[ mailman-Feature Requests-930969 ] Duplicated storage of
attachments with MM 2.1
SourceForge.net
noreply at sourceforge.net
Wed Apr 7 06:26:39 EDT 2004
Feature Requests item #930969, was opened at 2004-04-07 10:26
Message generated for change (Tracker Item Submitted) made by Item Submitter
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=350103&aid=930969&group_id=103
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Richard Barrett (ppsys)
Assigned to: Nobody/Anonymous (nobody)
Summary: Duplicated storage of attachments with MM 2.1
Initial Comment:
In Mailman 2.1.4 the process() function in $prefix/Mailman/
Handlers/Scrubber.py is called by instances of the ArchiverMailbox
class ($prefix/Mailman/Mailbox.py) and by the process() function
for the ToDigest handler ($prefix/Mailman/Handlers/ToDigest.py)
via the send_digests() and send_i18n_digests() functions.
In former case the function may extract attachments from a
message before that message is archived by pipermail. In the
latter case the function may extract attachments from a message
before that message is added to a text-only digest that is being
prepared.
If extraction is done in both situations for the same message, this
appears to lead to two copies of the same attachment being stored
in different files in the relevant $prefix/archives/private/
<listname>/attachments/ directory.
While not suggesting this is an easy problem to fix, the increase in
disc storage for busy lists whose messages carry large
attachments can be a problem. Optimising so that only one copy is
saved seems appropriate.
Incorporating a solution to this in Mailman 3 seems like a good
thing.
The problem can be ameliorated by setting the 'digestable'
attribute of affected lists to No.
As an aside, digests for digestable lists are constructed even if no
list subscriber is a digest member. Again, an optimisation to avoid
this wasted processing might be appropriate for Mailman 3.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=350103&aid=930969&group_id=103
More information about the Mailman-coders
mailing list