Messages posted to multiple mailing lists
Philip Kendall
emul (at) and mia (at) gns.getronics.nl.
[Both these lists are meant to be private, so please don't go around
spamming them...]
A message was posted to both these lists (ie it included the To: line
To: [...], <m...@g.g.n>, <em...@g.g.n>), and this message was then replied
to on the 'mia' list.
However, the original article is apparently available only in the 'emul'
list; in the 'mia' list, it appears as not available (with a '/' sign
when I expand the thread), and if I then try to open it, I get the 'This
article is already available! You can read it immediately' dialog box,
and the article appears. Also, it doesn't appear in the count of unread
articles for the 'mia' list (ie if I mark it as unread, leave Turnpike
and then re-enter, I get one unread article in 'emul' and none in
'mia').
Personally, this isn't too much of a problem for me, as I can always
read the article in 'emul', but if I had a (say) sister, who had a
different Turnpike seat, and was subscribed to 'mia', but not to 'emul',
she wouldn't (AFAICT) ever see the article, unless she happened to try
to open the 'unavailable' message, which I would claim is a bug :-)
The questions:
1) Is this known behaviour? (I can't remember seeing it mentioned
before, but then that isn't saying much...)
2) Is there anything I can do to make this message available by default
in the 'mia' list?
I've had a quick play around with setting up a couple of mailing lists
at kendalls.d.c.u, but can't replicate this behaviour; if no-one else
can, I'll post the headers and/or messages here, but I think this
message is getting long enough as it is...
TIA,
Phil
--
Philip Kendall (pa...@cam.ac.uk pa...@kendalls.demon.co.uk)
John Woodgate
>To: [...], <m...@g.g.n>, <em...@g.g.n>), and this message was then replied
>to on the 'mia' list.
>
>However, the original article is apparently available only in the 'emul'
>list; in the 'mia' list, it appears as not available (with a '/' sign
>when I expand the thread), and if I then try to open it, I get the 'This
>article is already available! You can read it immediately' dialog box,
>and the article appears. Also, it doesn't appear in the count of unread
>articles for the 'mia' list (ie if I mark it as unread, leave Turnpike
>and then re-enter, I get one unread article in 'emul' and none in
>'mia').
>
>
>1) Is this known behaviour? (I can't remember seeing it mentioned
>before, but then that isn't saying much...)
Probably not, but the situation is a somewhat rare one, I suppose.
>2) Is there anything I can do to make this message available by default
>in the 'mia' list?
>
The term 'article' suggests that you have these mailing lists configured
as newsgroups. Is that so? Anyway, my suggestion is that the relevant
'newsbase' or 'mailbase' spool file is slightly damaged. However, if you
rebuild it, the article may disappear for good.
There is, it seems, a potential problem inherent in the situation you
describe. If the article is sent simultaneously to both lists, TP stores
only one copy (if two copies are received with the same message ID, one
is dumped). If you now read it in one pseudo-newsgroup, it is marked
'read', so it must, since there is only one 'read' tag on each message,
appear as 'read' in the other p-ng, even though it hasn't been read
there. I suppose if you have the expire times of the p-ngs set
differently, it could even expire from one and not the other, because
the 'expired' flag is applied to the p-ng's internal list, not the
article itself. Why, though, the article is not then erased from the
hard disc is not clear. Rather than a real bug (programming error), this
seems the inevitable consequence of 'conditions not envisaged' (program
not infinitely versatile).
--
Regards, John Woodgate, OOO - Own Opinions Only.
Phone +44 (0)1268 747839 Fax +44 (0)1268 777124.
http://www/jmwa.demon.co.uk Did you hear about
the hungry genetic engineer who made a pig of himself?
PLEASE DO ****NOT**** MAIL COPIES OF NEWSGROUP POSTS TO ME!!!!
Paul Terry
>I am a recipient of two mailing lists (amongst others)
...
>A message was posted to both these lists (ie it included the To: line
>To: [...], <m...@g.g.n>, <em...@g.g.n>), and this message was then replied
>to on the 'mia' list.
>However, the original article is apparently available only in the 'emul'
>list; in the 'mia' list, it appears as not available (with a '/' sign
>when I expand the thread), and if I then try to open it, I get the 'This
>article is already available!
This is because Turnpike has spotted that the messages are duplicates.
It will have therefore discarded one copy, and will have marked the
other as appropriate for both lists.
>Personally, this isn't too much of a problem for me, as I can always
>read the article in 'emul',
That's the idea: TP saves you reading the same thing twice.
> but if I had a (say) sister, who had a
>different Turnpike seat, and was subscribed to 'mia', but not to 'emul',
>she wouldn't (AFAICT) ever see the article, unless she happened to try
>to open the 'unavailable' message, which I would claim is a bug :-)
It would be if it were true, but I don't think it is.
You see different behaviour to your sister, because you read the article
under "emul" - because of this TP didn't bother showing it to you again
under "mia".
Your sister hasn't read the article at all, so TP will show it to her as
unread in the "mia" mailing list.
>1) Is this known behaviour?
Not as you describe it - you are speculating incorrectly!
>2) Is there anything I can do to make this message available by default
>in the 'mia' list?
Read "mia" first - then you will find the duplicate won't get shown
when you move on to read "emul".
If you really must read the same article twice, then you will need to
mark it as "unread" when you leave "mia" - it will then be ready to read
all over again when you start on "emul" :)
--
Paul Terry
Philip Kendall
<pa...@musonix.demon.co.uk> writes
>In article <hR+uGDAY...@kendalls.demon.co.uk> Philip Kendall wrote:
>
>>I am a recipient of two mailing lists (amongst others)
[snip]
>>However, the original article is apparently available only in the 'emul'
>>list; in the 'mia' list, it appears as not available (with a '/' sign
>>when I expand the thread), and if I then try to open it, I get the 'This
>>article is already available!
[snip]
>Your sister hasn't read the article at all, so TP will show it to her
>as unread in the "mia" mailing list.
No, it doesn't. This is my entire point :-)
>Read "mia" first - then you will find the duplicate won't get shown
>when you move on to read "emul".
If I do that, the article does not appear in !mia, unless I specifically
click on the 'article unavailable' slash thing.
For example, if I exit TP, restart, select !mia and then mark all the
articles there as unread, the cross-posted article does *not* get marked
as unread.
[Just to complicate the issue further, I'm going to be away from my
newsfeed for the next week. If all replies could also be mailed to
pa...@cam.ac.uk, that would be appreciated. Ta...]
Philip Kendall
<j...@jmwa.demon.co.uk> writes
><hR+uGDAY...@kendalls.demon.co.uk>, Philip Kendall
>>To: [...], <m...@g.g.n>, <em...@g.g.n>), and this message was then replied
>>to on the 'mia' list.
>>
>>list; in the 'mia' list, it appears as not available (with a '/' sign
>>when I expand the thread), and if I then try to open it, I get the 'This
>>and the article appears. Also, it doesn't appear in the count of unread
>>articles for the 'mia' list (ie if I mark it as unread, leave Turnpike
>>and then re-enter, I get one unread article in 'emul' and none in
>>'mia').
>>
>>in the 'mia' list?
>>
>as newsgroups. Is that so?
Correct -- should have made that more clear originally. Sorry :-)
> Anyway, my suggestion is that the relevant
>'newsbase' or 'mailbase' spool file is slightly damaged. However, if you
>rebuild it, the article may disappear for good.
Rebuilt news. No effect :-(
(I don't want the article to _dis_appear though -- I want it to appear
in the 'mia' list)
>There is, it seems, a potential problem inherent in the situation you
>describe. If the article is sent simultaneously to both lists, TP stores
>only one copy (if two copies are received with the same message ID, one
>is dumped). If you now read it in one pseudo-newsgroup, it is marked
>'read', so it must, since there is only one 'read' tag on each message,
>appear as 'read' in the other p-ng, even though it hasn't been read
>there.
That isn't my problem -- I don't mind it being marked as read in both
pseudo-newsgroups when I read it in one; the problem is that the article
is not appearing correctly in the !mia list, so if I have another seat
who is subscribed to !mia, but not !emul, they wouldn't notice the
article at all if they just went through pressing space to read their
news.
Paul Terry
>In article <bZWjNaAB...@musonix.com>, Paul Terry
><pa...@musonix.demon.co.uk> writes
>>Your sister hasn't read the article at all, so TP will show it to her
>>as unread in the "mia" mailing list.
>No, it doesn't. This is my entire point :-)
I didn't read your article in that light at all - I assumed you were
*speculating* about what would happen since you said
>if I had a (say) sister, who had a different Turnpike seat, and was
>subscribed to 'mia', but not to 'emul', she wouldn't (AFAICT) ever see
>the article
If you are now saying that you have set this up, and your sister really
isn't seeing the article at all, then you have something wrong.
>>Read "mia" first - then you will find the duplicate won't get shown
>>when you move on to read "emul".
>
>If I do that, the article does not appear in !mia, unless I specifically
>click on the 'article unavailable' slash thing.
Umm ... you have snipped the most important part of my advice, which
leaves me wondering if you noticed it!
You MUST mark the article as unread in "mia" AFTER YOU HAVE READ IT.
If you don't mark it unread, then TP will operate as you describe.
Again, if you are doing this, but it isn't working, then something is
wrong, as it is not how Turnpike works here (I am assuming that you read
both mailings lists from the same seat, by the way).
>For example, if I exit TP, restart, select !mia and then mark all the
>articles there as unread, the cross-posted article does *not* get marked
>as unread.
Ah. Then there is definitely something wrong or something I have
misunderstood from your description, since that should not be happening
and, again, doesn't happen here on any test I can think of.
>[Just to complicate the issue further, I'm going to be away from my
>newsfeed for the next week. If all replies could also be mailed to
>pa...@cam.ac.uk, that would be appreciated. Ta...]
Copy sent by email.
--
Paul Terry
Chris Taylor
<mailto:pa...@kendalls.demon.co.uk> writes
on crossposting to "mailing lists as newsgroups"
>>>However, the original article is apparently available only in the 'emul'
>>>list; in the 'mia' list, it appears as not available (with a '/' sign
>>>when I expand the thread), and if I then try to open it, I get the 'This
>>>article is already available!
apologies if I'm wrong:
AFAIK, this happens when the article should have expired for the
newsgroup you are reading, but is still available in a group with a
longer expiry date.
the message stops you requesting it again.
>You can read it immediately' dialog box,
>>>and the article appears. Also, it doesn't appear in the count of unread
>>>articles for the 'mia' list (ie if I mark it as unread, leave Turnpike
>>>and then re-enter, I get one unread article in 'emul' and none in
>>>'mia').
>>>
>>>2) Is there anything I can do to make this message available by default
>>>in the 'mia' list?
it will, if you read it within its expiry time
[snip]
>That isn't my problem -- I don't mind it being marked as read in both
>pseudo-newsgroups when I read it in one; the problem is that the article
>is not appearing correctly in the !mia list, so if I have another seat
>who is subscribed to !mia, but not !emul, they wouldn't notice the
>article at all if they just went through pressing space to read their
>news.
they would - see above.
--
chris taylor chris @ cjt.co.uk
computers provide everyone with the ability to
create a big impact by making a simple mistake.
Philip Kendall
<pa...@musonix.demon.co.uk> writes
>In article <$TeJGPAb...@kendalls.demon.co.uk> Philip Kendall wrote:
>
>>In article <bZWjNaAB...@musonix.com>, Paul Terry
>><pa...@musonix.demon.co.uk> writes
>
>>subscribed to 'mia', but not to 'emul', she wouldn't (AFAICT) ever see
>>the article
>
>If you are now saying that you have set this up, and your sister really
>isn't seeing the article at all, then you have something wrong.
Or TP has it wrong. I still claim the second :-)
>>>Read "mia" first - then you will find the duplicate won't get shown
>>>when you move on to read "emul".
>>
>>If I do that, the article does not appear in !mia, unless I specifically
>>click on the 'article unavailable' slash thing.
>
>Umm ... you have snipped the most important part of my advice, which
>leaves me wondering if you noticed it!
I did; unfortunately, it's not relevant :-(
>You MUST mark the article as unread in "mia" AFTER YOU HAVE READ IT.
>If you don't mark it unread, then TP will operate as you describe.
But I'm not offered the chance to read the article in mia (unless I
explicitly 'request' the article, say by selecting the slash icon which
refers to that message and then pressing Enter).
This is where I think TP is wrong: it should not be giving me a slash
icon, as the original article *was* posted to mia.
John Woodgate
<pa...@kendalls.demon.co.uk> inimitably wrote:
>This is where I think TP is wrong: it should not be giving me a slash
>icon, as the original article *was* posted to mia.
I think you are right (or nearly): the article is shown as 'expired' in
mia when it hasn't actually expired. This *may* be a peculiarity of your
installation of TP, but even so, short of reinstalling, I don't see what
you can do to put it right.
--
Regards, John Woodgate, Elector of Rayleigh, Grand Four-cusped Astroid of the
First Order of Magnitude. OOO - Own Opinions Only. Phone +44 (0)1268 747839
Fax +44 (0)1268 777124. http://www.jmwa.demon.co.uk
***PLEASE DO NOT E-MAIL COPIES OF NEWSGROUP POSTS TO ME***
Paul Terry
>In article <33fIWcAmB$23E...@musonix.com>, Paul Terry
><pa...@musonix.demon.co.uk> writes
re: duplicate articles sent to "mia" and "emul" mailing lists
>>>if I had a (say) sister, who had a different Turnpike seat, and was
>>>subscribed to 'mia', but not to 'emul', she wouldn't (AFAICT) ever see
>>>the article
>>If you are now saying that you have set this up, and your sister really
>>isn't seeing the article at all, then you have something wrong.
>Or TP has it wrong. I still claim the second :-)
No. I took the trouble to test the above scenario before I posted. A
user who is subscribed to only one of the two mailing lists WILL be
presented with the unread article, and not with just a marker, as you
claimed.
The user subscribed to BOTH lists will, having read the duplicate
article in one list, not be presented with it again in the second list.
To avoid this problem I advised that you could mark the article as
unread in the first group. However, I said ...
>>you have snipped the most important part of my advice, which
>>leaves me wondering if you noticed it!
>
>I did; unfortunately, it's not relevant :-(
How so? It achieves what you want, which is to be presented with the
same article in both mailing lists!
>This is where I think TP is wrong: it should not be giving me a slash
>icon, as the original article *was* posted to mia.
Phil, we are now going around in circles: you think that Turnpike should
present duplicate articles twice. Turnpike does not do this, and I have
never seen anyone before argue that it should.
--
Paul Terry
Paul Terry
><NnzFbIAVf$43E...@kendalls.demon.co.uk>, Philip Kendall
><pa...@kendalls.demon.co.uk> inimitably wrote:
>>This is where I think TP is wrong: it should not be giving me a slash
>>icon, as the original article *was* posted to mia.
>I think you are right (or nearly): the article is shown as 'expired' in
>mia when it hasn't actually expired.
Indeed, but I think it is important to realise that the slash icon does
not mean expired. It simply means, as it says in the helpfile, "not
immediately available".
However, I think one could argue whether this symbol/definition is
really very helpful, since the same one is used for articles which
really aren't available other than by making a request and then logging
on to fetch the article.
In the case Philip has raised, the article is in a different category of
"very nearly immediately available" :)
Using the standard "article read" symbol might be more logical, and it
is certainly what happens in the case of cross-posted articles. It would
not, though, satisfy Philip's argument that he wants (AIUI) to read the
same article in both lists to which it was posted and therefore it
should appear unread in the second group even though it has been read in
the first group.
As I said previously, I don't think this would curry much support and -
if carried logically to making cross-posted news articles appear as
unread in every group to which they were posted - would be very
unpopular and unconventional.
--
Paul Terry
Philip Kendall
<pa...@musonix.demon.co.uk> writes
>In article <NnzFbIAVf$43E...@kendalls.demon.co.uk> Philip Kendall wrote:
>
>>In article <33fIWcAmB$23E...@musonix.com>, Paul Terry
>><pa...@musonix.demon.co.uk> writes
>
>re: duplicate articles sent to "mia" and "emul" mailing lists
>
>>>isn't seeing the article at all, then you have something wrong.
>
>>Or TP has it wrong. I still claim the second :-)
>
>No. I took the trouble to test the above scenario before I posted. A
>user who is subscribed to only one of the two mailing lists WILL be
>presented with the unread article, and not with just a marker, as you
>claimed.
Not here and with this specific article :-( I couldn't replicate this
problem using a couple of dummy pseudo-newsgroups, but it *is* happening
with the article on mia and emul.
See http://www.kendalls.demon.co.uk/pak21/turnpike.html for the
screenshots of this happening.
Paul Terry
>Not here and with this specific article :-( I couldn't replicate this
>problem using a couple of dummy pseudo-newsgroups, but it *is* happening
>with the article on mia and emul.
I can now replicate this - see below.
>See http://www.kendalls.demon.co.uk/pak21/turnpike.html for the
>screenshots of this happening.
In the top screen shot, Turnpike knows about the article from Worldwide
Ramsoft because the information about it is held in the header of the
article by Martijn van der Heide.
If you look in the headers of the van der Heide article you will, I
think, see that he has an "In Reply To" header which contains the
Message ID of the Ramsoft article - this is what produces the Ramsoft
reference in Turnpike, along with the slash symbol.
When you double-click the Ramsoft article, Turnpike checks its database
for the message ID given by Martijn van der Heide, finds it because the
article still exists in the "emul" mailing list, and says "The article
is already available".
Now, why is it already available - since it has clearly expired from
your "mia" list ?
The answer must be, I think, because you have auto-keep set on the
"emul" list - hence the little padlocks that appear - but probably not
on the "mia" list as well.
----
So, to get back to your original problem, you are not offered the choice
to read the article in "mia" because the article has, I suspect, expired
as far as "mia" is concerned.
Were this not the case, Turnpike still would not show you the article as
unread if you had actually read it before (e.g. in "emul"), but it would
have displayed it as a "read" article" with a tick, rather than as an
article "not immediately available".
To replicate this you need to import an article to "emul" (only) and you
need to import a different article to "mia" that references the emul
article in an In_reply-To header, thus:
In-Reply-To: <message ID of the emul article>
You should then find that this combination produces what you show in
your screen-shots.
One last point. If this is happening on only some - not all - articles
posted to the two lists, I suspect that the cause is faulty software
used by some of the posters to the list, that is allowing follow-ups to
one mailing list to be made to a different mailing list. I can't be sure
of this, but it is certainly something I have seen happen in other
lists.
----
If that still doesn't seem to be the explanation, I think we shall need
to see the headers of the Ramsoft and van der Heide messages and also
have some details about the expiry period you have set on the two
mailing lists and whether they are, in fact, different.
--
Paul Terry
Philip Kendall
<pa...@musonix.demon.co.uk> writes
>In article <iCHeSFAA...@kendalls.demon.co.uk> Philip Kendall wrote:
>
>>Not here and with this specific article :-( I couldn't replicate this
>>problem using a couple of dummy pseudo-newsgroups, but it *is* happening
>>with the article on mia and emul.
>
>I can now replicate this - see below.
>
>>See http://www.kendalls.demon.co.uk/pak21/turnpike.html for the
>>screenshots of this happening.
>
>think, see that he has an "In Reply To" header which contains the
>Message ID of the Ramsoft article - this is what produces the Ramsoft
>reference in Turnpike, along with the slash symbol.
Correct.
>When you double-click the Ramsoft article, Turnpike checks its database
>for the message ID given by Martijn van der Heide, finds it because the
>article still exists in the "emul" mailing list, and says "The article
>is already available".
Agreed.
>Now, why is it already available - since it has clearly expired from
>your "mia" list ?
>
>The answer must be, I think, because you have auto-keep set on the
>"emul" list - hence the little padlocks that appear - but probably not
>on the "mia" list as well.
Nope :-( Both emul and mia are set to 3 day expire, no auto-keep. (I
manually kept both the Ramsoft and van der Heide articles so they
wouldn't go walk-about whilst I was away for the past week).
>So, to get back to your original problem, you are not offered the choice
>to read the article in "mia" because the article has, I suspect, expired
>as far as "mia" is concerned.
This may now be the case, but this effect also happened within 10
minutes of me downloading the messages in the first place, and I
certainly got other messages which were posted only to mia at the same
time, so I didn't have anything like a 0 day expiry time set on mia at
that stage.
>Were this not the case, Turnpike still would not show you the article as
>unread if you had actually read it before (e.g. in "emul"), but it would
>have displayed it as a "read" article" with a tick, rather than as an
>article "not immediately available".
Agreed. (And this *is* what I think it should do, BTW :-) )
[snip method for reproducing my screenshots which works]
>One last point. If this is happening on only some - not all - articles
>posted to the two lists, I suspect that the cause is faulty software
>used by some of the posters to the list, that is allowing follow-ups to
>one mailing list to be made to a different mailing list.
AFAICR, this was the first message cross-posted to both lists. (But as
TP normally wouldn't show me the article in the second list, I may not
have noticed this sort of thing before).
>If that still doesn't seem to be the explanation, I think we shall need
>to see the headers of the Ramsoft and van der Heide messages and also
>have some details about the expiry period you have set on the two
>mailing lists and whether they are, in fact, different.
<fx: flash of light>
The routing rules I have set up for emul and mia are:
emul: /Reply-To:[ \t]+emul@(pampus\.)?gns\.getronics\.nl/h
mia: /Reply-To:[ \t]+mia@(pampus\.)?gns\.getronics\.nl/h
as the lists are set up so that any messages posted to them come up with
Reply-To: set to the list, and this works fine when messages are posted
separately to the two lists. However, when a message is cross-posted to
both lists, Turnpike routes the second copy to the Waste Bin as it has
an identical Message-ID to the first message. Unfortunately, it isn't an
identical message and this breaks my routing rules, as the copy in the
database has Reply-To: emul, not Reply-To: mia.
Should have realised this a while ago -- sorry, people :-)
The question here is, I suppose, how unique should a Message-ID be? I
can see possible problems with any definition I can think of...