gitlab-org--gitlab-foss/app/models/member.rb

323 lines
9 KiB
Ruby
Raw Normal View History

class Member < ActiveRecord::Base
include Sortable
include Importable
include Expirable
include Gitlab::Access
attr_accessor :raw_invite_token
belongs_to :created_by, class_name: "User"
belongs_to :user
belongs_to :source, polymorphic: true
2017-02-22 17:35:08 -05:00
delegate :name, :username, :email, to: :user, prefix: true
2015-04-10 09:09:37 -04:00
validates :user, presence: true, unless: :invite?
validates :source, presence: true
validates :user_id, uniqueness: { scope: [:source_type, :source_id],
2015-04-10 09:09:37 -04:00
message: "already exists in source",
allow_nil: true }
validates :access_level, inclusion: { in: Gitlab::Access.all_values }, presence: true
2015-11-17 09:49:37 -05:00
validates :invite_email,
presence: {
if: :invite?
},
email: {
2015-11-17 09:49:37 -05:00
allow_nil: true
},
uniqueness: {
scope: [:source_type, :source_id],
allow_nil: true
}
# This scope encapsulates (most of) the conditions a row in the member table
# must satisfy if it is a valid permission. Of particular note:
#
# * Access requests must be excluded
# * Blocked users must be excluded
# * Invitations take effect immediately
# * expires_at is not implemented. A background worker purges expired rows
scope :active, -> do
is_external_invite = arel_table[:user_id].eq(nil).and(arel_table[:invite_token].not_eq(nil))
user_is_active = User.arel_table[:state].eq(:active)
includes(:user).references(:users)
.where(is_external_invite.or(user_is_active))
.where(requested_at: nil)
end
scope :invite, -> { where.not(invite_token: nil) }
scope :non_invite, -> { where(invite_token: nil) }
scope :request, -> { where.not(requested_at: nil) }
scope :non_request, -> { where(requested_at: nil) }
scope :has_access, -> { active.where('access_level > 0') }
scope :guests, -> { active.where(access_level: GUEST) }
scope :reporters, -> { active.where(access_level: REPORTER) }
scope :developers, -> { active.where(access_level: DEVELOPER) }
scope :masters, -> { active.where(access_level: MASTER) }
scope :owners, -> { active.where(access_level: OWNER) }
scope :owners_and_masters, -> { active.where(access_level: [OWNER, MASTER]) }
scope :order_name_asc, -> { left_join_users.reorder(Gitlab::Database.nulls_last_order('users.name', 'ASC')) }
scope :order_name_desc, -> { left_join_users.reorder(Gitlab::Database.nulls_last_order('users.name', 'DESC')) }
scope :order_recent_sign_in, -> { left_join_users.reorder(Gitlab::Database.nulls_last_order('users.last_sign_in_at', 'DESC')) }
scope :order_oldest_sign_in, -> { left_join_users.reorder(Gitlab::Database.nulls_last_order('users.last_sign_in_at', 'ASC')) }
2015-04-10 09:09:37 -04:00
before_validation :generate_invite_token, on: :create, if: -> (member) { member.invite_email.present? }
after_create :send_invite, if: :invite?, unless: :importing?
2016-06-16 08:07:49 -04:00
after_create :send_request, if: :request?, unless: :importing?
after_create :create_notification_setting, unless: [:pending?, :importing?]
after_create :post_create_hook, unless: [:pending?, :importing?]
after_update :post_update_hook, unless: [:pending?, :importing?]
after_destroy :post_destroy_hook, unless: :pending?
Fix race conditions for AuthorizedProjectsWorker There were two cases that could be problematic: 1. Because sometimes AuthorizedProjectsWorker would be scheduled in a transaction it was possible for a job to run/complete before a COMMIT; resulting in it either producing an error, or producing no new data. 2. When scheduling jobs the code would not wait until completion. This could lead to a user creating a project and then immediately trying to push to it. Usually this will work fine, but given enough load it might take a few seconds before a user has access. The first one is problematic, the second one is mostly just annoying (but annoying enough to warrant a solution). This commit changes two things to deal with this: 1. Sidekiq scheduling now takes places after a COMMIT, this is ensured by scheduling using Rails' after_commit hook instead of doing so in an arbitrary method. 2. When scheduling jobs the calling thread now waits for all jobs to complete. Solution 2 requires tracking of job completions. Sidekiq provides a way to find a job by its ID, but this involves scanning over the entire queue; something that is very in-efficient for large queues. As such a more efficient solution is necessary. There are two main Gems that can do this in a more efficient manner: * sidekiq-status * sidekiq_status No, this is not a joke. Both Gems do a similar thing (but slightly different), and the only difference in their name is a dash vs an underscore. Both Gems however provide far more than just checking if a job has been completed, and both have their problems. sidekiq-status does not appear to be actively maintained, with the last release being in 2015. It also has some issues during testing as API calls are not stubbed in any way. sidekiq_status on the other hand does not appear to be very popular, and introduces a similar amount of code. Because of this I opted to write a simple home grown solution. After all, all we need is storing a job ID somewhere so we can efficiently look it up; we don't need extra web UIs (as provided by sidekiq-status) or complex APIs to update progress, etc. This is where Gitlab::SidekiqStatus comes in handy. This namespace contains some code used for tracking, removing, and looking up job IDs; all without having to scan over an entire queue. Data is removed explicitly, but also expires automatically just in case. Using this API we can now schedule jobs in a fork-join like manner: we schedule the jobs in Sidekiq, process them in parallel, then wait for completion. By using Sidekiq we can leverage all the benefits such as being able to scale across multiple cores and hosts, retrying failed jobs, etc. The one downside is that we need to make sure we can deal with unexpected increases in job processing timings. To deal with this the class Gitlab::JobWaiter (used for waiting for jobs to complete) will only wait a number of seconds (30 by default). Once this timeout is reached it will simply return. For GitLab.com almost all AuthorizedProjectWorker jobs complete in seconds, only very rarely do we spike to job timings of around a minute. These in turn seem to be the result of external factors (e.g. deploys), in which case a user is most likely not able to use the system anyway. In short, this new solution should ensure that jobs are processed properly and that in almost all cases a user has access to their resources whenever they need to have access.
2017-01-22 12:22:02 -05:00
after_commit :refresh_member_authorized_projects
2015-04-10 09:09:37 -04:00
default_value_for :notification_level, NotificationSetting.levels[:global]
class << self
def search(query)
joins(:user).merge(User.search(query))
end
def sort(method)
case method.to_s
when 'access_level_asc' then reorder(access_level: :asc)
when 'access_level_desc' then reorder(access_level: :desc)
when 'recent_sign_in' then order_recent_sign_in
when 'oldest_sign_in' then order_oldest_sign_in
when 'last_joined' then order_created_desc
when 'oldest_joined' then order_created_asc
else
order_by(method)
end
end
def left_join_users
users = User.arel_table
members = Member.arel_table
member_users = members.join(users, Arel::Nodes::OuterJoin).
on(members[:user_id].eq(users[:id])).
join_sources
joins(member_users)
end
2016-07-25 09:21:55 -04:00
def access_for_user_ids(user_ids)
2016-08-04 01:59:14 -04:00
where(user_id: user_ids).has_access.pluck(:user_id, :access_level).to_h
2016-07-25 09:21:55 -04:00
end
def find_by_invite_token(invite_token)
invite_token = Devise.token_generator.digest(self, :invite_token, invite_token)
find_by(invite_token: invite_token)
end
def add_user(source, user, access_level, current_user: nil, expires_at: nil)
user = retrieve_user(user)
access_level = retrieve_access_level(access_level)
# `user` can be either a User object or an email to be invited
member =
if user.is_a?(User)
source.members.find_by(user_id: user.id) ||
source.requesters.find_by(user_id: user.id) ||
source.members.build(user_id: user.id)
else
source.members.build(invite_email: user)
end
return member unless can_update_member?(current_user, member)
member.attributes = {
created_by: member.created_by || current_user,
access_level: access_level,
expires_at: expires_at
}
if member.request?
::Members::ApproveAccessRequestService.new(
source,
current_user,
id: member.id,
access_level: access_level
).execute
else
member.save
end
member
end
def access_levels
Gitlab::Access.sym_options
end
private
# This method is used to find users that have been entered into the "Add members" field.
# These can be the User objects directly, their IDs, their emails, or new emails to be invited.
def retrieve_user(user)
return user if user.is_a?(User)
User.find_by(id: user) || User.find_by(email: user) || user
end
def retrieve_access_level(access_level)
access_levels.fetch(access_level) { access_level.to_i }
end
def can_update_member?(current_user, member)
2015-11-17 09:49:37 -05:00
# There is no current user for bulk actions, in which case anything is allowed
!current_user || current_user.can?(:"update_#{member.type.underscore}", member)
end
def add_users_to_source(source, users, access_level, current_user: nil, expires_at: nil)
users.each do |user|
add_user(
source,
user,
access_level,
current_user: current_user,
expires_at: expires_at
)
end
end
2015-04-10 09:22:31 -04:00
end
def real_source_type
source_type
end
2015-04-10 09:09:37 -04:00
def invite?
self.invite_token.present?
end
def request?
requested_at.present?
end
def pending?
invite? || request?
2015-04-10 09:09:37 -04:00
end
def accept_request
return false unless request?
updated = self.update(requested_at: nil)
after_accept_request if updated
updated
end
2015-04-10 09:09:37 -04:00
def accept_invite!(new_user)
2015-04-10 09:22:31 -04:00
return false unless invite?
2015-04-10 09:09:37 -04:00
self.invite_token = nil
self.invite_accepted_at = Time.now.utc
self.user = new_user
saved = self.save
after_accept_invite if saved
saved
end
2015-04-10 10:37:02 -04:00
def decline_invite!
return false unless invite?
destroyed = self.destroy
after_decline_invite if destroyed
destroyed
end
2015-04-10 09:09:37 -04:00
def generate_invite_token
raw, enc = Devise.token_generator.generate(self.class, :invite_token)
@raw_invite_token = raw
self.invite_token = enc
end
def generate_invite_token!
generate_invite_token && save(validate: false)
end
def resend_invite
return unless invite?
generate_invite_token! unless @raw_invite_token
send_invite
end
def create_notification_setting
user.notification_settings.find_or_create_for(source)
end
def notification_setting
@notification_setting ||= user.notification_settings_for(source)
end
2015-04-10 09:09:37 -04:00
private
def send_invite
# override in subclass
end
def send_request
notification_service.new_access_request(self)
2015-04-10 09:09:37 -04:00
end
def post_create_hook
system_hook_service.execute_hooks_for(self, :create)
end
def post_update_hook
Fix race conditions for AuthorizedProjectsWorker There were two cases that could be problematic: 1. Because sometimes AuthorizedProjectsWorker would be scheduled in a transaction it was possible for a job to run/complete before a COMMIT; resulting in it either producing an error, or producing no new data. 2. When scheduling jobs the code would not wait until completion. This could lead to a user creating a project and then immediately trying to push to it. Usually this will work fine, but given enough load it might take a few seconds before a user has access. The first one is problematic, the second one is mostly just annoying (but annoying enough to warrant a solution). This commit changes two things to deal with this: 1. Sidekiq scheduling now takes places after a COMMIT, this is ensured by scheduling using Rails' after_commit hook instead of doing so in an arbitrary method. 2. When scheduling jobs the calling thread now waits for all jobs to complete. Solution 2 requires tracking of job completions. Sidekiq provides a way to find a job by its ID, but this involves scanning over the entire queue; something that is very in-efficient for large queues. As such a more efficient solution is necessary. There are two main Gems that can do this in a more efficient manner: * sidekiq-status * sidekiq_status No, this is not a joke. Both Gems do a similar thing (but slightly different), and the only difference in their name is a dash vs an underscore. Both Gems however provide far more than just checking if a job has been completed, and both have their problems. sidekiq-status does not appear to be actively maintained, with the last release being in 2015. It also has some issues during testing as API calls are not stubbed in any way. sidekiq_status on the other hand does not appear to be very popular, and introduces a similar amount of code. Because of this I opted to write a simple home grown solution. After all, all we need is storing a job ID somewhere so we can efficiently look it up; we don't need extra web UIs (as provided by sidekiq-status) or complex APIs to update progress, etc. This is where Gitlab::SidekiqStatus comes in handy. This namespace contains some code used for tracking, removing, and looking up job IDs; all without having to scan over an entire queue. Data is removed explicitly, but also expires automatically just in case. Using this API we can now schedule jobs in a fork-join like manner: we schedule the jobs in Sidekiq, process them in parallel, then wait for completion. By using Sidekiq we can leverage all the benefits such as being able to scale across multiple cores and hosts, retrying failed jobs, etc. The one downside is that we need to make sure we can deal with unexpected increases in job processing timings. To deal with this the class Gitlab::JobWaiter (used for waiting for jobs to complete) will only wait a number of seconds (30 by default). Once this timeout is reached it will simply return. For GitLab.com almost all AuthorizedProjectWorker jobs complete in seconds, only very rarely do we spike to job timings of around a minute. These in turn seem to be the result of external factors (e.g. deploys), in which case a user is most likely not able to use the system anyway. In short, this new solution should ensure that jobs are processed properly and that in almost all cases a user has access to their resources whenever they need to have access.
2017-01-22 12:22:02 -05:00
# override in sub class
2015-04-10 09:09:37 -04:00
end
def post_destroy_hook
system_hook_service.execute_hooks_for(self, :destroy)
end
Fix race conditions for AuthorizedProjectsWorker There were two cases that could be problematic: 1. Because sometimes AuthorizedProjectsWorker would be scheduled in a transaction it was possible for a job to run/complete before a COMMIT; resulting in it either producing an error, or producing no new data. 2. When scheduling jobs the code would not wait until completion. This could lead to a user creating a project and then immediately trying to push to it. Usually this will work fine, but given enough load it might take a few seconds before a user has access. The first one is problematic, the second one is mostly just annoying (but annoying enough to warrant a solution). This commit changes two things to deal with this: 1. Sidekiq scheduling now takes places after a COMMIT, this is ensured by scheduling using Rails' after_commit hook instead of doing so in an arbitrary method. 2. When scheduling jobs the calling thread now waits for all jobs to complete. Solution 2 requires tracking of job completions. Sidekiq provides a way to find a job by its ID, but this involves scanning over the entire queue; something that is very in-efficient for large queues. As such a more efficient solution is necessary. There are two main Gems that can do this in a more efficient manner: * sidekiq-status * sidekiq_status No, this is not a joke. Both Gems do a similar thing (but slightly different), and the only difference in their name is a dash vs an underscore. Both Gems however provide far more than just checking if a job has been completed, and both have their problems. sidekiq-status does not appear to be actively maintained, with the last release being in 2015. It also has some issues during testing as API calls are not stubbed in any way. sidekiq_status on the other hand does not appear to be very popular, and introduces a similar amount of code. Because of this I opted to write a simple home grown solution. After all, all we need is storing a job ID somewhere so we can efficiently look it up; we don't need extra web UIs (as provided by sidekiq-status) or complex APIs to update progress, etc. This is where Gitlab::SidekiqStatus comes in handy. This namespace contains some code used for tracking, removing, and looking up job IDs; all without having to scan over an entire queue. Data is removed explicitly, but also expires automatically just in case. Using this API we can now schedule jobs in a fork-join like manner: we schedule the jobs in Sidekiq, process them in parallel, then wait for completion. By using Sidekiq we can leverage all the benefits such as being able to scale across multiple cores and hosts, retrying failed jobs, etc. The one downside is that we need to make sure we can deal with unexpected increases in job processing timings. To deal with this the class Gitlab::JobWaiter (used for waiting for jobs to complete) will only wait a number of seconds (30 by default). Once this timeout is reached it will simply return. For GitLab.com almost all AuthorizedProjectWorker jobs complete in seconds, only very rarely do we spike to job timings of around a minute. These in turn seem to be the result of external factors (e.g. deploys), in which case a user is most likely not able to use the system anyway. In short, this new solution should ensure that jobs are processed properly and that in almost all cases a user has access to their resources whenever they need to have access.
2017-01-22 12:22:02 -05:00
# Refreshes authorizations of the current member.
#
# This method schedules a job using Sidekiq and as such **must not** be called
# in a transaction. Doing so can lead to the job running before the
# transaction has been committed, resulting in the job either throwing an
# error or not doing any meaningful work.
def refresh_member_authorized_projects
Fix race conditions for AuthorizedProjectsWorker There were two cases that could be problematic: 1. Because sometimes AuthorizedProjectsWorker would be scheduled in a transaction it was possible for a job to run/complete before a COMMIT; resulting in it either producing an error, or producing no new data. 2. When scheduling jobs the code would not wait until completion. This could lead to a user creating a project and then immediately trying to push to it. Usually this will work fine, but given enough load it might take a few seconds before a user has access. The first one is problematic, the second one is mostly just annoying (but annoying enough to warrant a solution). This commit changes two things to deal with this: 1. Sidekiq scheduling now takes places after a COMMIT, this is ensured by scheduling using Rails' after_commit hook instead of doing so in an arbitrary method. 2. When scheduling jobs the calling thread now waits for all jobs to complete. Solution 2 requires tracking of job completions. Sidekiq provides a way to find a job by its ID, but this involves scanning over the entire queue; something that is very in-efficient for large queues. As such a more efficient solution is necessary. There are two main Gems that can do this in a more efficient manner: * sidekiq-status * sidekiq_status No, this is not a joke. Both Gems do a similar thing (but slightly different), and the only difference in their name is a dash vs an underscore. Both Gems however provide far more than just checking if a job has been completed, and both have their problems. sidekiq-status does not appear to be actively maintained, with the last release being in 2015. It also has some issues during testing as API calls are not stubbed in any way. sidekiq_status on the other hand does not appear to be very popular, and introduces a similar amount of code. Because of this I opted to write a simple home grown solution. After all, all we need is storing a job ID somewhere so we can efficiently look it up; we don't need extra web UIs (as provided by sidekiq-status) or complex APIs to update progress, etc. This is where Gitlab::SidekiqStatus comes in handy. This namespace contains some code used for tracking, removing, and looking up job IDs; all without having to scan over an entire queue. Data is removed explicitly, but also expires automatically just in case. Using this API we can now schedule jobs in a fork-join like manner: we schedule the jobs in Sidekiq, process them in parallel, then wait for completion. By using Sidekiq we can leverage all the benefits such as being able to scale across multiple cores and hosts, retrying failed jobs, etc. The one downside is that we need to make sure we can deal with unexpected increases in job processing timings. To deal with this the class Gitlab::JobWaiter (used for waiting for jobs to complete) will only wait a number of seconds (30 by default). Once this timeout is reached it will simply return. For GitLab.com almost all AuthorizedProjectWorker jobs complete in seconds, only very rarely do we spike to job timings of around a minute. These in turn seem to be the result of external factors (e.g. deploys), in which case a user is most likely not able to use the system anyway. In short, this new solution should ensure that jobs are processed properly and that in almost all cases a user has access to their resources whenever they need to have access.
2017-01-22 12:22:02 -05:00
# If user/source is being destroyed, project access are going to be
# destroyed eventually because of DB foreign keys, so we shouldn't bother
# with refreshing after each member is destroyed through association
return if destroyed_by_association.present?
UserProjectAccessChangedService.new(user_id).execute
end
2015-04-10 09:09:37 -04:00
def after_accept_invite
post_create_hook
end
2015-04-10 10:37:02 -04:00
def after_decline_invite
# override in subclass
end
def after_accept_request
2015-04-10 09:09:37 -04:00
post_create_hook
end
def system_hook_service
SystemHooksService.new
end
def notification_service
NotificationService.new
end
end