Re: Newsgroups files

Hi Wolfgang,

On E-S (both reader and transit server):
 fr.bienvenue            L'accueil des nouveaux venus dans leurs premiers pas sur Usenet.

This is indeed the right one :)

I seem to remember that this started when the checkgroups messages for
fr.* were changed to UTF-8 (that doesn't account for the "Moderated"
flag, though). Julien Élie might shed some light on this, as he is the
current issuer of control messages for fr.*.

There was a change of encoding in December 2011 when I convinced the former sender of control articles, Ollivier Robert, to switch to UTF-8. Control articles were previously sent using ISO-8859-15 (differing from ISO-8859-1 with a couple of special French characters like "œ", as in œuf, an egg in French, or œil, an eye).
And looking back at my archives, I see that there was no encoding declared at all before January 2009 in fr.* checkgroups.
It doesn't make me any younger, but it corresponds to when I was way more active in INN development and greatly improved, among other things, the handling of control articles, implementing most of the new RFCs. According to the changelog:
   Changes in 2.5.0 (2009-05-21)
     controlchan recognizes the new application/news-groupinfo entity
     described in USEPRO and can handle character set conversions of
     newsgroup descriptions.  The MIME::Parser and Encode modules are
     used.  Processing control messages has been greatly improved,
     especially checkgroups: the active and newsgroups files are now
     properly updated when they are processed, and all matching lines in
     control.ctl for a given checkgroups are honoured (which for instance
     allows using both drop and doit actions for the same checkgroups
     message).
     A new control.ctl.local file has also been added in pathetc.  Rules
     set in that file override rules in control.ctl, allowing
     administrators to specify local rules for some control messages
     without modifying the control.ctl configuration file that comes with
     INN.  It also specifies encodings to use for the newsgroups file.
     By default, UTF-8 will be used for newsgroup descriptions, as
     strongly recommended by RFC 3977.

PS: Please refer to
 http://usenet.trigofacile.com/hierarchies/fr.html
 for the change from "Moderated" to unmoderated and the current content
of the official checkgroups file.

The changes mentioned in the web page explain the 3 different descriptions found by Nigel:
2011-12-19 23:30:02 changegroup fr.bienvenue from m to y (by control article)
2020-12-25 21:50:02 changedesc fr.bienvenue (by checkgroups)
2023-10-28 18:20:02 changedesc fr.bienvenue (by checkgroups)
Before 2011-12-19, when it was moderated:
   Aider les nouveaux venus dans leurs premiers pas sur Usenet. (Moderated)
Then from 2011-12-19 to 2020-12-25:
   Aider les nouveaux venus dans leurs premiers pas sur Usenet.
Then from 2020-12-25 to 2023-10-28:
   Aide aux nouveaux venus dans leurs premiers pas sur Usenet.
Finally, since 2023-10-28:
   L'accueil des nouveaux venus dans leurs premiers pas sur Usenet.
This description only has ASCII characters, thus was not affected by the change from ISO-8859-15 encoding to UTF-8.  There were just wording changes.  The last one in 2023 was to globally homogenize the style of the descriptions in the whole checkgroups (notably with an article: le/la/l').
--
Julien ÉLIE
« Felix qui potuit rerum cognoscere causas. » (Virgile)