Re: [9fans] Usenet Gateway and Mailing List Archive Woes.

Re: [9fans] Usenet Gateway and Mailing List Archive Woes.

[rob pike]

> The funny thing about the message rob sent was that the attachment
> was labeled as being of type message/rfc822, when in fact it was not.

This is acme/mail's fault. it's got a terrible heuristic for deciding the
type of the attachment.   it's hardly a heuristic at all.  suggestions
welcome (in private mail to rob@plan9.bell-labs.com). what is a file
type in our world anyway?

-rob

Re: [9fans] Usenet Gateway and Mailing List Archive Woes.

[Boyd Roberts]

rob pike wrote:
> This is acme/mail's fault. it's got a terrible heuristic for deciding the
> type of the attachment.

The real problem is the MINEfield (sic) that is the 4 MIME RFCs.

Re: [9fans] Usenet Gateway and Mailing List Archive Woes.

[Lucio De Re]

On Tue, Feb 19, 2002 at 01:05:21PM +0100, Boyd Roberts wrote:
> 
> The real problem is the MINEfield (sic) that is the 4 MIME RFCs.

I don't think MIME is all _that_ bad.  What's better?

++L

Re: [9fans] Usenet Gateway and Mailing List Archive Woes.

[Matt H]

On Tue, 19 Feb 2002 13:05:21 +0100
"Boyd Roberts" <boyd@strakt.com> wrote:

> rob pike wrote:
> > This is acme/mail's fault. it's got a terrible heuristic for deciding
> > the type of the attachment.
> 
> The real problem is the MINEfield (sic) that is the 4 MIME RFCs.
> 

watch out for multipart/related too, it's for HTML
formatted mail but many MUA's don't seem to know about it (as I've
recently find out after making my html email generator produce it. OE5 was
fine with it, OE5.5+ ignores it) 

M

Re: [9fans] Usenet Gateway and Mailing List Archive Woes.

[Boyd Roberts]

Lucio De Re wrote:
> I don't think MIME is all _that_ bad.  What's better?

Read the 4th RFC.  It's not a question of what's better.

Re: [9fans] Usenet Gateway and Mailing List Archive Woes.

[Lucio De Re]

On Tue, Feb 19, 2002 at 01:35:24PM +0100, Boyd Roberts wrote:
> 
> Lucio De Re wrote:
> > I don't think MIME is all _that_ bad.  What's better?
> 
> Read the 4th RFC.  It's not a question of what's better.

Oh, OK, by the third revision they could have lost all sense of
purpose and/or direction.

Still, where does one go from (t)here?

++L

Re: [9fans] Usenet Gateway and Mailing List Archive Woes.

[Boyd Roberts]

Lucio De Re wrote:
> Oh, OK, by the third revision they could have lost all sense of
> purpose and/or direction.

No, they are not revisions.  You have to read _all four_.

By starting with the 4th it makes the exercise a lot quicker
and less painful.

> Still, where does one go from (t)here?

Good question.  Throw it away and start again?

Re: [9fans] Usenet Gateway and Mailing List Archive Woes.

[Ralph Corderoy]

Hi Scott,

> > Jon Snader made me aware that Rob had posted -mpm-related material.
> > I read comp.os.plan9 and didn't see the post.  Google Groups has it
> > without the attachments.
> 
> The funny thing about the message rob sent was that the attachment
> was labeled as being of type message/rfc822, when in fact it was not.
> This confused mh, and I wouldn't be surprised if other things were
> confused too.

Yes, you're right.  That could well be what made some things drop the
attachment and others, perhaps, the whole post.

nmh here has problems, and I heard from Dennis Davis that trn+metamail
doesn't like it either.

It did reach Bakul Shah successfully on Usenet so it looks like the
mail-to-Usenet gateway isn't the problem.  His Path header was

    Path: typhoon.sonic.net!feed.news.sonic.net!unlnews.unl.edu!headwall
    .stanford.edu!newsfeed.stanford.edu!nntp.cs.ubc.ca!logbridge.uoregon
    .edu!server3.netnews.ja.net!south.jnrs.ja.net!server2.netnews.ja.net
    !bath.ac.uk!ccsis

I'll ask my ISP if they got any sign of it and if so, why they didn't
accept.

> Because of that, I mirror the archives at
>       http://bio.cse.psu.edu/~schwartz/9fans/

Fantastic!  I've made a note for future use.  I see it's also in the
FAQ along with the https site.  (I already got the post privately from
Jon Snader.)

> >     I suspect that the web server at lists.cse.psu.edu has been
> >     configured to not accept export-grade cryptography.
> 
> It's configured to be retarded, that's all.
> 
> See, like all broken software, Mailman relies on sending cleartext
> passwords across the net.  So you have to hide the cleartext in
> cryptographic tunnels.  

But you don't *have* to because, in Mailman's case, it makes perfectly
clear the password isn't meant to give any security, just stop a very
casual attack.  To re-inforce this it emails you your password, in
plain text, once a month.  So they've no good reason to hide it for web
accesses AFAICS.

Thanks for the reply,


Ralph.

Re: [9fans] Usenet Gateway and Mailing List Archive Woes.

[Scott Schwartz]

| Jon Snader made me aware that Rob had posted -mpm-related material.  I
| read comp.os.plan9 and didn't see the post.  Google Groups has it
| without the attachments.

The funny thing about the message rob sent was that the attachment
was labeled as being of type message/rfc822, when in fact it was not.
This confused mh, and I wouldn't be surprised if other things were
confused too.

| It's pointless having an archive for a public list hidden behind HTTPS,
| and in addition demanding recent encryption algorithms.  Anyone got a
| public archive of the list?

I agree, and I've complained to the people who run lists.cse.psu.edu,
but to no avail.  Because of that, I mirror the archives at
	http://bio.cse.psu.edu/~schwartz/9fans/

There are three sets of files there.  The ones of the form 9fans.nnnnnn.*
are from the days when Majordomo ran the list.  The ones of the form
nnnn-Month.* are since Mailman (spit!) took over.  To work around a
bug in Mailman, whereby all the "From:" lines are totally wrong, I also
mirror a copy of 9fans.mbox.bz2, which is the raw submissions to the list.

|     I suspect that the web server at lists.cse.psu.edu has been
|     configured to not accept export-grade cryptography.

It's configured to be retarded, that's all.

See, like all broken software, Mailman relies on sending cleartext
passwords across the net.  So you have to hide the cleartext in
cryptographic tunnels.  

[9fans] Usenet Gateway and Mailing List Archive Woes.

[Ralph Corderoy]

Hi,

Jon Snader made me aware that Rob had posted -mpm-related material.  I
read comp.os.plan9 and didn't see the post.  Google Groups has it
without the attachments.

    http://groups.google.com/groups?selm=0ce9a17866461d0a3e0a61628c4149a9%40plan9.bell-labs.com&output=gplain

    Newsgroups: comp.os.plan9
    From: rob@plan9.bell-labs.com (rob pike)
    Subject: Re: [9fans] [troff] Is there a macro package called -mpm ?
    Approved: plan9mod@bath.ac.uk
    Content-Type: multipart/mixed;
            boundary="upas-wnnfbneirzpjzzrtpxqsivpjsg"
    Reply-To: 9fans@cse.psu.edu
    Sender: ccsis@bath.ac.uk (Icarus Sparry)
    Organization: Plan 9 mailing list
    Message-ID: <0ce9a17866461d0a3e0a61628c4149a9@plan9.bell-labs.com>
    Mime-Version: 1.0
    Date: Thu, 14 Feb 2002 17:58:52 GMT
    Lines: 3850


    Enjoy.


    --

My guess is the mailing list to Usenet gateway stripped off the
attachments but didn't adjust the optional Lines header.  My ISP's news
server correctly dropped this post when it received it because it
appears corrupt.  Since no better version ever arrives from another
peer I never see the post on comp.os.plan9.

Can whoever runs the gateway confirm this supposition is correct and do
something about it, e.g. delete any Lines headers since they're
optional.

[Moderator's notes:

    (1) The mailing list to Usenet gateway doesn't strip off
    attachments.  It's only concerned with adjusting headers to
    suit.

    (2) The mailing list to Usenet gateway runs through the News
    system here.  That is neutral as far as the Lines: header is
    concerned.  It doesn't generate one for articles that pass
    through it.  If one is present in the incoming article, it is
    left untouched.

    (3) The mailing list to Usenet gateway is also neutral as far
    as the Lines: header is concerned.  If one is present in the
    incoming message it is passed through.  If one isn't present,
    the s/w doesn't add one.

    In practice none of the above should cause problems.  I don't
    know of of any email client or MTA that adds the Lines header
    to mail messages.  Certainly Outlook Express can differentiate
    between posting a News article and a mail message. It adds a
    Lines: header to the former but not the latter.

    So News articles that are injected from the mailing list
    generally leave here *without* a Lines: header.  Certainly that
    is the case for the article in question.  I believe this header
    was added further downstream of here.  There's not much I can do
    about that.

    As noted in private email I have trouble reading this article
    with the trn Newsreader + metamail-2.7.  I get the error
    message:

    metamail: Incomplete multipart message -- unexpected EOF

    I suspect that this is more likely to be the source of
    confusion.  However the contents should be easily extracted
    manually.]

Knowing it existed on the mailing list I went to the archives.
Unfortunately, http://lists.cse.psu.edu/archives/9fans re-directs you
to a HTTPS connection as has previously been mentioned.  And
https://lists.cse.psu.edu/archives/9fans doesn't work with this
Netscape which complains `Netscape and this server cannot communicate
securely because they have no common encryption algorithm(s)'.

It's pointless having an archive for a public list hidden behind HTTPS,
and in addition demanding recent encryption algorithms.  Anyone got a
public archive of the list?

[Moderator's note:

    I suspect that the web server at lists.cse.psu.edu has been
    configured to not accept export-grade cryptography.  And this
    may be the source of the problem.  Versions of Netscape since
    4.73 have been available with full-strength cryptography from
    the download page at:

    http://www.netscape.com/download/

    Earlier versions of Netscape can be brough up to full-strength
    cryptography by "fortifying" them.  See:

    http://www.fortify.net/

    for details.]

Cheers,


Ralph.