nodupes doesn't seem to be working with aliases
I'm not sure if this counts as suppressable dupes. All users have nodupes set.
bin/export.py | grep nodupes | wc -l 183 bin/export.py | grep nodupes | grep -v True | wc -l 0
I have a postfix virtual-alias of the form:
a b, c, d c e, f d g, h g i, j h k, l
If I send mail to a set of mailing lists then dupes get suppressed but not if I send to an alias that expands to the same set of mailing lists. Mailman gets two copies from postfix but the message-id is the same for both. There is a to= and orig_to= in maillog but the messages should be otherwise identical including the timestamps.
Is there a way to suppress dupes in this situation?
I have 5 such aliases for convenience but if I can't suppress dupes its an annoyance since some people are on multiple mailing lists. For some reason I thought this used to work (but might have imagined it).
I'd rather not send a bunch of cross posted test messages, though if more info is needed I can create some new "just for test" mailing lists and aliases (on a just for test server if need be).
Curtis
On 05/05/2016 04:33 PM, Curtis Villamizar wrote:
I have a postfix virtual-alias of the form:
a b, c, d c e, f d g, h g i, j h k, l
If I send mail to a set of mailing lists then dupes get suppressed but not if I send to an alias that expands to the same set of mailing lists. Mailman gets two copies from postfix but the message-id is the same for both. There is a to= and orig_to= in maillog but the messages should be otherwise identical including the timestamps.
Is there a way to suppress dupes in this situation?
Mailman's avoid duplicates setting does nothing about posts to multiple lists. It only suppresses the list copy to a recipient who is also directly addressed in To: or Cc:.
If I understand correctly, there's nothing you can do in Mailman about this. Mailman has a Non-digest options -> regular_exclude_lists feature.
Your "If I send mail to a set of mailing lists then dupes get suppressed" statement tells me you may already be using this. However, it depends on the "sibling list" being explicit in To: or Cc: which will not be the case when the post comes via an alias.
I have 5 such aliases for convenience but if I can't suppress dupes its an annoyance since some people are on multiple mailing lists. For some reason I thought this used to work (but might have imagined it).
If what I say above doesn't satisfactorily explain what you see, I'll try further, but
avoid dups whon't affect this.
regular_exclude_lists can help but only if lists are explicitly addressed. For example, if both c@domain and d@domain are in b@domain's regular_exclude_lists and d@domain is in c@domain's regular_exclude_lists, and a post is explicitly addressed to all three of b@domain c@domain and d@domain, then a regular member of all three lists will receive the post only from d@domain, a member of the two lists c@domain and d@domain will receive the post only from d@domain, a member of the two lists b@domain and d@domain will receive the post only from d@domain and a member of the two lists b@domain and c@domain will receive the post only from c@domain
However if the post is addressed to a@domain and forwarded to b, c and d by an alias, none of b@domain, c@domain or d@domain will be an explicit To: or Cc: addressee of the post so everyone will receieve a copy from each list of which they are a member.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
trimmed.
In message <572C0016.1080505@msapiro.net> Mark Sapiro writes:
Your "If I send mail to a set of mailing lists then dupes get suppressed" statement tells me you may already be using this. However, it depends on the "sibling list" being explicit in To: or Cc: which will not be the case when the post comes via an alias.
That explains it. My mistake in setting up this way.
I have 5 such aliases for convenience but if I can't suppress dupes its an annoyance since some people are on multiple mailing lists. For some reason I thought this used to work (but might have imagined it).
I'll have to rework my local solution. Since all mailman subscriptions are generated from a database maintained elsewhere, I just need to replace the aliases with new mailing lists that are supersets. A bit of work but doable. Until then, I'll just manually expand on the Cc to the set of lists covered by an alias to avoid sending dups. These lists go to a set of non-technical volunteers so best not to annoy them. Volunteers are already hard to come by.
Curtis
participants (2)
-
Curtis Villamizar -
Mark Sapiro