97422 - Cache issues when network connection gets dropped.

Status: RESOLVED WORKSFORME

(MailNews Core :: Networking: NNTP, defect)

Description

[[e-mail] (gone - use [e-mail] instead)]

Build ID:  2001-08-28-03, Windows 2000.

Summary: Cache issues when network connection gets dropped.

Steps to Reproduce:

1.  Connect to a newsserver (I used news://poisonivy.mcom.com) and subscribe to
a group.
2.  Read a couple of postings, doesn't matter what you read.
3.  Drop your connection to the network (if on dialup, disconnect, if on LAN,
pull RJ-45 cable or do ipconfig/release on a win2k box).
4.  Try to read a message (you get 2 errors, both connection-related [why 2?
might be a seperate bug]).
5.  Now, reconnect to the network by doing ipconfig/renew or dialing up.
6.  Attempt to read messages.

Expected Results:

Messages, whether read or not, will display 100%.

Actual Results:

Especially in threads it seems, we are not reading from the cache correctly, or
invalidating from the cache.  Screenshots to follow.  One before shot, one after
shot.

Comment 1

[[e-mail] (gone - use [e-mail] instead)]

Attachment: Before

Comment 2

[[e-mail] (gone - use [e-mail] instead)]

Attachment: After

Comment 3

[[e-mail] (gone - use [e-mail] instead)]

*** Bug 97025 has been marked as a duplicate of this bug. ***

Comment 4

[Scott MacGregor]

not an emojo stopper.

Comment 5

[[e-mail] (gone - use [e-mail] instead)]

Attachment: NNTP logfile

Comment 6

[[e-mail] (gone - use [e-mail] instead)]

Hopefully, the log is helpful.  Here are my steps to produce that log:

Messages "test123" and "test456" were unread messages.  Message "there can only 
be one" was a message I have already read.

1.  In netscape.test, I read "test123" with my cable connection still alive.
2.  Disconnected my connection, and attempted to read a read message ("there 
can only be one").
3.  Waited for timeout
4.  Read "test456" (this didn't display correctly)
5.  Read "test123", which finally loaded correctly.

Comment 7

[[e-mail] (gone - use [e-mail] instead)]

Seems bug 59449's fix corrects this, or helps it tremendously.  But I'll keep
testing.

Comment 8

[(not reading, please use [e-mail] instead)]

it makes sense that the fix for #59449 fixes this, but I'll have to debug to
confirm.

I'd like to leave this bug open because I bet that when the connection gets
dropped we're leaking the protocol instance.

Comment 9

[(not reading, please use [e-mail] instead)]

sorry for the spam.  making bugzilla reflect reality as I'm not working on these bugs.  filter on FOOBARCHEESE to remove these in bulk.

Comment 10

[Wayne Mery (:wsmwk)]

has a log file

Comment 11

[Joshua Cranmer [:jcranmer]]

Can't reproduce -> WFM.