'compose_views' is used inside an outer db transaction created in
'update_user_group_backend'.
`transaction.atomic()` block in 'compose_views' resulted in
savepoint creation.
This commit adds `savepoint=False` to avoid that.
This commit adds 'durable=True' to the outermost transactions
of the following functions:
* do_create_multiuse_invite_link
* do_revoke_user_invite
* do_revoke_multi_use_invite
* sync_ldap_user_data
* do_reactivate_remote_server
* do_deactivate_remote_server
* bulk_handle_digest_email
* handle_customer_migration_from_server_to_realm
* add_reaction
* remove_reaction
* deactivate_user_group
It helps to avoid creating unintended savepoints in the future.
This is as a part of our plan to explicitly mark all the
transaction.atomic decorators with either 'savepoint=False' or
'durable=True' as required.
* 'savepoint=True' is used in special cases.
Previously, the `group_id` was not returned in the success
response of the user group creation API.
This commit updates the API to return a success response
containing the unique ID of the user group with the key
`group_id`. This enhancement allows clients to easily reference
the newly created user group.
Fixes: #29686
This commit adds support to add subgroups to a group while
creating it.
User can add the subgroups to group irrespective of permissions
like user can add members during creating it.
This commit updates code to allow users with permission
to add members to add subgroups as well. And only users
with permission to manage the group can remove subgroups.
Also updated tests to check permissions in separate tests
and removed them from the existing test.
Users with permission to manage the group have all the permissions
including joining/leaving the group, adding others group which also
have a separate setting to control them.
So, it makes sense to just check managing permissions first in
access_user_group_for_update and then check the specific permission.
There is no behavioral change in this commit, it only changes the
order of checking permissions.
We want to allow the user, who can add others to group, to
join the group as well irrespective of can_join_group setting.
Previously, the permission to add others (or say anyone) was
controlled by can_manage_group setting, but now it is controlled
by can_add_members_group setting. This commit fixes the code to
use can_add_members_group setting to check permission for joining
the group.
This commit also improves the tests for checking permission to
join the group such that different settings are tested in isolation.
Removing members will be controlled by `can_manage_group` until we add
`can_remove_members_group` in the future.
Users with permission to manage a group can add members to that group by
default without being present in `can_add_members_group`.
This commit updates backend code to not allow adding deactivated
users to groups including when creating groups and also to not
allow removing deactivated users from groups.
This commit adds access_user_group_to_read_membership function
so that we can avoid calling get_user_group_by_id_in_realm with
"for_read=True" from views functions, which is better for security
since that function does not do any access checks.
This commit refactors the code to check permission for
accessing user group in such a way that we can avoid
duplicate code in future when we will have different
settings controlling the permissions for editing group
details and settings, joining the group, adding others
to group, etc.
This commit renames "allow_deactivated" parameter in
"GET /user_groups" endpoint to "include_deactivated_groups", so
that we can have consistent naming here and for client capability
used for deciding whether to send deactivated groups in register
response and how to handle the related events.
We only allow updating name of a deactivated group, and not
allow updating description, members, subgroups and any setting
of a deactivated user group.
Deactivated user groups cannot be a a subgroup of any group
or used as a setting for a group.
Earlier there was only a realm level setting for configuring
who can edit user groups. A new group level setting is also added
for configuring who can manage that particular group.
Now, a user group can be edited by a user if it is allowed from
realm level setting or group level setting.
This commit make changes to also use group level setting
in determining whether a group can be edited by user or not.
Also, updated tests to use api_post and api_delete helpers instead
of using client_post and client_delete helpers with different users
being logged in.
Earlier there was a single decorator function to check whether
user can create and edit user groups. This commit adds a new
decorator function to check whether user has permissions to
create user groups.
This was done because in future commits we will be adding a
realm level setting for configuring who can create user groups.
The database operations in 'access_user_group_for_setting' and
'check_add_user_group' used in 'add_user_group' view should be
collectively atomic.
This commit adds transaction.atomic decorator for that purpose.
Earlier, we were using 'send_event' in 'edit_user_group' codepath
which can lead to a situation where we enqueue events but the
function fails at a later stage.
Events should not be sent until we know we're not rolling back.
Fixes part of #30489.
Migrate the following endpoints from @has_request_variables
to @typed_endpoint:
- get_user_group()
- delete_user_group()
- update_user_group_backend()
- update_subgroups_of_user_group()
- get_is_user_group_member()
- get_user_group_members()
- get_subgroups_of_user_group()
With tweaks from tabbott to avoid calling thunks unnecessarily.
This commit adds a server level setting which controls whether the setting
can be set to anonymous user groups. We only allow it in the tests for
now because the UI can only handle named user groups.
This commit fixes the code store correct old value in audit
log data when changing can_mention_group setting from a
anonymous group to another anonymous group. The bug was
because the old value was being computed after updating
the UserGroup object with new members and subgroups and
is fixed by computing the old value for all the cases
and passing it to do_change_user_group_permission_setting.
This commit moves validate_group_setting_value_change,
are_both_setting_values_equal and parse_group_setting_value
functions, which are used for updating the group settings, to
"zerver.lib.user_groups" as these functions will also be used for
group based realm and stream settings and "zerver.lib.user_groups"
file seems a better place to place such functions which are used
at multiple places.
For same reasons, we also move GroupSettingChangeRequest dataclass
to "zerver.lib.user_groups" file.
This commit adds support to pass object containing both old and new
values of the can_mention_group setting, as well as detailed API
documentation for this part of the API system.
Co-authored-by: Tim Abbott <tabbott@zulip.com>
Co-authored-by: Greg PRice <greg@zulip.com>
We now pass the complete configuration object for a setting to
access_user_group_for_setting instead of passing the configuration
object's fields as different variables.
Adds support for bulk-adjusting a single user's membership in multiple
user groups in a single transaction in the low-level actions
functions, for future use by work on #9957.
This is important because the "guests" value isn't one that we'd
expect anyone to pick intentionally, and in particular isn't an
available option for the similar/adjacent "email invitations" setting.
This commit adds id_field_name field to GroupPermissionSetting
type which will be used to store the string formed by concatenation
of setting_name and `_id`.
**Background**
User groups are expected to comply with the DAG constraint for the
many-to-many inter-group membership. The check for this constraint has
to be performed recursively so that we can find all direct and indirect
subgroups of the user group to be added.
This kind of check is vulnerable to phantom reads which is possible at
the default read committed isolation level because we cannot guarantee
that the check is still valid when we are adding the subgroups to the
user group.
**Solution**
To avoid having another transaction concurrently update one of the
to-be-subgroup after the recursive check is done, and before the subgroup
is added, we use SELECT FOR UPDATE to lock the user group rows.
The lock needs to be acquired before a group membership change is about
to occur before any check has been conducted.
Suppose that we are adding subgroup B to supergroup A, the locking protocol
is specified as follows:
1. Acquire a lock for B and all its direct and indirect subgroups.
2. Acquire a lock for A.
For the removal of user groups, we acquire a lock for the user group to
be removed with all its direct and indirect subgroups. This is the special
case A=B, which is still complaint with the protocol.
**Error handling**
We currently rely on Postgres' deadlock detection to abort transactions
and show an error for the users. In the future, we might need some
recovery mechanism or at least better error handling.
**Notes**
An important note is that we need to reuse the recursive CTE query that
finds the direct and indirect subgroups when applying the lock on the
rows. And the lock needs to be acquired the same way for the addition and
removal of direct subgroups.
User membership change (as opposed to user group membership) is not
affected. Read-only queries aren't either. The locks only protect
critical regions where the user group dependency graph might violate
the DAG constraint, where users are not participating.
**Testing**
We implement a transaction test case targeting some typical scenarios
when an internal server error is expected to happen (this means that the
user group view makes the correct decision to abort the transaction when
something goes wrong with locks).
To achieve this, we add a development view intended only for unit tests.
It has a global BARRIER that can be shared across threads, so that we
can synchronize them to consistently reproduce certain potential race
conditions prevented by the database locks.
The transaction test case lanuches pairs of threads initiating possibly
conflicting requests at the same time. The tests are set up such that exactly N
of them are expected to succeed with a certain error message (while we don't
know each one).
**Security notes**
get_recursive_subgroups_for_groups will no longer fetch user groups from
other realms. As a result, trying to add/remove a subgroup from another
realm results in a UserGroup not found error response.
We also implement subgroup-specific checks in has_user_group_access to
keep permission managing in a single place. Do note that the API
currently don't have a way to violate that check because we are only
checking the realm ID now.
Prakhar Pratyush
Aditya Kumar Kasaudhan
Sahil Batra
Shubham Padia
Tim Abbott
Ujjawal Modi
bedo
Anders Kaseorg
Vector73
Zixuan James Li