It feels I'm the only one coming by to clean every now and then - feels bad :-/ The change was revno: 1433
For the sake of any progress I made a ppa with the fix available at .
I thought lacking the experience to set it up correctly I can't drive a SRU  But actually, this is easy enough just by looking at the code:
That should be enough I'd think: $ grep -A 2 'change_header(' Mailman/Handlers/Tagger.py; grep -A 2 def Mailman/Handlers/CookHeaders.py | grep 'change_header(' change_header('X-Topics', NLTAB.join(hits.keys()), mlist, msg, msgdata, Delete=False) def change_header(name, value, mlist, msg, msgdata, delete=True, repl=True):
@Mark - if you have not given up on this bug, if you can provide a "steps to reproduce" please feel free to add.
** Description changed:
- Code added to Tagger.py in 2.1.16 to support the from_is_list Wrap - Message option contained a misspelling. + [Impact] + + * Using Mailman Lists with topics enabled can fail due to a typo + + * Fix is trivial and a backport active in everything >Trusty for quite a + while now + + [Test Case] + + * I beg your pardon for not knowing a "live" case to set up, but I think + it is obvious. With packages installed check the argument "Delete" vs + "delete": + + # grep -A 2 'change_header(' /usr/lib/mailman/Mailman/Handlers/Tagger.py + change_header('X-Topics', NLTAB.join(hits.keys()), + mlist, msg, msgdata, Delete=False) + + # grep -A 2 def /usr/lib/mailman/Mailman/Handlers/CookHeaders.py | grep 'change_header(' + def change_header(name, value, mlist, msg, msgdata, delete=True, repl=True): + + [Regression Potential] + + * It fixes an obvious broken variable access, but nothing else - so I'd + consider it safe. I don't think any code reaching this line ever worked + but if one had a setup awkwardly working while accepting this breakage + he might now get a different behavior. + + [Other Info] + + * embarrassingly old + + --- + + + Code added to Tagger.py in 2.1.16 to support the from_is_list Wrap Message option contained a misspelling.