When sending an email to an external system, I AWL the address (outbound email is scanned, and txrep_whitelist_out is set to 20).
Step 1. Pre condition: zero rows:
mysql> select * from awl where email = 'example@gmail.com';
Empty set (0.02 sec)
Step 2. Send email to example@gmail.com (changed to protect address):
mysql> select * from awl where email = 'example@gmail.com';
+-----------+-----------------------+------+-------+----------+----------+---------------------+
| username | email | ip | count | totscore | signedby | lastupdate |
+-----------+-----------------------+------+-------+----------+----------+---------------------+
| all_users | example@gmail.com | none | 1 | -20 | | 2015-01-23 13:10:51 |
+-----------+-----------------------+------+-------+----------+----------+---------------------+
1 row in set (0.03 sec)
Step 3. Upon replying from gmail:
mysql> select * from awl where email = 'example@gmail.com';
+-----------+-----------------------+------+-------+----------+-----------+---------------------+
| username | email | ip | count | totscore | signedby | lastupdate |
+-----------+-----------------------+------+-------+----------+-----------+---------------------+
| all_users | example@gmail.com | none | 1 | -20 | | 2015-01-23 13:10:51 |
| all_users | example@gmail.com | none | 2 | -2.178 | gmail.com | 2015-01-23 13:11:40 |
+-----------+-----------------------+------+-------+----------+-----------+---------------------+
2 rows in set (0.03 sec)
(count is 2 because inbound email gets scanned twice -- once at the SMTP layer, where messages can't be modified, and a second time by the user account, to add the spamassassin headers).
As further proof, the spam assassin header lists:
* 0.0 TXREP TXREP: TXREP
Expected behavior is that the AWL row will get used.
When sending an email to an external system, I AWL the address (outbound email is scanned, and txrep_whitelist_out is set to 20).
Step 1. Pre condition: zero rows:
Step 2. Send email to example@gmail.com (changed to protect address):
Step 3. Upon replying from gmail:
(count is 2 because inbound email gets scanned twice -- once at the SMTP layer, where messages can't be modified, and a second time by the user account, to add the spamassassin headers).
As further proof, the spam assassin header lists:
* 0.0 TXREP TXREP: TXREP
Expected behavior is that the AWL row will get used.