This documentation is for Dovecot v2.x, see wiki1 for v1.x documentation.
Differences between revisions 2 and 3
Revision 2 as of 2014-05-20 16:10:16
Size: 4036
Editor: cpc46-slam6-2-0-cust675
Comment:
Revision 3 as of 2014-05-20 18:01:28
Size: 4183
Editor: cpc46-slam6-2-0-cust675
Comment:
Deletions are marked like this. Additions are marked like this.
Line 94: Line 94:
As with the flat file, the database query will need to return the Sieve script all in one line, otherwise the subsequent lines will be ignored.

Dict Lookup for Sieve Script

As per the Sieve configuration documentation, Sieve scripts can be obtained from a number of different locations. This page shows how to retrieve them from a dict lookup, which can have either a file or database backend.

This page contains a number of examples. For full details, please read the script-location-dict manual page.

The script location syntax is specialized as follows:

sieve = dict:<dict-uri>[;<option>[=<value>][;...]]

Using a flat file backend

Flat file example 1

To retrieve the Sieve script named "keep" from the dict file /etc/dovecot/sieve.dict:

sieve = dict:file:/etc/dovecot/sieve.dict;name=keep

The file /etc/dovecot/sieve.dict might look like this. Note that with the above configuration, only the "keep" script will be used.

priv/sieve/name/keep
1
priv/sieve/name/discard
2
priv/sieve/data/1
keep;
priv/sieve/data/2
discard;

Flat file example 2

Following on from example 1, a more advanced script. This notifies an external email address when new mail has arrived. Note that the script all needs to be on one line.

priv/sieve/name/notify
5
priv/sieve/data/5
require ["enotify", "variables"]; if header :matches "From" "*" { set "from" "${1}";} notify :importance "3" :message "New email from ${from}" "mailto:other@domain.com?body=New%20email%20has%20arrived.";

Using a SQL backend

For greater flexibility, it's possible to use a SQL backend for your dict scripts. First, set up a configuration file (such as /etc/dovecot/dict-sieve-sql.conf) with your database configuration. This should consist of the following parts:

# The database connection params
connect = host=localhost dbname=dovecot user=dovecot password=password

# The name mapping that yields the ID of the Sieve script
map {
    pattern = priv/sieve/name/$script_name   # The name of the script, as per the "sieve" config parameter
    table = user_sieve_scripts               # The database table
    username_field = username                # The field in the table to query on
    value_field = id                         # The field which contains the return value of the script ID
    fields {
        script_name = $script_name           # FIXME: The other database field to query?
    }
}

# The name mapping that yields the script content from ID
{
    pattern = priv/sieve/data/$id            # The ID, obtained from above
    table = user_sieve_scripts               # The database table
    username_field = username                # The field in the table to query
    value_field = script_data                # The field which contains the script
    fields {
        id = $id                             # FIXME: The other database field to query?
    }
}

Next, create a dict proxy service. Normally in /etc/dovecot/dovecot.conf:

dict {
    sieve = pgsql:/etc/dovecot/dict-sieve-sql.conf.ext
}

Finally, configure Sieve to check the dict (e.g. in /etc/dovecot/conf.d/90-sieve.conf). This looks up a script called "active" in the database.

plugin {
    sieve = dict:proxy::sieve;name=active
}

As with the flat file, the database query will need to return the Sieve script all in one line, otherwise the subsequent lines will be ignored.

Note: you might need to configure the proxy permissions

Caching the compiled Sieve binaries

With the configuration described above, the Sieve binaries will be compiled each time they are called. To improve performance, it is preferable to cache them, which can be done using the bindir parameter, which is added to the Sieve configuration. For example:

{
    sieve = dict:file:/etc/dovecot/sieve.dict;name=keep;bindir=~/.sieve-bin
}

Or:

{
    sieve = dict:file:/etc/dovecot/sieve.dict;name=keep;bindir=/var/sieve-scripts/%u
}

Note: Sieve uses the ID number as its cache index and to detect the need to compile. Therefore, if a script is changed, then its ID must also be changed for it to be reloaded.

None: Pigeonhole/Sieve/Configuration/Dict (last edited 2022-02-04 23:08:03 by TimoSirainen)