historically, Libervia (SàT before) was using camelCase as allowed by PEP8 when using a
pre-PEP8 code, to use the same coding style as in Twisted.
However, snake_case is more readable and it's better to follow PEP8 best practices, so it
has been decided to move on full snake_case. Because Libervia has a huge codebase, this
ended with a ugly mix of camelCase and snake_case.
To fix that, this patch does a big refactoring by renaming every function and method
(including bridge) that are not coming from Twisted or Wokkel, to use fully snake_case.
This is a massive change, and may result in some bugs.
the caching is triggered by `li.blog.get`, but it is done in parallel without blocking the
command, thus there is a possibility that the data is not yet in cache when the following
test is done.
A loop with delay is done, to be sure that cache is filled before doing the assertion.
items are stored in chronological order in Pubsub, and as we want the more recent one, we
need to set `before=''`. Otherwise, we can get a wrong item from a previous test.
RSM `max` alone will return first (oldest) items, while `max_items` will return latest
ones. As Libervia Pubsub as been fixed in this regard, tests are now fixed too.
- e2e tests are launched inside the new docker e2e test environment
- `run_e2e.py` launch the docker container, mount the current code base in it, launch the
e2e tests and print report in real time
- `conftest.py` are pytest fixtures managing many things such as account creation, fake files
management, JSON or Domish.Element parsing, fake editor, etc.
- `test_jp.py` are end-to-end test done with `jp`. `sh` library is used to make tests
writting as user-friendly as possible. The `SAT_TEST_ENV_E2E` environment variable is
checked, and tests will be skipped if it's not set.
