Re: Bug? Extra whitespace in split subject lines in summary buffer

[Katsumi Yamaoka]

>>>>> In <m3ejxq82ge.fsf@kfs-l.imdomain.dk> Kim F. Storm wrote:

> I use Gnus from emacs CVS.

> If an article header contains a split subject line like this (where the second
> line is indented by a TAB):

> Subject: patch: bugfixes in rcirc.el (rcirc-default-user-full-name,
>       rcirc-process-list)

> the summary buffer replaces the TAB by 3-4 spaces, like this:

>     31-May [  84: Daniel Brockman     ] patch: bugfixes in rcirc.el 
> (rcirc-default-user-full-name,   rcirc-process-list)

I often look at such ones in the nntp groups.  It seems to be
due to the INN server, that replaces LWSPs with a couple of
spaces when generating NOV databases.  I believe it never
happens with the other back ends which generate NOV, such as
nnml.  Though it might sometimes happen between ASCII words and
RFC2047-encoded words.

> Is this intentional?

Maybe yes, in the sense that Gnus assumes that data in NOV have
already been cooked, since the summary lines generation should
be fast anyway.  It will be better that Gnus does nothing with
data in NOV if possible, except for RFC2047-decoding.

> Is there any way to fix it.

There are several ways to do that if you don't mind that it
might slow Gnus.  Here are two examples:

;; 1
(setq gnus-decode-encoded-word-function
      (byte-compile
       (lambda (string)
         (replace-regexp-in-string
          "  +" " "
          (mail-decode-encoded-word-string string)))))

;; 2
(eval-after-load "gnus-sum"
  '(progn
     (setcar
      (cdr (assq ?S gnus-summary-line-format-alist))
      `(replace-regexp-in-string
        "  +" " "
        ,(macroexpand '(mail-header-subject gnus-tmp-header))))
     (gnus-update-format-specifications t 'summary)))

> I tried setting gnus-simplify-subject-functions, but it has no effect.

It is used for other purposes: to test whether two subjects are
alike, for instance.