Closed Bug 185972 Opened 22 years ago Closed 16 years ago

Last message from group range missing is never updated

Categories

(MailNews Core :: Networking: NNTP, defect)

x86
Linux
defect
Not set
major

Tracking

(Not tracked)

RESOLVED DUPLICATE of bug 79130

People

(Reporter: 3.14, Unassigned)

Details

Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3a) Gecko/20021122

I am observing this bug for a very long time, but never worked it out. Now I did
investigate and found the following:

When you have a newsgroup where the last message (as given by the group command)
is missing (nothing returned by xover), there will alway be an unread count of
at least one. If you mark the group as read this goes down to zero, but the next
update brings it back.

Here are the details. Let me first show you the situation as observed with telnet:

[3.14@pi ~]$ telnet news.univie.ac.at 119
Trying 131.130.1.118...
Connected to news.univie.ac.at.
Escape character is '^]'.
200 news.univie.ac.at NNRP Service Ready - news-adm@news.univie.ac.at (posting ok).
MODE READER
200 news.univie.ac.at NNRP Service Ready - news-adm@news.univie.ac.at (posting ok).
GROUP de.comm.provider.usenet
211 666 11167 11832 de.comm.provider.usenet
xover 11832
224 data follows
.
quit
205 Transferred 236 bytes in 0 articles, 1 group.  Disconnecting.
Connection closed by foreign host.

So you see that the group contains messages 11167-11832, but the last does not
exist anymore (probably cancelled).

Now let's see what Mozilla does (from nntp log, I reduce to the interesting part):

First I open (expand) this news server:

1024[80896e0]: (8dff020) Receiving: 200 news.univie.ac.at NNRP Service Ready -
news-adm@news.univie.ac.at (posting ok).
1024[80896e0]: (8dff020) Next state: NNTP_LOGIN_RESPONSE
1024[80896e0]: (8dff020) Next state: NNTP_SEND_MODE_READER
1024[80896e0]: (8dff020) Sending: MODE READER
1024[80896e0]: (8dff020) Next state: NNTP_RESPONSE
1024[80896e0]: (8dff020) Receiving: 200 news.univie.ac.at NNRP Service Ready -
news-adm@news.univie.ac.at (posting ok).
1024[80896e0]: (8dff020) Next state: NNTP_SEND_MODE_READER_RESPONSE
1024[80896e0]: (8dff020) Next state: SEND_FIRST_NNTP_COMMAND
1024[80896e0]: (8dff020) Next state: NEWS_DISPLAY_NEWS_RC
[various groups]
1024[80896e0]: (8dff020) Sending: GROUP de.comm.provider.usenet
1024[80896e0]: (8dff020) Next state: NNTP_RESPONSE
1024[80896e0]: (8dff020) Receiving: 211 666 11167 11832 de.comm.provider.usenet
1024[80896e0]: (8dff020) Next state: NEWS_DISPLAY_NEWS_RC_RESPONSE
1024[80896e0]: (8dff020) Next state: NEWS_DISPLAY_NEWS_RC
1024[80896e0]: (8dff020) Next state: NEWS_DONE
1024[80896e0]: (8dff020) Next state: NEWS_FREE

OK, now I click on the group de.comm.provider.usenet, at this time one new
message is displayed:

1024[80896e0]: (8dff020) ParseURL
1024[80896e0]: (8dff020) fullPath = /de.comm.provider.usenet
1024[80896e0]: (8dff020) m_messageID = (null)
1024[80896e0]: (8dff020) group = de.comm.provider.usenet
1024[80896e0]: (8dff020) commandSpecificData = (null)
1024[80896e0]: (8dff020) m_key = -1
1024[80896e0]: (8dff020) Next state: SEND_FIRST_NNTP_COMMAND
1024[80896e0]: (8dff020) Sending: GROUP de.comm.provider.usenet
1024[80896e0]: (8dff020) Next state: NNTP_RESPONSE
1024[80896e0]: (8dff020) Receiving: 211 666 11167 11832 de.comm.provider.usenet
1024[80896e0]: (8dff020) Next state: SEND_FIRST_NNTP_COMMAND_RESPONSE
1024[80896e0]: (8dff020) Next state: SETUP_NEWS_STREAM
1024[80896e0]: (8dff020) Next state: NNTP_XOVER_BEGIN
1024[80896e0]: (8dff020) SetCurrentGroup to de.comm.provider.usenet
1024[80896e0]: (8dff020) Next state: NNTP_FIGURE_NEXT_CHUNK
1024[80896e0]: (8dff020) Next state: NEWS_PROCESS_XOVER
1024[80896e0]: (8dff020) Next state: NEWS_DONE
1024[80896e0]: (8dff020) Next state: NEWS_FREE

Too bad the log file does not show the actual dialog, but it must have sent
XOVER to the server and received nothign for the last message. Anyhow, the
unread count is still at one, but no message is found as unread. So I press mark
group read. Then I collapse and expand the server again. The log looks exactly
like the the first time above.

Here is what newsrc-news.univie.ac.at has to say about the group in question
(after closing Mozilla, just to be sure it is finally written):
de.comm.provider.usenet: 1-11831

As you can see, the last number is too small. It should have been updated on
XOVER or at the latest on mark group read.

There is nothing I can see to repair the unread count for the user. The only
solution is to manually repair the .newsrc. Hence marking major.

pi
Product: MailNews → Core
sorry for the spam.  making bugzilla reflect reality as I'm not working on these bugs.  filter on FOOBARCHEESE to remove these in bulk.
Assignee: sspitzer → nobody
Filter on "Nobody_NScomTLD_20080620"
QA Contact: stephend → networking.news
Boris, do you still see this problem in TB 2.0.0.* or recent Trunk TB builds? If no response in 3 weeks, I will close this bug as INCO.
Whiteboard: [jcranmer:unconfirmed] closeme 2008-07-17
I no longer use SeaMonkey for news. But I have given clear explanation how you can reproduce the exact situation which causes the problem. So you can test it.

pi
It's difficult to actually test problems with expired messages. In lieu of any better information, duping to 79130.
Status: NEW → RESOLVED
Closed: 16 years ago
Resolution: --- → DUPLICATE
Whiteboard: [jcranmer:unconfirmed] closeme 2008-07-17
It does not depend on a particular message, but on a specific situation as described. Having said that the dupe sound right.

pi
Product: Core → MailNews Core
You need to log in before you can comment on or make changes to this bug.