Commit Graph

6734 Commits

Author SHA1 Message Date
picapi_ e2c14724d9 mypy: Use Python 3 type syntax in zerver/views/messages.py. 2017-12-03 13:48:18 -08:00
Rhea Parekh 460e9b4634 Scripts: Add script for resetting postgres sequences for imports. 2017-12-01 06:38:08 -08:00
Alena Volkova 3be8e95d13 settings: Send bot types from the backend instead of hardcoding them. 2017-11-30 15:28:53 -08:00
Tim Abbott 2310794451 bots: Don't allow embedded bots bot type in production.
This is important for making it easy for us to merge code related to
UI for embedded bots now, without enabling them in the production UI.
2017-11-30 15:28:53 -08:00
Rishi Gupta 7d1c88f0fb invitations: Limit realms to 100 invites per day.
To guard against using zulip invites as a vector for spam. Stopgap measure
until we figure out something better.
2017-11-29 22:18:05 -08:00
Rishi Gupta 7c9694077e confirmation: Move check_prereg_key_and_redirect to registration.py. 2017-11-29 22:18:05 -08:00
Tim Abbott 6e8f4ffc77 test_signup: Expand tests for invalid confirmation links.
We'll need the expanded test coverage when we move
check_prereg_key_and_redirect to zerver/views/registration.py to avoid
test failures, and these are also tests we should really have anyway.
2017-11-29 22:18:05 -08:00
Rishi Gupta 331a9bee6b confirmation: Add Confirmation.REALM_CREATION.
Does not change any behavior. Groundwork for adding a realm field to
Confirmation.
2017-11-29 22:03:09 -08:00
Greg Price 4926228071 rabbitmq: Do a better job of retrying failed connections.
Empirically, the retry in `_on_connection_closed` didn't actually work
-- if a reconnect failed, that was it, and the exception handler
didn't get run.  A traceback would get logged, but all its frames were
in Tornado or Pika, not our own code; presumably something magic and
async was happening to the exception.

Moreover, though we would make one attempt to reconnect if we had a
connection that got closed, we didn't have any form of retry if the
original attempt at connecting failed in the first place.

Happily, upstream offers a perfectly reasonable bit of API that avoids
both of these problems: the on-open-error callback.  So use that.
2017-11-29 16:56:29 -08:00
Greg Price 7ac2b58584 rabbitmq: Reorder a bit to group our reconnect logic together. 2017-11-29 16:56:29 -08:00
Greg Price c32b16715d tornado: Use spiffy new `call_later` rather than `add_timeout`.
This method was new in Tornado 4.0.  It saves us from having to get
the time ourselves and do the arithmetic -- which not only makes the
code a bit shorter, but also easier to get right.  Tornado docs (see
http://www.tornadoweb.org/en/stable/ioloop.html) say we should have
been getting the time from `ioloop.time()` rather than hardcoding
`time.time()`, because the loop could e.g. be running on the
`time.monotonic()` clock.
2017-11-29 16:56:29 -08:00
Greg Price 73886f57d2 zerver/lib/queue: Clean up import order. 2017-11-29 16:56:29 -08:00
Greg Price 3c4e4c14c9 rabbitmq: Add on-close callback atomically in creating the connection.
Adding it afterward is inherently racy, and upstream's API is quite
reasonable for avoiding that -- just like we can pass an on-open
callback up front, we can do the same with the on-close callback.

This is a more thorough version of 4adf2d5c2 from back in 2013-04.
2017-11-29 16:56:29 -08:00
Greg Price e88c2a7ee4 rabbitmq: Cut redundant `stop_ioloop_on_close` parameter.
The default value of this parameter is already False upstream.
(It was already False in pika version 0.9.6, which we were
supposedly using when we introduced this in 4baeaaa52; not sure
what the story was there.)
2017-11-29 16:56:29 -08:00
Tim Abbott be832387c5 user_groups: Fix reporting of empty user groups.
Previously, we weren't doing a proper left join in
user_groups_in_realm_serialized, resulting in empty user groups being
excluded from the query.  We want to leave decisions about excluding
empty user groups to the UI layer, so we include these here.
2017-11-29 16:13:55 -08:00
Greg Price fe979af822 purge_queue: Allow purging Tornado-consumed queues.
These are the exceptions to the rule that our queues correspond to
queue-processor workers.

Purging `notify_tornado` in particular is a useful workaround right
now for some error spew in the dev environment.
2017-11-29 14:44:53 -08:00
Tim Abbott b2cb443d24 subs: Fix clearing unread counts when leaving private streams.
Because we use access_stream_by_id here, and that checks for an active
subscription to interact with a private stream, this didn't work.

The correct fix to add an option to active_stream_by_id to accept an
argument indicating whether we need an active subscription; for this
use case, we definitely do not.
2017-11-29 14:40:08 -08:00
Tim Abbott 6bc1a682de notifications: Separate push and email notifications checks.
This is an early step in a larger refactor to properly decouple the
email and push notification code paths.
2017-11-28 17:51:18 -08:00
Sarah ba70b542d3 notification settings: Add stream email notifications setting.
Adds a field to UserProfile for enabling emails about stream
messages and a field to Subscriptions about email
notification preferences.
2017-11-28 17:51:18 -08:00
rht ee546a33a3 zerver/lib: Use python 3 syntax for typing.
Edited by tabbott to improve various line-wrapping decisions.
2017-11-28 17:15:14 -08:00
rht 229a8b38c0 zerver/lib: Use Python 3 syntax for typing for several files.
This adds a number of annotations that had been missed in previous
passes.
2017-11-28 17:02:24 -08:00
rht 115f7e6055 zerver/tornado: Use Python 3 syntax for typing (part 2). 2017-11-28 17:02:24 -08:00
rht a1cc720860 zerver: Use Python 3 syntax for typing.
Tweaked by tabbott to fix some minor whitespace errors.
2017-11-28 16:49:36 -08:00
Greg Price 0ec2a9d259 auth: Try switching to register even if user exists on another realm.
For example, this means that if a user already has an account on one
realm and they try to make an account on another by hitting "Sign in
with Google" (rather than following the little "Register" link to a
"Sign up with Google" button instead), they'll get to make an account
instead of getting an error.

Until very recently, if the user existed on another realm, any attempt
to register with that email address had to fail in the end, so this
logic gave the user a useful error message early.  We introduced it in
c23aaa178 "GitHub: Show error on login page for wrong subdomain"
back in 2016-10 for that purpose.  No longer!  We now support reusing
an email on multiple realms, so we let the user proceed instead.

This function's interface is kind of confusing, but I believe when its
callers use it properly, `invalid_subdomain` should only ever be true
when `user_profile` is None -- in which case the revised
`invalid_subdomain` condition in this commit can never actually fire,
and the `invalid_subdomain` parameter no longer has any effect.  (At
least some unit tests call this function improperly in that respect.)
I've kept this commit to a minimal change, but it would be a good
followup to go through the call sites, verify that, eliminate the use
of `invalid_subdomain`, then remove it from the function entirely.
2017-11-28 16:38:41 -08:00
Vishnu Ks d329636079 tests: Add test for user signup with already-used email.
[Test simplified a bit by greg.]
2017-11-28 16:23:10 -08:00
Greg Price 50e451cd7b populate_db: Create an email-reusing Cordelia in the "lear" realm. 2017-11-28 16:23:10 -08:00
Tim Abbott 586e18b237 auth: Allow accounts with the same email in different realms.
[Modified by greg to (1) keep `USERNAME_FIELD = 'email'`,
(2) silence the corresponding system check, and (3) ban
reusing a system bot's email address, just like we do in
realm creation.]
2017-11-28 16:23:10 -08:00
Tim Abbott e6f460f511 auth: Replace user_email_is_unique validator.
As we migrate to allow reuse of the same email with multiple realms,
we need to replace the old "no email reuse" validators.  Because
stealing the email for a system bot would be problematic, we still ban
doing so.

This commit only affects the realm creation logic, not registering an
account in an existing realm.
2017-11-28 16:23:10 -08:00
Greg Price d6cfa56bc1 test_signup: Test that cloning a system bot's email is not allowed.
Just now this is largely redundant with `test_signup_already_active`;
but very soon when we allow reusing an email across realms, the logic
will diverge.
2017-11-28 16:23:10 -08:00
Greg Price 5f38b374b6 register: Activate existing mirror dummy only if on same realm.
The one thing this bit of logic is used for is to decide whether
there's an existing user which is a mirror dummy that we should
activate.  This change causes us to ignore such an existing user if
it's on some other realm, and go straight into `do_create_user`.
2017-11-28 16:23:10 -08:00
Greg Price eb8b05b2b5 test_email_mirror: Remove a brittle bit of logic.
This code appears to exist to cover a few extra lines in
zerver/lib/digest.py.  But it's rather brittle, tucked as it is into
the middle of a different test's loop, and with the upcoming
introduction of the `lear` realm in testing, this test code itself
loses coverage.

For now, rather than fix this test code up just delete it; we don't
have 100% coverage on `zerver/lib/digest.py`, while we do on this test
file, so that avoids breaking coverage in CI.  As a followup, we
should add back some logic like this but in a more robust way,
probably as its own separate test method.
2017-11-28 16:23:10 -08:00
Greg Price 453585abb1 test_messages: Clean up some error spew by updating a test.
This test would produce a bunch of log messages with tracebacks,
complaining that `welcome-bot@zulip.com` tried to send cross-realm PMs
and can't.  The issue is that the test overrides
`settings.CROSS_REALM_BOT_EMAILS`, and hasn't kept up with additions
to the normal value for that setting.  Update it so that welcome-bot
is permitted as usual.
2017-11-28 16:23:10 -08:00
Greg Price b6cc21b438 debug: Add facility to dump tracemalloc snapshots.
Originally this used signals, namely SIGRTMIN.  But in prod, the
signal handler never fired; I debugged fruitlessly for a while, and
suspect uwsgi was foiling it in a mysterious way (which is kind of
the only way uwsgi does anything.)

So, we listen on a socket.  Bit more code, and a bit trickier to
invoke, but it works.

This was developed for the investigation of memory-bloating on
chat.zulip.org that led to a331b4f64 "Optimize query_all_subs_by_stream()".

For usage instructions, see docstring.
2017-11-28 15:52:07 -08:00
derAnfaenger 6dd639454e integrations: Add topic parameter for Appfollow.
Tweaked by tabbott to document the integration.
2017-11-28 15:49:19 -08:00
Umair Khan 1acdfef13c two_factor: Disable prefix in OurAuthenticationForm.
In two factor authentication every step adds a unique prefix to the fields,
due to this the name of the form fields differs from the HTML fields. If
we do not do this we will have to change the name in the HTML, which
will cause the change in tests.
2017-11-28 15:27:44 -08:00
Tim Abbott 4f5a5a8547 auth: Extract add_dev_login_context.
This simple refactor cleans up the code for the dev_auth_enabled code
path to be a bit less deeply in the login_page() logic.
2017-11-28 15:17:33 -08:00
Tim Abbott d5bb45db60 webhooks/gci: Fix missing mypy annotation. 2017-11-28 14:08:44 -08:00
Sampriti Panda bea653fabc webhooks/gci: Improve message formatting.
* Use Student Name instead of Task Name in subject.
* Use Task Instance URL instead of Task Definition URL (and workaround
  for a bug in the API).
2017-11-28 13:53:42 -08:00
Robert Hönig 1e66bd1c02 tests: Add newlines to test_service_bot_system.test_internal_endpoint(). 2017-11-27 21:05:34 -08:00
derAnfaenger c8a5ae753c embedded bots: Consistently use 'storage' instead of 'state.' 2017-11-27 21:05:34 -08:00
Tim Abbott 9645c8b31f integrations: Remove disabled googlesearch integration.
This was moved to another part of the python-zulip-api codebase
because it was unfinished.
2017-11-27 20:50:37 -08:00
Eeshan Garg c45517f544 python-zulip-api: Upgrade to PyPI package release 0.3.8.
There's one migration required by this release:

* queue_processors: Stop passing state_handler to handle_message.

  state_handler is now a property of bot_handler and thus, does
  not need to be passed to bot_handler.handle_message().

  The commit responsible is:
  2a74ad11c5
2017-11-27 20:31:37 -08:00
rht e538f4dd44 zerver/views: Use Python 3 syntax for typing.
Edited by tabbott to remove state.py and streams.py, because of
problems with the original PR's changes, and wrap some long lines.
2017-11-27 17:10:39 -08:00
Tim Abbott bd8196c3a5 lint: Fix too-long line in auth.py. 2017-11-27 15:28:46 -08:00
Vishnu Ks d4ee3023b2 registration: Require an explicit realm on PreregistrationUser.
This completes the last commit's work to fix CVE-2017-0910, applying
to any invite links already created before the fix was deployed.  With
this change, all new-user registrations must match an explicit realm
in the PreregistrationUser row, except when creating a new realm.

[greg: rewrote commit message]
2017-11-27 14:59:34 -08:00
Vishnu Ks 985768b2fd registration: Check realm against PreregistrationUser realm.
We would allow a user with a valid invitation for one realm to use it
on a different realm instead.  On a server with multiple realms, an
authorized user of one realm could use this (by sending invites to
other email addresses they control) to create accounts on other
realms. (CVE-2017-0910)

With this commit, when sending an invitation, we record the inviting
user's realm on the PreregistrationUser row; and when registering a
user, we check that the PregistrationUser realm matches the realm the
user is trying to register on.  This resolves CVE-2017-0910 for
newly-sent invitations; the next commit completes the fix.

[greg: rewrote commit message]
2017-11-27 14:58:26 -08:00
Tim Abbott 8b935f4e99 settings: Add setting for SYSTEM_BOT_REALM.
This fixes some subtle JavaScript exceptions we've been getting in
zulipchat.com, caused by the system bot realm there not being "zulip"
interacting with get_cross_realm_users.
2017-11-27 14:46:07 -08:00
Tim Abbott 25fd4c5508 bulk_get_users: Edit the cache keys to make them more unique.
While at this point I was to rewrite this function, this at least
plugs the issues for now.
2017-11-27 14:41:31 -08:00
Tim Abbott 339b67f7af get_cross_realm_dicts: Harden against caching bug in last commit.
This should help protect us from future issues with the way that
`bulk_get_users` does caching.

It's likely that we'll want to further restructure `bulk_get_users` to
not have this base_query code path altogether (since it's kinda
buggy), but I'm going to defer that for a time when we have another
user.
2017-11-27 14:35:50 -08:00
Tim Abbott 646ba5b9e5 bulk_get_users: Fix issues with users in multiple realms.
The previous implementation had a subtle caching bug: because it was
sharing its cache with the `get_user_profile_by_email` cache, if a
user happened to have an email in that cache, we'd return it, even
though that user didn't match `base_query`.

This causes `get_cross_realm_users` to no longer have a problematic
caching bug.
2017-11-27 14:34:45 -08:00