Before this, the user's authentication token would still be stored in
the session even when it's found to be invalid. This caused a login
action to fail, but not in such a way that we would redirect to the login
page of Blender ID. Rather, it would keep you not logged in. By clearing
the session we're sure that the invalid token is forgotten, and the next
request will handle the login properly.
This uses the orphan-files.txt file output by find_orphan_files() to
mark those files as deleted. This allows for a two-stage approach, where
file IDs are found on one machine (against a read-only MongoDB slave, for
example) and soft-deleted on another machine (against a writable master).
This is quite a heavy thing to run, since it goes over all files of a
project, and then goes over every document in (almost) every collection
which has a property 'project' that's set to the project ID. It then goes
over every document to find all ObjectIDs and removes those from the set
of file ObjectIDs for that project. The remaining ObjectIDs are considered
orphans.
This is a very thorough search, but it doesn't require any knowledge of
the document and collection structure, so it should be future-proof.
User experience
===============
For users it means we can provide localized web-sites to enrich their
overall experiences.
Although for the Blender Cloud this doesn't make much sense (since the
content is in English), Flamenco and Attract can really benefit from
this.
New configuration settings
==========================
There are two new parameters in config.py:
* DEFAULT_LOCALE='en_US'
* SUPPORT_ENGLISH=True
They are both properly documented in the `config.py` file.
Technicall details
==================
We are using the 'Accept-Languages' header to match the
available translations with the user supported languages.
If an extension has a `translations` folder, it's used for translations.
However the main application (e.g., Blender Cloud) is the one that
determines the supported languages based on its `languages` folder.
How to mark strings for translation
===================================
See the documentation in README.md.
But as an example, 404.pug and pillar/__init__.py::handle_sdk_resource_invalid
have marked up strings that will be extracted once you install pillar,
or run any of the translations commangs.
Remember to **gulp** after you update the template files.
How to setup translations
=========================
You will need to create translation for the main project, and for each
extension that you want to see translated. I added a new entry-point to
the installation of Pillar.
So all you need is to use the `translations`
script to initialize, update and compile your translations.
Pending tasks
=============
Aside from marking more strings for extraction and start the translation
effort it would be interesting to replace the pretty_date routine with
momentjs.
Acknowledgement
===============
Many thanks for Sybren Stüvel for the suggestions and throughout code
review. Thanks also to Francesco Siddi for the original documentation
and suggesting me to tackle this. And Kudos for Pablo Vazquez for the
motivational support and for the upcoming "strings mark up" task force!
The core of the implementation is based on Miguel Grinberg i18n chapter
of his great 'The Mega Flask Tutorial'.
Reviewers: sybren
Differential Revision: https://developer.blender.org/D2826
This should be hidden in the UI as well, but the backend should support this too.
We also want to set initial rating of 1 for contents that need it.
This commit includes a new unittest for this case.
Reviewers: sybren
Differential Revision: https://developer.blender.org/D2825
This mimicks the behaviour of flask_login. In our case, it only redirects
when redirect_to_login=True and the user is anonymous. Otheriwse it still
results in a 403 Forbidden response.
The original file isn't always accessible, and can be of completely
different size than the downloadable variations. This is mostly appliccable
to videos.
Now settings live in a dedicated space, the settings blueprint can be used by Pillar applications, and the templates can be extended or overridden. Moved subscription and email settings to the blender-cloud repository.
Setting flask_login.current_user ourselves was a bad idea, and messed up
flask_login's internal administration. Our code now just manages
g.current_user in these specific instances, which works fine.
- No more direct access to g.current_user, unless unavoidable.
- Using pillar.auth.current_user instead of g.current_user or
flask_login.current_user.
- p.a.current_user is never checked against None.
- p.a.current_user.is_authenticated or is_anonymous is used, and never
together with a negation (instead of 'not is_anon' use 'is_auth').
- No more accessing current_user a a dict.
- No more checks for admin role, use capability check instead.
Also added SERVER_NAME in config_testing and pre-populated the keys of OAUTH_CREDENTIALS, since the implementation of providers is part of the application.
