duplicate emails etc

2 views
Skip to first unread message

joe mc cool

unread,
May 8, 2000, 3:00:00 AM5/8/00
to
SCO OSE 5.05.

I have to write some scripts that extract data from emails arriving on
the above system.

Things are going well. The files arrive as /usr/spool/mail/username.
But I find that quite often the mails are duplicated. Also the
attachments contain spurious "=" signs.

Yes, the sender is using Outlook Express to send flat files as
attachments. Does this say it all ?

Any comments.
--
joe mc cool
========================================================================
Tangent Computer Research BT71 7LN (www.tangent-research.com)
voice:044-2837-548074fax:(44)-870-0520185 The more you say the less the better.


Jeff Liebermann

unread,
May 8, 2000, 3:00:00 AM5/8/00
to
On 08 May 2000 22:38:50 +0000, joe mc cool <j...@benburb.demon.co.uk> wrote:

>SCO OSE 5.05.

Yet another mutated per-version. That should be "The Santa Cruz Operation,
Open Server Enterprise 3.2v5.0.5".

>I have to write some scripts that extract data from emails arriving on
>the above system.
>
>Things are going well. The files arrive as /usr/spool/mail/username.
>But I find that quite often the mails are duplicated. Also the
>attachments contain spurious "=" signs.
>
>Yes, the sender is using Outlook Express to send flat files as
>attachments. Does this say it all ?

1. The duplicated messages are feature of some versions of Windoze Outlook
and Outlook Express. If the outgoing SMTP mail server does not produce a
recognizeable response to successfully send an outgoing message, Outlook
will continue to send and resend even though the message has been delivered.
MS thinks this is a feature, not a bug. The usual problem is that Outlook
times out before the SMTP server returns a success message.

2. The surplus = signs and =20 crap at the end of lines are the effects of
sending a message as a MIME attachment instead of plain text. Since
Microsoft does not control the consortium or standards group that defines
plain text, users of Microsoft products are discouraged from using plain
text. Instead, HTML, MIME, uuencode, or PDFicated text is the norm. Be
thankful that it wasn't sent in one of the numerous RichText mutations.


--
Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
(831)421-6491 pgr (831)426-1240 fax (831)336-2558 home
http://www.cruzio.com/~jeffl WB6SSY
je...@comix.santa-cruz.ca.us je...@cruzio.com

Jean-Pierre Radley

unread,
May 8, 2000, 3:00:00 AM5/8/00
to
Jeff Liebermann propounded (on Mon, May 08, 2000 at 03:24:37PM -0700):

| On 08 May 2000 22:38:50 +0000, joe mc cool <j...@benburb.demon.co.uk> wrote:
|
| >SCO OSE 5.05.
|
| Yet another mutated per-version. That should be "The Santa Cruz Operation,
| Open Server Enterprise 3.2v5.0.5".

Not really 'Open Server', two words, was correct prior to the release of
'OpenServer 5'


--
JP

Jeff Liebermann

unread,
May 8, 2000, 3:00:00 AM5/8/00
to

Sorry. I was a bit spacey.

Actually, it's:
"The Santa Cruz Operation, Inc., OpenServer Enterprise 3.2v5.0.5"
I forgot the "Inc.". See:
http://www.sco.com/copyright/
What's an "SCO Nihongo OpenServer"? Never mind. I don't wanna know.

Also, not to be confused with OpenServe products from eLinux.com:
http://www.elinux.com

Mike Kenyon

unread,
May 9, 2000, 3:00:00 AM5/9/00
to
Jeff Liebermann wrote:
> Actually, it's:
> "The Santa Cruz Operation, Inc., OpenServer Enterprise 3.2v5.0.5"

I'm sure you've missed an "r4" in there somewhere...

Is there any logic to version numbers nowadays?

--
Mike Kenyon <mke...@promtek.com> Software Engineer for Promtek Ltd

Bill Vermillion

unread,
May 9, 2000, 3:00:00 AM5/9/00
to
In article <3917D9AD...@promtek.com>,

Mike Kenyon <mke...@promtek.com> wrote:
>Jeff Liebermann wrote:
>> Actually, it's:
>> "The Santa Cruz Operation, Inc., OpenServer Enterprise 3.2v5.0.5"

>I'm sure you've missed an "r4" in there somewhere...

>Is there any logic to version numbers nowadays?

You mean like SCO moving from Unixware 2.x to Unixware 7, so Sun
moved from Solaris 2.6 to Solaris 7 and Solaris 8? I wonder if the
marketing people really think users are that dumb.

--
Bill Vermillion bv @ wjv.com

Jeff Liebermann

unread,
May 9, 2000, 3:00:00 AM5/9/00
to
On Tue, 09 May 2000 10:26:05 +0100, Mike Kenyon <mke...@promtek.com> wrote:

>Jeff Liebermann wrote:
>> Actually, it's:
>> "The Santa Cruz Operation, Inc., OpenServer Enterprise 3.2v5.0.5"
>
>I'm sure you've missed an "r4" in there somewhere...

Nope. Sys V Release 4 probably sounded a bit like an AT&T product. Since
Novell was selling Unixware as an AT&T product, this was deemed a bad thing.
There was also massive confusion when both AT&T and SCO released, Release
4.2 at approximately the same time. Obviously, this was a coincidence.

>Is there any logic to version numbers nowadays?

Yes. SCO is an equal opertunity version numberer. Each department at SCO
gets to add its contribution to the version number. "The Santa Cruz
Operation" was the original name of the company. This is to placate the
traditionalists that refuse to acronymify the name. The word "The" was
prepended to insure that SCO would be lost in alphabetic searches. The
"Inc" was added by the legal department to protect the stockholders and to
give the impression of bigness. The ".," was added by documentation, which
specializes in such detail work.

"OpenServer" as mutated from "Open Server" was the exemplary contribution of
the marketting department in order to differentiate the product from
previous versions that lacked 13,000 symlinks and actually worked somewhat
better.

The "3.2" is what remains of the original AT&T legal departments
contribution that required all AT&T licensees to adhere to a common version
numbering scheme while AT&T learned to abuse Roman numerals. At the time,
to avoid litigation and to differentiate the SCO version of Unix from the
AT&T incantation, SCO began using the term "version" instead of "release"
thus initiating the use of the "v". The 5.0.5 is the contribution of
product development which translates into:
Version version 5
Major Release 0
Minor Release 5
Please note that only programmers count starting at zero instead of one.
Marketting is not infested with programmers so you are assured that we will
never see "Version 0" or any such abomination. However, since programmers
are in control of the rest of the numbering, the major and minor release
numbers start at zero.

The actual version number is 3.2v5.0.5Eb as disgorged by various obscure
utilities. The "Eb" is the contribution of the support department. The
exact meaning is a bit obscure as support is apparently required to use
letters of the alphabet instead of numbers, which are the exclusive domain
of marketting and development. I suspect it really stands for the initials
of someones name, but I cannot be certain.

Note how all the various departments at SCO, work harmoniously together to
produce an unworkable conglomeration. However, this is considerably better
than if any one department were in control. For example, if marketting
gained an upper hand, the product and possibly the company would derive it's
name from something something generated by the allegedly pronounceable
password generator and end in a vowel. If engineering were in control, the
company name would be an acronym of other acronyms, and the version number
would look like an SNMP OID. Were support in control, it would probably be
something like "Unix, 05/01/2000 edition". While the current name and
numbers are a bit awkward, they are significantly better than the potential
alternatives.

Nic Sarginson

unread,
May 10, 2000, 3:00:00 AM5/10/00
to
The Unixware 7 numbering is arrived at because it is the way forward
from OpenServer 5 and UnixWare 2, i.e 5+2 = 7

Sad but true

Cheers

Nic

-----Original Message-----
From: bi...@wjv.com.REMOVEME (Bill Vermillion)
[mailto:bi...@wjv.com.REMOVEME]
Posted At: Tue 09 May 2000 15:12
Posted To: misc
Conversation: duplicate emails etc
Subject: Re: duplicate emails etc


In article <3917D9AD...@promtek.com>,


Mike Kenyon <mke...@promtek.com> wrote:
>Jeff Liebermann wrote:
>> Actually, it's:
>> "The Santa Cruz Operation, Inc., OpenServer Enterprise 3.2v5.0.5"

>I'm sure you've missed an "r4" in there somewhere...

>Is there any logic to version numbers nowadays?

You mean like SCO moving from Unixware 2.x to Unixware 7, so Sun

Bill Vermillion

unread,
May 10, 2000, 3:00:00 AM5/10/00
to
In article <E47927FC0A9CD3119D090000216DAFF50A0AD8@thoth>,

Nic Sarginson <ni...@amonra-insight.com> wrote:
>The Unixware 7 numbering is arrived at because it is the way forward
>from OpenServer 5 and UnixWare 2, i.e 5+2 = 7

>Sad but true

Hm. Well since Sun and SCO appear to be running "My version is
higher than yours" now that Solaris 8 is out does Unixware go
to 9 or does it add 7 to 8 and come up with 15 :-)

Of course that mean that HP will have to rename theirs - since that
would be larger than their 10.

Reminds me of the Detroit horsepower wars of many year ago.

Roger Binns

unread,
May 10, 2000, 3:00:00 AM5/10/00
to
Bill Vermillion <bi...@wjv.com.removeme> wrote:

: Of course that mean that HP will have to rename theirs - since that


: would be larger than their 10.

Except they been at revision 11 for quite a while. For a while with
the 10.x series they made each point release binary incompatible
if your program did password checking. At one point we had to
ship 4 different binaries for VisionFS. That of course meant
that our supported platform list had rather a lot of HP/UX listed.

Excuse me while I return to my email petition for Tarantella
Infinity plus 17.

Roger
--
Roger Binns rog...@sco.com http://tarantella.sco.com
Business Development Manager, Tarantella
SCO, Vision Park, Cambridge, CB4 9ZR, United Kingdom
Tel +44 1223 518046 Fax +44 1223 518001

Charlie Gibbs

unread,
May 10, 2000, 3:00:00 AM5/10/00
to
In article <u5eghs0u03cjkq69e...@4ax.com>
je...@comix.santa-cruz.ca.us (Jeff Liebermann) writes:

>The actual version number is 3.2v5.0.5Eb as disgorged by various
>obscure utilities. The "Eb" is the contribution of the support
>department. The exact meaning is a bit obscure as support is
>apparently required to use letters of the alphabet instead of
>numbers, which are the exclusive domain of marketting and development.

However, the use of a trailing lower-case letter suggests a
possible influence from marketing (cf. BMW model numbers).

>I suspect it really stands for the initials of someones name,
>but I cannot be certain.

He's a musician. While working on this version he heard something
in E flat.

--
cgi...@sky.bus.com (Charlie Gibbs)
Remove the first period after the "at" sign to reply.


Geoff Johnson

unread,
May 11, 2000, 3:00:00 AM5/11/00
to
Roger Binns wrote:
>
> Bill Vermillion <bi...@wjv.com.removeme> wrote:
>
> : Of course that mean that HP will have to rename theirs - since that
> : would be larger than their 10.
>
> Except they been at revision 11 for quite a while. For a while with
> the 10.x series they made each point release binary incompatible
> if your program did password checking. At one point we had to
> ship 4 different binaries for VisionFS. That of course meant
> that our supported platform list had rather a lot of HP/UX listed.
>

If you got your passwd checking working think yourself very lucky.
We had a 15000 user /etc/passwd file (nearly 1 MB) that would end up
scrambled or 0MB long several times a month.
NIS on HPUX sucks big time.

> Excuse me while I return to my email petition for Tarantella
> Infinity plus 17.
>
> Roger
> --
> Roger Binns rog...@sco.com http://tarantella.sco.com
> Business Development Manager, Tarantella
> SCO, Vision Park, Cambridge, CB4 9ZR, United Kingdom
> Tel +44 1223 518046 Fax +44 1223 518001

--

Geoff Johnson

Jeff Liebermann

unread,
May 11, 2000, 3:00:00 AM5/11/00
to
On 10 May 00 09:04:02 -0800, "Charlie Gibbs" <cgi...@sky.bus.com> wrote:

>However, the use of a trailing lower-case letter suggests a
>possible influence from marketing (cf. BMW model numbers).

Long ago, some human factors research book I read in college mentioned that
one should end nomenclature with a letter that is pronounced while smiling.
It works, which explains the popularity of model numbers ending in B C E G P
T V Z letters.

The trailing vowel has a similar effect on diction causing the speaker to
pause momentarily on the trailing vowel. This gives more "air time" and
attention to the product name. This may not seem like much of an advantage
until you consider that most display advertisements are considered
successful if they get a total of 5 seconds of reader attention. Really
great ads only get 10 seconds. The trailing vowel is also fairly easy to
pronounce when compared to various Slavic terms that are grossly deficient
in vowels.

A recent thread in another newsgroup bemoaned the demise of properly named
companies and products and suggested that the rediculously creative company
names, that are currently in fashion, is the result of copyright, trademark,
and imagination deficiency. The former is not true as a quick search of the
Patent and Trademark Bureaucracy database at:
http://trademarks.cnidr.org/access/search-mark.html
for the term "Unix" will reveal manufacturers of baby diapers, sunglasses,
steel, and respiratory pharmaceuticals that use the Unix trademark. (For
additional entertainment, try "bunnypeople" to see what Intel had in mind).
For those with a lack of imagination, it was noted that the VMS allegedly
pronounceable password generator produced words with an uncanny semblance to
some of the new Internet company names. This is obviously a coincidence as
a market research company that charges $60,000 per name, would never stoop
to such methods.

>>I suspect it really stands for the initials of someones name,
>>but I cannot be certain.

>He's a musician. While working on this version he heard something
>in E flat.

So that's where the Eb came from. It probably goes well with the major and
minor revision levels in an equal tempered licensing scheme. However, I
doubt that it was the result of hearing music in Eb. SCO is both tone and
feature request deaf. There were some hints that the next release would
include background music while waiting for Custom+ to install the operating
system.

joe mc cool

unread,
May 11, 2000, 3:00:00 AM5/11/00
to
Sorry to try all your patience:

Duplicate emails (sometimes with the same date and time on the
header), are arriving on my customer's OSR5 3.2v5.0.5 box.

Can I be absolutely positive that this problem is not being caused by
the receiving end (ie the SCO box). How can I reasure my customer ?

The sender is adamant that when she sends to another (admittedly
Windoze) box, that everything is ok and there is no duplication. "The
problem _must_ be at your end", she says !

joe mc cool

unread,
May 12, 2000, 3:00:00 AM5/12/00
to

> >But I find that quite often the mails are duplicated. Also the
> >attachments contain spurious "=" signs.

First problem solved:

The ISP concerned has admitted that they were forwarding the mail
twice !

My customer has a 2X64K link. As each link came up the ISP was
throwing it the mail !

Ye Gods !

--
joe mc cool SMIEEE


========================================================================
Tangent Computer Research BT71 7LN (www.tangent-research.com)

voice:(44)2837-548074fax:(44)-870-0520185 The more you say the less the better.

Jeff Liebermann

unread,
May 12, 2000, 3:00:00 AM5/12/00
to
On 11 May 2000 19:27:07 +0000, joe mc cool <j...@benburb.demon.co.uk> wrote:

>Sorry to try all your patience:

Grumble, growl, snarl, etc...

>Duplicate emails (sometimes with the same date and time on the
>header), are arriving on my customer's OSR5 3.2v5.0.5 box.
>
>Can I be absolutely positive that this problem is not being caused by
>the receiving end (ie the SCO box). How can I reasure my customer ?

If the duplicated email is coming from multiple sources and servers, then
it's highly likely that the duplication is being cause locally on the OSR5
box. However, it all the duplicates seem to come from one user or one ISP,
then it's probably not the OSR5 box.

One of the more entertaining way of simulating the duplicate problem is if
your customer has "leave messages on server" set on their mail reader. One
version of Eurora (3.03???) would show every message twice when polled for
the 2nd time. Therefore, look at the mail spool (/usr/spool/mail/username)
and check if there really are duplicated messages.

>The sender is adamant that when she sends to another (admittedly
>Windoze) box, that everything is ok and there is no duplication. "The
>problem _must_ be at your end", she says !

The first step to solving a problem is to assign the blame. Make this user
part of the solution. Ask her to send email to herself and see if it shows
as a duplicate. Have her poll for email twice. If the test message shows
up twice, it must be her mail clients fault.

If this user has a clue, ask her to run:

telnet mail.cruzio.com 110
+OK QPOP (version 2.52) at mail.cruzio.com starting.
user jeffl
+OK Password required for jeffl.
pass xxxxxx
+OK jeffl has 0 messages (0 octets).
list
+OK 0 messages (0 octets)
quit

Argh. I just polled for email and don't have any messages waiting. Anyway,
the list command will show the number and size of each message. If there's
a bunch of duplicates, they will show up as pairs of identical messages. If
several hundred messages magically appear, she's been leaving her email on
the server and has a broken mail reader.

Lukasz Wiechec

unread,
May 14, 2000, 3:00:00 AM5/14/00
to
Bill Vermillion <bi...@wjv.com.removeme> wrote:
: In article <E47927FC0A9CD3119D090000216DAFF50A0AD8@thoth>,

: Nic Sarginson <ni...@amonra-insight.com> wrote:
:>The Unixware 7 numbering is arrived at because it is the way forward
:>from OpenServer 5 and UnixWare 2, i.e 5+2 = 7

:>Sad but true

: Hm. Well since Sun and SCO appear to be running "My version is
: higher than yours" now that Solaris 8 is out does Unixware go
: to 9 or does it add 7 to 8 and come up with 15 :-)

: Of course that mean that HP will have to rename theirs - since that


: would be larger than their 10.

: Reminds me of the Detroit horsepower wars of many year ago.

What horsepower wars ?

--
------------------------
MAGIC aka Lukasz Wiechec
Internet mailto:ma...@bet.po.opole.pl
for PGP public key: finger ma...@bet.po.opole.pl
tel/fax +(48 77) 555 113

Bill Vermillion

unread,
May 15, 2000, 3:00:00 AM5/15/00
to
In article <09ilf8...@ns.linux.home>,

Lukasz Wiechec <ma...@bet.po.opole.pl> wrote:
>Bill Vermillion <bi...@wjv.com.removeme> wrote:
>: In article <E47927FC0A9CD3119D090000216DAFF50A0AD8@thoth>,
>: Nic Sarginson <ni...@amonra-insight.com> wrote:
>:>The Unixware 7 numbering is arrived at because it is the way forward
>:>from OpenServer 5 and UnixWare 2, i.e 5+2 = 7
>
>:>Sad but true
>
>: Hm. Well since Sun and SCO appear to be running "My version is
>: higher than yours" now that Solaris 8 is out does Unixware go
>: to 9 or does it add 7 to 8 and come up with 15 :-)
>
>: Of course that mean that HP will have to rename theirs - since that
>: would be larger than their 10.

>: Reminds me of the Detroit horsepower wars of many year ago.

>What horsepower wars ?

Maybe I should have said many many years ago.

Each years more horsepower, more cubic inches. The first oil
crisis and bad press reigned it in.

Seven liter engines were quite common then - over 400 cubic inches
for the metrically impaired. 350-400 horsepower from the showroom
floor was possible. 5 to 6 miles per gallon too.

Pre-pollution control, pre-everything. Pre-mandatory seat belts.
Crash helmets for race drivers were optional. In those days sex was
safe and racing was dangerous.

If a manufacturer did nothing else they added more horsepower or
more cubic inches than the preceding year. Most topped out in
the 427cu/in range

Reply all
Reply to author
Forward
0 new messages