bayes bayes_sql_override_username

[Nigel Frankcom <ni...@blue-canoe.com>]

Hi All,

Apologies if this has been mentioned and I've missed it, but....

I noted today that since upgrading from 3.2.0 to 3.2.1 bayes stopped
working, or rather, bayes started ignoring the username etc passed
from the local.cf.

Using  bayes_sql_override_username has resolved the issue. My worry is
that this is a complete reversal of all previous versions of SA. I am
not sure if it's something I've done wrong (entirely possible) or if
it's an inherent problem. Either way, for those running 3.2.1 check
your bayes stuff.

To complicate matters further, it will --lint correctly. I only
noticed after a few spam got through and noticed they had no bayes
test header.

Hope that helps someone else.

If I've been a muppet I'm sure someone will let me know soon enough
:-D

Kind regards

Nigel

Re: bayes bayes_sql_override_username

[Duane Hill <d....@yournetplus.com>]

On Sat, 7 Jul 2007 at 16:04 +0100, nigel@blue-canoe.com confabulated:

> Hi All,
>
> Apologies if this has been mentioned and I've missed it, but....
>
> I noted today that since upgrading from 3.2.0 to 3.2.1 bayes stopped
> working, or rather, bayes started ignoring the username etc passed
> from the local.cf.
>
> Using  bayes_sql_override_username has resolved the issue. My worry is
> that this is a complete reversal of all previous versions of SA. I am
> not sure if it's something I've done wrong (entirely possible) or if
> it's an inherent problem. Either way, for those running 3.2.1 check
> your bayes stuff.
>
> To complicate matters further, it will --lint correctly. I only
> noticed after a few spam got through and noticed they had no bayes
> test header.
>
> Hope that helps someone else.
>
> If I've been a muppet I'm sure someone will let me know soon enough
> :-D
>
> Kind regards
>
> Nigel

Works fine here. I upgraded from 3.2.0 to 3.2.1 without any issues. The 
following local.cf config has never changed since I first installed SA 
back on version 3.1.8:

bayes_store_module              Mail::SpamAssassin::BayesStore::MySQL
bayes_sql_dsn                   DBI:mysql:spamass:localhost:3306
bayes_sql_username              username
bayes_sql_password              password

I have bayes turned off in local.cf and leave the option up to our 
customers if they want it turned on or not. The same holds true for 
autolearning. I see several logs where bayes has autolearned messages as 
ham/spam. I also have logs that show customers reporting messages to be 
learned as ham/spam.

-----
  _|_
(_| |

RE: bayes bayes_sql_override_username

["James D. Rallo" <jr...@uppermohawkinc.com>]

Slight correction, username passed via the SpamD/C protocol was ignored.

Loading the settings from the userpref table did use the proper name as
passed from the SpamC client.

Thanks,
James

> -----Original Message-----
> From: Nigel Frankcom [mailto:nigel@blue-canoe.com]
> Sent: Saturday, July 07, 2007 11:05 AM
> To: users@spamassassin.apache.org
> Subject: bayes bayes_sql_override_username
> 
> Hi All,
> 
> Apologies if this has been mentioned and I've missed it, but....
> 
> I noted today that since upgrading from 3.2.0 to 3.2.1 bayes stopped
> working, or rather, bayes started ignoring the username etc passed
> from the local.cf.
> 
> Using  bayes_sql_override_username has resolved the issue. My worry is
> that this is a complete reversal of all previous versions of SA. I am
> not sure if it's something I've done wrong (entirely possible) or if
> it's an inherent problem. Either way, for those running 3.2.1 check
> your bayes stuff.
> 
> To complicate matters further, it will --lint correctly. I only
> noticed after a few spam got through and noticed they had no bayes
> test header.
> 
> Hope that helps someone else.
> 
> If I've been a muppet I'm sure someone will let me know soon enough
> :-D
> 
> Kind regards
> 
> Nigel