| Age | Commit message (Collapse) | Author |
|---|
|
As a class variable multiple account instances share
mailboxes which is bad if its different users or tests
|
|
|
|
|
|
The key extract should check first for attached keys and if this fails
then should try the OpenPGP header.
- Resolves: #7480
|
|
|
|
|
|
|
|
We can't import attached keys with different email address than the
sender. Now we don't fail in this case, just log it.
- Resolves: #7454
|
|
|
|
in this way, we're using twisted reactor instead of having another
thread with zmq's own copy of tornado ioloop.
Resolves: #7274
|
|
|
|
So now it:
- Accepts credentials in a file pointed by environment variable.
- Allows to specify the mailbox to select as a command line flag.
- Allows to select a given message by subject.
For example:
BITMASK_CREDENTIALS=/tmp/bm.secrets ./getmail --mailbox INBOX --subject
'test mail
The two flags are case-insensitive.
This is intended to be used as a helper in end-to-end tests.
Getting a message by subject it's suboptimal, but I think it's good
enough for our testing purposes right now.
Related: #7427
|
|
|
|
This way we can tell if a message was originally encrypted, so that we
can show that information to the end user.
|
|
|
|
|
|
|
|
|
|
In the constructor values already is normalized (i.e. with underscores),
while kwargs contains items that are not normalized (i.e. with dashes).
Joining the dicts resulted in two entries that only differed by dash or
underscores. The setattr then set the value that occurred later in
items, thereby sometimes overriding the correct value with the default
one.
|
|
The get_raw_docs method accesses header field content-transfer-encoding
using the string 'content-transfer-type' so the raw doc dict always ends
up with that value set to empty string.
|
|
--use-leap-wheels sets --trusted-host (remove it when we have a proper
cert) and WHEELHOUSE to https://ftp.lizard.leap.se
Until we get ftp.lizard cname, use lizard as the wheels server.
- Related: #7339
|
|
|
|
This test was not updated for a while and it doesn't make sense
in this context, when we create acceptance tests we can make sure
we cover this feature
The issue for acceptance tests is:
https://leap.se/code/issues/7341
|
|
Functionality that is not implemented should not fail, instead
it should be skipped until it is implemented.
Added a return to test_imap_store_fetch setup to make sure the deferred
is being handled
|
|
Added requirements-latest to make sure we always use the
latest develop of all dependencies in an automated way
|
|
generate_wheels uses $WHEELHOUSE to generate and store the wheels for
requirements.pip and requirements-testing.pip (if it exists).
pip_install_requirements.sh installs requirements.pip from them if
possible (if not, then it fetches them from pypi) or, if passed the
--testing flag, it installs requirements-testing.pip.
Related: #7327
|
|
|
|
The lambdas take two args, so it needs to be a tuple.
Furthermore filter needs a collection.
|
|
- update pip
- install base reqs
|
|
|
|
this is part of a process to make the setup of the development mode less
troublesome. from now on, setting up a virtualenv in pure development
mode will be as easy as telling pip to just install the external
dependencies::
pip install -r pkg/requirements.pip
and traversing all the leap repos for the needed leap dependencies doing::
python setup.py develop
- Related: #7288
|
|
|
|
of pep8
|
|
sync_hooks to lambdas to real functions
|
|
|
|
|
|
This is quite manual for the moment being, and it's not integrated into
the unittests. But it is useful to have it documented, with some luck we
can automate the process even more and add it to the CI quite soon.
|
|
in order to start with a fresh folder each time.
current test script has some troubles with dirty state.
|
|
|
|
For #7244, this is needed for regression tests
|
|
|
|
|
|
this is just the bare minimum implementation of MSN (message sequence
numbers). It is not enough feature-wise, but I'm doing it now just to
support testing with the default imap client that we're using with the
regression tests.
|
|
In the case of a message with just one non-text attachment, something
has to be returned for the body payload. (For instance, a message with
only one image).
|
|
|
|
The decission to index the pending_inserts dict by message-id is a bit
brittle. I assume any well-formed message in the RealWorld (tm) will
have this header, but many samples used in the tests will break this
assumption.
|
|
For some reason that I haven't discovered yet, nested multipart is
hitting an off-by-one error (that had been wrongly ammended in a
previous commit, breaking many other cases but fixing the particular
Mail.app sample I was working with). This is just a temporary hack
to make all the current regression tests happy, but further
investigation is needed to discover the cause of the off-by-one part
retrieval solved and correctly documented.
|
|
This commit fix a very simplistic and until now broken handling of
nested multipart that went undetected due to the structure of the mails
used in tests until now. Incidentally, the way that Mail.app structures
attachments made this bug noticeable.
There was also an off-by-one indexing error when retrieving the subpart
message for a given subpart.
Be aware that the current implementation will only handle correctly 2
levels of multipart nesting. Extending beyond in a more generic way will
need further work.
Closes: #7244
|
|
|
|
|
