Forum Post
Index > Scribe > InScribe v1.89 Test20

Author/Date InScribe v1.89 Test20
Nora
12/08/2007 1:16pm
After I figured out how to upgrade InScribe instead of downgrading back to i.Scribe (thanks, Fret!), it appears that Test20 has introduced a new bug. Now all incoming emails are landing in the Spam folder. I checked and all my filters are still there and should be working, but they're not. And my Bayesian filtering options have not changed.

fret
13/08/2007 3:00am
I'd switch off the bayesian filtering until we sort the issue out. I'd also try rebuilding the word databases (takes a while, esp if you've got lots of email).

If you select a random email in your inbox, and click Filters -> Analyse Selected Email what does it say? (on XP you can copy the contents of the msgbox using ctrl+c)
Nora
13/08/2007 10:14pm
Here are the Analyze Selected mail results for two emails from the same sender. The first one received into my inbox before Test20, and the second one in the Spam box today:

---------------------------
From my Inbox - received last week prior to Test20
---------------------------
Message Spam Probability:

[0] 'It'=0.500000
[1] 's'=0.500000
[2] 'John'=0.500000
[3] 'from'=0.500000
[4] 'Owner'=0.500000
[5] 'Locker'=0.500000
[6] 'john@ownerslocker.com'=0.500000
[7] 'D'=0.500000
[8] 'Van'=0.500000
[9] 'Meter'=0.500000
[10] 'Hi'=0.500000
[11] ' '=0.500000
[12] 'Brian'=0.500000
[13] 'has'=0.500000
[14] 'shared'=0.500000
[15] 'your'=0.500000

Result=0.500000

---------------------------
From the Spam folder - received today (same sender as above)
---------------------------
Message Spam Probability:

[0] 'and I'=0.990000
[1] 'are nuts'=0.990000
[2] ' Yeah'=0.990000
[3] 'try!!'=0.990000
[4] 'tangled'=0.990000
[5] 'but I would keep'=0.990000
[6] 'duplicate'=0.990000
[7] 'so I'=0.990000
[8] 'hell!'=0.990000
[9] 'short-sleeve'=0.990000
[10] 'chilly'=0.990000
[11] 'sweatshirt'=0.990000
[12] 'misjudge'=0.990000
[13] 'Ewwww!'=0.990000
[14] 'canned'=0.990000
[15] 'RE'=0.500000

Result=1.000000

---------------------------
OK
---------------------------

I'll turn off the Bayesian filter for now until this is resolved.

Thanks.
Reply