| Commit message (Collapse) | Author | Age | Files | Lines |
| | |
|
| | |
|
| |
|
|
| |
Closes SF bug # 703941 by Stuart Bishop, who also suggested the basic fix.
|
| | |
|
| |
|
|
|
|
| |
Mailman/cmd_confirm.py (but doesn't really use it :)
2.1 backport candidate.
|
| |
|
|
|
|
|
| |
- this is a subscription being confirmed and send_welcome_msg is true
- this is an unsub being confirmed and send_goodbye_msg is true
We don't need redundant responses.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
confirmation string was invalid" messages. The problem was if the
responder left the confirmation string in the Subject: and also copied
it to the body, the second processing of the confirmation would find
no cookie.
So now we always stop after processing the first confirmation command
(yes, that means no other commands in the message will be processed,
which is fine). We also trim the unprocessed commands in the results
to not include identical confirmation strings -- otherwise they'll see
their confirmation as unprocessed.
|
| |
|
|
|
| |
process(): Catch NotAMemberError instead of MMNoSuchUserError. Also,
slight rewrite of the result message.
|
|
|
unmaintainable) MailCommandHandler.py file, we've now got a framework
where each command is implemented in a separate file. This means it's
both more extensible and more flexible:
- you can easily add new commands for things I haven't thought of
<wink>, and the `help' command will automatically adjust
- you can disable commands entirely by removing the appropriate file
- you can disable, change, or add commands on a per-list (or even
per-message or per-sender) basis
CommandRunner.py is the module that calls into this framework. Each
command is implemented as a cmd_<command>.py file. The `set' command
is the most complicated. The help text is currently implemented as
module docstrings (for most commands), so the i18n catalogs must be
updated. Also the help.txt files will be updated.
|