Re: message.el dependency in mm-uu.el

[Reiner Steib <reinersteib+gmane@imap.cc>]

[ Cc-ing ding ]

On Sat, Dec 10 2005, Reiner Steib wrote:

> On Fri, Dec 09 2005, Romain Francoise wrote:
>
>>> +    (message-marks
>>> +     ,(concat "^" (regexp-quote message-mark-insert-begin))
>>> +     ,(concat "^" (regexp-quote message-mark-insert-end))
>>> +     (lambda () (mm-uu-verbatim-marks-extract 0 -1 1 -2))
>>
>> This causes a problem with MH-E, which uses mm-uu.el for one of its
>> functions, but doesn't load message.el first.  See the following Debian
>> bug:
>>
>>     <URL: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=342521>
>>
>> A possible solution would be to make mm-uu.el require message.el.
>> I'm not too comfortable with that because the dependency is needed for a
>> very small part of mm-uu and could probably be avoided.  What do you
>> think?
>
> Thanks for pointing this out.
>
> MML (mml*.el and mm-*.el) is supposed to be independent from Gnus.
> Most probably also independent from message.el (which should in turn
> be independent from Gnus).
>
> I'm trying to document the dependencies in `texi/gnus-coding.texi'.
> Any help with improving this document and remove existing dependencies
> would be very welcome.
>
> The most simple solution would be to hard-code the values of
> message-mark-insert-begin/end (and add a comment to both places
> referring to each other).  Or do you have a better solution in mind?
>
> --- message.el	27 Nov 2005 18:15:17 +0100	7.112
> +++ message.el	10 Dec 2005 00:12:25 +0100	
> @@ -335,6 +335,9 @@
>  
>  ;;; marking inserted text
>  
> +;; If `message-mark-insert-begin' or `message-mark-insert-end' are changed, we
> +;; should also change the `message-marks' element of `mm-uu-type-alist' in
> +;; `mm-uu.el'.
>  (defcustom message-mark-insert-begin
>    "--8<---------------cut here---------------start------------->8---\n"
>    "How to mark the beginning of some inserted text."
>
> --- mm-uu.el	06 Dec 2005 15:47:44 +0100	7.20
> +++ mm-uu.el	10 Dec 2005 00:10:39 +0100	
> @@ -165,8 +165,16 @@
>       nil
>       mm-uu-diff-test)
>      (message-marks
> -     ,(concat "^" (regexp-quote message-mark-insert-begin))
> -     ,(concat "^" (regexp-quote message-mark-insert-end))
> +     ;; Don't use `message-mark-insert-begin' and `message-mark-insert-end'
> +     ;; here to avoid dependency on `message.el'.
> +     ,(concat
> +       "^"
> +       (regexp-quote
> +	"--8<---------------cut here---------------start------------->8---\n"))
> +     ,(concat
> +       "^"
> +       (regexp-quote
> +	"--8<---------------cut here---------------end--------------->8---\n"))
>       (lambda () (mm-uu-verbatim-marks-extract 0 -1 1 -2))
>       nil)
>      ;; Omitting [a-z8<] leads to false positives (bogus signature separators

Maybe the following mm-uu.el patch would be better:

#v+
--- mm-uu.el	06 Dec 2005 15:47:44 +0100	7.20
+++ mm-uu.el	10 Dec 2005 10:38:29 +0100	
@@ -165,8 +165,10 @@
      nil
      mm-uu-diff-test)
     (message-marks
-     ,(concat "^" (regexp-quote message-mark-insert-begin))
-     ,(concat "^" (regexp-quote message-mark-insert-end))
+     ;; Don't use `message-mark-insert-begin' and `message-mark-insert-end'
+     ;; here to avoid dependency on `message.el'.
+     "^-\\{2,\\}[8<>]*-\\{9,\\}[a-z ]+-\\{9,\\}[a-z ]+-\\{9,\\}[8<>]*-\\{2,\\}$"
+     "^-\\{2,\\}[8<>]*-\\{9,\\}[a-z ]+-\\{9,\\}[a-z ]+-\\{9,\\}[8<>]*-\\{2,\\}$"
      (lambda () (mm-uu-verbatim-marks-extract 0 -1 1 -2))
      nil)
     ;; Omitting [a-z8<] leads to false positives (bogus signature separators
#v-

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