| Age | Commit message (Collapse) | Author |
|---|
|
After we receive one document from the target database, we have to update the
target metadata or else we will not be able to succesfully return the new
generation and transaction id of the target when receiving exactly one
document during a sync.
|
|
Tests actually expect a tuple instead of a list on the return value of
get_sync_info().
|
|
|
|
Previous to this modification, the initialization of the sync decrypter pool
could happen concurrently with other database operations. That could cause the
pool to hang because it could be waiting for something that was mistakenly
deleted because of the wrong order of database operations.
This commit implements a standard which we already use in leap.keymanager and
leap.mail which makes some methods wait for the initialization operation
before they are actually called.
Closes: #7386
|
|
The old strategy for having Soledad sync running asynchronously with other
API calls was to have the sync running in its own threadpool. This is not
needed now that all sync code uses deferreds and will not block. This commit
removes what's left from the old threadpool.
|
|
Previous to this modification, the post-sync hooks were run after the sync
lock was released. In that scenario, the next sync process could start before
the previous sync's post-sync hooks were run. In general, we want the hooks to
run while the current sync lock is still locked, even though for some plugins
this might not make a difference.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
* Add possibility of passing sync_db and sync_enc_pool to target creation.
* Ensure defer_encryption and defer_decryption are True when setting up the
test environment.
|
|
* Remove test to assert database autocreation. We don't have that feature in
soledad and so don't need that test here.
* Remove redundant test without token authentication.
* Fix server setup.
* Skip interruptable sync test case, for now.
|
|
|
|
When you tried to start a local sqlcipher that was created
before, with the wrong passphrase, the code was raising
a sqlcipher DatabaseError, there were tests covering this
but they were expecting a WrongMacError that was never raised.
I added code to wrap the DatabaseError and raise a new exception
DatabaseAccessError that is specific to soledad and adapted the
tests to expect it
|
|
test_db_sync_autocreate was using yield but didn't have the
inlineCallbacks decorator, that made it a generator but the
trial tests don't support generators. Fixed this
|
|
multiprocessing.Queue is suitable for process communication, but its not
the ideal for a reactor model. This commit changes it to DeferredQueue,
where consumers and producers doesnt block and Twisted can handle them
better.
|
|
It was inheriting from twisted testcase in both u1db_tests.TestCase
and in BaseSoledadTest, removed u1db testcase and everything is ok now
|
|
Fixed pep8 warnings after latest PR
|
|
* redefine the BaseLeapTest class so it inherits from
twisted.trial.unittest.TestCase.
* allow passing the source_replica_uid when creating the sync target. This
is needed because the SoledadHTTPSyncTarget needs to be instantiated with
the source_replica_uid passed as a parameter.
* close the sync target in tests that use it.
* raise and U1DBError instead of an Exception on failure_and_retry test,
because otherwise the u1db http app server will re-raise the Exception and
twisted's WSGI server will raise the exception
* set defer_decryption=False in some sync_exchange() calls so the call
is consistent with soledad instantiation in tests and the new signature of
that method.
* yield get_sync_info() in tests that use this method.
|
|
When we started to use the twisted http agent, we forgot to intercept http
response and raise the appropriate u1db errors based on the response status
code and messages. This commit implements that by redefining the http body
reader used by the http agent.
|
|
While #7370 is being researched, it doesnt make sense to have this test
running for now. Will be enabled again after investigation finishes.
|
|
As suggested by drebs, the callback chaining can put None to sync,
bringing undesired behavior.
|
|
Some calls to soledad_sync_target were being made with
a full path, but the the method itself handles that, so I
removed the extraneous calls those problems
|
|
The first test to be run in the TestSoledadParseReceivedDocResponse
suite was erroring with connection refused, that is because it
actually connects to the server, but wasn't setting it up, I changed
it to inherit the simpler SoledadWithCouchServerMixin and adapted the
setUp and tearDown, now it works everytime with no problems
|
|
The encryption pool could be stopped twice and would break
on the second attempt because it deletes the encryption queue
variable. Added a condition to make sure it only deletes the
encryption queue if it exists, making it more idempotent
|
|
Test sync now doesn't inherit from u1db tests, it's
tests will now be run and were failing, so I fixed them.
Sync deferred will not be correctly handled in these tests.
I also pulled over a method from the u1db test and there seems
to be a pattern, I'll try to deduplicate these methods as soon
as it gets clearer
|
|
Change locking to be class based and each lock generated by db file
path.
|
|
Pulled the requirement for the test scenario in sqlcipher sync so
we could remove the dependency on the u1db test_sync
|
|
Sqlcipher sync tests were not being run because they were inheriting
from a u1db class that was having the tests skipped, I removed that
dependency and fixed the tests (they were expecting the synchronous,
old http_target api).
|
|
Test_http_target was using u1db test_sync in a lot of places,
I removed those references on the last changes and am removing
the dependency at once in this one
|
|
Removed inherit from u1db test that was causing these tests
to be ignored, the tests were broken with the latest changes
to the sync_target, so I fixed them, no tests are being skipped
on test_sync_target anymore
|
|
Sync very large files will now fail early because it was exploding
the memory, it needs some love, but for now I'll just make it fail
and open an issue to take a better look at it
|
|
Changes threading.lock to DeferredLock and checks syncing attribute by
looking into the lock state.
Also, applies more of startTwistedServer on tests that relies on
HTTP/1.1.
Fixes mock for events
|
|
Two spaces before class, not one
|
|
Some mixins were being called from u1db_tests but not used,
what was really needed was pulled from there and the remaining
dependencies were identified and used from the base
(u1db_tests.__init__.py)
|
|
Started looking at what we can do to remove the u1db test
dependencies directly, the parse was calling another class
setup that was in fact the u1db testcase, changed that to
testcase directly and it still works
|
|
U1DB test_http_app file was not being required anywhere and it's tests were
being skipped, removing it to clean up
|
|
Some tests still rely on python WSGI logic while others expect HTTP/1.1
from Twisted WSGI. While we can't use a single implementation for all,
both are needed.
startServer -> python default wsgi (http 1.0)
startTwistedServer -> Twisted WSGI (http 1.1 'production-like')
|
|
Soledad has a close method that wasn't calling http_target close. The
reference to sync exchange was being deleted without proper closing of
underlying resources.
|
|
Soledad was being served with python's default wsgi server, which uses
HTTP 1.0. Since Soledad server doesn't send length or chunked headers,
Twisted HTTP Client gets partial downloads when served with HTTP1.0.
Soledad is meant to be used with Twisted WSGI container on HTTP/1.1 and
the tests should reflect that.
|
|
|
|
- ability to disable sending (to get raw receive times)
- ability to specify size and number of payloads
- ability to point to a file to be used as raw source of payloads
|
|
|
|
|
|
Updating the profile-sync script to the latest deferred-based sync.
- Added a couple of options to have finer control about what output to
get.
- Add support for the plop profiler https://pypi.python.org/pypi/plop/
- To get meaningful plop profiles, make sure to disable the system stats
collection, like this::
./profile-sync.py --no-stats --plop -b /tmp/syncdata/ -p sikret user@provider
A good practice for this is having a pre-seeded soledad account
(currently you have to do that through the wizard, a cli will be very
nice to have in the coming future) with a known amount of data (for
instance, sending some mails with known payload weight).
It is VERY IMPORTANT that you *NEVER* process the data in this account
(ie, do not ever log in with a mail client, for instance, since that
will alter the original documents).
In order to have comparable results, you should always run this sync
script in similar conditions. Ideally, on a machine with runlevel 1.
Also, make sure of deleting the contents in the folder if you are
not using a temporal dir.
|
|
|
|
--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
|
