For realms that don't have any presence-active users, we know for a
fact that there aren't any active clients that will be reloading just
after the server restarts, so we can skip filling the cache with data
related to that realm.
For zulipchat.com, this results in a significant performance
optimization for the recipient and stream caches, and a moderate
performance improvement for the user caches as well.
Private messages make up the bulk of Recipient objects. While private
messages are ~50% of messages, if you weight by messages received
(which is what is important for message-loading performance), it's
pretty strongly balanced towards stream messages.
We don't need to include long-term idle or other inactive users here,
since fetching them consumed to vast majority of the time.
(On chat.zulip.org, this decreased the runtime for populating the user
cache by 5x, removing only users we're unlikely to need to access).
This doesn't seem to have a huge performance downside (less than 1s
extra time for loading / on chat.zulip.org), and it means the
possibility of users having so many unreads that we get weird/buggy
behavior is much more unlikely to exist.
We'll still want a better experience for users who somehow go over
this limit, but it can be pretty firmly "you need to go mark some
things as read".
This renames Realm.show_digest_email field to
digest_emails_enabled, for greater clarity as to what it does
just from seeing the setting name, without having to look it up.
Fixes part of #10042.
We were getting event-handling exceptions in JS in production if a new
user was created and then went and set a custom profile field, because
there was no `.profile_data` on their user object. We were able to
trace the issue down to the fact that our events didn't include that
field when creating a new user.
This renames Realm.restricted_to_domain field to
emails_restricted_to_domains, for greater clarity as to what it does
just from seeing the setting name, without having to look it up.
Fixes part of #10042.
The previous error messages for this were written for a tool only to
be used by a couple people, and didn't make clear what potential
causes were. Tweak these to provide greater clarity about what's
going on.
The main cause of these errors appearing in practice was fixed in
7ea5987e5d, but nothing strongly
prevents a similar issue from being introduced in the future.
Fixes#10078.
Apparently, our old unminify logic relied on the fact that the
filenames displayed in tracebacks were of the form "app.js" (and the
`app.js` copy of the source map in the appropriate
`/home/zulip/deployments/`). The correct behavior is to just look up
the source map for the appropriate hash-named
`app.a40806b10565c1dee5bf.js` type file.
We fix this with a few small tweaks to the regular expressions. I
wish this file had reasonable unit tests.
We already had a setting for whether these logs were enabled; now it
also controls which stream the messages go to.
As part of this migration, we disable the feature in dev/production by
default; it's not useful for most environments.
Fixes the proximal data-export issue reported in #10078 (namely, a
stream with nobody ever subscribed to having been created).
The is_private flag is intended to be set if recipient type is
'private'(1) or 'huddle'(3), otherwise i.e if it is 'stream'(2), it
should be unset.
This commit adds a database index for the is_private flag (which we'll
need to use it). That index is used to reset the flag if it was
already set. The already set flags were due to a previous removal of
is_me_message flag for which the values were not cleared out.
For now, the is_private flag is always 0 since the really hard part of
this migration is clearing the unspecified previous state; future
commits will fully implement it actually doing something.
History: Migration rewritten significantly by tabbott to ensure it
runs in only 3 minutes on chat.zulip.org. A key detail in making that
work was to ensure that we use the new index for the queries to find
rows to update (which currently requires the `order_by` and `limit`
clauses).
As part of our effort to change the data model away from each user
having a single API key, we're eliminating the couple requests that
were made from Django to Tornado (as part of a /register or home
request) where we used the user's API key grabbed from the database
for authentication.
Instead, we use the (already existing) internal_notify_view
authentication mechanism, which uses the SHARED_SECRET setting for
security, for these requests, and just fetch the user object using
get_user_profile_by_id directly.
Tweaked by Yago to include the new /api/v1/events/internal endpoint in
the exempt_patterns list in test_helpers, since it's an endpoint we call
through Tornado. Also added a couple missing return type annotations.
This is a preparatory refactor for adding
UserProfile.can_subscribe_other_users.
Although there existed a test for limiting users from creating
streams at `test_subs.test_user_settings_for_adding_streams`,
it did not test the logic inside can_add_streams, tests have
been added to solve that issue.
Extracting this helper library will help us avoid an import loop
between notifications.py and message.py (with bugdown in between).
But in addition to that, it's a more natural model, since some of the
uses for these functions weren't part of the notifications code
anyway.
It's sorta an unusual state to get into, to have a user own a
deactivated bot, when they can't create a bot of that type, but
definitely a valid possibility that we should be checking for.
Fixes#10087.
This is a follow-up in response to Tim's comments on #9951.
In instances where all messages from a BitBucket integration are
grouped under one user specified topic (specified in the URL), we
should include the title of the PR in the message body, since
the availability of a user-specified topic precludes us from
including it in the topic itself (which was the default behaviour).
This is a follow-up in response to Tim's comments on #9951.
In instances where all messages from a Gogs integration are
grouped under one user specified topic (specified in the URL), we
should include the title of the PR in the message body, since
the availability of a user-specified topic precludes us from
including it in the topic itself (which was the default behaviour).
This is a follow-up in response to Tim's comments on #9951.
In instances where all messages from a GitHub integration are
grouped under one user specified topic (specified in the URL), we
should include the title of the issue/PR in the message body, since
the availability of a user-specified topic precludes us from
including it in the topic itself (which was the default behaviour).
This is a follow-up in response to Tim's comments on #9951.
In instances where all messages from a Gitlab integration are
grouped under one user specified topic (specified in the URL), we
should include the title of the issue/MR in the message body, since
the availability of a user-specified topic precludes us from
including it in the topic itself (which was the default behaviour).
Implement this function in 'bulk_import_model'
and 'update_model_ids'.
This lets us save on redundant-feeling arguments in these
frequently-called helper functions.
This deletes the unused Subscription.notifications field and removes
it from some testing and analytics code (which should not have been
using it in the first place).
Fixes#10042.
One of the code examples for GET /users was using the raw call_endpoint
method from our Python bindings, rather than get_members, which has been
specifically designed to interact with this endpoint.
Now all the examples here use the appropiate method.
Whenever a parameter for an endpoint in our REST API has a default
value, it is displayed under the "Description" section of the
arguments table in the docs.
This way, we don't need to explicitly indicate the default values in the
description, thus avoiding duplicate information in the OpenAPI source.
This has the benefit that we now get the usual data about the
user/request/etc. in error emails related to bugdown exceptions;
previously we were just getting the traceback in the emails (since our
`mail_admins` template was very simple) and no other debugging
details.
Comments tweaked by tabbott to help make clear exactly what's going on
here, since it's a little subtle and a little hacky.
Fixes#8843.
This makes a few tweaks from a pass through the file following up
on the recent 1eb67e4bf, which added documentation comments and
docstrings to a number of fields and models. Also reorder a few
more fields, where that makes things clearer.
There are a lot of specific edits, made as I read through and
particularly where I didn't immediately understand things, or
imagined a reader might have questions. A couple of themes:
* Use blank lines everywhere to set off a given comment and the
items it applies to.
* In a few places where we didn't already, specify concretely how
the meaning is represented in the value (answers include: it's a
number of messages; it's a name from such-and-such namespace;
it's a JSON blob.)
This reorganizes the field definitions to be more readable, and adds
descriptive comments for a number of fields whose purpose might be
otherwise unclear.
Our nested code block processor wasn't using the correct test for
whether a paragraph was empty of other content; first, we need to
confirm no children, and second, we need to confirm there is no text
before/after the code block element inside the p tag.
None of the file types here are actually processed by our static asset
pipeline in a way that would result in the hash-named versions of the
files (stored in staticfiles.json) being accessed. (If they were,
we'd be using something like `render_bundle` to access their paths).
So, we should not be generating/shipping these hash-named versions of
image, audio, and locale files.
This change decreases the size of a Zulip release tarball from 153MB
to 93MB, by removing all of the duplicated copies of various asset
files.
This may also help with #10038, but I'm not marking it as completing
that issue yet, because part of #10038 is that the non-hash-named
image files in prod-static/generated/emoji do not seem to have been
properly overwritten on upgrade, and it's unclear why.
Fixes#5971.
A stream created in the last few hours likely won't be in StreamCount
(since that gets updated once a day), and hence won't be in the
recent_traffic dict.
However, get_average_weekly_stream_traffic should be None in this case,
not 0.
This commit closes a long pending issue which involved moving the
`EMOTICON_CONVERSION` mapping to build_emoji infrastructure so
that there is only one source of truth. This was pending from the
time when this feature was implemented.
This setting isn't intended to exist long term, but instead to make it
possible to merge our search pills code before we're ready to cut over
production environments to use it.
The function 'update_model_ids' should be used on
the models BotStorageData and BotConfigData.
It is wrongly added here for UserGroup model.
Also the sequence name for BotStorageData and
BotConfigData is 'zerver_botuserstatedata_id_seq' and
'zerver_botuserconfigdata_id_seq' respectively, which
should be specifically mentioned in the function
'allocate_ids'.
This fixes some nondeterministic test failures.
The gitter mentions are in the format '@usermention'
and the mentions are included in the export data as:
"mentions": [
{
"screenName": "usermention",
"userId": "54d7876c15522ed4b3dbbefb",
"userIds": []
}]
We extract this data and map this mention to @**usermention**
for Zulip.
Messages can be bulky, and storing them in a single
data structure can cause a memory error.
In this commit, the messages are written to a file
batch-wise, thus avoiding the memory error.
Similar to commit 6b7b6b38ad
This will be used while for any ManyToMany field which
is being imported.
We add an internal function which takes in the old ID list
of the ManyToMany field and return the new updated ID list.
Various pieces of our thumbor-based thumbnailing system were already
merged; this adds the remaining pieces required for it to work:
* a THUMBOR_URL Django setting that controls whether thumbor is
enabled on the Zulip server (and if so, where thumbor is hosted).
* Replaces the overly complicated prototype cryptography logic
* Adds a /thumbnail endpoint (supported both on web and mobile) for
accessing thumbnails in messages, designed to support hosting both
external URLs as well as uploaded files (and applying Zulip's
security model for access to thumbnails of uploaded files).
* Modifies bugdown to, when THUMBOR_URL is set, render images with the
`src` attribute pointing /thumbnail (to provide a small thumbnail
for the image), along with adding a "data-original" attribute that
can be used to access the "original/full" size version of the image.
There are a few things that don't work quite yet:
* The S3 backend support is incomplete and doesn't work yet.
* The error pages for unauthorized access are ugly.
* We might want to rename data-original and /thumbnail?size=original
to use some other name, like "full", that better reflects the fact
that we're potentially not serving the original image URL.
This adds support to the event queue system for triggering
missed-message notifications (whether push or email) to support the
stream push notifications feature.
This modifies the logic for formatting outgoing missed-message emails
to support the upcoming stream email notifications feature (providing
a new format for the subject, etc.).
Because we're passing through the trigger for notifications to
do_send_missedmessage_events_reply_in_zulip, we don't need to go back
to the database to determine which messages actually mentioned the
user.
This change converts our logic for determining whether the current
user was mentioned in a group of messages from the implicit "if it was
sent to a stream, it's a mention" to the explicit "we actually know
there was a mention in the message". This is an important
prerequisite for our upcoming feature to support getting email
notifications for streams always (even without a mention).
Previously, maybe_enqueue_notifications had this very subtle logic,
where it set the notice variable only inside the block for push
notifications, but then also used it inside the block for email
notifications.
This "worked", because previously the conditions for push
notifications were always true if the conditions for email
notifications were, but the code was unnecessarily confusing. The
only good reason to write it this way is if build_offline_notification
was expensive; in fact, the most expensive thing it does is calling
time.time(), so that reason does not apply here.
This was further confusing, in that in the original logic, we relied
on the fact that push notification code path edited the "notice"
dictionary for further processing.
Instead, we just call it separately and setup the data separately in
each code path.
This data will be required for correctly implementing the upcoming
stream_push_notify feature; it also helps support cleaning up the code
for the existing stream mentions logic.
Because in upcoming commits, we'll want to pass additional per-message
data into do_send_missedmessage_events_reply_in_zulip, we need to
expand the format for how we represent messages to account for that.
This refactors the generate_topic_history_from_db_rows function to not
depend upon the assumption of rows passed as parameter to be sorted in
reverse order of max_message_id field.
Additionally, we add sorting and some tests that verify correct
handling of these cases.
In this commit we add a new endpoint so as to have a way of fetching
topic history for a given stream id without having to be logged in.
This can only happen if the said stream is web public otherwise we
just return an empty topics list. This endpoint is quite analogous
to get_topics_backend which is used by our main web app.
In this commit we also do a bit of duplication regarding the query
responsible for fetching all the topics from DB. Basically this
query is exactly the same as what we have in the
get_topic_history_for_stream function in actions.py. Basically
duplicating now is the right thing to do because this query is
really gonna change when we add another criteria for filtering
messages which is:
Only topics for messages which were sent during the period the
corresponding stream was web public should be returned.
Now when we will do this, the query will change and thus it won't
really be a code duplication!
We already include the issue title in the topic. But if one chooses
to group all gitlab notifications under one topic, the message body
is misleading in the sense that only the Issue ID and the description
are displayed, not the title, which isn't super helpful if the topic
doesn't tell you the title either.
I think we should err on the side of always including the title in
the main message body, which is what this commit does.
Fixes#9913.
This migrates Zulip to use a dramatically better set of names and
aliases for our emoji set, defined in emoji_names.py (which is in turn
manually generated from our hand-curated CSV file).
This should significantly improve the experience of using Zulip's
emoji picker and emoji typeahead for finding what one is looking for.
Fixes#7665
In case of invitation events, 'invites_changed' event without
any real payload is sent to all the realm admins and the user.
The event is handled by reloading the list to view recent changes.
Commit tweaked by shubhamdhama:
* Send an `invite_changed` event when an user accept an invite.
Also, added the test for the same.
* No need to delete the invite list in frontend, current logic
handles the case when the invite data is changed properly.
* Extracted the common logic for sending an event into
`notify_invites_changed`.
Some of the arguments in our REST API have to be sent as JSON objects,
which only accept double quotes for strings.
If we display the examples as normal Python objects, the syntax would be
quite similar but it would use simple quotes, which is invalid JSON (and
isn't accepted by the server).
That's why all the examples should be JSON-serialized in order to comply
with the API's requirements.
Until now, we were displaying an empty "Arguments" section in the REST
API docs whenever an endpoint didn't use input arguments.
In the case of OpenAPI-based docs, that was also annoying because it
required removing the {generate_api_arguments_table|...} template tag or
leaving an empty "parameters" field in zulip.yaml.
After this, we show a paragraph indicating that the endpoint doesn't
need arguments under the "Arguments" section.
If the argument table generator isn't able to reach a file that is
supposed to read, the two most likely causes are:
- The source .md documentation file that is requesting the table has a
typo in the path.
- The file with the arguments isn't there, for some reason.
In either case, we don't want the server to fail silently-ish and
display the docs as if there was no arguments for that endpoint. That's
why the most logic thing to do is to raise an exception and let the
admins know that there's something wrong.
<script charset=…>, <script type=…>, and <style type=…> are “obsolete
but conforming” in HTML5. They make the validator.nu output noisier
and real problems a little harder to find.
(type was required in HTML 4, which is not relevant to us.)
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
POST and DELETE operations in /users/me/alert_words may leave the
user's list of alert words in an unknown state: POSTing adds words to a
list that the client may not know from the begining, and the same with
DELETE.
Replying with the current status of the alert words list is the best way
of letting the client alter the list and knowing its contents after
being updated with a single query.
This is especially useful taking into account that POSTing words that
were already present and DELETing non-existing words both produce a
successful response.
An extra test has been added to avoid leaving GET /users/me/alert_words
too untested.
Querying an endpoint with no information (thus a noop) and it producing
a successful response doesn't seem to be expected.
Given the case that the client makes such query with no content it will
probably be unintentional and the API should let them know about it.
For importing huddles we have to have unique huddle hashes.
Huddle hashes are extracted from the list of users participating
in a huddle. So to extract these user ids, we first use huddle
id to getting the matching recipient, and then we use subscription
to get the user ids from the recipient id.
Added tests for the same (tests slightly tweaked by tabbott).
The tests for GET /users were looking for a specific user, asuming that
it would always be in the same position. Since the users' sorting isn't
guaranteed in any way, this can lead to errors in the tests.
Now we make sure the user we grab from the list is the one we need by
checking its email address.
This is just a hotfix that addresses the short-term problem: we have
already made some efforts to make sure these tests are more
deterministic, and now we only need to finish the migration of the old
enpoints to the new system as a long-term solution.
This is all the plumbing that makes it possible to enable the
stream_email_notifications setting via the Zulip API. The flag still
doesn't do anything yet, but this is a nice checkpoint along the way
to implementing this feature.
This commit creates a new field called delivery_email. For now, it is
exactly the same as email upon user profile creation and should stay
that way even when email is changed, and is used only for sending
outgoing email from Zulip.
The purpose of this field is to support an upcoming option where the
existing `email` field in Zulip becomes effectively the user's
"display email" address, as part of making it possible for users
actual email addresses (that can receive email, stored in the
delivery_email field) to not be available to other non-administrator
users in the organization.
Because the `email` field is used in numerous places in display code,
in the API, and in database queries, the shortest path to implementing
this "private email" feature is to keep "email" as-is in those parts
of the codebase, and just set the existing "email" ("display email")
model field to be something generated like
"username@zulip.example.com" for display purposes.
Eventually, we'll want to do further refactoring, either in the form
of having both `display_email` and `delivery_email` as fields, or
renaming "email" to "username".
This commit adds a Markdown tree-processor extension that renders
multi-line code blocks that are nested inside lists with the
formatting. Note that the code block could be nested inside multiple
list levels and would still get rendered correctly.
Tim: This fixes the need for unpleasant workarounds like
f5bfa4e793 and makes nested code blocks
in our documentation look exactly how users would expect them to.
Given that we allow adding emoji reactions by only using the
emoji_name, we should offer the same possibility for removing
reactions to make the experience for API clients not require looking
up emoji codes.
Since this is an additional optional parameter, this also preserves
backward compatibility.
Complete, correct implementations of Zulip's emoji reactions API need
to send both emoji_code and emoji_name in order to add a reaction;
this is important for corner cases around clicking on a reaction in a
message that was first reacted to a year ago, when the emoji
name->code mappings have changed for the given code point in the
intervening time.
However, for folks building tools using the Zulip API, that corner
case is not particularly common; as a result, it makes sense to offer
an interface that allows adding a reaction by only specifying the
emoji name.
This is why the only field that needs to be required is emoji_name,
which can now be mapped to a single emoji. Both fields will be
necessary when "voting" an old reaction, but since we stil allow
specifying the two of them, these changes offer retrocompatibility.
This adds a new settings, SOCIAL_AUTH_SUBDOMAIN, which specifies which
domain should be used for GitHub auth and other python-social-auth
backends.
If one is running a single-realm Zulip server like chat.zulip.org, one
doesn't need to use this setting, but for multi-realm servers using
social auth, this fixes an annoying bug where the session cookie that
python-social-auth sets early in the auth process on the root domain
ends up masking the session cookie that would have been used to
determine a user is logged in. The end result was that logging in
with GitHub on one domain on a multi-realm server like zulipchat.com
would appear to log you out from all the others!
We fix this by moving python-social-auth to a separate subdomain.
Fixes: #9847.
* If `zerver_realmauditlog` is present in the exported data,
`RealmAuditLog` would be imported normally.
* If it is not present, `create_subscription_events`
function in would create the `subscription_created`
events for RealmAuditLog. The reason this function
is in `import_realm` module and not in the individual
export tool scripts (like Slack) is because this
function would be common for all export tools.
This fixes#9846 for users who have not already done an import of
their organization from Slack.
Fixes#9846.
Tim Abbott
Rhea Parekh
Vishnu Ks
Roman Godov
Rishi Gupta
Cynthia Lin
Shubham Padia
Aditya Bansal
Rohitt Vashishtha
Shubham Dhama
Eeshan Garg
Yago González
Harshit Bansal
Greg Price
Aman Agrawal
Joshua Pan
Anders Kaseorg
Steve Howell
Sarah
Aastha Gupta