http://bugzilla.spamassassin.org/show_bug.cgi?id=4019





------- Additional Comments From [EMAIL PROTECTED]  2004-12-06 12:35 -------
Subject: Re:  BayesSQL token column type for MySQL may end up with
 semi-bogus data

> ------- Additional Comments From [EMAIL PROTECTED]  2004-12-06 12:17
> 
> 'foo' == 'foo  ' could cause some sort of problem I think, but maybe
> I'm over analyzing things.

Given that the tokens are (conceptually in the SQL database) a fixed 
length of 5 characters this shouldn't be a problem since we'll never be 
trying to match a string less the 5 characters (we can therefore 
assuming shorter strings are space padded).

i.e. 'foo' in the database had to have come from a token that was really 
'foo  '.  Any other token, such as 'foo e' would appear as 'foo e' in 
the database.


Of course, this is wrong if I'm mistaken in that the tokens are fixed 
length and are really variable length with a maximum of 5 characters.


Daryl





------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

Reply via email to