weird error message about X-Face:

weird error message about X-Face:

[Didier Verna]

        Hi !

I'm suddenly (after an upgrade) getting a strange error message when posting
news. Gnus tells me

"Couldn't send message via news: 441 437 No colon-space in X-Face: header".

The *nntpd* buffer however contains this:

340 Ok, recommended ID <9hvet6$nh8$4@kualalumpur.lrde.epita.fr>

but the news can't be posted. I have to remove my X-Faces :-(

This seems unrelated to my recent patch against nntp: I get the same behavior
with former nntp.el versions. Having only one X-Face makes no difference. My
X-Face settings are exactly the same as in this message.

Ideas ?

--
Didier Verna, didier@lrde.epita.fr, http://www.lrde.epita.fr/~didier

EPITA / LRDE, 14-16 rue Voltaire   Tel.+33 (1) 53 14 59 47
94276 Le Kremlin-Bicêtre, France   Fax.+33 (1) 44 08 01 99   didier@xemacs.org

Re: weird error message about X-Face:

[Kai Großjohann]

On Wed, 04 Jul 2001, Didier Verna wrote:

> "Couldn't send message via news: 441 437 No colon-space in X-Face:
> header".

I also get this.  (Got it before I removed X-Face.)

kai
-- 
~/.signature: No such file or directory

Re: weird error message about X-Face:

[Simon Josefsson]

Didier Verna <didier@lrde.epita.fr> writes:

>         Hi !
> 
> I'm suddenly (after an upgrade) getting a strange error message when posting
> news. Gnus tells me
> 
> "Couldn't send message via news: 441 437 No colon-space in X-Face: header".
> 
> The *nntpd* buffer however contains this:
> 
> 340 Ok, recommended ID <9hvet6$nh8$4@kualalumpur.lrde.epita.fr>
> 
> but the news can't be posted. I have to remove my X-Faces :-(
> 
> This seems unrelated to my recent patch against nntp: I get the same behavior
> with former nntp.el versions. Having only one X-Face makes no difference. My
> X-Face settings are exactly the same as in this message.
> 
> Ideas ?

Is it related to this one?

2001-06-24  Simon Josefsson  <jas@extundo.com>

	* rfc2047.el (rfc2047-fold-region): The check to skip WSP
	insertion when breaking lines looked for " \t" instead of "[ \t]".
	(rfc2047-encode-message-header): Fold lines even if
	no QP encoding is done.

If not, maybe you could try nailing it down by going backwards in the
CVS history to see what patch caused it.

Re: weird error message about X-Face:

[Colin Walters]

Simon Josefsson <jas@extundo.com> writes:

> Is it related to this one?
> 
> 2001-06-24  Simon Josefsson  <jas@extundo.com>
> 
> 	* rfc2047.el (rfc2047-fold-region): The check to skip WSP
> 	insertion when breaking lines looked for " \t" instead of "[ \t]".
> 	(rfc2047-encode-message-header): Fold lines even if
> 	no QP encoding is done.

I believe so.  The following patch seems to fix the problem for me.

As I see it, the problem was that rfc2047-fold-region was destroying
the space following the colon in the header.  I think it makes more
sense to just fold the header _value_.  Just (skip-chars-forward "^:")
is sufficient to skip over the header name, right?

2001-07-05  Colin Walters  <walters@cis.ohio-state.edu>

	* rfc2047.el (rfc2047-encode-message-header): Don't include the
	header name when folding.

--- rfc2047.el.~6.14.~	Fri Jun 29 02:51:27 2001
+++ rfc2047.el	Thu Jul  5 17:30:11 2001
@@ -138,7 +138,8 @@
 		     (point-min) (point-max)
 		     (car message-posting-charset)))
 		;; No encoding necessary, but folding is nice
-		(rfc2047-fold-region (point-min) (point-max)))
+		(rfc2047-fold-region (save-excursion
+				       (skip-chars-forward "^:")) (point-max)))
 	    ;; We found something that may perhaps be encoded.
 	    (setq method nil
 		  alist rfc2047-header-encoding-alist)

Re: weird error message about X-Face:

[Colin Walters]

Colin Walters <walters@cis.ohio-state.edu> writes:

> Just (skip-chars-forward "^:") is sufficient to skip over the header
> name, right?

Nope.  I think this patch actually works, but it's uglyyy...maybe
someone has a better solution.

--- rfc2047.el.~6.14.~	Fri Jun 29 02:51:27 2001
+++ rfc2047.el	Thu Jul  5 18:02:37 2001
@@ -138,7 +138,16 @@
 		     (point-min) (point-max)
 		     (car message-posting-charset)))
 		;; No encoding necessary, but folding is nice
-		(rfc2047-fold-region (point-min) (point-max)))
+		(rfc2047-fold-region (save-excursion
+				       ;; Skip over the header.
+				       (goto-char (point-min))
+				       (skip-chars-forward "^:")
+				       (unless (eobp)
+					 (forward-char))
+				       (unless (eobp)
+					 (forward-char))
+				       (point))
+				     (point-max)))
 	    ;; We found something that may perhaps be encoded.
 	    (setq method nil
 		  alist rfc2047-header-encoding-alist)

Re: weird error message about X-Face:

[Simon Josefsson]

Colin Walters <walters@cis.ohio-state.edu> writes:

> I believe so.  The following patch seems to fix the problem for me.

Thanks.  I massaged it a little and committed it.

Re: weird error message about X-Face:

[Paul Jarc]

Colin Walters <walters@cis.ohio-state.edu> writes:
>+				       ;; Skip over the header.
>+				       (goto-char (point-min))
>+				       (skip-chars-forward "^:")
>+				       (unless (eobp)
>+					 (forward-char))
>+				       (unless (eobp)
>+					 (forward-char))

This skips over the header field name, the colon, and space, right?
Technically, the space is part of the field's value, and it need not
be present.  So it's best to check for a space before skipping it.  (I
haven't been paying much attention to this thread, though, so maybe we
know that it will always be present here?)


paul

Re: weird error message about X-Face:

[Simon Josefsson]

prj@po.cwru.edu (Paul Jarc) writes:

> Colin Walters <walters@cis.ohio-state.edu> writes:
>>+				       ;; Skip over the header.
>>+				       (goto-char (point-min))
>>+				       (skip-chars-forward "^:")
>>+				       (unless (eobp)
>>+					 (forward-char))
>>+				       (unless (eobp)
>>+					 (forward-char))
> 
> This skips over the header field name, the colon, and space, right?
> Technically, the space is part of the field's value, and it need not
> be present.  So it's best to check for a space before skipping it.  (I
> haven't been paying much attention to this thread, though, so maybe we
> know that it will always be present here?)

Right, something along that line has been committed.