| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
InitVars(): Fix the permissions on the empty archive index. Fixes SF bug
#835012.
CheckHTMLArchiveDir(): Only make the mbox link public if PUBLIC_MBOX is true.
|
| |
|
|
|
|
| |
Also bump DATA_FILE_VERSION to pick up the new header_filter_rules variable.
|
| |
|
| |
|
|
|
|
| |
database.
|
|
|
|
|
|
|
|
|
| |
filters.
ProcessConfirmation(): A fix for bug # 833384, where the incorrect admin
password given in a the response to a hold message's confirm attachment would
end up discarding the message. Now, an error message is returned to the
sender and the held message is re-pended.
|
|
|
|
| |
rejection message, make sure it ends with at least two newlines.
|
|
|
|
|
|
| |
access to all list mbox files (this is not a per-list configuration).
HeaderFilter: a new widget type to support the expanded spam filtering.
|
|
|
|
|
| |
LC_DESCRIPTIONS. This is mostly useful for debugging since I
occasionally add languages for testing and then disable them.
|
|
|
|
|
| |
(Note that Croatian is added but not yet supported. I'm still waiting
for Nino Katic to give me a blessed translation.)
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
described here:
http://mail.python.org/pipermail/mailman-developers/2003-November/016119.html
The issue is that if we get an SMTPResponseException with a permanent
error code, we don't want to add all the recipients to the failures
dict, otherwise they'll all score a bounce. That's not appropriate
because the message is being rejected by the local smtpd for reasons
having to do with the content of the message, not because there's some
delivery problem to the recipient.
Log all problems now to logs/smtp-failure.
|
|
|
|
|
|
|
|
| |
address globally, it is possible the new address is already a member
of this (or one of the other) lists. In that case, the old address is
removed.
Closes SF # 835036.
|
|
|
|
|
| |
the valid postingaddr. If not given (the default), use
DEFAULT_EMAIL_HOST as before.
|
|
|
|
|
|
|
|
|
|
| |
calculated host name, and if VIRTUAL_HOST_OVERVIEW is true, that host
name must match one of our known virtual hosts. Otherwise we'll
refuse to create the list. If VIRTUAL_HOST_OVERVIEW is false, we'll
do no checking (meaning the old behavior of being able to create a
list with a non-fqdn hostname still exists).
Also, pass the email host name into the MailList.Create() call.
|
| |
|
|
|
|
|
|
|
|
| |
fix for SF bug # 846681, where the confirmation cookie emailed in a
bounce disabled notification was always out of date. The problem was
that we created the cookie at the time of the first bounce. This is
very likely more than three days before the time of the last bounce,
which is when we should have (and now do) create the cookie.
|
| |
|
|
|
|
| |
value to the whence argument.
|
|
|
|
| |
ApprovedAddMember().
|
|
|
|
|
|
|
|
|
|
|
| |
descriptive string explaining where the member got added from.
Improves subscription logging.
AddMember(): Pass 'remote' to ApprovedAddMember()'s whence argument.
ProcessConfirmation(): We know whether the confirmation came from
email or via the web, so pass this information to
ApprovedAddMember()'s whence argument.
|
| |
|
|
|
|
| |
user's realname and the user's language were added.
|
|
|
|
|
|
| |
was trying to use the setMemberOption() interface. This doesn't work
for delivery settings -- it needed to use the setDeliveryStatus()
interface.
|
|
|
|
|
|
|
| |
ignored for the unsub button on the member login page. Report and
original patch by Pasi Sjoholm, modify by Barry.
Forward port candidate.
|
| |
|
| |
|
|
|
|
| |
before closing the config.pck file. Should improve robustness.
|
|
|
|
|
| |
Promote SYNC_AFTER_WRITE to a Defaults.py/mm_cfg.py variable after
all.
|
|
|
|
|
| |
all, since we have the same potential problem (and solution) for the
MailList config.pck files that we did for the msg.pck files.
|
| |
|
| |
|
|
|
|
| |
of self.obscure_addresses.
|
|
|
|
|
|
| |
exploits found by Ned Dawes. Also, whitespace normalization.
Backported from HEAD.
|
|
|
|
|
|
|
| |
ENOENT but log all other IOErrors instead of propagating them up.
Closes a tiny exploit found by Ned Dawes.
Backported from HEAD.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
| |
'Release_2_1-maint'.
|
|
|
|
|
| |
I18ners please double check (except Tokio). It's always so painful to
resolve conflicts here, I hope I got them right.
|
| |
|