Original commit 3b59d3ee9a
Breaking commit 32e25ce129
The breaking commit was actually meant to remove the need for this
hook logic entirely, by relying on a custom validator instead.
This works for nodes, but it currently does not work for projects.
The issue needs to be further investigated via T63006.
Use this decorator on Flask endpoints that should respond with CORS
headers. These headers are sent in a reply when the browser sends an
`Origin` request header; for more info see [1].
This commit rolls back the previous commit (0ee1d0d3), as this new
approach with a separate decorator is both easier to use and less
error-prone.
[1] https://developer.mozilla.org/en-US/docs/Web/HTTP/CORS
This makes the `require_login` decorator always return a Flask response.
Previously it could also raise a `Forbidden` exception; now it returns a
403 Forbidden response in that case too.
For the user:
Ctrl + L-Mouse to select multiple tasks/shots/assets and then edit
the nodes as before. When multiple items are selected a chain icon
can be seen in editor next to the fields. If the chain is broken
it indicates that the values are not the same on all the selected
items.
When a field has been edited it will be marked with a green background
color.
The items are saved one by one in parallel. This means that one item
could fail to be saved, while the others get updated.
For developers:
The editor and activities has been ported to Vue. The table and has
been updated to support multi select.
MultiEditEngine is the core of the multi edit. It keeps track of
what values differs and what has been edited.
Notifications for when someone posted a comment on your node
was not created.
Root cause was that default values defined in schema was not set,
resulting in activity subscriptions not being active.
There were 2 bugs preventing them to be set:
* The way the caching of markdown as html was implemented caused
default values not to be set.
* Eve/Cerberus regression causes nested default values to fail
https://github.com/pyeve/eve/issues/1174
Also, a 3rd bug caused nodes without a parent not to have a
subscription.
Migration scripts:
How markdown fields is cached has changed, and unused properties
of attachments has been removed.
./manage.py maintenance replace_pillar_node_type_schemas
Set the default values of activities-subscription
./manage.py maintenance fix_missing_activities_subscription_defaults
These are used by Flamenco to store task logs as gzipped text files, but to
send them to the browser with such HTTP headers that the browser can gunzip
them and display directly (rather than having to download & gunzip yourself).
For now this returns a bucket in the default storage backend, since
individual projects do not have a 'storage backend' setting (this is
set per file, not per project).
* Drag and drop files to comment editor to add a file attachment
* Using Vue to render comments
Since comments now has attachments we need to update the schemas
./manage.py maintenance replace_pillar_node_type_schemas
Changed how and what we store in elastic to unify it with how we store
things in mongodb so we can have more generic javascript code
to render the data.
Elastic changes:
Added:
Node.project.url
Altered to store id instead of url
Node.picture
Made Post searchable
./manage.py elastic reset_index
./manage.py elastic reindex
Thanks to Pablo and Sybren
project.
Javascript tutti.js and timeline.js is needed, and then the following to
init the timeline:
$('.timeline')
.timeline({
url: '/api/timeline'
});
# Javascript Notes:
## ES6 transpile:
* Files in src/scripts/js/es6/common will be transpiled from
modern es6 js to old es5 js, and then added to tutti.js
* Files in src/scripts/js/es6/individual will be transpiled from
modern es6 js to old es5 js to individual module files
## JS Testing
* Added the Jest test framework to write javascript tests.
* `npm test` will run all the javascript tests
Thanks to Sybren for reviewing
This adds a new Blinker signal `user_logged_in` that is only sent when
the user logs in via the web interface (and not on every token
authentication and every API call).
Asset nodes now have a new field called "properties.duration_seconds". This holds a copy of the duration stored on the referenced video file and stays in sync using eve hooks.
To migrate existing duration times from files to nodes you need to run the following:
./manage.py maintenance reconcile_node_video_duration -ag
There are 2 more maintenance commands to be used to determine if there are any missing durations in either files or nodes:
find_video_files_without_duration
find_video_nodes_without_duration
FFProbe is now used to detect what duration a video file has.
Reviewed by Sybren.
Synchronisation is performed in the background by the Celery Beat, every
10 minutes. It has a time limit of 9 minutes to prevent multiple refresh
tasks from running at the same time.
Synchronisation is also possible with the `manage.py badges sync` CLI
command, which can sync either a single user or all users.
This also changes the way we treat Blender ID tokens. Before, the Blender ID
token was discarded and a random token was generated & stored. Now the
actual Blender ID token is stored.
The Facebook and Google OAuth code still uses the old approach of generating
a new token. Not sure what the added value is, though, because once the
Django session is gone there is nothing left to authenticate the user and
thus the random token is useless anyway.
This partially reverts commit c57aefd48b.
The code to check against hashed tokens remains, because existing tokens
should still work.
The unhashed tokens are necessary for fetching badges from Blender ID.
Video progress updates:
- Mark as 'done' when 90% or more is watched.
- Keep 'done' flag when re-watching.
The video progress is stored on three events, whichever comes first:
- Every 30 seconds of video.
- Every 10% of the video.
- Every pause/stop/navigation to another page.
- When we detect the video is looping.
