[Pan-users] Re: Re: Re: odd newsgroup behaviour

[Duncan]

R Kimber posted <address@hidden>, excerpted
below,  on Fri, 01 Jul 2005 17:21:00 +0100:

> I've discovered that fetchnews sometimes complains, e.g.
> article /var/spool/news/uk/politics/misc/362 contained illegal headers:
> References: does not end with ">"
> I suppose this is a server error (?)
> 
> I don't know if this is the same as the previous, Pan, error but anyway
> I changed to Knode and this seems to be working OK.

That might indeed do it.  Your PAN error was the references header, so if
it's not ending as it's supposed to, PAN might be parsing the next line as
a "folded" continuance of the same header, and therefore coming up with
unexpected characters.  ... Or... it could simply be balking at the
unexpected CRLF sequence, when the next line is obviously a new header,
not a folded continuation.

It'd be interesting to trace down what useragent was responsible for the
problem, and possibly file a bug with the developers.  We all like to bash
OE, generally for good reason, but if it was OE except in some extreme
corner case, there should be a /lot/ more of them, given how common OE is.
While OE might not be perfect with usenet conventions, it DOES do fairly
good on standards-track RFC compliance (at least in what it sends, there's
at least one corner case where it's not fully compliant in how it
interprets server replies as it downloads, but that's different).  That's
one thing in its favor.  Actually, that's one of the reasons I know what I
know about the news RFCs, and indeed, RFCs in general, because a long time
ago in a far away virtual land (before I defected from slaveryware to the
land of freedom software), I used OE, and ended up doing some research on
how it treated line endings, in MIME/Quoted-printable output mode, due to
someone claiming it wasn't RFC compliant, but I was able to prove it was.

Anyway, now's not the best time to file a PAN bug due to the fact it's
under major rewrite right now, but knowing what client is causing the
issue with what it sends would give Charles and Chris something to look
at, and test their new version against once it begins to settle down to
the point where testing is again sensible.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman in
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html