Fixes
[
File "/srv/zulip/zerver/lib/actions.py", line 605, in recipient_for_emails
if not (normalized_emails & admin_realm_admin_emails or normalized_emails & settings.CROSS_REALM_BOT_EMAILS):
TypeError: unsupported operand type(s) for &: 'set' and 'list'
(imported from commit f39a95dad7b3207e9188fc03926cd116061ef3f3)
Include new field on Realm to control whether e-mail invitations are required
separately from whether the e-mail domain must match.
Allow control of these fields from admin panel.
Update logic in registration page to use these fields.
(imported from commit edc7f0a4c43b57361d9349e258ad4f217b426f88)
Also increase the number of messages sent as context from 5 to 10 and
look up to 15 minutes in to the past for context.
(imported from commit bfaed9bcff1ee2047fc3b7a63acf93cd2d47cc7d)
Now we have 2 different Zulip apps out there, and they are signed with
two certs: Zulip and Dropbox. The Dropbox-signed apps are going to need
to be sent APNS notifications from the appropriate APNS connection
(imported from commit 6db50c5811847db4f08e5c997c7bbb4b46cfc462)
Missed message email were including the context messages in the number
of messages you were mentioned in.
(imported from commit 1749c5d272d2e17d6e28456ace932f80715103a3)
* Fixes a few bugs with missed message address for PMs and huddles.
* Uses missed message address for all missed message reply-to headers on
the zulip.com realm.
(imported from commit 61dd09386e1bbdf9a5096e2400984d31e73a5b74)
The one time use address are a unique token which maps to stored stated
in redis. We store the user_id, recipient_id, and subject. When an email
is received at this address it is sent to the stored recipient by the
stored user. Anyone with this address can send a single message as this
user.
(imported from commit 4219417bdc30c033a6cf7a0c7c0939f7d0308144)
Send a different missed message email for each recipient. This allows us
to set a different reply to address for each one. PMs and huddles use
the existing logic, replies will be sent to all parties via email.
Missed @-mention emails will have the reply to address set to the
stream's email address.
(imported from commit bfb7cf7c1382adbf3720caa74cbb927c10dea267)
One common place that this happens (for us) is on a local
Dropbox .dev.corp.dropbox.com instance, which can't be reached
by the Zulip servers.
This commit also:
* Fixes the test suite
* Properly previews /photos/ links
(imported from commit b4788b6236e7a9d390e1efc4673be34d9ba5e091)
Truthfully, the actual way to do this is going to be a bit
more involved and also involves changing Realm.NOTIFICATION_STREAM_NAME,
probably on a realm-by-realm basis.
(imported from commit b6a05849d215e07ee6716d116ff5e2c819d5b4be)
When you are at mentioned in a stream we will now send you up to the
last five messages which were sent in the past 5 minutes on the same
topic and stream.
(imported from commit 6df6c1cf868722a7bf76e54710e38741a7ac8f31)
URLs with a realm of "unk" will be queried against the new bucket to
determine the relevant realm of the uploading user.
(imported from commit 5d39801951face3cc33c46a61246ba434862a808)
This way if two browsers are disagreeing about your active status, the
active one wins. The active browser continues to update your timestamp,
and the idle browser's changes are discarded until the timestamp on your
active status expires.
(imported from commit dc29e013d045c4b72793097f611ba6802c58e57a)
We still don't show this in the frontend, aside from our usual "Not
delivered" message that we also show when you send to a non-existent
user.
Addresses #2349
(imported from commit 2f348b15a4d539987ddbcccbbf40e2be87c1f92d)
In a test run with a hand-constructed query, this sped up the query time from
280ms to 50ms.
(imported from commit 8cbe199ca50a487491d13d6d6ef940ea668c1038)
See #2357. We now support `~~~ .py ` with that trailing space.
Note that the test coverage is Python-side only due to
bugdown_matches_marked being set to false, since we don't yet
support language syntax on the client side.
(imported from commit ccd5fcb0eee01478d349161400103480678d7486)
Adds APIs edit a bot's default_to_stream, default_events_register_stream
and default_all_public_streams.
(imported from commit c848a94b7932311143dad770c901d6688c936b6d)
Support setting default_to_stream, default_events_register_stream, and
default_all_public_streams during in the bot creation API.
(imported from commit bef484dd8be9f8aacd65a959594075aea8bdf271)
This allows bot owners to configure which streams messages are delivered
to without needing to change webhook URLs or configuration files.
(imported from commit 32a0c26657c145b001cd8cb3ce0a0364d48902ce)
Before saving a Message object, call update_calculated_fields()
to set the has_attachment/has_image/has_link fields.
Note that the pre_save hook we added here does not get called
if you call bulk_create, hence the explicit call to
update_calculated_fields() in do_send_messages().
(imported from commit 1d60ae5908ef186aa5ff1e39277dbb2b765e60d4)
A stream is vacant when it has no subscribers and occupied when it has at least
one subscriber.
We have a slightly odd model where stream creation is conflated with
subscription creation. Streams are created by attempting to subscribe to a
stream that doesn't exist. We also hide streams with no subscribers from users
to make it seem like they've gone away. However, we can't actually remove those
streams because we want to preserve history.
This commit moves us towards a separation of these two concepts. By sending
events for stream creation, occupation, vacancy, and deletion, we allow clients
to directly observe the global state of streams rather than indirectly observing
subscription information. A more complete solution would involve adding a view
for explicitly creating streams without subscribing to them.
This commit does not handle the intricacies of invite-only streams. We
currently simply do not send these events for invite-only streams.
(imported from commit 5430e5a5eecefafcdba4f5d4f9aa665556fcc559)
Previously, the email mirror queue worker used the API bindings to send
messages to Zulip, as if it were any other API client.
This is inefficient since we're running the worker inside the Django
context on a machine with database access; we can instead just use the
internal message-sending functions we use elsewhere. This also resolves
potential issues with SSL certificates, etc. that might occur when we
were previously making a HTTPS connection.
(imported from commit 6de8015829bec440f1af0199a2138828e86ed2a4)
Previously, digest emails provided links to Zulip that didn't correctly
encode "/" if it occurred in a stream name or topic. By explicitly
specifying «safe=""», we can request that urllib.quote escape such
slashes.
Closes trac #2294.
(imported from commit 2e6334672969d4cf4032d2ea5dc80091af96d672)
We now allow the list of recipients to be sent as a
comma-delimited string with optional JSON encoding.
(imported from commit e928b037bbd258348eb5b2ecca486d0bb77f593e)
We now will match an alert word even if it is used at the boundry of
bolding, backtick escaping, or caret quoting.
Closes trac #2186.
(imported from commit 984bc63eb621772c95a01ca5c5bfeb190767f71f)
Before we deploy this commit, we must migrate the data from the staging redis
server to the new, dedicated redis server. The steps for doing so are the
following:
* Remove the zulip::redis puppet class from staging's zulip.conf
* ssh once from staging to redis-staging.zulip.net so that the host key is known
* Create a tunnel from redis0.zulip.net to staging.zulip.net
* zulip@redis0:~$ ssh -N -L 127.0.0.1:6380:127.0.0.1:6379 -o ServerAliveInterval=30 -o ServerAliveCountMax=3 staging.zulip.net
* Set the redis instance on redis0.zulip.net to replicate the one on staging.zulip.net
* redis 127.0.0.1:6379> slaveof 127.0.0.1 6380
* Stop the app on staging
* Stop redis-server on staging
* Promote the redis server on redis0.zulip.net to a master
* redis 127.0.0.1:6379> slaveof no one
* Do a puppet apply at this commit on staging (this will bring up the tunnel to redis0)
* Deploy this commit to staging (start the app on staging)
* Kill the tunnel from redis0.zulip.net to staging.zulip.net
* Uninstall redis-server on staging
The steps for migrating prod will be the same modulo s/staging/prod0/.
(imported from commit 546d258883ac299d65e896710edd0974b6bd60f8)
Apparently the "inline" treeprocessor is what runs the inline
patterns. Also re-enable the rewriting-to-https support.
(imported from commit 2fde2c1f15217a784f26b16db25ee745f424f2f0)
Have the server send down the stream's id for removal
events, and have the client use that id to look up the
stream in its internal data structures. This sets the
stage for eventually just sending the stream id (and not
the stream name) down to clients, once all our clients
are ready to use the stream id.
(imported from commit 922516c98fb79ffad8ae7da0396646663ca54fd0)
Here, we don't want to check the uploading users' realm when determining
message privacy, because that'll prevent non-Zulip users from having
email-mirror-uploaded images. Instead, we just pass along the target
realm for the message explicitly to upload_message_image()
(imported from commit 6891261552135b1f41ff9da55ffe963ee5000556)
Our overall guideline is the type names for events are singular, and the list of
events of that type are plural. 'subscriptions' was not following this guideline
and (potentially as a result) had a bug where it was impossible for clients to explicitly
subscribe to subscription change events properly.
(imported from commit 7b3162141fd673746e0489199966c29ea32ee876)
For EventsRegisterTest that test updates to streams and
subscriptions, we now validate the events generated by
the actions under test conform to predicted schemas.
We define the schemas with help from the validators code
that is also sometimes used to validate incoming request
parameters for our views.
(imported from commit b4222b920a588e15cccee4a2349c074ca9697448)
This change also makes it so that the test_rename_stream()
test exercises the code path. We need to subscribe the user
to the stream in order to generate events.
(imported from commit 77f965efbf5a766eb8de23486e303fa135b2e638)
We now show the module name (e.g. "tests or test_hooks") in the
test output. This change also eliminates the intermediate use
of slashes in the test_name var, which was passed to
bounce_key_prefix_for_testing().
(imported from commit 58e73301037a0b07d7e437514c247f7cb559420e)
Instead of having home() set page_params.realm_name directly from
the user_profile object, have fetch_initial_state_data() set it.
This is more consistent with how we treat other data, and it protects
us against a race condition where realm name updates arrive during
the DB fetching.
(imported from commit 545e3bd73f150438126e3f941e9bebc7aa1d0614)
In particular, make the stream history inaccessible and free up the
name to be re-used.
(imported from commit 6063b7a484ed0ba0279a17d2b3e9a92b5ef1f762)
The file test_runner.py has our subclass of DjangoTestSuiteRunner
and various methods that help it work.
(imported from commit 8eca39a7ed3f8312c986224a810d4951559e7a8b)
The function update_user_profile_caches now operates on a list
of user_profiles, so callers like flush_realm() can benefit from
having a single cache_set_many() call. This slightly complicates
the call from flush_user_profile().
(imported from commit e064871d849b873c6ca388f00d4f7afaba1bf222)
For the realm-wide caches of active user dicts and alert words, just
make a single call to cache_delete() when you are deactivating a
realm. Before this change, we were doing O(N) cache_deletes as
part of the code path through flush_user_profile(). Now we just
call update_user_profile_caches() directly to clear the user_profile
caches.
This change also sets us up to turn flush_realm() into a post-save hook.
(imported from commit 699b4ea226ae15fc8c402cb4bc64ff6bdc041fc2)
This is a slight behavior change, as we now flush user_profile
caches for bots as well as humans.
(imported from commit 24c72c44d851ee4c66a67a4728cd6c548faeedcd)
This function updates all the user_profile-related caches
that are keyed on a per-user basis.
(This had some test coverage already.)
(imported from commit 37979400514a7b46a6dcb7e36665b0fee2f3c525)
Stream name and descriptions updates were being sent to all of the
active users on a realm. They are now only send to users who would have
information about that stream.
(imported from commit 2621ee8029f7356bf44ec493d7b5361bd546a8f5)
This is a lot simpler and eliminates a possible failure mode in the
data transfer path.
(imported from commit 19308d2715bbd12dc9385234f1d9156f91bdfae0)
To mutate the state for removing subscribers, the previous
code was essentially adding in event['subscriptions'] to
state['unsubscribed'], but that was a naive approach, since
the event object only has the name of the subscriber, and not
the full subscription info.
We instead effectively copy records over from state['subscriptions']
to state['unsubscribed'], and we also do surgery on the subscribers
that made me need to add the user_profile parameter to apply_events().
With the code apparently working now, I was able to remove the
match_except() test helper and use a more thorough matcher in
the test on do_remove_subscription().
Part of fixing the "remove" case was cleaning up the "add" case,
since they aren't quite symmetric opposites of each other, although
under this refactoring they now share the new name() helper.
(imported from commit 0deab67d0c7b08b3ad962493efae3762a835fd29)
Because full_name and is_admin changes go through many similar,
generic codepaths, it is almost more work at this point to keep
is_admin out of page_params as it is to just put it in. So
I put it in. This should pave the path for showing admins in
the GUI.
This commit actually starting by my adding a test
that calling apply_events() with the notification you get
when calling do_change_is_admin() updates
state['realm_users'] to be similar to what you would get
out of fetch_initial_state_data(). We didn't have test coverage
there before. Making that test pass forced my hand to
either add is_admin to page_params or to special-case
apply_events() not to update page_params with is_admin. I
chose the former approach.
(imported from commit 1e49d59c66540014284529c29d5007224be6a0c6)
A description was added to the streams and it is now displayed on the
subscriptions page. It can not be set in the UI yet.
(imported from commit 81d08b65eee42dba87cd99dd5bd30106c4eb6c6a)
While we're at it, lets comment up the function so I know what this is
doing next time :)
(imported from commit e745be75fcd6dbce9997e1d73464619fc8b73996)
If a name change event arrived during the call to
fetch_initial_state_data(), we would call apply_events() to
update the data structure that eventually becomes page_params.
Our update code, instead of surgically updating the fields, was
just overwriting the record, which led to is_bot being taken
out of the record when only full_name was in the event.
Apart from fixing the "update" case to do the right thing,
this commit also does a bit of cleanup on the code handling
"realm_user" events to make it more generic in how it does
add/remove/update. If we could standardize our events a bit
more, this could eventually lead to DRYing up some of the
apply_events() code.
(imported from commit 772e2fcd6a5605ccb6e8d1bc499b5f336934cf3c)
When new streams are created we now send a message with a custom
markdown tag that renders a subscribe button.
(imported from commit 9dfba280b3b4ff4f32f6431ef9227867c8bf4b40)
Added a default_desktop_notifications boolean to userprofile with a UI
in Zulip Labs. This flag is used to default the notification flag on new
subscriptions.
(imported from commit a25223cc5ecf09980cf877991e25034bb3fd4046)
Google Groups won't let you add an email address that has a '+' in it
as a member of a group, so we allow '.' as well.
This commit also fixes a current issue with email mirroring, where it
doesn't work if you had a + in the stream name.
This fixes Trac #2102.
(imported from commit 9a7a5f5d16087f6f74fb5308e170a6f04387599e)
Cross-realm private messages are only used to respond to support
requests. So now cross-realm private messages are only allowed if
exactly two realms are in the private message and one of them is
zulip.com.
(imported from commit f01a2824e214682acb22a6995714a9d1b0d0c66f)
This does result in a few more rabbitmq events to be processed (though
a negligible number compared to what we already do), but it saves a
database query from inside Tornado whenever we occasionally have a
cache miss looking up the UserProfile, which is far more important.
(imported from commit a553a00a3004ba27bfb54ffbc3e9c9b170ebae4d)
When we edit a message, send out UserMessage flags to the recipients.
This sets the stage for making sure that changes related to
user-specific alert words or mentions get sent out to users.
(imported from commit bce1de19acef44b5e106352f261203352ece02b9)
Image and video links in the twitter API are media and need to be
handed on separately. We also include a preview image if the media link
is a to a picture.
(imported from commit 2bd00d267e51b29ad0ba681195b2bfea9b991d8c)
This converts links in tweets to a tags. We also convert the displayed
text to the target of the twitter short URL. Mentions are linked to the
users twitter page.
(imported from commit 192d5546a7eea82759f9ae30d82c102aed15ff71)
Previously, we were processing in Tornado mirroring dummy users (and
deactived users) as recipients -- resulting in bugs where the missed
message hooks would fire for these nonexistent users.
Our Tornado real-time delivery system only needs the list of active
users both for delivery and for presence information updates.
(imported from commit b81143f106a4d0eefa4b838e7c074b2963259746)
Before this is deployed to prod, we need to manually frob our database
to set the is_mirror_dummy=True bit for all existing mirror users.
(imported from commit 39f1938cef091cf1d7d97307f76b137fe1d92b6c)
This contains the various fixes that needed to be made in order to get
accurate statistics.
Most notably, the active_users_between function in the previous
version of zerver/lib/statistics.py was broken for end dates in the
past, because it used the UserActivity table to get its data -- so in
fact it really was querying "users last active between".
This commit isn't super clean, but I figure we're probably better off
having our latest code for historical usage data in git so it doesn't
bitrot and anyone can improve on it.
(imported from commit 24ff2f24a22e5bdc004ea8043d8da12deb97ff2f)
Features:
* Only shows messages in the narrow
* New messages in the narrow will arrive as they are sent
* Works even for streams you're not subscribed to
* Automatically subscribes you to a stream on send
* Doesn't update your pointer
* All searches etc. automatically have the narrow added
(imported from commit 2e12b76849f6ca0f53dda5985dad477a04f7bbac)
This sets up a scheme to validate complex data structures and
give specific error messages for improperly typed parameters.
(imported from commit 33b2f070d993da4ee929119dd41503bd0128c8eb)
* Deal with shorter tweet IDs
(some old tweets don't have a full 18-character ID)
* Allow trailing slash
* Deal with old-style #! syntax
* Deal with links that link to a photo
(imported from commit 008a98c806f3b8dddd9e2f18a8f002af6932766f)
These images at least load now, but that's because Camo redirects
the browser to the origin server, so the only effect is an extra
round-trip time.
(imported from commit 0d6b9c888a5cdfaa9299272d74a085e872dfa434)
In plaintext e-mails these will be simple links.
In HTML e-mails these become a <link> and <img>, which some web mail
clients may inline.
(imported from commit b1242dfd917008a019981eb2224c1c7f5f84739f)
You can't have unread PMs sent by you, so we weren't explicitly
checking this, but when testing locally we often ignore the unread
check. Filter PMs sent by you to reduce confusion when testing
locally.
(imported from commit 0205c4a3ed67790b9d60d4f2b927e4cb9e720bf3)
Since we changed the initial subscription data to include
user_profile_ids rather than emails, we need to preserve that when
adding in events generated during the page load.
(imported from commit 4f4071b8ba30e57c6f64c9e7b54c1cc754e8f010)
This will allow us to substantially decrease the server-side work that
we do to support our Mirroring systems (since the personal mirrors can
request only messages that user sent) and also is what we need to
support a single-stream Zulip widget that we embed in webpages.
(imported from commit 055f2e9a523920719815181f8fdb44d3384e4a34)
This caused problems with our tests suite where we were using a logged
in browser session and actually acting as a different user.
(imported from commit 73b8cb39d5d669e682fbacf2f7e574c228885c2f)
This replaces the AppleDeviceToken table with a generic
PushDeviceToken with a `kind` field to make it easier to add functionality
like per-device/per-stream settings that share code between Android and
iOS devices.
The schema must continue to work on prod with the old table name, so we
add the new table in parallel and can drop the old table once this code
hits prod and any necessary data is copied.
(imported from commit 0209a7013f2850ac6311f23c3d6f92c65ffd19e3)
We have observed additional exceptions being thrown from zulip_finish and we
need to make sure that the handler is disconnected from the queue, or else the
event queue will keep throwing exceptions due to the handler being closed.
(imported from commit 59273aa14495216430b9eb1525b2cce230d8913d)
This should dramatically improve the speed of the dump/load part of
our restart process, especially with large long-lived event queues.
(imported from commit ae4ae20ba2ca4433e25a5e7beeb4fa4882c53972)
Previously, we had an issue with the ACKing protocol, where if a
virtualizable event (like a "read" flag) was dispatched to a queue
client immediately, we would not properly ACK the change because it
had been made a virtual event.
(imported from commit ea09812f8a5ba1d5aad65f536022e3dbc77b0f9e)
Unbundle the push notifications from the missed message queue processors
and handlers. This makes notifications more immediate, and sets things up
for better badge count handling, and possibly per-stream filtering.
(imported from commit 11840301751b0bbcb3a99848ff9868d9023b665b)
This should dramatically improve the speed of the dump/load part of
our restart process, especially with large long-lived event queues.
(imported from commit cc493fa50b4c339257e060b3f0c0956c682e449d)
Now that we support email aliases, we have to be careful when going from
an email address to a domain that we assume we can use to get a Realm
object for. When we care about the Realm's domain, we want to follow
any RealmAliases that exist for a certain domain.
When we just care about the original email address domain itself,
for comparison or other purposes, use split_email_from_domain
This removes the ambiguity of having to decide when to use
email_to_domain + RealmAlias or just email_to_domain
(imported from commit 0e199495502d946ce2e1aae56263e7e8665be4ed)
Now we can nest fenced code/quote blocks inside of quote
blocks down to arbitrary depths. Code blocks are always leafs.
Fenced blocks start with at least three tildes or backticks,
and the clump of punctuation then becomes the terminator for
the block. If the user ends their message without terminators,
all blocks are automatically closed.
When inside a quote block, you can start another fenced block
with any header that doesn't match the end-string of the outer
block. (If you don't want to specify a language, then you
can change the number of backticks/tildes to avoid amiguity.)
Most of the heavy lifting happens in FencedBlockPreprocessor.run().
The parser works by pushing handlers on to a stack and popping
them off when the ends of blocks are encountered. Parents communicate
with their children by passing in a simple Python list of strings
for the child to append to. Handlers also maintain their own
lists for their own content, and when their done() method is called,
they render their data as needed.
The handlers are objects returned by functions, and the handler
functions close on variables push, pop, and processor. The closure
style here makes the handlers pretty tightly coupled to the outer
run() method. If we wanted to move to a class-based style, the
tradeoff would be that the class instances would have to marshall
push/pop/processor etc., but we could test the components more
easily in isolation.
Dealing with blank lines is very fiddly inside of bugdown.
The new functionality here is captured in the test
BugdownTest.test_complexly_nested_quote().
(imported from commit 53886c8de74bdf2bbd3cef8be9de25f05bddb93c)
This reverts commit 1147814b22fb9737a807057ddbdbe0e9554086e0.
This seems to with some probability screw up our Zephyr mirroring
script.
(imported from commit 4f82452f1b0ca98e6b895db020e071d2daa325e4)
This requires a puppet apply on each of staging and prod0 to update
the nginx configuration to support the new URL when it is deployed.
(imported from commit a35a71a563fd1daca0d3ea4ec6874c5719a8564f)
There will be browser errors on staging when this is deployed due to the socket
protocol changing.
(imported from commit f1eda5b5c2ec9c60c23b3ca96277a61debadf5bb)
I believe there may also be others. I'm still not sure why clients would be
sending open requests without session or csrf values in their cookies, though.
(imported from commit 7e9660c1c4d5c2abf55ff21b433ba0117180eb82)
CUSTOMER13 doesn't want it, and there's currently no nginx config
or configurable Camo URI, so it wouldn't work if image preview
were enabled.
(imported from commit 615d4a32acbc4d4d590f88cf4e7d45d8f49db1d3)
ScheduledJobs with type Email displace the usual mandrill codepaths
in the Zulip Enterprise deploys
* Email-specific helper functions will appear in deliver_email.py
* 0058_auto__add_scheduledjob.py
(imported from commit 8db08d8a279600322acfdbed792dc1a676f7a0ab)
The !gravatar markdown no longer hard codes to Gravatar, but
instead it serves up our generic avatar URL.
(imported from commit 4e3e2baeb3374bcf025a18ff27a8452b975c22b7)
We also now separate out the times for the socket overhead, the
request service time, and the queuing delays.
(imported from commit e1683f7f28b968b86ebb701b0ac29b00ac6d67c3)
One quirk here is that the Request object is built in the
message_sender worker, not Tornado. This means that the request time
only counts time taken for the actual sending and does not account
for socket overhead. For this reason, I've left the fake logging in
for now so we can compare the two times.
(imported from commit b0c60a3017527a328cadf11ba68166e59cf23ddf)
This logging is kinda excessive since it adds like 4 log lines per
recipient, so I expect we'll end up reverting it once we've debugged
the proximal issue.
(imported from commit 5e6ab3e230f32b65ad9cf0d95f20ffbc0fe7397e)
This will hopefully help with the send dialog being stuck on
"sending" as well as allowing us to not show errors to the user on
reconnect.
(imported from commit 31ee889853f348e486863073dc130cdfb4e1338d)
Anders Kaseorg
David Roe
Tim Abbott
Kate Buckner
acrefoot
Yoyo Zhou
Luke Faraone
acrefoot
Jason Michalski
Leo Franchi
Tim Abbott
Steve Howell
Waseem Daher
Paul Ruan
Luke Faraone
Waseem Daher
Tolga Tezel
Zev Benjamin
Leo Franchi
Jessica McKellar
Kevin Mehall