[Mailman-Users] Bounce message confusion

[Mark Sapiro]

Robin Rowe wrote:
>
>I got 11 messages like this today. All timestamped 7:00 AM:


Is your Mailman server in a different time zone? The standard cron runs
cron/disabled at 9:00 a.m. This is the script that sends warnings and
does unsubscribes.


>  Subject: ScreenplayLab unsubscribe notification
>  From: mailman-bounces at movieeditor.com
>  Date: 7:00 AM
>  To: ScreenplayLab-owner at movieeditor.com
>
>  <email addrees> has been removed from ScreenplayLab.
>
>where "email address" is a subscriber's address, mostly from yahoo, 
>hotmail, or aol.


Yes. These messages do not include a bounce message. The occur because
the last warning was sent bounce_you_are_disabled_warnings_interval
days ago, and the member still hasn't visited her options page and
re-enabled delivery.

The message you need to look at is the notification that the member has
been disabled due to bounces. That contains the triggering bounce.


>> The MTA should be returning a status code with the refusal. If it's a
>> 500 series code other than 552. it will be treated as a bounce. If it
>> is 552 or a 400 series code, Mailman will retry delivery later.
>
>Thanks, will look for those the next time I send out a post. There's no 
>diagnostic # in the unsubscribe messages. I don't have any earlier 
>messages because I unknowingly had it pointed to an invalid email address.
>
>> No. You can't do this. The recipient list is in a 'hash table' order
>> which is subsequently grouped into bins by top level domain. bin 1
>> contains all .com, bin 2 - all .net and .org, bin 3 - all .edu, .us
>> and .ca, and bin 0 - all the rest. These bins are also processed in a
>> hashed order, so you can't make an arbitrary address be at the end.
>
>Both the bins and their contents are processed in hash order?
>Why wouldn't the bins be processed sequentially?


Because both the member lists and the bins are Python dictionaries
(actually the members are a simple list at this point, but they were
retrieved from a dictionary) which are data structures relating values
and keys and the actual storage of key-value pairs is hashed by key.
The dictionaries are processed sequentially by location, but because
the keys are stored by hash, the retrieved keys are in hash sequence,
not in key sequence.

-- 
Mark Sapiro <msapiro at value.net>       The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan