[Spambayes] Feedback on sb_dbexpimp.py

[Christophe Deprez]

Hi there,

Thanks for spambayes. We've been using our own server-side adaptation of 
spambayes and the amount of spam received in the Institute has 
drastically dropped... for free!

Recently, I've been using the sb_dbexpimp.py utility to merge databases
and I am questionning the "feature" introduced in recent versions, which 
tries different western encodings in the uunquote function:
    for encoding in ("utf-8", "cp1252", "iso-8859-1"):
My point is that when non-ASCII characters are converted in an encoding 
different from utf-8, it is irreversible!
So when exporting and re-importing a database, I don't end up with the 
initial database, as I first expected.
Is this intentional? What is the effect on the filtering efficiency of 
the new filter?

Cheers

-- 
Christophe Deprez                         christophe.deprez at bri.nrc.ca
----------------------------------------------------------------------
Institut de Recherche en Biotechnologies / Biotech. Research Institute
6100 Royalmount, Montréal (QC) H4P 2R2, Canada     Tel: (514) 496-6164