Opened 6 years ago

Closed 13 months ago

#7073 closed Bug/Something is broken (fixed)

mailman personalize settings

Reported by: Daniel Kahn Gillmor Owned by: JaimeV
Priority: Medium Component: Tech
Keywords: mailman maintenance Cc:
Sensitive: no

Description

I just noticed from reviewing #652 and mfplsvn/trunk/mailman that we seem to be carrying a diff in mailman on leslie:

0 leslie:~# diff -u /usr/lib/mailman/Mailman/MailList.py{.orig,}
--- /usr/lib/mailman/Mailman/MailList.py.orig	2011-02-16 15:14:57.000000000 -0500
+++ /usr/lib/mailman/Mailman/MailList.py	2011-10-29 21:18:09.000000000 -0400
@@ -374,7 +374,7 @@
         self.filter_action = mm_cfg.DEFAULT_FILTER_ACTION
         # Analogs to these are initted in Digester.InitVars
         self.nondigestable = mm_cfg.DEFAULT_NONDIGESTABLE
-        self.personalize = 0
+        self.personalize = 1
         # New sender-centric moderation (privacy) options
         self.default_member_moderation = \
                                        mm_cfg.DEFAULT_DEFAULT_MEMBER_MODERATION
1 leslie:~# 

Assata does not have this change.

This hacky change is also mentioned in our version of mm_cfg.py.

This seems like a really brittle way to carry around a default config change. Is there some reason that we've made the change on leslie but not on assata?

is there some better way that we can convince mailman to keep this change than having to remember to edit MailList.py after every upgrade?

If someone wants to fix this, they should probably read the mailman source and try to figure out if there's some other way to adjust the default self.personalize value. it may be that we need to make a patch that provides a more generic way to make the change, so that the patch can be accepted upstream, and then we can just set something in a config file.

Change History (9)

comment:1 Changed 6 years ago by Ross

Owner: set to Jamie McClelland
Status: newassigned

I suspect only Jamie knows the answers to the questions you ask about this change dkg.

comment:2 Changed 6 years ago by Jamie McClelland

Owner: Jamie McClelland deleted

No good reason it's only on leslie. And the reason it's so hacky is because I made the change years ago before I was more knowledgeable about how to submit things upstream. Maintaing/submitting this patch could definitely use an improvement.

I'm un-assigning it from me because I probably won't be able to get it very quickly.

jamie

comment:3 Changed 3 years ago by JaimeV

In Mailman/Defaults.py I see the following variable and description.

# And finally, we can VERP normal, non-personalized deliveries.  However,
# because it can be a significant performance hit, we allow you to decide how
# often to VERP regular deliveries.  This is the interval, in number of
# messages, to do a VERP recipient address.  The same variable controls both
# regular and digest deliveries.  Set to 0 to disable occasional VERPs, set to
# 1 to VERP every delivery, or to some number > 1 for only occasional VERPs.
VERP_DELIVERY_INTERVAL = 0

More details here: http://mailman.readthedocs.io/en/release-3.0/src/mailman/runners/docs/outgoing.html This would appear to be what we want.

comment:4 Changed 3 years ago by JaimeV

Owner: set to JaimeV

So with VERP_DELIVERY_INTERVAL set in theory having personalization activated for each list is not important, we shouldn't have to hack Mailman/MailList.py or set personalization variable on individual lists.

I've set VERP_DELIVERY_INTERVAL = 1 in puppet "modules/mayfirst/templates/mailman/mm_cfg.py.erb" and pushed to leslie. We could set that to a higher number if load on leslie becomes too much.

comment:5 Changed 3 years ago by Jamie McClelland

That's great! not hacking mailman is much better....

comment:6 Changed 3 years ago by JaimeV

Do we need to remove the application of that diff from puppet?

comment:7 Changed 3 years ago by Jamie McClelland

I don't think the diff is in puppet - I think it pre-dated our move to puppet.

comment:8 Changed 13 months ago by JaimeV

Resolution: fixed
Status: assignedfeedback

I don't think we've seen any downsides to this change. I'm inclined to say this is now fixed.

comment:9 Changed 13 months ago by automatic

Status: feedbackclosed

No news is good news (we hope)! Given the lack of feedback, we think this ticket can be closed.

Please login to add comments to this ticket.

Note: See TracTickets for help on using tickets.