summaryrefslogtreecommitdiff
path: root/Mailman/MTA/Manual.py
diff options
context:
space:
mode:
authorbwarsaw2006-08-29 02:58:35 +0000
committerbwarsaw2006-08-29 02:58:35 +0000
commit85eba63c38c990f4b5687fb8f6693c0778444fa1 (patch)
tree1f615f474424c5d33e09d7ae568da72bec00acde /Mailman/MTA/Manual.py
parent391c15b319a2d310c79a595f3a4d639607453012 (diff)
downloadmailman-85eba63c38c990f4b5687fb8f6693c0778444fa1.tar.gz
mailman-85eba63c38c990f4b5687fb8f6693c0778444fa1.tar.zst
mailman-85eba63c38c990f4b5687fb8f6693c0778444fa1.zip
Fix several problems reported by Mark. First, Python 2.3 doesn't have
built-in sets and sadly we still have to support it. Utils.list_names() therefore now uses a sets.Set if it finds the built-in set type missing. main() in list_lists.py also has to list()-ify the set in order to sort it. There's a bootstrapping problem when bin/update tries to re-situate a non-qualified list name to a qualified list name. The problem is that the MailList ctor will call Load() with the nonqual name, which in turn will call CheckVersion() by default. The latter will then turn around and try to call Load() with the fqdn list name, but that list name won't yet exist since it hasn't been situated. The solution (but maybe not the best one) then is to pass the check_version flag to the MailList ctor, which gets passed down to the Load() call. Only situate_list() will call this to avoid the version checking before the list has been situated.
Diffstat (limited to 'Mailman/MTA/Manual.py')
0 files changed, 0 insertions, 0 deletions