Changed:
- ♻️(backend) reset collaboration connection in cascade for all children
Fixed:
- 💄(frontend) fix some UI/UX in the left panel
- 🐛(frontend) fix tree dnd firefox
- Collapse panel button is now:
- fully clickable,
- the title is hidden
- when the doc title is visible
- when small mobile the title is hidden
- Add button login when logout whatever the device
- Fix flickering logo
- New icon name to avoid cache
With Firefox, the DND was not working properly anymore.
From a previous refacto, a item is now a link,
by default a link has a draggable attribute set
to true, it was conflicting with the DND library,
so we need to set it to false.
When an access is updated or deleted, or a link_configuration is
changed, the collaration_service is used to reset the connection in the
collaboration server. As accesses and link_configuration are inherited
in a Docs tree, if the user is connected to a child, the connection is
not reset. This commit fix this issue by calling the reset on every
children in the tree.
In the settings, the logging has set the propagate property to False.
While writing tests it wasn't possible to assert log was made using the
caplog fixtures because this propagate is set to False. Changing it to
True allow to correctly tests that a log is made using the caplog
fixture.
The y-provider `/api/convert` handler catches every writer/reader
exception and returns a generic `{"error":"An error occurred"}` 500. The
real JS exception is silently dropped: `logger()` only prints to stdout
when `COLLABORATION_LOGGING=true`, and Sentry's
`setupExpressErrorHandler` only sees unhandled errors — never this one.
Result: Sentry issues `DOCS-4ZY` / `DOCS-4ZZ` / `DOCS-6DK` all fire on
every 500, but none of them carry the underlying JS stack, making
root-cause diagnosis impossible from Python alone.
This patch adds `Sentry.captureException` in the outer catch with
`handler:convert` tag and `contentType` / `accept` / `bodyBytes` extras.
Behavior otherwise unchanged.
Signed-off-by: Stephan Meijer <me@stephanmeijer.com>
Added:
- ✨(y-provider) preserve callouts, PDFs, page breaks, interlinking
links and commented text on HTML/markdown export
- ✨(frontend) add a user menu
- ✨(frontend) new header and responsive harmonization
- ✨(backend) add management command to reset a Document
Changed:
- ♿️(frontend) hide mobile left panel from screen readers when collapsed
- ♿️(frontend) enable blocknote heading ids for toc anchors
- ♿️(frontend) focus export modal on format select
- ♿️(frontend) configurable legal submenu in HelpMenu, remove Crisp
- ♻️(frontend) new create button for docs
- ♿️(frontend) align search modal field label with placeholder
- 🚚(frontend) move Waffle to bottom left
- ♿️(frontend) remove redundant aria-label on table of contents links
- ♻️(core) fix typo in settings COLLABORATION_WS_NOT_CONNECTED_READY_ONLY
- ♻️(backend) scope document search by document id instead of path
Fixed:
- 🐛(backend) prevent owner from leaving a soft-deleted document
- 🐛(frontend) fix removed item in the tree
- 🐛(frontend) fix service worker causing reload on tab focus
- 🐛(backend) update restore ability for inherited deletion
- 🔧(dev) make the dev stack domain-agnostic
- 🐛(frontend) stop force index redirect when delete doc
- 🐛(frontend) fix CTA on move modal on mobile
Removed:
- 🔥(backend) remove unused default authentication backend
We decided to hide the import button on a
document because it is confusing for the user to
have the possibility to import a document when he
is already in a document. The import button is
now only available on the document list page.
We will add back the button soon, but a import
will create a new sub-document instead.
The search endpoint does not accept anymore the usage of the path
parameter but it is using the document id directly. We have to reflect
this changes in the frontend application.
The search in a document tree was triggered by the usage of the document
path. The path is something guessable by incrementing it you can
discover public documents. We decided to change this to use the document
id which is not guessable and prevent discovering public documents.
Thanks to @maboukerfa for discovering it.
Following the recommendations of the design team,
we lightly improved some UI/UX elements to be
more consistent and user-friendly before the
next release.
It includes the following changes:
- add a background to the grid
- add a fadein to the header buttons
- reduce the generic size icon
- add a title to the doc item
On some mobile devices, the CTA button on the
move modal was not visible due to a CSS issue.
This commit fixes the issue by adjusting the modal's
layout and ensuring that the CTA buttons are visible.
Keycloak's hostname, its client's redirect/web
origins, and Django's CSRF trusted origins were
hardcoded to localhost, so the dev stack was
only reachable from that exact origin.
It is now possible to run the dev stack depending
what is configured in the env files.
Depends the state, after deleted a doc, we were
automatically redirecting to the index page, which is
not always the desired behavior. This change stops
the automatic redirect after deleting a document,
allowing for more flexible navigation based on user
context.
The settings COLLABORATION_WS_NOT_CONNECTED_READY_ONLY contains a typo.
We don't want to see READY_ONLY but READ_ONLY. This commit fix the typo
and change everywhere in the codebase the settings name.
We want to animate the header floating bar with
a Lottie animation. This commit adds the Lottie
dependency and implements the animation in the
HeaderFloatingBar component.
We change the way that the left panel is displayed
in mobile and tablet views. Now, the left panel will
be over the content when mobile, and will be displayed
as a side panel when tablet. We simplify the
overall structure of the left panel, and the way
the responsive behavior is handled.
We replace the header by a floating header, to have more
space for the content and in fine to have more
harmony between the doc and the doc-grid, in mobile view
or not.
The layout have to be adapted, as well the legacy home
page.
We want to use the floating bar in multiple places, so
we need to make it more generic and reusable.
This commit refactors the floating bar component
and its related components to achieve this goal.
We need a management command to reset a Document to an initial state and
deletes everything related to it. This command can be usefull to reset a
demo for example.
The authentication backend
mozilla_django_oidc.contrib.drf.OIDCAuthentication
is present in the default authentication classes for the REST_FRAMEWORK
settings. This backend should not be used by our application and can
lead to the usage of our main api with an access_token instead of the
cookie session.
We need to override the drf SessionAuthentication backend to implement
the authenticate_header method. Without this, a 403 status code is
returned, but it is not valid. It must a be 401
We are now using the UserMenu component from
the ui-kit. It is a dropdown displaying the user's
name and email, along with a logout button and a
language picker. To fit the design of the user menu,
we adapted the LanguagePicker adn are now using the
ui-kit language picker component.
In the yprovider deployment we now need to mount the cacert containing
the certificates generated by mkcert. Then, we have to declare the
environment variable NODE_EXTRA_CA_CERTS with the path where this cacert
is mounted. With this new config we don't have tls issue anymore.
Updated the restore ability so that only directly deleted documents can
be restored. This prevents the restore action from being exposed for
child documents whose parent is deleted. This PR addresses #2127.
Owners could leave trashed documents via the leave endpoint, deleting
their `DocumentAccess` and all subtree access. When the last owner
leaves a document it creates an orphan as no owners remains in the doc.
Signed-off-by: BOUKERFA Mohamed El Amine <boukerfa.ma@gmail.com>
The floating waffle menu is positioned only on the bottom
left of the trigger button, so when positioned
on the bottom left, the floating menu is not visible.
This commit fixes the positioning.
We also added a patch to the ui-kit, so when
the patch is released, we can remove this patch
from the codebase.
During rolling deploys, multiple replicas serve
different builds of service-worker.js
(each build has a random buildId baked in).
This causes a controllerchange loop: the new SW
activates via skipWaiting, the page reloads, then the
next request may hit an older replica,triggering another
update and another reload — surfacing as a reload
every time the tab becomes visible.
We added a sessionStorage guard so a controllerchange
triggers at most one automatic reload per session; subsequent
flaps between versions are silently ignored, while legitimate
updates still cause a single reload automatically.
We added a sub button to create sub-documents
from the New button. This sub button is only visible
when the user is in a document, and it allows them to
create a new document that is a child of the current
document.
New create button for docs.
It is now a link button instead of a pure button,
meaning we can navigate to this new created doc
and open it in another tab by doing ctrl + click.
This new button has a dropdown menu that allows
users to import a docx file or a markdown file.
We removed the old import button.
The 0.51.4 markdown serializer no longer emits the optional link title
attribute, so the exported link is `[text](url)` without ` "title"`. The
title is still present in the HTML export. Update the expectation to match.
The "comment" mark must exist in the conversion editor schema, otherwise
y-prosemirror silently drops every commented run of text when reading the
Yjs document, turning any block holding a comment into an empty block.
Register the CommentsExtension on the conversion editor so commented text
round-trips like the other custom blocks.
Wire the docs BlockNote schema (callout, pdf, uploadLoader, interlinking
link, page break) into the conversion editor so /api/convert no longer
drops or mangles these blocks.
The tree rerenders crazily, creating performance
issues, especially on big trees. To fix it properly
we will need a big refacto, on the uikit side.
Before doing that, we try to optimize the current
code by memoizing the tree view, and by using the
selected node from the url instead of the one
from the context.
We improve the treeitem link, we can now easily
opening a new tab by doing CTRL+Click.
react-arborist's scrollTo calls react-window's scrollToItem, which mutates
the internal scrollOffset state. When navigating to a deep item in a large
tree, this causes all items above the target to be removed from the DOM
(virtualized away), making the tree appear empty above the selected node.
We no-op it to prevent that — the panel's own overflow-y handles scrolling.
Added
- ✨(backend) add limit on distinct reactions per comment #1978
- ✨(frontend) leave a document #2410
- ✨(frontend) add top parent on sub docs search #1952
- ✨(frontend) unauthenticated users can search #2407
- ✨(backend) specific user delete method to delete its relations #2437
Changed
- 👷(CI) remove test-e2e-other-browser job #2404
- ♿️(frontend) use heading element for pinned documents section title #2380
- ♿️(frontend) use anchor links for table of contents entries #2390
- ♿️(frontend) improve presenter mode screen reader and keyboard support #2383
- ♿️(frontend) link export modal name to its heading #2422
Fixed
- 🐛(frontend) overlap of block menu dropdown #2406
- ⚡️(backend) fix N+1 queries when serializing thread comments #2415
