GNKSA (Good Net-Keeping Seal of Approval) compliance

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

Hi,

I noticed that Gnus still fails to comply with GNKSA (Good Net-Keeping
Seal of Approval)

The last evaluation of Gnus was in 2000, Gnus 5.8, which failed on
several strict and soft requirements:

,----[ http://www.xs4all.nl/~js/gnksa/Evaluations/gnus-5.8.txt ]
| Regarding the strict GNKSA requirements (MUSTs), Gnus 5.8
| FAILS on the following points:
| 
|     11b  Will not refuse to post articles without subject
|     12b  Fails to refuse to post messages with syntactically incorrect From:
| 
| Gnus 5.8 violates `soft' GNKSA requirements (SHOULDs)
| in the following way:
| 
|     3c   Fails to encourage setting a Followup-To on large crossposts
|     7e   Fails to omit `broken' Message-IDs.
|     16b  Does not refuse posting an empty article
|     16d  Does not refuse to post quoted text only
|     16f  Does not try to prevent posting multiple copies entirely
`----

Checks for several requirements have been added in Oort (5.10), but
current CVS still failed on these,
cf. <http://www.newsreaders.com/gnksa/cgi/gnksa-form.cgi>:

   7)  Make sure followups contain valid References
   e.  Does not propagate broken Message-IDs in original References

  10)  Provides adequate quotation and attribution facilities
   d.  Omits correctly (`-- ') delimited signatures from quoted text by default

  15) Separate signatures correctly, and don't use excessive ones
   b. Warns against or refuses to use excessive signatures 

  16) Tries to prevent obvious user errors
   c. Warns when post articles containing quoted material only
   d. Refuses posting quoted-text-only articles 

In order to comply with 7e, I changed the regexp in
`message-shorten-references' to be more strict.

For 10d, I changed `message-signature-separator' to the strict value
"^-- " and changed `message-cite-function' to
`message-cite-original-without-signature'.

For 15b, there already was `(message-check 'signature ...) in
`message-check-news-body-syntax', but it always passed because of this
bogus test:

      (goto-char (point-max))
      (if (> (count-lines (point) (point-max)) 5)

For 16c and 16 Gnus failed when using posting styles because then the
buffer was marked as modified and the `message-checksum' check in
`message-yank-original' was skipped.

I don't know if anyone still cares about GNKSA compliance, but
nevertheless most of these items are bugs and needed to be fixed.  If
these changes cause any problems, please report.

I intend to backport them to v5-10/Emacs_22 after some testing in the
trunk.  It is probably too late for inclusion in Emacs 22.2 / Gnus
5.10.10, though.

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