This commit updates the API to check the permission to subscribe other
users while inviting. The API will error if the user passes the
"stream_ids" parameter (even when it contains only default streams)
and the calling user does not having permission to subscribe others to
streams.
For users who do not have permission to subscribe others, the
invitee will be subscribed to default streams at the time of
accepting the invite.
There is no change for multiuse invites, since only admins are allowed
to send them, and admins always have the permission to subscribe
others to streams.
Since 74dd21c8fa in Zulip Server 2.1.0, if:
- ZulipLDAPAuthBackend and an external authentication backend (any aside
of ZulipLDAPAuthBackend and EmailAuthBackend) are the only ones
enabled in AUTHENTICATION_BACKENDS in /etc/zulip/settings.py
- The organization permissions don't require invitations to join
...then an attacker can create a new account in the organization with
an arbitrary email address in their control that's not in the
organization's LDAP directory.
The impact is limited to installations which have the specific
combination of authentication backends described above, in addition to
having the "Invitations are required for joining this organization
organization" permission disabled.
This argument was added with the default incorrectly set to `True` in
bb0eb76bf3 - despite
`maybe_send_to_registration` only ever being called in production code
in a single place, with `password_required=False` explicitly. And then
it just got carried forward through refactors.
`maybe_send_to_registration` was/is also called twice in tests, falling
back to the default, but the `password_required` value is irrelevant to
the tests - and if anything letting it use the `True` has been wrong,
due to not matching how this function is actually used.
Servers that had upgraded from a Zulip server version that did not yet
support the user_uuid field to one that did could end up with some
mobile devices having two push notifications registrations, one with a
user_id and the other with a user_uuid.
Fix this issue by sending both user_id and user_uuid, and clearing
The existing `except subprocess.CalledProcessError` only catches if
there are syntax errors which prevent the `lastrun` file from being
written; it does not handle if there are properly-defined resources
which fail to evaluate (e.g. due to a missing dependency or file).
Check the `failed` resource count, and exit 2 if there are any such
resources. This will cause `zulip-puppet-apply --force --noop` (which
is used as a pre-flight check during upgrades) to properly detect and
signal on more types of invalid puppet configurations. In turn, this
will cause `upgrade-zulip` to not attempt to power through upgrades it
knows are destined to fail.
We now add a checkbox in "Steams they should join" to directly
select all default streams in the realm for the invite.
We hide the stream list if that option is selected.
We now allow users to invite without specifying any stream to join.
In such cases, the user would join the default streams, if any, during
the process of account creation after accepting the invite.
It is also fine if there are no default streams and user isn't
subscribed to any stream initially.
We do not add user to the default streams if the streams list passed
while sending the invite (both email and multi-use) was empty since
invite explicitly selected to not subscribe the user to default
streams.
The previous implementation leaked database connections, as a new
thread (and thus a new thread-local database connection) was made for
each timer execution. While these connections were relatively
lightweight in Python, they also incur memory overhead in the
PostgreSQL server itself. The logic for managing the timer was also
unclear, and the unavoidable deadlock in the stopping logic was rather
unfortunate.
Rewrite with one explicit worker thread which handles the delayed
message sending. The RabbitMQ consumer creates the database rows, and
notifies the worker to start its 5s timeout. Because it is controlled
by a condition variable, it does not hold the lock while waiting, and
can be notified to exit.
(cherry picked from commit 3efc0c9af3)
Previously, `QuerySet` does not support isinstance check since it is
defined to be generic in django-stubs. In a recent update, such check is
possible by using `QuerySetAny`, a non-generic alias of `QuerySet`.
Signed-off-by: Zixuan James Li <p359101898@gmail.com>
(cherry picked from commit 24f24d236d)
Black 23 enforces some slightly more specific rules about empty line
counts and redundant parenthesis removal, but the result is still
compatible with Black 22.
(This does not actually upgrade our Python environment to Black 23
yet.)
Signed-off-by: Anders Kaseorg <anders@zulip.com>
(cherry picked from commit df001db1a9)
Fixes#23588
When a new message arrives in an inactive stream, it will lead
to stream being marked as active from inactive which leads to
stream list being rearranged which zooms out the user since the
active stream is reset momentarily. To avoid this, we delay the
stream list redraw until user zooms out and only update the topics
list.
Having active streams greyed out can be confusing to the user. This
is especially useful when the app is still fetching messages on
a reload and the active stream has no messages and is thus marked
as `inactive`.
Zulip Server 4.0 is now 22 months old, which is more than 18 months.
Per the general policy in the "Client apps" section below, that
means it's time to drop support for older versions.
We released 5.0 near the end of 2022-03, so near the end of 2023-09
we can update this further to say 5.0.
‘exit’ is pulled in for the interactive interpreter as a side effect
of the site module; this can be disabled with python -S and shouldn’t
be relied on.
Also, use the NoReturn type where appropriate.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
The `pg_upgrade` tool uses `pg_dump` as an internal step, and verifies
that the version of `pg_upgrade` is the same exactly the same as the
version of the PostgreSQL server it is upgrading to. A mismatch (even
in packaging versions) leads to it aborting:
```
/usr/lib/postgresql/14/bin/pg_upgrade -b /usr/lib/postgresql/13/bin -B /usr/lib/postgresql/14/bin -p 5432 -P 5435 -d /etc/postgresql/13/main -D /etc/postgresql/14/main --link
Finding the real data directory for the source cluster ok
Finding the real data directory for the target cluster ok
check for "/usr/lib/postgresql/14/bin/pg_dump" failed: incorrect version: found "pg_dump (PostgreSQL) 14.6 (Ubuntu 14.6-0ubuntu0.22.04.1)", expected "pg_dump (PostgreSQL) 14.6 (Ubuntu 14.6-1.pgdg22.04+1)"
Failure, exiting
```
Explicitly upgrade `postgresql-client` at the same time we upgrade
`postgresql` itself, so their versions match.
Fixes: #24192
We will hopefully be able to just this in #16208 to document what
users need to configure in order to do this manually, but the content
here will be useful for anyone who hasn't set that up regardless.
The previous logic incorrectly tried to map elements of the list of
user IDs beyond 4 senders to their `.id` fields, which were undefined;
the correct thing to do is just use the list of user IDs that we
already have.
This is the behaviour inherited from Django[^1]. While setting the
password to empty (`email_password = `) in
`/etc/zulip/zulip-secrets.conf` also would suffice, it's unclear what
the user would have been putting into `EMAIL_HOST_USER` in that
context.
Because we previously did not warn when `email_password` was not
present in `zulip-secrets.conf`, having the error message clarify the
correct configuration for disabling SMTP auth is important.
Fixes: #23938.
[^1]: https://docs.djangoproject.com/en/4.1/ref/settings/#std-setting-EMAIL_HOST_USER
(cherry picked from commit 015a10637b)
- Renames "Customize Zulip" to "Server configuration".
- Cross-links "Server configuration" with "System and deployment
configuration".
Fixes part of #23984.
(cherry picked from commit 538801c651)
When the email mirror gateway is sending messages "as" a user (as
triggered by having access to the missed-message email address),
attachments were still created as the Email Gateway bot. Since the
sender (the end-user) was not the owner of those attachments (the
gateway bot), nor were they referenced yet anywhere, this resulted in
the attachments being "orphaned" and not allowed to be accessed by
anyone -- despite the attachment links being embedded in the message.
This was accompanied by the error:
```
WARN [] User 12345 tried to share upload 123/3LkSA4OcoG6OpAknS2I0SFAQ/example.jpf in message 123456, but lacks permission
INFO [zerver.lib.email_mirror] Successfully processed email from user 12345 to example-stream
```
We solve this by creating attachment objects as the users the message
will be sent from.
(cherry picked from commit 608c787c52)
Using curl to POST to the CircleCI workflow endpoint on CZO:
- Doesn't work on zulip/zulip@main (CZO runs a revert)
- Sets a bad example for other orgs
- Robs us of an opportunity to dogfood our own zulip/github-actions-zulip
Refactor the Actions workflows in this repo to report failure states
using the Zulip Action, and reimplement the related helper scripts in
Python, since they'd previously mostly shelled out to Python anyway.
(cherry picked from commit 3a0620a40c)
If `invite_as` is passed as a number outside the range of a PostgreSQL
`SMALLINT` field, the database throws an exception. Move this exception
to the glass as a validation error to allow better client-side error
handling and reduce database round-trips.
(cherry picked from commit ebc6ab4744)
The English word "at" was manually appended to the string output of
datetime-related functions to generate the string shown in the tooltip
when hovering over the timestamp of a message. Use the translated form
"{date} at {time}" instead, as found elsewhere in the codebase.
(cherry picked from commit 465d0b2710)
An HTTP HEAD response with a non-empty message body is not compliant
with the standard.
Signed-off-by: Zixuan James Li <p359101898@gmail.com>
(cherry picked from commit 33716f6156)
Intercom sends a HEAD request to validate the webhook URL on their side,
which was not expected in the previous implementation.
This fixes the problem that we send out a confusing error message for it.
Fixes#23912.
Signed-off-by: Zixuan James Li <p359101898@gmail.com>
(cherry picked from commit 1e36d79b83)
During installation on a new host, `create-database` attempts to
verify that there isn't a bunch of data already in the database which
is it about to drop and recreate. In the most common case, this
statement emits a scary-looking warning, since the database does not
exist yet:
```
+ /home/zulip/deployments/current/scripts/setup/create-database
+ POSTGRES_USER=postgres
++ crudini --get /etc/zulip/zulip.conf postgresql database_name
++ echo zulip
+ DATABASE_NAME=zulip
++ crudini --get /etc/zulip/zulip.conf postgresql database_user
++ echo zulip
+ DATABASE_USER=zulip
++ cd /
++ su postgres -c 'psql -v ON_ERROR_STOP=1 -Atc '\''SELECT COUNT(*) FROM zulip.zerver_message;'\'' zulip'
psql: error: connection to server on socket "/var/run/postgresql/.s.PGSQL.5432" failed: FATAL: database "zulip" does not exist
```
Because we are attempting to gracefully handle the case where the
database does not exist yet, we also continue (and drop the database)
in other, less expected cases -- for instance, if database contains a
schema we do not expect.
Explicitly check for the database existence first, and once we verify
that, allow any further failures in the `SELECT COUNT(*)` to abort
`create-database`. This serves the dual purpose of hiding the "FATAL"
error for the common case when the database does not exist, as well as
preventing dropping the database if anything else goes awry.
(cherry picked from commit a1151aef8b)
This greatly improves the readability of the diffs and in-codebase
translation strings over using ASCII sequences for unicode in the JSON.
We've previously noticed [^1] some JSON translation files ending up with
escaped Unicode sequences on disk, which Transifex indicates is expected
behavior [^2], though it is sometimes fixed by `manage.py
compilemessages` [^3]. Further, as noted in #23932 [^4], some JSON
translation files include HTML-escaped entities like quotation marks.
This script will ingest valid JSON files and output them as proper UTF-8
files with appropriately unescaped (unless otherwise necessary, like
double quotes being backslash-escaped) sequences, except when the key
itself contains HTML escape sequences (as it's presumed the value of
such entries must be pre-escaped before being passed to consumers).
[^1]: https://chat.zulip.org/#narrow/stream/58-translation/topic/Transifex.20client/near/1479205
[^2]: https://chat.zulip.org/#narrow/stream/58-translation/topic/an.20email.20for.20Transifex.20support/near/1481287
[^3]: https://chat.zulip.org/#narrow/stream/58-translation/topic/an.20email.20for.20Transifex.20support/near/1481908
[^4]: Which is not end-to-end fixed yet by this commit: that will
require a new release of Zulip Server.
gitlint-ignore: B1, title-trailing-punctuation, body-min-length, body-is-missing
Fixes#23625.
The U+FE0F character was appearing in the data supplied by
emoji-datasource-google, but isn't in the CLDR data. By
removing that variation selector character, all remaining
emoji (that were previously marked as "not found in custom
emoji name maps, but also not found in CLDR data. Skipping.")
are now supported.
The authenticate_by_username limit of 5 attempts per 30 minutes can get
annoying in some cases where the user really forgot their password and
should be allowed to keep trying with admin approvial - so we should
document the command that allows unblocking them.
transifex-client went EOL on November 30, 2022, replaced by
transifex/cli [^1].
Swap this in-place, since per the upstream README [^2]:
> The current version of the client maintains backwards compatibility
> for the tx push and tx pull commands. So, if you have a CI setup that
> uses them, you should not have to change anything.
As the mobile team found out, this is a partial truth if one previously
used some of the more advanced CLI flags, but all workflows referenced
in tools/ and docs/ use forwards-compatible flags to the new version.
[^1]: https://github.com/transifex/transifex-client/
[^2]: a0f28a1cf3/README.md
If a previous attempt at an upgrade failed for some reason, the new
PostgreSQL may be installed, and the conversion will succeed, but the
new PostgreSQL daemon will not be running (Puppet does not force it to
start). This causes the upgrade to fail when analyzing statistics,
since the daemon isn't running.
Explicitly start the new PostgreSQL; this does nothing in most cases,
but will provider better resiliency when recovering from previous
partial upgrades.
The class for "tr" element was changed in c22de76af7, so
get_stream_id function was not able to get the correct stream-id
from the target element. This commit fixes get_stream_edit
function to also check for closest ".stream-notifications-row"
ancestor element and thus it returns the correct stream id.
Fixed a bug in expanded compose box preview mode,
which lead to buttons at the bottom of the compose box
being pushed off the screen and text overflowing.
Manually tested for different screen sizes.
Fixes: #23493.
When this code was moved from being in zerver in 21a2fd482e, it kept
the `if ZILENCER_ENABLED` blocks. Since ZILENCER and CORPORATE are
generally either both on or both off, the if statement became
mostly-unnecessary.
However, because tests cannot easily remove elements from
INSTALLED_APPS and re-determine URL resolution, we switch to checking
`if CORPORATE_ENABLED` as a guard, and leave these in-place.
The other side effect of this is that with e54ded49c4, most Zulip
deployments started to 404 requests for `/apps` instead of redirecting
them to `https://zulip.com/apps/` since they no longer had any path
configured for `/apps`. Unfortunately, this URL is in widespread use
in the app (e.g. in links from the Welcome Bot), so we should ensure
that it does successfully redirect.
Add the `/apps` path to `zerver`, but only if not CORPORATE_ENABLED,
so the URLs do not overlap.
This fixes a visual regression in newer Electron builds (without
regressing Firefox) for which I still don't entirely know the root
cause, where extra "borders" were being applied to messages in both
streams and PMs. Applying a negative "spread radius" to the box-shadow
properties of these elements, and moving that pixel to the "horizontal
shadow" aspect of the property (which is used to create the left-side
"ruler" effect), restores the expected look and feel.
Tested in qutebrowser (Chromium 87-based), Electron v18+v19, and Firefox
107.
Refs (and should unblock) zulip/zulip-desktop#1251
Add in a check to `recent_topics_data.process_message` so that we
know if any conversation data was updated and can rerender the
based on that information.
The script will fail if we process with a droplet object with no
.ip_address, which can sometimes happen if we just droplet.load().
Sometimes that gives droplet.ip_address set to None. If we wait in a
loop, we eventuall get the .ip_address set.
This fixes an error in the `inplace_rerender` code when we
are trying to remove a rendered row and correcting the `row_focus`
if it is on the row being removed. If the table is not in focus,
we don't need to do anything.
The documentation included the full policy for the file uploads
bucket, but only one additional statement for the avatars bucket; the
reader needed to assemble the full policy themselves.
Switch to explicitly providing the full policy for both.
Fixes#23110.
This fixes a bug where when all the streams in a section
in the left sidebar are muted, the section divider disappears.
The bug is caused by not considering muted streams when writing
the logic for when to show/not show section divider this commit
fixes it.
fixes#23241.
The reason utcToZonedTime would return an invalid Date (i.e. ‘new
Date(NaN)’) is because we passed in an invalid timezone. Fix the
comments and tests to reflect this; remove unnecessary mocking and the
misleading ‘instanceof Date’ check.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
`current_date` being invalid here could result in `user card`
not being displayed to users. So, it is important to verify that
the date is valid so that we don't run into any errors.
This fixes (unverified) an issue where user card was not being
displayed to some users on Zulip Desktop app because `current_date`
was not valid.
In 16c45aa, the check for an empty narrow was updated to no longer
confirm if no messages were returned by the message fetch request.
Updates the empty narrow check to confirm that the messages fetched,
and being processed are for the current / narrow message list before
showing an empty narrow.
Fixes an issue introduced in 6b4ab21 when we started using the
user's full name for empty narrow titles in a single operator
narrow of either "pm-with", "group-pm-with" or "sender".
For these empty narrow titles, any ', & or < characters in the
user's full name are not escaped.
Some terminals (e.g. ssh from OS X) set an invalid locale, which
causes the `pg_upgradecluster` call late in the upgrade to fail.
Force a known locale, for consistency. This mirrors the settings in
upgrade-zulip-stage-2, set in 11ab545f3b, and its subsequent
cleanups in 64c608a51a, ee0f4ca330, and eda9ce2364.
The custom input was not being shown when changing user email
notification batching period setting dropdown to "Custom" option.
This was because parse_time_limit was undefined in case organization
settings was not loaded till then, and the code raised an error.
This commit changes the code to define parse_limit_function at
module-level instead of defining it in
register_save_discard_widget_handlers which is called only when
build_page is called which happens only when organization settings
is opened before changing the notification batching setting.
Fixes#23674.
Processing 1k messages takes roughly 1s on a live server like
chat.zulip.org, which is a good enough duration after which
user will be looking to get some feedback on the request. Once
we show the in-progress banner, it makes sense to do bigger
batches to speed up the process.
As detailed in the comment, the new "mark all as read" bankrupcty code
path would fail to mark as read muted stream messages that are older
than the oldest unread non-muted message.
This was because of the semantics of the `first_unread` anchor value
being actually "fire unmuted unread".
We can just change this to "oldest", because we're already using an
`is:unread` for the narrow query; doing so likely is a small
optimization, since we save the query to compute the first unread.
Adjusts the check in `message_filter.process_results` to happen
after the messages have been processed so that the empty narrow
banner is shown if there are no messages in the current message
list, which covers the case of there being no messages returned
from the server (e.g. a keyword search, no private messages with
a user) and the case of all existing stream topics being muted
by the current user.
This matches the current bulk-operation behavior of the message lists,
and avoids spending an enormous amount of time adjusting unread counts
once per message in a loop over thousands of messages.
Also add some comments on adjacent code paths.
As of 550a32b, when private messages were added to recent
conversations, `recent_topics_data.process_message` will
always return true.
Updates `recent_topics_data.process_message` for no return
value. Also, removes the `topic_data_changed` logic from
`recent_topics_ui.process_messages` and instead checks for
messages to process before updating the data and calling the
rerender.
`recent_topics_ui.complete_rerender` first checks for whether
the recent conversations view is visible before rerendering.
This commit reverts the changes for replace_syntax in 1ca4566eb2
and we again use JS replace instead of replace from text-field-edit.
We do this change because replace from text-field-edit leaves the
replaced text selected, which we don't want. This change is
temporary and we can probably use replace method from
text-field-edit once this issue is fixed in upstream.
Fixes#23406.
Fixes#14053.
This also uses the new alert banner system.
It is not clear if we want to scroll user to the end of the narrow
after this update is complete. There is a "scroll to bottom" button
for users who want to be placed at the end of the narrow but there
is no reverse option for users who don't. So, maybe not changing
the scroll position is preferred especially since we now use
banners for the update where a user could have been reading something
while we were updating the message flags in the background.
We do not handle the case of `mark as read` and `mark as unread`
both running in batches at the same (This is a somewhat rare case
given the high cap on `NUM_OF_MESSAGES_UPDATED_PER_BATCH` but
possible.).
`unread.declare_bankruptcy` is no longer used in the code but
we keep it since it is used extensively in tests.
Previous algorithm was not correct if we didn't have full data for
the current narrow loaded from the server. #23512 adds the support
to mark messages in a narrow unread after a give message_id.
Added a new alert banner to indicate loading and success states of
an ongoing request. This is useful when requests can take a long
time to complete.
Doing rapid pace mark-as-unread in the Zulip web application, one
observed assertion failures showing that the server would send an
event containing multiple message IDs but only one of the messages
present in the message_details side data structure.
The cause of this was the "virtual events" compression system; two
flags/remove/read events were being combined by simply concatenating
the lists of events, without any attempt to merge the
`message_details` field on those objects.
The immediate fix is to disable virtual events compression for this
event class, but it's not unlikely we'll need to just eliminate the
virtual_events system entirely, because it seems difficult to make it
soundly handle a message whose state for a given flag changes back and
forth while the client is offline.
But we'll leave that for later, since removing that optimization
deserves more discussion than fixing this event corruption bug.
The naive solution #23465 creates situations where the same user can have
multiple reactions as the base emojis are not unique, e.g. +1::skin2
and +1::skin4 would both reduce to +1 but the userlists are separate.
This solution handles the reduction, merges the same-base reactions,
and deduplicates the userlist.
Co-authored-by: Alex Vandiver <alexmv@zulip.com>
Co-authored-by: rht <rhtbot@protonmail.com>
Following b563fdc537, the z-index of the
compose box (2) tied with that of unread date dividers within a topic;
this could lead to the dividers incorrectly appearing on top of the
compose box.
We don't have other elements that can overlap with compose with a
z-index between 2 and 4, so just increase it to 4.
This commit removes comments related to stream-admin option from
wildcard_mention_allowed.
We have already removed stream-admin option from wildcard mention
policy setting in 83383090f9 since we are no longer planning to
implement stream admin feature and instead are working on new
permissions model based on user groups.
return inside finally blocks causes exceptions to be silenced.
Although these blocks follow blanket ‘except Exception’ handlers, they
do not seem to have a goal of silencing BaseException and exceptions
thrown by the exception handler, so rewrite them to avoid it.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
We do not want to import any of our modules into this file, so
it has an independent implementation of the fix for #22062. Also,
using -13 as a keyCode helps avoid having to make changes to all
the comparisons with e.keyCode that the module relies on.
Both of these compatibility blocks can be deleted, since you can't
upgrade directly to any supported release from the versions where the
old event formats would be used.
This solves the problem that resolving a topic with a long name (>60
characters) will cause the topic name to be truncated, and thus the edit
message code path thinks that the topic is being moved in addition to
being resolved.
We store the pre-truncation topic and use it to check against the
original topic when determining whether a topic is being moved while
getting (un)resovled or not.
Fixes#23482
Signed-off-by: Zixuan James Li <p359101898@gmail.com>
We intended to send both the "topic was resolved" and the "topic was
moved here" notification when resolving and moving a topic at the same
time in #22312.
The previous implementation did not work as expected and it was only
sending the "topic was moved here" notification.
This removes the check for old_topic and new_topic that have
RESOLVED_TOPIC_PREFIX stripped in maybe_send_resolve_notifications, so
that the notification will be sent regardless if the topic name without
the prefix stays the same or not.
Note that weird topic handling ("✔ ✔✔ some topic") in the comments
was added in e231a03eff is unaffected. In case of confusion, the lstrip
check is not essential to detecting topic being unresolved/resolved.
As we mainly have that handled in the latter part of the helper.
Signed-off-by: Zixuan James Li <p359101898@gmail.com>
These suffixes suppress some checks in the process, but still generate
and upload a tarball, push a tag, and make a Github prerelease.
`upload-release` already understands that anything with a suffix never
becomes the "latest" release.
CircleCI has updated its webhook format[1] for CircleCI Cloud,
Server version 3.x and 4.x. This commit rewrites the CircleCI
integration to parse the new webhook structure. The tests have also
been rewritten for the new format.
With this commit, we support webhooks from projects that use GitHub,
BitBucket and GitLab as VCS providers. The CircleCI integration doc
has been updated to mention the same. The doc has also been updated
with the latest instructions for configuring a webhook on the CircleCI
interface, and the new output screenshots.
References:
[1]: https://circleci.com/docs/webhooks
Previously, emoji.json was read from
"$ZULIP_PATH/node_modules/emoji-datasource-google/emoji.json".
This path doesn't exist in production when installing from scratch from
a release tarball. And so, we ensure emoji.json exists by copying it to
`static/generated/emoji`.
With tweaks to comments by tabbott.
Fixes: #23469
We already re-rendered the realm-logo when the user selects either
the light or dark color scheme, but on `color-scheme-automatic`,
we didn't have any event handlers to trigger a render. This commit
adds a handler to listen for changes to `prefers-color-scheme` property.
Fixes#23548.
This commit adds call_function_periodically helper function
which will be used to call functions periodically using
setTimeout. Currently, this new function is used to send
presence requests and trying reload.
We retry reloading repeatedly at an interval of 30 seconds,
to handle the case where window.location.reload has no
immediate affect.
Previously, setInterval was used for this, but this commit
replaces it with nested setTimeout calls.
This change will help us in avoiding a large number of
requests to `/` in case when browser tries to "catch up"
pending calls after unsuspend.
This is a follow-up to d201229df8.
do_get_invites_controlled_by_user queries for Confirmations when finding
multiuse invites controlled by a user. This means that a revoked
multiuse invite cannot really be fetched here, because
do_revoke_multi_use_invite deletes the Confirmation object when revoking
the invitations. However, having a defensive assert here should be
useful to make this doesn't secretly break in the future if the query
used changes or if there are unexpected revoked multiuse invites with an
existing Confirmations for any (buggy) reason.
This allows us to revoke MultiUseInvites by changing their .status
instead of deleting them (which has been deleting the helpful tracking
information on PreregistrationUsers about which MultiUseInvite they came
from).
This commit adds support for Grafana's new alerting system, Grafana
Alerting. The existing Grafana integration has been modified to
detect the version of the notification through the structure of the
payload body, since the the structure varies by version. Support for
legacy alerting is been continued. Example fixtures have been added
for Grafana Alerting's webhooks.
Tests updated.
The tweaks I made as part of merging
f204a4c592 apparently were wrong; the
CSS class .popover_user_presence had been shared between both the user
card and the profile.
Fix this by referencing the right CSS classes. Additionally, fix the
margins to have a square hover region around the presence dot.
Fixes#23588.
Instead of using setInterval to call send_presence_to_server
repeatedly, we now simluate setInterval using setTimeout.
This change helps us to avoid the condition likely related to resume
from suspend, when the browser tries to catch up by running a large
number of "missed" copies of the setInterval hook, causing the server
to rate-limit the client.
Fixes#23438.
Fixes#23517.
While this feature was added to Zulip very early, it has been troubled
for most of that time; it never looked great visually, had a lot of
implementation complexity around resize.js, and has a weird model (a
setting that changes the UI only in certain window sizes).
This option is not commonly used; while a significant portion of users
have it enabled, many of them just don't use window sizes where it
actually has an effect. So it's not clear that it will be missed if
removed; we got very few bug reports when it was completely broken for
a few days after we first integrated the new left sidebar private
messages design.
Even with it no longer being broken, it does not work very well with
the addition of the new PMs section in the left sidebar. (Having two
scrollbars in the sidebar looks quite awkward.) The new private
messages section in the left sidebar also addresses some of the use
cases for always keeping the Users list always visible, even in narrow
windows.
This option is only removed from frontend for now. To make this
decision easily reversible, the backend code of this feature
is still kept.
Some Safari users users observed an issue where these modal buttons
appeared with an ugly white-on-grey format. Removing the CSS line with
`appearance: button` appears to fix this for affected users.
That line seems unlikely to have been constructive; as documented in
https://developer.mozilla.org/en-US/docs/Web/CSS/appearance, this is a
deprecated value equivalent to `auto`, and a bit of a misfeature.
Additionally, the "Log in" button was incorrectly underlined on hover;
override that style.
Increase z-index of unread-marker to always appear above the message and
focused outline. Also, z-index of date_row is increased to hide
unread marker for it in special conditions.
Modals are closed when selecting something inside the modal if mouse
is dragged outside the modal. This happens as leaving the mouse after
dragging it outside the modal induces a click even outside the modal
and the current behavior is to close the modal on clicking outside it.
We remove the "data-micromodal-close" attribute from "modal__overlay"
element which is responsible for closing the modal. And we add
click handler which closes the modal on clicking outside the modal but
does not closes it when the mouse is dragged outside while selecting
something.
Fixes#23350.
This doesn't make sense if the realm is active and will fail as soon as
do_reactivate_realm is fixed in the next commit to be a noop and not
create confused RealmAuditLog entries when the realm is active.
There was the following bug here:
1. Send an email invite to a user.
2. Have the user sign up via social auth without going through that
invite, meaning either going via a multiuse invite link or just
straight-up Sign up if the org permissions allow.
That resulted in the PreregistrationUser that got generated in step (1)
having 2 Confirmations tied to it - because maybe_send_to_registration
grabbed the object and created a new confirmation link for it. That is a
corrupted state, Confirmation is supposed to be unique.
One could try to do fancy things with checking whether a
PreregistrationUser already have a Confirmation link, but to avoid races
between ConfirmationEmailWorker and maybe_send_to_registration, this
would require taking locks and so on - which gets needlessly
complicated. It's simpler to not have them compete for the same object.
The point of the PreregistrationUser re-use in
maybe_send_to_registration is that if an admin invites a user, setting
their initial streams and role, it'd be an annoying experience if the
user ends up signing up not via the invite and those initial streams
streams etc. don't get set up. But to handle this, we can just copy the
relevant values from the pre-existing prereg_user, rather than re-using
the object itself.
This line was added in 94e099eaab,
presumably because of the
del request.session["multiuse_object_key"]
line that was just above it.
Looks like it should have been removed in
868a763cec, which eliminated that `del`
operation.
This is primarily for administrators needing to provide message
history for compliance or auditing purposes. Search terms can be
pulled from a file, one per line, or from arguments provided on the
command line.
In 1fce1c3c73, we added logic to parse
the User-Agent in /register requests; this logic crashed if an HTTP request
was missing that header.
Includes a test for `/register` with no user agent passed; this should catch
similar regressions in the future.
Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
The notifications headings were not consistent between the personal stream settings panel
and /help/desktop-notifications.
This changes notifications headings from "Audio" to "Audible", since that seems like the
better choice of word.
Fixes#23515.
Fixes#23284
The basic approach used by `text-area-edit` is same as we were
using, so there is no real change. There are some nice checks
in `text-area-edit` which we don't do that helps us avoid
common bugs.
This commit refactors the code for realm message retention
setting to make it more similar to code for other time-limit
realm settings. There are still separate functions for the
message retention setting but they are similar to that of
other time-limit settings and can probably help us to
deduplicate the code in future. This refactor also helps
us in reusing some code for stream message retention
setting for changes in #23013.
This commit also fixes a bug where save-discard widget
not being hidden when resetting the setting to original
value. This can only be reproduced when the original
setting value is "retain_forever".
Following changes have been done in this commit -
- Added data-setting-widget-type attribute which is set to
"message-retention-setting" similar to "time-limit" value for
message edit and delete limit settings.
- Added get_message_retention_setting_value function similar
to get_time_limit_setting_value function which is used by
get_input_element_value.
- Removed "message-retention" block from
get_complete_data_for_subsection and the message retention
setting is now added to request data in populate_data_for_request
itself using data-setting-widget-type attribute.
- Updated id and class of elements to support the above changes.
When resetting a setting in "Other settings" section to its original
value, the save-discard widget for the section is not hidden. This can
be reproduced when no language is set for code-block default language
setting. This was because check_property_changed incorrectly returned
true as the original value for code-block language setting was null
but the widget element returned empty string.
To implement this, we are rendering users in PMs with
the status icon (via `user_with_status_icon.hbs`) and
then passing rendered HTML to the recent topics row.
Fixes#23262
We have bound the `mouseenter` event with a circle
status parent element to render the tooltip. Instead of passing
the tooltip content directly in `data-tippy-content`, we
are rendering it via JS because when showing the bot
tooltip we want a multiline tooltip.
We have also made some minor changes to `do_render_buddy_list_tooltip`,
so we can append the tooltip to the target element and avoid adding
a observer to it.
This commit also sets the font weight for tippy content to
400 to avoid bold content.
When processing an "update_message_flags" event to mark messages
as unread, check the current message list once messages are updated
to see if the `mark_as_read_turned_off_banner` should be visible.
This means that other clients in a view that also does not mark
messages as read (for example a sender or keyword search narrow)
will also load the banner when the view is updated.
Fixes#23502.
Moving the last active status icon to the modal header, we obtain a
cleaner visual design for the user profile modal which is more
consistent with other parts of UI.
We can reuse the content from multiple sections of the Slack article
to update "Import from Mattermost/Gitter/Rocket.Chat" by making small
wording tweaks.
- Renames the page to "Usage statistics" (title & left sidebar).
- Adds an intro section that outlines the types of information
that can be found on the stats page.
Fixes#23449.
Showing section separators in the left sidebar ("Pinned streams",
"Active streams", etc.) is unnecessary when there is only one section,
and can feel confusing. We should show the separators only when there
is more than one section present.
Fixes#22843.
This fixes a race in certain intermediate sizes, with the
left_side_userlist setting, the sidebar sizes would have been computed
with invalid information about the size of the two regions to divide
space between (because the sidebar was closed when the calculation was
done).
Structurally, the goal of this logic is to just hide all popovers
(including potential user popovers) and not the sidebars; use the
function designed for that purpose rather than hackily
hand-implementing it.
In window width ranges where the left sidebar is fully open and the
right sidebar is toggleable, the existing logic to hide all popovers
and then reopen the left sidebar stream list.
We have a function designed to close all popovers that aren't
sidebars; we should just use it.
See 2d414fa897 for a previous incomplete
fix for this issue.
The `size_blocks` code was cool looking but wasn't working well.
Since we only have to place streams list and users list, we can
fix their `max-height` more easily without trying to solve the
`multi-block` case which `size_blocks` was trying to do.
This also fixes users list max-height being incorrectly set,
resulting in user unable to scroll to some user list items.
Adds "View your mentions" page to document how to access @-mentions
and the @ indicator for unread @-mentions on desktop/web and mobile.
Fixes#23422.
Co-authored-by: Alya Abbott <alya@zulip.com>
The stream-specific notifications table incorrectly used the
.stream-row CSS class. As a result, live-update code for the "stream
settings" UI when changing a stream's permissions model would
incorrectly replace rows of this table with a "Manage streams" style
row.
In 2f547ea, the custom `has_error` logic in BaseDocumentationSpider
was removed in favor of checking whether any errors were logged.
One reference to that method was missed in UnusedImagesLinterSpider
and went unremarked as no images were added to the documentation
that were not used.
Replaces that reference with an error log.
Fixes a warning caused by using the deprecated `2.6` value
of the `REQUEST_FINGERPRINTER_IMPLEMENTATION` setting.
Upgrades `settings.py` to what would have been generated
by starting the documentation_crawler project using the
`startproject` command of scrapy v2.7.
This provides a hook for users to get to these /help/ pages, which contain
details on exactly how these notifications/settings work, as well as
troubleshooting advice.
Fixes: #23369.
For PATCH requests to `/realm/profile_fields`, the field_type is
determined via the field.id, which is in the URL. Therefore,
`field_type` does not need to be passed as part of the request
data in tests.
Instead of using "these people", we use "these users".
Note that the "pm-with" single user case has already been updated
to show the user's full name instead of "this person".
Fixes small error in the example at the top of the article, and
adds a short statement about narrows searching the current user's
personal message history with a link to the help center article.
Gives the final section a subheader and adds the two other narrow
operators that take an ID.
Updates the relevant help center articles to reflect that the
sections for Pinned and Inactive streams in the left sidebar
now have labels.
Fixes#23423.
We do not create historical UserMessage rows, for messages that didn't
have one, while marking messages as read and simply ignore those messages.
We do so because there is no user of creating UserMessage rows and it just
wastes storage.
Note that we still allow to mark messages from unsubscribed streams as
read but only those which have UserMessage rows for them to handle the
case when the unread messages were not marked as read while unsubscribing
from the stream due to some race condition. In such cases, messages
will not be included in the unread count shown in "All messages" menu
(and stream is anyways not present in the left sidebar), but the message
border on the left is green if viewing the stream after unsusbcribing it.
So, to avoid the confusion for users, the messages will be marked as read
when user scrolls down.
This unifies the length of the shortened SHA our integrations generate,
and ensures that they are long enough for projects of various sizes with
a chosen value defined in get_short_sha.
Fixes#23475
Signed-off-by: Zixuan James Li <p359101898@gmail.com>
7 characters are not enough for large projects, so we change
it to reasonably longer. As an example, The Linux kernel needs
at least 11 characters of sha in its shortened form to identify
a revision. We pick 11 so it should work for most of the projects.
Signed-off-by: Zixuan James Li <p359101898@gmail.com>
We use `templates/zerver/api/api-doc-template.md` as a base template
for the documented API endpoints in `zerver/openapi/zulip.yaml`.
Previously, if this template was called as an endpoint, then it
would fail an assertion check and send server error. Now we check
for specifically for that potential path and return a 404 error
response for no existing article.
Fixes#21876.
It is better not to use a tab in the instructions for how to
"Toggle whether names of reacting users are displayed", as this setting
applies to both web/desktop and mobile apps.
The instructions themselves still indicate that the setting is controlled
via the web/desktop app.
Prior to 53231aa, the `ignore_unhashable_lru_cache` decorator had
a check for the development environment so that changes could be
seen on refresh.
Puts that check back in IgnoreUnhashableLruCacheWrapper class.
Due to incorrect heights set for user-group creation and stream
creation body we had a UI bug because of which bottom part of
user group and stream creation forms were slightly visible
through the bottom.
It was due to somewhat fussy css height rules and incorrect
height calculations for that UI.
This function is only called to calculate the size of stream and
user list. Using 250px as minimum height for these blocks gives
them enough space to be usable regardless of their height ratio.
If the buddy list is displayed in the left sidebar, height of
both these blocks needs to be changed to accommodate for the
changes.
The fixes the bug that stream list and user list don't fill
left sidebar and have 80px height regardless of viewport height.
When building search suggestions for stream topics, instead of
assuming that the user has access to the stream's topic history
from the server, we check whether the user has access to the
stream's topics.
In the outgoing webhook handler, there is potentially several seconds
of trying between when a message triggering an outgoing webhook
arrives, and when it fails. In the meantime, the stream the
triggering message was on may have been deleted, causing the
"Failure!" message to have no valid stream to be sent to.
Rather than raise an exception in the outgoing webhook worker, ignore
the exception and move on.
The lambda passed to `queue_json_publish` is used if
`settings.USING_RABBITMQ` is unset -- which is only true in tests. As
such, this pattern causes failures to never actually retry within
tests.
This behaviour has existed ever since the outgoing webhook code was
introduced in 53a8b2ac87, with no explanation. Not passing that
argument allows tests to verify the retry behaviour when webhooks
fail.
Previously, test cases or clients accessing /json/ views using HTTP
Basic Auth would be accepted, while we intended to only allow clients
authenticated with a session cookie to access these views.
This adds a check on the accessed path to avoid this possibility.
It seems unlikely that any API clients clients were taking advantage
of this unintended quirk; so we're not going to bother documenting
this bug fix as an API change. In any case, it should be trivial for
anyone affected to consult the documentation and then switch their
/json/foo URL to a correct /api/v1/foo URL.
Signed-off-by: Zixuan James Li <p359101898@gmail.com>
Updates the check in compose validate for the organization's
policy on sending private messages to use the code/value in
settings_config, instead of the number value.
Private messages between a user and a bot are not restricted by
the organization's policy setting for sending private messages.
So, for the empty banner text, we set the default title for
a single bot for the "pm-with" narrow.
Group private messages that include a bot are restricted by
this policy setting, so those messages stay the same.
Also, since bot's aren't people, we update the text for valid
"pm-with", "sender" and "group-pm-with" operands to include the
user's full name in the title.
When there are multiple people in the "pm-with" narrow, we still
send a generic "these people" message instead of listing all the
users in the narrow.
Finally, we make the default message end in a period instead of
an exclamation mark.
For narrows that focus on private messages ("is:private", "pm-with",
"group-pm-with"), we want to check the organization private message
policy and set an empty narrow title that matches that policy.
For these narrows with invalid users, we prioritize that message
for the empty narrow title, instead of the private message policy
message.
Fixes#21889.
Mobile clients older than v27.192 do not support PRONOUNS type
custom profile fields, so we instead change the type of it to
SHORT_TEXT in the data sent with register response and also in
the events sent to those clients.
A user could trigger a "pm-with" narrow with an invalid user by moving
back/forward in the session, which would trigger the compose box to
open via a hash change. Updates the check to pass in any "pm-with"
narrow with an invalid user.
Also, if the compose box was active/started in the previous view, but
without any content, then we want to actively cancel/close it in the
invalid user "pm-with" narrow.
If an organization has disabled sending private messages, we do not
want to load the compose box automatically for "pm-with" narrows.
We still open the compose box for private messages narrows with a
single bot user as this is not limited by this organization setting.
Also, if the compose box was active/started in a narrow with a bot
user, but had no content, then we want to close/cancel that in a
new narrow with a person or group.
Fixes#23332
Check if col we are trying to focus exists for the row inside
`set_table_focus`.
We call `revive_current_focus` after `topics_widget` is defined
since it can be used inside it.
Fixes#23331
Combine checks for last row between `vim_dowm` and `down_arrow`.
Check for the presence of unread counter in `set_table_focus`
directly so that it can be used in other cases as well. This is
especially useful when marking the last row in the table as read.
We used to hide and show topic rows in the DOM when topics are
updated. This resulted in incorrect calculations in the length of
visible topics. As a consequence, focus is sometimes set to hidden topic.
Removing hidden topics from DOM helps us keep
the calculations correct.
The fixes bugs related to focus being lost when trying to mute
or mark as read the last row.
Instead of topic filter box being a part of the list when keeps
updating, we move it out and fix its position.
This should reduce rendering time of topics list and provide
a smoother experience to users when waiting for topics list to
load.
Updates the first part of the article to discuss the basics of
the search feature. Adds sub-headers / sections to the list of
search operators. Makes the exluding messages a section header
instead of a sub-section of the list.
Updates the help center article on searching for messages to have
the correct format for the "group-pm-with" operator.
Also, reorders the list so that the private message searches are
all together, and the `streams:public` operator is last so that
it stands out a bit more.
On multi-realm systems this results in traversal of all messages in
all realms and returns a massive payload of 1 row per stream on
the server, not the intended one row per realm.
As explained in 158287f998,
wantMessagesSigned can't be enabled globally (as it'll break setups with
IdPs that sign SAMLResponse assertions) - but is needed for
LogoutRequests, and will be for LogoutResponses in the SP-initiated SLO
flow in future commits.
We extract a function with the necessary hacky logic for re-use in the
SP-initiated SLO implementation.
2022-11-02 21:58:46 -07:00
1264 changed files with 33515 additions and 22316 deletions
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
