Age | Commit message (Collapse) | Author |
|
this is simplistic, but adds a minimal protection against trivial
DoS. the call to the ps command should be fast, but could use some
profiling for the case of some ten of thousands files.
- Resolves: #8778
|
|
|
|
use a powers of two ceiling for the reported size in the preamble.
for this to be effective against a passive adversary, cover traffic
should be used in the uploads too.
This is just a first-stop measure; proper research should be done to
determine a good tradeoff between avoiding information leakage and
saving some storage and bandwidth..
- Documentation: #8759
- Related: #8759
|
|
|
|
|
|
Motivation is that I need to pass partial data to the decryptor, mainly.
|
|
|
|
- Resolves: #8757, #8771, #8772
|
|
In order to do this from CI, please create a secret variable
(at i.e. https://0xacab.org/varac/soledad/settings/ci_cd)
called PYTEST_OPTS and use this example:
--benchmark-autosave --benchmark-storage="elasticsearch+https://USER:PASSWORD@moose.leap.se:9200"
|
|
If an application tries to delete None, by accident, raising earlier
would help to show the caller stack trace instead of inner sqlite
backend stack and make the error more friendly.
- Resolves: #8791
|
|
|
|
with @tayane
|
|
|
|
|
|
|
|
With this commit all tests on py34 tox environment
are collected.
|
|
|
|
At this point tests are not yet collectable
in the new environment. But no regressions observed.
|
|
On debian stable the `python3-tox` and `python-tox` packages
come with a really old version of `pip` and `setup_tools`.
This makes unusable PEP496 that comes handy when dependencies
are "py-version-related".
As a workaround we can "inject" an upgraded version of `pip` just
before running `tox`.
|
|
|
|
|
|
|
|
|
|
The offline flag is not needed and rendered the soledad client prone to
bugs because it would need to be toggled and that is succeptible to
timing problems.
|
|
|
|
|
|
Resolves: https://0xacab.org/leap/soledad/issues/8788
|
|
|
|
- use subprocess.check_call() to ensure any errors during twistd startup
will properly show up on test reports.
- use SIGTERM instead of SIGKILL to gracefully terminate twistd.
|
|
|
|
We were previously not using an empty local db for download benchmark
tests, so there was actually nothing to sync. This commit fixes that by
adding a way to force an empty local db on soledad client instantiation.
|
|
|
|
It needs a defer.Deferred
|
|
Handle it only if self.deferred wasnt called yet, otherwise that's just
an out-of-sync call from a scheduled deferred. Since it was already
logged, it's ok to ignore.
|
|
|
|
|
|
- Resolves: #8777
|
|
After refactor, the client secret bootstrap logic was flawed, and remote
secret was not being saved properly. This commit fixed that and tries to
improve the bootstrap code to make it more clear.
|
|
|
|
|
|
|
|
In order to be able to change passphrase, token and offline status of
soledad from the bitmask client api, the secrets api also has to be able
to use up-to-date values when encrypting/decrypting secrets and
uploading/downloading them to the server. This commit makes public some
soledad attributes that were previously "private" (i.e. used to start
with "_" and were not meant to be accessed from outside), and passes the
whole soledad object to the client secrets api. This makes the code
cleaner and also allows for always getting newest values of soledad
attributes.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
it seems evident that the functions were thought to pass a threadpool
along, but it finally wasn't properly passed and so there was a new
threadpool created to handle every resource.
I have removed the creation from the factory because I don't think it
makes sense to create a threadpool on the fly, it's prone to errors.
- Resolves: #8774
|
|
it doesn't make sense to create the resources for every request, we can
reuse the same resource and create it in the constructor.
- Resolves: #8770
|