Problems with to-address

Problems with to-address

[Norbert Koch]

Hi!

- Gnus v5.10.3
- 21.4 (patch 13) "Rational FORTRAN" XEmacs Lucid

Is anybody else seeing this, too?  I can't 'F'ollow-up to a message in
groups that have to-address set in their group parameters:

,----
| ((timestamp 16085 49371)
|  (to-address . "xemacs-winnt@xemacs.org")
|  (to-list . "xemacs-winnt@xemacs.org")
|  (subscribed . t)
|  (total-expire . t)
|  (expiry-wait . 30)
|  (gcc-self . none))
`----

The error message is

Signaling: (void-variable to-address)
  gnus-post-news(nil "nnml:xemacs-winnt" [73 "Re: \"packages/sasl-1.12-pkg.tar.gz too short\" " "Darryl Okahata <darrylo@soco.agilent.com>" "Wed, 28 May 2003 16:58:33 -0700" "<200305282358.QAA01059@mina.soco.agilent.com>" "<vz4r3ergtj.fsf@viteno.dyns.net>" 600 18 "viteno.dyns.net xemacs-winnt:73" ((To . "Norbert Koch <viteno@xemacs.org>"))] "*Article*" (73) nil nil)
  gnus-summary-followup((73) nil)
  gnus-summary-followup-with-original(nil)
  call-interactively(gnus-summary-followup-with-original)

which I don't really understand, because to my watering eyes
to-address is bound in gnus-post-news.

Thanks,
norbert.

Re: Problems with to-address

[Norbert Koch]

Norbert Koch <nk@viteno.net> writes:

Hi!

I've edebugged the function and now, I don't have by any means more
clues :-(

It seems, the behaviour does not depend on the fact that to-address is
set.  I can reproduce this with any group.

[...]
> The error message is
>
> Signaling: (void-variable to-address)
>   gnus-post-news(nil "nnml:xemacs-winnt" [73 "Re: \"packages/sasl-1.12-pkg.tar.gz too short\" " "Darryl Okahata <darrylo@soco.agilent.com>" "Wed, 28 May 2003 16:58:33 -0700" "<200305282358.QAA01059@mina.soco.agilent.com>" "<vz4r3ergtj.fsf@viteno.dyns.net>" 600 18 "viteno.dyns.net xemacs-winnt:73" ((To . "Norbert Koch <viteno@xemacs.org>"))] "*Article*" (73) nil nil)
>   gnus-summary-followup((73) nil)
>   gnus-summary-followup-with-original(nil)
>   call-interactively(gnus-summary-followup-with-original)
>
> which I don't really understand, because to my watering eyes
> to-address is bound in gnus-post-news.

But it gets unbound by (gnus-msg.el, rev 6.135, line 947):

	    (set-buffer gnus-article-copy)

Any pointers?
norbert ... bailing out.

Re: Problems with to-address

[Dave Love]

Norbert Koch <viteno@xemacs.org> writes:

> But it gets unbound by (gnus-msg.el, rev 6.135, line 947):
>
> 	    (set-buffer gnus-article-copy)
>
> Any pointers?

I haven't checked the code, but if I understand correctly, that's
expected if XEmacs doesn't have rms and gerd's changes to the binding
mechanism for buffer-local variables.  See Info `(elisp)Intro to
Buffer-Local' in Emacs, though I think that text is obsolete in Emacs
21.

Re: Problems with to-address

[Norbert Koch]

Dave Love <d.love@dl.ac.uk> writes:

> I haven't checked the code, but if I understand correctly, that's
> expected if XEmacs doesn't have rms and gerd's changes to the binding
> mechanism for buffer-local variables.  See Info `(elisp)Intro to
> Buffer-Local' in Emacs, though I think that text is obsolete in Emacs
> 21.

Thanks for this hint.  I'm still wondering why this happens right now,
ie what's been changed in the code base to trigger this effect.

I ask on xemacs-beta.

norbert.