These models should be one-to-one. Various bugs in the past have
leaked violations in both directions; we clean them up such that every
PreregistrationUser which is not a multi-use invite has exactly one
Confirmation object associated with it.
Fixes: #22025
The "invites" worker exists to do two things -- make a Confirmation
object, and send the outgoing email. Making the Confirmation object
in a background process from where the PreregistrationUser is created
temporarily leaves the PreregistrationUser in invalid state, and
results in 500's, and the user not immediately seeing the sent
invitation. That the "invites" worker also wants to create the
Confirmation object means that "resending" an invite invalidates the
URL in the previous email, which can be confusing to the user.
Moving the Confirmation creation to the same transaction solves both
of these issues, and leaves the "invites" worker with nothing to do
but send the email; as such, we remove it entirely, and use the
existing "email_senders" worker to send the invites. The volume of
invites is small enough that this will not affect other uses of that
worker.
Fixes: #21306Fixes: #24275
This prevents users from hammering the invitation endpoint, causing
races, and inviting more users than they should otherwise be allowed
to.
Doing this requires that we not raise InvitationError when we have
partially succeeded; that behaviour is left to the one callsite of
do_invite_users.
Reported by Lakshit Agarwal (@chiekosec).
This only every did a single extra fetch, and now that this fetch is
in the same direction as our main fetching sequence, no longer is
meaningfully different from just specifying a different value of
consts.maximum_initial_backfill_size.
Previously, when a message is edited, it is locally echoed with its
pre-edit content.
This is because previously, when we tried to render the edited
message of the edit box during local echo, in order to update
the content, flags, and is_me_message properties of the message
object with that of those returned is markdown.render(), we used
the spread operator and created a new message object, and updated
the existing message object with this new one.
This was misconverted, since edit_locally() method already has a
fully-rendered message object to start with, and is just doing a
rerendering, it should be mutating what message it received, rather
than constructing a new local variable.
As a follow up to commit ad013a9, this updates the /help/dark-theme
docs to describe the newer way to switch themes via the theme switcher,
located in the Personal Menu.
This replaces the documentation on switching themes via
"Personal Settings".
Fixes#29550.
This removes the common min-width being used across all popovers and
instead sets the min-width according to a popover's requirements.
This allows for greater control over the popovers since we have a
variety of use cases for them — which a single common min-width cannot
accommodate.
Gitter broke their older API as part of being integrated
into Matrix.
Their announcement blog says:
"Anything left using the Gitter APIs will need to be
updated to use the Matrix API"
This commit drops the legacy Gitter import tool and
we plan to build a new one for Matrix in future.
For onboarding banners, we replace the close button with
"Got it" button.
Also, the banner is marked as read for the user only after
the "Got it" button is clicked. Earlier it was marked as read
as soon as it was displayed.
Register our `onTextInputHook` to be called on "input" events so that
the hook receives the updated text content of the input unlike the "keydown"
event which does not update the text content before running the hook.
Adds tools/i18n/create-legacy-stream-translations to create a
`legacy_stream_translations.json` file for every non-English
language locale that will serve as the legacy translations for
the stream to channel rename.
Adds tools/i18n/update-for-legacy-translations to manage adding
the legacy translations for any stream/channel strings that we'd
like to maintain existing translations for, which are defined in
LEGACY_STRINGS_MAP.
Updates the base hash for the streams setting overlay to be
"channels" instead of "streams".
Because there are Welcome Bot and Notification Bot messages that
would have been sent with the "/#streams" hash, we will need to
support parsing those overlay hashes as an alias for "/#channels"
permanently.
Part of the stream to channels rename project.
Updates the streams settings URL in the message view header to be
generated with hash_util.stream_edit_url, which correctly encodes
the stream name in the URL.
Additionally, the text colors have been updated for both light and dark
themes, it starts showing when 900 or less characters are left, as 999
was too soon, and has a tooltip to show the maximum characters limit.
Fixes: #28706.
We now "first_message_id" of the stream on the deletion of the first
message that was sent to it. This results in 1 extra query when any
stream message is deleted and 3 extra queries when the first message
sent to any stream is deleted.
Fixes#28877.
Mateusz Mandera
N-Shar-ma
Tim Abbott
Alex Vandiver
roanster007
nimishmedatwal
Karl Stolley
Alya Abbott
Sayam Samal
Prakhar Pratyush
Anders Kaseorg
luska44
Lalit
Lauryn Menard
Aman Agrawal
Vector73