| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This merges the SubscriptionManager and UnsubscriptionManager into a
single SubscriptionManager implementation that handles both register()
and unregister(). This allows us to use direct class-based adaptation
instead of the more clunky getAdapter() API. We can also eliminate the
funky _get_workflow() implementation detail.
This has a couple of side-effects. .confirm() must lookup the token in
the pendings database and pull out the pending type, dispatching to the
proper class depending on the type, or raising a LookupError if the
token is None or there is no pendable associated with the given token.
This feels like an acceptable trade-off.
However, this *also* means that IWorkflowStateManager must lose its
'name' argument in its methods. That's because we won't actually know
the name until its too late. Honestly, the name wasn't providing much
value anyway (it was always the subclass's name), so losing that seems
fine too. The complication here is that the name was a primary key in
the 'workflowstate' table, so we need to add its removal in the database
migration.
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
| |
use len() and slicing on SQLAlchemy query results. We also don't need to
list()-ify the results in the tests. This isn't perfect, but at least it
doesn't introduce yet another layer violation.
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* Make sure Registrar.discard() removces any workflow state manager state
associated with the token, and that this is flushed to SA.
* Adjust the email commands to the new IRegistrar API.
* Update the IRegistrar interface.
* Add IWorkflowStateManager.discard() and make `count` an attribute/property.
* Mark two tests as expected failures due to LP: #1444184.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
| |
load_external() now always opens in utf-8 mode.
More test repair.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* New events:
- ConfirmationNeededEvent is triggered when a pendable requiring
confirmation is created. This allows us to define an event handler for
this event which sends the user notification.
- SubscriptionEvent is triggered when a member is added to a mailing list.
This lets us define an event handler which sends the welcome message.
* send_welcome_message() now takes a member parameter instead of an address,
which lets us directly access the member's delivery mode and user display
name (if the member has a user, which it might not in some cases).
* Use the list id in the pendable record instead of the list name for
robustness (the latter can change but the former is permanent).
* Test more registration conditions.
* In the bin/runner command line switch handling, default `verbose` to None
instead of False. This makes it work better with nose's -E switch (log to
stderr).
* In call_api(), if a POST, PUT, or PATCH method is used and data is None,
encode the empty dictionary; seems like the behavior of urlencode() has
changed, so this is safer.
* Fix style and pyflakes warnings.
|
| | |
|
| | |
|
| |
|
|
|
| |
The code base is now also `python2.7 -3` clean, although there are still
some warnings in 3rd party dependencies. LP: #1073506
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- email.iterators.body_line_iterator() cannot handle unicodes in Python 2.6,
because it uses cStringIO.StringIO under the covers, and *that* can't handle
unicode. This works fine in Python 2.7, so I override this for the tests
only under 2.6 (the code itself isn't affected).
- AddressError needs to str() its IAddress attribute explicitly in the
__str__() method, otherwise under Python 2.6, you'll get unprintable reprs
in the doctests. Again, this works correctly in 2.7, but EIBTI, so it can't
hurt either way.
- EmailError: a new exception, not related to AddressError. The reason for
this it to conform to current nomenclature: "address" means an IAddress
while "email" means a text email address. So InvalidEmailAddressError
now derives from EmailError instead of AddressError because it gets passed a
text email address, and because that is invalid, it never gets turned into
an IAddress. The __str__() of this new base exception class does some
tricky encoding to keep it compatible between Python 2.6 and 2.7.
- UnverifiedAddressError derives from AddressError instead of the more generic
MailmanError.
- A few random code cleanups are included.
|
| |
|
|
|
|
|
|
|
| |
`digest=` argument now accepts the following values: `no` (for regular
delivery), `mime`, or `plain`.
Also:
- Fix get_queue_messages() to properly sort on the string representation of a
header, since it could be a Header instance.
|
| |
|
|
| |
are already members of the list.
|
| |
|
