/gnus/lisp/gnus.el,v
retrieving revision 7.6
diff -u -u -r7.6 gnus.el
--- lisp/gnus.el	10 Jan 2004 04:26:59 -0000	7.6
+++ lisp/gnus.el	15 Jan 2004 09:30:15 -0000
@@ -3909,6 +3909,9 @@
       (setq valids (cdr valids)))
     outs))
 
+(eval-when-compile
+  (autoload 'message-y-or-n-p "message" nil 'macro))
+
 (defun gnus-read-group (prompt &optional default)
   "Prompt the user for a group name.
 Disallow invalid group names."

-- 
|---<Steve Youngs>---------------<GnuPG KeyID: A94B3003>---|
|              Ashes to ashes, dust to dust.               |
|      The proof of the pudding, is under the crust.       |
|------------------------------<sryoungs@bigpond.net.au>---|

[-- Attachment #2: Type: application/pgp-signature, Size: 256 bytes --]

Re: [PATCH] Autoload `message-y-or-n-p' macro in gnus.el at compile

[Simon Josefsson]

Steve Youngs <sryoungs@bigpond.net.au> writes:

> Over at XEmacs.org, they've got a couple of awk scripts to help them
> find mis-compiled macros and stuff.  It turned up this
> one... `message-y-or-n-p', which is defined in message.el isn't
> getting picked up in gnus.el (where it's used in `gnu-read-group').

I think gnus-read-group should be fixed.  There is a `gnus-y-or-n-p'.

> Anyway, I was just about to commit this when I had a thought (yep,
> occasionally I do have em).  Should this patch go into gnus.el, or
> into dgnushack.el?

If gnus-read-group is fixed, is it still required?

Re: [PATCH] Autoload `message-y-or-n-p' macro in gnus.el at compile

[Steve Youngs]

[-- Attachment #1: Type: text/plain, Size: 1436 bytes --]

* Simon Josefsson <jas@extundo.com> writes:

  > Steve Youngs <sryoungs@bigpond.net.au> writes:
  >> Over at XEmacs.org, they've got a couple of awk scripts to help them
  >> find mis-compiled macros and stuff.  It turned up this
  >> one... `message-y-or-n-p', which is defined in message.el isn't
  >> getting picked up in gnus.el (where it's used in `gnu-read-group').

  > I think gnus-read-group should be fixed.  There is a `gnus-y-or-n-p'.

Who says it's broken?  And besides, `gnus-y-or-n-p' and
`message-y-or-n-p' are _very_ different animals.  The former is not
much more that `y-or-no-p' with a different name, but the latter has
the option of displaying a "help" buffer as well as the prompt.

  >> Anyway, I was just about to commit this when I had a thought (yep,
  >> occasionally I do have em).  Should this patch go into gnus.el, or
  >> into dgnushack.el?

  > If gnus-read-group is fixed, is it still required?

By the time you "fixed" `gnus-read-group', _maintaining it's current
intended functionality_, you will have re-invented `message-y-or-n-p'.
So what'd be the point, especially when a quick (eval-when-compile
(autoload macro)) will do the job just as well?


-- 
|---<Steve Youngs>---------------<GnuPG KeyID: A94B3003>---|
|              Ashes to ashes, dust to dust.               |
|      The proof of the pudding, is under the crust.       |
|------------------------------<sryoungs@bigpond.net.au>---|

[-- Attachment #2: Type: application/pgp-signature, Size: 256 bytes --]

Re: [PATCH] Autoload `message-y-or-n-p' macro in gnus.el at compile

[Reiner Steib]

On Thu, Jan 15 2004, Steve Youngs wrote:

> By the time you "fixed" `gnus-read-group', _maintaining it's current
> intended functionality_, you will have re-invented `message-y-or-n-p'.
> So what'd be the point, especially when a quick (eval-when-compile
> (autoload macro)) will do the job just as well?

I don't know which solution is preferable:

(1) Your suggestion:

    (eval-when-compile
      (autoload 'message-y-or-n-p "message" nil 'macro))

(2) Adding `message-y-or-n-p' to the existing `eval-and-compile' form:

,----
| ;; Define some autoload functions Gnus might use.
| (eval-and-compile
| 
|   ;; This little mapcar goes through the list below and marks the
|   ;; symbols in question as autoloaded functions.
|   [...]
`----

--8<---------------cut here---------------start------------->8---
--- gnus.el.~7.6.~	Sun Jan 11 18:44:40 2004
+++ gnus.el	Thu Jan 15 13:11:48 2004
@@ -2556,7 +2556,7 @@
      ("qp" quoted-printable-decode-region quoted-printable-decode-string)
      ("ps-print" ps-print-preprint)
      ("message" :interactive t
-      message-send-and-exit message-yank-original)
+      message-send-and-exit message-yank-original message-y-or-n-p)
      ("babel" babel-as-string)
      ("nnmail" nnmail-split-fancy nnmail-article-group)
      ("nnvirtual" nnvirtual-catchup-group nnvirtual-convert-headers)
--8<---------------cut here---------------end--------------->8---

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

Re: [PATCH] Autoload `message-y-or-n-p' macro in gnus.el at compile

[Simon Josefsson]

Steve Youngs <sryoungs@bigpond.net.au> writes:

> * Simon Josefsson <jas@extundo.com> writes:
>
>   > Steve Youngs <sryoungs@bigpond.net.au> writes:
>   >> Over at XEmacs.org, they've got a couple of awk scripts to help them
>   >> find mis-compiled macros and stuff.  It turned up this
>   >> one... `message-y-or-n-p', which is defined in message.el isn't
>   >> getting picked up in gnus.el (where it's used in `gnu-read-group').
>
>   > I think gnus-read-group should be fixed.  There is a `gnus-y-or-n-p'.
>
> Who says it's broken?  And besides, `gnus-y-or-n-p' and
> `message-y-or-n-p' are _very_ different animals.  The former is not
> much more that `y-or-no-p' with a different name, but the latter has
> the option of displaying a "help" buffer as well as the prompt.

Ah, sorry.  Then I agree with your patch, but shouldn't there be
another nil in there?  (autoload 'message-y-or-n-p "message" nil
'macro) makes the INTERACTIVE argument 'macro, which seems weird.

(autoload FUNCTION FILE &optional DOCSTRING INTERACTIVE TYPE)
Fifth arg TYPE indicates the type of the object:
   nil or omitted says FUNCTION is a function,
   `keymap' says FUNCTION is really a keymap, and
   `macro' or t says FUNCTION is really a macro.

>   >> Anyway, I was just about to commit this when I had a thought (yep,
>   >> occasionally I do have em).  Should this patch go into gnus.el, or
>   >> into dgnushack.el?
>
>   > If gnus-read-group is fixed, is it still required?
>
> By the time you "fixed" `gnus-read-group', _maintaining it's current
> intended functionality_, you will have re-invented `message-y-or-n-p'.
> So what'd be the point, especially when a quick (eval-when-compile
> (autoload macro)) will do the job just as well?

None, I was mislead by the function names.  I assumed message-y-or-n-p
and gnus-y-or-n-p was yet another duplicated function definition.

OTTH, I guess the functionality provided by message-y-or-n-p might be
a candidate for moving into a "utilities" library, it seems generally
useful.  Especially now that both Message and Gnus use it.

Re: [PATCH] Autoload `message-y-or-n-p' macro in gnus.el at compile

[Steve Youngs]

[-- Attachment #1: Type: text/plain, Size: 628 bytes --]

* Simon Josefsson <jas@extundo.com> writes:

  > Ah, sorry.  Then I agree with your patch, but shouldn't there be
  > another nil in there?  (autoload 'message-y-or-n-p "message" nil
  > 'macro) makes the INTERACTIVE argument 'macro, which seems weird.

Yep, you're right.  I must have been in too much of a hurry when I
sent the patch.  A correct patch has been applied to CVS.

-- 
|---<Steve Youngs>---------------<GnuPG KeyID: A94B3003>---|
|              Ashes to ashes, dust to dust.               |
|      The proof of the pudding, is under the crust.       |
|------------------------------<sryoungs@bigpond.net.au>---|

[-- Attachment #2: Type: application/pgp-signature, Size: 256 bytes --]

Re: [PATCH] Autoload `message-y-or-n-p' macro in gnus.el at compile

[Xavier Maillard]

[-- Attachment #1: Type: text/plain, Size: 449 bytes --]

On 15 Jan 2004, Simon Josefsson said:

 <snip>

> OTTH, I guess the functionality provided by message-y-or-n-p might be
> a candidate for moving into a "utilities" library, it seems generally
> useful.  Especially now that both Message and Gnus use it.

Yep I agree with that and since we have discussed about that before, it
makes me think, it is more than needed ;)

zeDek
-- 
.o.   Xavier Maillard            Tel: +33 6 68 04 64 37
..o  
ooo   


[-- Attachment #2: Type: application/pgp-signature, Size: 188 bytes --]