426838 - Baysian Junkfilter: Junk filter runs despite being disabled

Status: RESOLVED INCOMPLETE

(MailNews Core :: Filters, defect)

Description

[David A. Cobb]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9b4pre) Gecko/2008021504 Minefield/3.0b4pre
Build Identifier: 2008-03-24-03-trunk

Possibly related to Bug#2133758, which is for IMAP; I'm using POP3.
Background: I have two POP addresses on the same server (if that matters). Because I thought it was implicated in another bug, I have disabled junk detection on both.

The less-active account is downloaded automatically on startup. Immediately after reading a message from the account, T-bird spends an inordinate amount of time and CPU running the junkfilter.  It simply should not be running.

Reproducible: Always

Steps to Reproduce:
1. Set account to not do "adaptive junk filtering;"
2. Let mail download
3.
Actual Results:  
SEE the following, greatly snipped, NSPR_LOG excerpt
-1224194272[8bb69a0]: Entering NET_ProcessPop3 28
-1224194272[8bb69a0]: POP3: Entering state: 3
-1224194272[8bb69a0]: RECV: +OK DACobb is welcome here
-1224194272[8bb69a0]: POP3: Entering state: 31
-1224194272[8bb69a0]: POP3: Entering state: 7
-1224194272[8bb69a0]: SEND: STAT

-1224194272[8bb69a0]: Entering NET_ProcessPop3 13
-1224194272[8bb69a0]: POP3: Entering state: 3
-1224194272[8bb69a0]: RECV: +OK 1 34873
-1224194272[8bb69a0]: POP3: Entering state: 8
-1224194272[8bb69a0]: POP3: Entering state: 9
-1224194272[8bb69a0]: SEND: LIST

-1224194272[8bb69a0]: Entering NET_ProcessPop3 25
-1224194272[8bb69a0]: POP3: Entering state: 3
-1224194272[8bb69a0]: RECV: +OK 1 messages
-1224194272[8bb69a0]: POP3: Entering state: 10
-1224194272[8bb69a0]: RECV: 1 34873
-1224194272[8bb69a0]: POP3: Entering state: 10
<-- ********************************* SNIP ********************************* -->
-1224194272[8bb69a0]: SEND: UIDL

-1224194272[8bb69a0]: Entering NET_ProcessPop3 70
-1224194272[8bb69a0]: POP3: Entering state: 3
-1224194272[8bb69a0]: RECV: +OK 1 messages
-1224194272[8bb69a0]: POP3: Entering state: 12
-1224194272[8bb69a0]: RECV: 1 <[e-mail]>
-1224194272[8bb69a0]: POP3: Entering state: 12
<-- ********************************* SNIP ********************************* -->
-1224194272[8bb69a0]: Entering NET_ProcessPop3 3
-1224194272[8bb69a0]: POP3: Entering state: 12
-1224194272[8bb69a0]: RECV: .
-1224194272[8bb69a0]: POP3: Entering state: 15
-1224194272[8bb69a0]: POP3: Entering state: 18
-1224194272[8bb69a0]: SEND: RETR 1

<-- ********************************* SNIP ********************************* -->
-1224194272[8bb69a0]: Entering NET_ProcessPop3 1448
-1224194272[8bb69a0]: POP3: Entering state: 3
-1224194272[8bb69a0]: RECV: +OK 34873 octets
-1224194272[8bb69a0]: POP3: Entering state: 19
-1224194272[8bb69a0]: Opening message stream: MSG_IncorporateBegin
-1224194272[8bb69a0]: Done opening message stream!
-1224194272[8bb69a0]: RECV: Return-Path: <[e-mail]>
<-- ********************************* SNIP ********************************* -->
-1224194272[8bb69a0]: RECV: from: [e-mail]
-1224194272[8bb69a0]: RECV: to: [e-mail]
-1224194272[8bb69a0]: RECV: date: 31 Mar 2008 20:05:08 -0500
-1224194272[8bb69a0]: RECV: subject: New Jobs
<-- ********************************* SNIP ********************************* -->
-1224194272[8bb69a0]: RECV: 
-1224194272[8bb69a0]: RECV: .
-1224194272[8bb69a0]: Invalidate (rect) [94f24b8]: 0 36 386 19 (sync: 0)
-1224194272[8bb69a0]: Invalidate (rect) [94f24b8]: 0 36 386 19 (sync: 0)
-1224194272[8bb69a0]: Invalidate (rect) [94f24b8]: 0 36 386 19 (sync: 0)
-1224194272[8bb69a0]: Invalidate (rect) [94f24b8]: 0 36 386 19 (sync: 0)
-1224194272[8bb69a0]: RECV: (null)
-1224194272[8bb69a0]: POP3: Entering state: 20
-1224194272[8bb69a0]: SEND: DELE 1

<-- ********************************* SNIP ********************************* -->
-1224194272[8bb69a0]: POP3: Entering state: 3
-1224194272[8bb69a0]: RECV: +OK
-1224194272[8bb69a0]: POP3: Entering state: 21
-1224194272[8bb69a0]: POP3: Entering state: 15
-1224194272[8bb69a0]: junk probability threshold: 0.900000
-1224194272[8bb69a0]: add word: regions, (count=1)
-1224194272[8bb69a0]: adding word to tokenizer: regions, (len=8) (count=1)
-1224194272[8bb69a0]: add word: predicting (count=2)
-1224194272[8bb69a0]: adding word to tokenizer: predicting (len=10) (count=2)
-1224194272[8bb69a0]: add word: cantu (count=1)
-1224194272[8bb69a0]: adding word to tokenizer: cantu (len=5) (count=1)
-1224194272[8bb69a0]: add word: [151b] (count=2)
-1224194272[8bb69a0]: adding word to tokenizer: [151b] (len=6) (count=2)
<-- ********************************* SNIP ********************************* -->
-1224194272[8bb69a0]: POP3: Entering state: 22
-1224194272[8bb69a0]: SEND: QUIT

<-- ********************************* SNIP ********************************* -->
-1224194272[8bb69a0]: Entering NET_ProcessPop3 47
-1224194272[8bb69a0]: POP3: Entering state: 3
-1224194272[8bb69a0]: RECV: +OK DACobb InterMail POP3 server signing off.
-1224194272[8bb69a0]: POP3: Entering state: 40
-1224194272[8bb69a0]: POP3: Entering state: 23
-1224194272[8bb69a0]: POP3: Entering state: 25
<-- ********************************* SNIP ********************************* -->
<-- ****************************** END OF LOG EXCERPT ********************** -->

Expected Results:  
Program should proceed to do something useful, like wait to download the other account.

I could attach the full NSPR_LOG, but only if you absolutely need it -- it's really big.

Comment 1

[Wayne Mery (:wsmwk)]

this sounds serious.

(bug 213758 is close to being marked incomplete)

Comment 2

[Wayne Mery (:wsmwk)]

on second read, maybe not so serious since you don't say it affects your UI response. 

what is amount of CPU usage?
do you still see thie problem using an early release?
http://www.mozillamessaging.com/en-US/thunderbird/early_releases/
(backup your system first)

Comment 3

[Ludovic Hirlimann [:Usul]]

David could you try again with a recent version of TB and comment on the bug if you still have the issue please ?

Comment 4

[Joshua Cranmer [:jcranmer]]

RESO INCO due to lack of response to last comment. If you feel this change was made in error, please respond to this bug with your reasons why.

Comment 5

[David A. Cobb]

See how often I get to read my mail these days!  If one only didn't have to work!

Yeah, INCO is fine.  I don't see this happening on recent nightlies.