We change the import order to import UserGroup objects before
Stream such that we can set can_remove_subscribers_group correctly.
We do not import UserGroupMembership objects here along with
UserGroup since UserProfile objects are not imported and
GroupGroupMembership are also imported later as these are not
required before.
Earlier, if the content of the modal (apart from the header and footer)
overflowed, the whole modal would become scrollable which would hide
the modal header and footer on scrolling. This commit makes only
the modal content scrollable and keeps the modal header and
footer static.
This was added to avoid the `owner` dropdown field from being
only partially visible in the `Manage bot` modal. But, now the
`owner` field is moved up in the modal so this isn't needed.
cachify has been removed in 9d448e73d2.
We don't need to keep its tests.
TODO: functools.lru_cache can be replaced by functools.cache when we
drop Python 3.8.
Signed-off-by: Zixuan James Li <p359101898@gmail.com>
Updates the base titles so that they begin with the page content,
and end with "| Zulip" + information about the type of doc: "help
center", "API documentation", "terms and policies", or "integrations".
Adds a META_CATEGORY dict for categories that are not best described
as groups of 'tools', so that in the subsequent commit the PAGE_TITLE
can be set accordingly.
Also, removes 'tools' from the 'Miscellaneous' category text and
spells out 'Human resources' instead of using 'HR'.
Since we are updating the help center documentation to use `<kbd>`
HTML elements when documenting keyboard keys and shortcuts, we no
longer need to support `<code>` HTML elements in
`adjust_mac_shortcuts`.
Updates the tutorial in "Writing help center articles: Writing
style" for using `<kbd>` HTML elements when documenting keyboard
shortcuts in help center articles.
Also, adds section about the keyboard tip macro/syntax.
Updates references to keyboard keys in the help center docs to use
`<kbd>` HTML elements, which also means updating them to be as the
key would appear on a keyboard.
Previously, uppercase and lowercase letters were used to indicate
when/if the `Shift` key was being used, and even that was not
consistent throughout the documentation.
For CSS styling, adds a similar rule for `<kbd>` elements that is
used in `/static/styles/app_components.css`. And updates the CSS
class used in `/static/js/portico/help.js` for `adjust_mac_shortcuts`
accordingly.
Also, takes advantage of revising these pages for making small
updates for current help center documentation practices.
We do not need direct_members and direct_subgroups field of
UserGroup objects in the export data since we already have
UserGroupMembership and GroupGroupMembership object data.
While importing we keep these fields empty when creating
UserGroup objects and direct_members and direct_subgroups
fields will get set when UserGroupMembership and
GroupGroupMembership objects are created.
This change will also help us in further changes when we
will change the order of importing to import UserGroup
objects just after Realm objects.
Although our POST /messages handler accepts the ‘to’ parameter with or
without JSON encoding, there are two problems with passing it as an
unencoded string.
Firstly, you’d fail to send a message to a stream named ‘true’ or
‘false’ or ‘null’ or ‘2022’, as the JSON interpretation is prioritized
over the plain string interpretation.
Secondly, and more importantly for our tests, it violates our OpenAPI
schema, which requires the parameter to be JSON-encoded. This is
because OpenAPI has no concept of a parameter that’s “optionally
JSON-encoded”, nor should it: such a parameter cannot be unambiguously
decoded for the reason above.
Our version of openapi-core doesn’t currently detect this schema
violation, but after the next upgrade it will.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
The previous schema incorrectly prohibited the string, integer, and
string-array forms that we do in fact accept.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
The previous schema incorrectly prohibited the two-element array form
that we do in fact accept, and didn’t specify anything about the
contents of the object form.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
Adds a legend to the "org-registration" fieldset where the user
inputs info about the new organization. Previously, there was only
a legend for the "user-registration" fieldset. Also, increases
the opacity of CSS rule for the legend color so that they are
more visible.
Removes the hint text about what the organization URL value is
used for since it is pretty clear from the field's name.
Disable "External account type" select field from External account type
profile field edit form, doing it because this select field is not
editable and it is incorrectly looks like editable.
change the names of "github" and "twitter" external account fields to
"GitHub username" and "Twitter username" respectively and remove the
hints of them.
Removes the prefilled support email content in `500.html` and
`unsubscribe_link_error.html` templates since both cases are rather
rare and the prefilled content is not useful for organizations that
do not use English as their main communication language.
In 5c49e4ba06, we neglected to include
the CSRF and caching decorators required for all API views in the new
remote_server_dispatch function.
I'm not sure why our automated tests didn't catch this, but this made
the remote server API endpoints nonfunctional in a production
environment.
Until now, whenever typeahead autocompleted the spoiler syntax, there
was no indication if and how a visible header to the hidden content
could be added.
Now when autocompleting, the word "Header" is added as a placeholder
and highlighted, hinting at the format.
Fixes: #20868.
This makes the implementation more readable, and also dependant only
on the `highlight` object, not the detail that previously only slash
commands, which have an `item.placeholder`, used this code path.
This may have originally made sense as a class that managed the
browser state, but it has since turned into a dumping ground for
mostly pure functions that don’t make sense to instantiate.
Signed-off-by: Anders Kaseorg <anders@zulip.com>