gitlab-org--gitlab-foss/spec/models/project_team_spec.rb

475 lines
16 KiB
Ruby
Raw Normal View History

# frozen_string_literal: true
2013-01-04 21:35:31 +00:00
require "spec_helper"
describe ProjectTeam do
let(:maintainer) { create(:user) }
let(:reporter) { create(:user) }
let(:guest) { create(:user) }
let(:nonmember) { create(:user) }
2013-01-04 21:35:31 +00:00
context 'personal project' do
let(:project) { create(:project) }
before do
project.add_maintainer(maintainer)
project.add_reporter(reporter)
project.add_guest(guest)
end
describe 'members collection' do
it { expect(project.team.maintainers).to include(maintainer) }
it { expect(project.team.maintainers).not_to include(guest) }
it { expect(project.team.maintainers).not_to include(reporter) }
it { expect(project.team.maintainers).not_to include(nonmember) }
end
describe 'access methods' do
it { expect(project.team.maintainer?(maintainer)).to be_truthy }
it { expect(project.team.maintainer?(guest)).to be_falsey }
it { expect(project.team.maintainer?(reporter)).to be_falsey }
it { expect(project.team.maintainer?(nonmember)).to be_falsey }
it { expect(project.team.member?(nonmember)).to be_falsey }
it { expect(project.team.member?(guest)).to be_truthy }
it { expect(project.team.member?(reporter, Gitlab::Access::REPORTER)).to be_truthy }
it { expect(project.team.member?(guest, Gitlab::Access::REPORTER)).to be_falsey }
it { expect(project.team.member?(nonmember, Gitlab::Access::GUEST)).to be_falsey }
end
end
context 'group project' do
let(:group) { create(:group) }
let!(:project) { create(:project, group: group) }
before do
group.add_maintainer(maintainer)
group.add_reporter(reporter)
group.add_guest(guest)
# If user is a group and a project member - GitLab uses highest permission
# So we add group guest as maintainer and add group maintainer as guest
# to this project to test highest access
project.add_maintainer(guest)
project.add_guest(maintainer)
end
describe 'members collection' do
it { expect(project.team.reporters).to include(reporter) }
it { expect(project.team.maintainers).to include(maintainer) }
it { expect(project.team.maintainers).to include(guest) }
it { expect(project.team.maintainers).not_to include(reporter) }
it { expect(project.team.maintainers).not_to include(nonmember) }
end
describe 'access methods' do
it { expect(project.team.reporter?(reporter)).to be_truthy }
it { expect(project.team.maintainer?(maintainer)).to be_truthy }
it { expect(project.team.maintainer?(guest)).to be_truthy }
it { expect(project.team.maintainer?(reporter)).to be_falsey }
it { expect(project.team.maintainer?(nonmember)).to be_falsey }
it { expect(project.team.member?(nonmember)).to be_falsey }
it { expect(project.team.member?(guest)).to be_truthy }
it { expect(project.team.member?(guest, Gitlab::Access::MAINTAINER)).to be_truthy }
it { expect(project.team.member?(reporter, Gitlab::Access::MAINTAINER)).to be_falsey }
it { expect(project.team.member?(nonmember, Gitlab::Access::GUEST)).to be_falsey }
end
end
describe '#fetch_members' do
context 'personal project' do
let(:project) { create(:project) }
it 'returns project members' do
user = create(:user)
project.add_guest(user)
Use CTEs for nested groups and authorizations This commit introduces the usage of Common Table Expressions (CTEs) to efficiently retrieve nested group hierarchies, without having to rely on the "routes" table (which is an _incredibly_ inefficient way of getting the data). This requires a patch to ActiveRecord (found in the added initializer) to work properly as ActiveRecord doesn't support WITH statements properly out of the box. Unfortunately MySQL provides no efficient way of getting nested groups. For example, the old routes setup could easily take 5-10 seconds depending on the amount of "routes" in a database. Providing vastly different logic for both MySQL and PostgreSQL will negatively impact the development process. Because of this the various nested groups related methods return empty relations when used in combination with MySQL. For project authorizations the logic is split up into two classes: * Gitlab::ProjectAuthorizations::WithNestedGroups * Gitlab::ProjectAuthorizations::WithoutNestedGroups Both classes get the fresh project authorizations (= as they should be in the "project_authorizations" table), including nested groups if PostgreSQL is used. The logic of these two classes is quite different apart from their public interface. This complicates development a bit, but unfortunately there is no way around this. This commit also introduces Gitlab::GroupHierarchy. This class can be used to get the ancestors and descendants of a base relation, or both by using a UNION. This in turn is used by methods such as: * Namespace#ancestors * Namespace#descendants * User#all_expanded_groups Again this class relies on CTEs and thus only works on PostgreSQL. The Namespace methods will return an empty relation when MySQL is used, while User#all_expanded_groups will return only the groups a user is a direct member of. Performance wise the impact is quite large. For example, on GitLab.com Namespace#descendants used to take around 580 ms to retrieve data for a particular user. Using CTEs we are able to reduce this down to roughly 1 millisecond, returning the exact same data. == On The Fly Refreshing Refreshing of authorizations on the fly (= when users.authorized_projects_populated was not set) is removed with this commit. This simplifies the code, and ensures any queries used for authorizations are not mutated because they are executed in a Rails scope (e.g. Project.visible_to_user). This commit includes a migration to schedule refreshing authorizations for all users, ensuring all of them have their authorizations in place. Said migration schedules users in batches of 5000, with 5 minutes between every batch to smear the load around a bit. == Spec Changes This commit also introduces some changes to various specs. For example, some specs for ProjectTeam assumed that creating a personal project would _not_ lead to the owner having access, which is incorrect. Because we also no longer refresh authorizations on the fly for new users some code had to be added to the "empty_project" factory. This chunk of code ensures that the owner's permissions are refreshed after creating the project, something that is normally done in Projects::CreateService.
2017-04-24 15:19:22 +00:00
expect(project.team.members).to contain_exactly(user, project.owner)
end
it 'returns project members of a specified level' do
user = create(:user)
project.add_reporter(user)
expect(project.team.guests).to be_empty
expect(project.team.reporters).to contain_exactly(user)
end
it 'returns invited members of a group' do
group_member = create(:group_member)
project.project_group_links.create!(
group: group_member.group,
group_access: Gitlab::Access::GUEST
)
2017-06-21 13:48:12 +00:00
expect(project.team.members)
.to contain_exactly(group_member.user, project.owner)
end
it 'returns invited members of a group of a specified level' do
group_member = create(:group_member)
project.project_group_links.create!(
group: group_member.group,
group_access: Gitlab::Access::REPORTER
)
expect(project.team.guests).to be_empty
expect(project.team.reporters).to contain_exactly(group_member.user)
end
end
context 'group project' do
let(:group) { create(:group) }
let!(:project) { create(:project, group: group) }
it 'returns project members' do
group_member = create(:group_member, group: group)
expect(project.team.members).to contain_exactly(group_member.user)
end
it 'returns project members of a specified level' do
group_member = create(:group_member, :reporter, group: group)
expect(project.team.guests).to be_empty
expect(project.team.reporters).to contain_exactly(group_member.user)
end
end
end
describe '#find_member' do
context 'personal project' do
Use CTEs for nested groups and authorizations This commit introduces the usage of Common Table Expressions (CTEs) to efficiently retrieve nested group hierarchies, without having to rely on the "routes" table (which is an _incredibly_ inefficient way of getting the data). This requires a patch to ActiveRecord (found in the added initializer) to work properly as ActiveRecord doesn't support WITH statements properly out of the box. Unfortunately MySQL provides no efficient way of getting nested groups. For example, the old routes setup could easily take 5-10 seconds depending on the amount of "routes" in a database. Providing vastly different logic for both MySQL and PostgreSQL will negatively impact the development process. Because of this the various nested groups related methods return empty relations when used in combination with MySQL. For project authorizations the logic is split up into two classes: * Gitlab::ProjectAuthorizations::WithNestedGroups * Gitlab::ProjectAuthorizations::WithoutNestedGroups Both classes get the fresh project authorizations (= as they should be in the "project_authorizations" table), including nested groups if PostgreSQL is used. The logic of these two classes is quite different apart from their public interface. This complicates development a bit, but unfortunately there is no way around this. This commit also introduces Gitlab::GroupHierarchy. This class can be used to get the ancestors and descendants of a base relation, or both by using a UNION. This in turn is used by methods such as: * Namespace#ancestors * Namespace#descendants * User#all_expanded_groups Again this class relies on CTEs and thus only works on PostgreSQL. The Namespace methods will return an empty relation when MySQL is used, while User#all_expanded_groups will return only the groups a user is a direct member of. Performance wise the impact is quite large. For example, on GitLab.com Namespace#descendants used to take around 580 ms to retrieve data for a particular user. Using CTEs we are able to reduce this down to roughly 1 millisecond, returning the exact same data. == On The Fly Refreshing Refreshing of authorizations on the fly (= when users.authorized_projects_populated was not set) is removed with this commit. This simplifies the code, and ensures any queries used for authorizations are not mutated because they are executed in a Rails scope (e.g. Project.visible_to_user). This commit includes a migration to schedule refreshing authorizations for all users, ensuring all of them have their authorizations in place. Said migration schedules users in batches of 5000, with 5 minutes between every batch to smear the load around a bit. == Spec Changes This commit also introduces some changes to various specs. For example, some specs for ProjectTeam assumed that creating a personal project would _not_ lead to the owner having access, which is incorrect. Because we also no longer refresh authorizations on the fly for new users some code had to be added to the "empty_project" factory. This chunk of code ensures that the owner's permissions are refreshed after creating the project, something that is normally done in Projects::CreateService.
2017-04-24 15:19:22 +00:00
let(:project) do
create(:project, :public, :access_requestable)
Use CTEs for nested groups and authorizations This commit introduces the usage of Common Table Expressions (CTEs) to efficiently retrieve nested group hierarchies, without having to rely on the "routes" table (which is an _incredibly_ inefficient way of getting the data). This requires a patch to ActiveRecord (found in the added initializer) to work properly as ActiveRecord doesn't support WITH statements properly out of the box. Unfortunately MySQL provides no efficient way of getting nested groups. For example, the old routes setup could easily take 5-10 seconds depending on the amount of "routes" in a database. Providing vastly different logic for both MySQL and PostgreSQL will negatively impact the development process. Because of this the various nested groups related methods return empty relations when used in combination with MySQL. For project authorizations the logic is split up into two classes: * Gitlab::ProjectAuthorizations::WithNestedGroups * Gitlab::ProjectAuthorizations::WithoutNestedGroups Both classes get the fresh project authorizations (= as they should be in the "project_authorizations" table), including nested groups if PostgreSQL is used. The logic of these two classes is quite different apart from their public interface. This complicates development a bit, but unfortunately there is no way around this. This commit also introduces Gitlab::GroupHierarchy. This class can be used to get the ancestors and descendants of a base relation, or both by using a UNION. This in turn is used by methods such as: * Namespace#ancestors * Namespace#descendants * User#all_expanded_groups Again this class relies on CTEs and thus only works on PostgreSQL. The Namespace methods will return an empty relation when MySQL is used, while User#all_expanded_groups will return only the groups a user is a direct member of. Performance wise the impact is quite large. For example, on GitLab.com Namespace#descendants used to take around 580 ms to retrieve data for a particular user. Using CTEs we are able to reduce this down to roughly 1 millisecond, returning the exact same data. == On The Fly Refreshing Refreshing of authorizations on the fly (= when users.authorized_projects_populated was not set) is removed with this commit. This simplifies the code, and ensures any queries used for authorizations are not mutated because they are executed in a Rails scope (e.g. Project.visible_to_user). This commit includes a migration to schedule refreshing authorizations for all users, ensuring all of them have their authorizations in place. Said migration schedules users in batches of 5000, with 5 minutes between every batch to smear the load around a bit. == Spec Changes This commit also introduces some changes to various specs. For example, some specs for ProjectTeam assumed that creating a personal project would _not_ lead to the owner having access, which is incorrect. Because we also no longer refresh authorizations on the fly for new users some code had to be added to the "empty_project" factory. This chunk of code ensures that the owner's permissions are refreshed after creating the project, something that is normally done in Projects::CreateService.
2017-04-24 15:19:22 +00:00
end
let(:requester) { create(:user) }
before do
project.add_maintainer(maintainer)
project.add_reporter(reporter)
project.add_guest(guest)
project.request_access(requester)
end
it { expect(project.team.find_member(maintainer.id)).to be_a(ProjectMember) }
it { expect(project.team.find_member(reporter.id)).to be_a(ProjectMember) }
it { expect(project.team.find_member(guest.id)).to be_a(ProjectMember) }
it { expect(project.team.find_member(nonmember.id)).to be_nil }
it { expect(project.team.find_member(requester.id)).to be_nil }
end
context 'group project' do
let(:group) { create(:group, :access_requestable) }
let(:project) { create(:project, group: group) }
let(:requester) { create(:user) }
before do
group.add_maintainer(maintainer)
group.add_reporter(reporter)
group.add_guest(guest)
group.request_access(requester)
end
it { expect(project.team.find_member(maintainer.id)).to be_a(GroupMember) }
it { expect(project.team.find_member(reporter.id)).to be_a(GroupMember) }
it { expect(project.team.find_member(guest.id)).to be_a(GroupMember) }
it { expect(project.team.find_member(nonmember.id)).to be_nil }
it { expect(project.team.find_member(requester.id)).to be_nil }
end
end
describe '#members_in_project_and_ancestors' do
context 'group project' do
it 'filters out users who are not members of the project' do
group = create(:group)
project = create(:project, group: group)
group_member = create(:group_member, group: group)
old_user = create(:user)
ProjectAuthorization.create!(project: project, user: old_user, access_level: Gitlab::Access::GUEST)
expect(project.team.members_in_project_and_ancestors).to contain_exactly(group_member.user)
end
end
end
describe '#add_users' do
let(:user1) { create(:user) }
let(:user2) { create(:user) }
let(:project) { create(:project) }
it 'add the given users to the team' do
project.team.add_users([user1, user2], :reporter)
expect(project.team.reporter?(user1)).to be(true)
expect(project.team.reporter?(user2)).to be(true)
end
end
describe '#add_user' do
let(:user) { create(:user) }
let(:project) { create(:project) }
it 'add the given user to the team' do
project.team.add_user(user, :reporter)
expect(project.team.reporter?(user)).to be(true)
end
end
describe "#human_max_access" do
2018-05-22 11:55:07 +00:00
it 'returns Maintainer role' do
user = create(:user)
group = create(:group)
project = create(:project, namespace: group)
group.add_maintainer(user)
2018-05-22 11:55:07 +00:00
expect(project.team.human_max_access(user.id)).to eq 'Maintainer'
end
it 'returns Owner role' do
user = create(:user)
group = create(:group)
project = create(:project, namespace: group)
group.add_owner(user)
expect(project.team.human_max_access(user.id)).to eq 'Owner'
end
end
describe '#max_member_access' do
let(:requester) { create(:user) }
context 'personal project' do
Use CTEs for nested groups and authorizations This commit introduces the usage of Common Table Expressions (CTEs) to efficiently retrieve nested group hierarchies, without having to rely on the "routes" table (which is an _incredibly_ inefficient way of getting the data). This requires a patch to ActiveRecord (found in the added initializer) to work properly as ActiveRecord doesn't support WITH statements properly out of the box. Unfortunately MySQL provides no efficient way of getting nested groups. For example, the old routes setup could easily take 5-10 seconds depending on the amount of "routes" in a database. Providing vastly different logic for both MySQL and PostgreSQL will negatively impact the development process. Because of this the various nested groups related methods return empty relations when used in combination with MySQL. For project authorizations the logic is split up into two classes: * Gitlab::ProjectAuthorizations::WithNestedGroups * Gitlab::ProjectAuthorizations::WithoutNestedGroups Both classes get the fresh project authorizations (= as they should be in the "project_authorizations" table), including nested groups if PostgreSQL is used. The logic of these two classes is quite different apart from their public interface. This complicates development a bit, but unfortunately there is no way around this. This commit also introduces Gitlab::GroupHierarchy. This class can be used to get the ancestors and descendants of a base relation, or both by using a UNION. This in turn is used by methods such as: * Namespace#ancestors * Namespace#descendants * User#all_expanded_groups Again this class relies on CTEs and thus only works on PostgreSQL. The Namespace methods will return an empty relation when MySQL is used, while User#all_expanded_groups will return only the groups a user is a direct member of. Performance wise the impact is quite large. For example, on GitLab.com Namespace#descendants used to take around 580 ms to retrieve data for a particular user. Using CTEs we are able to reduce this down to roughly 1 millisecond, returning the exact same data. == On The Fly Refreshing Refreshing of authorizations on the fly (= when users.authorized_projects_populated was not set) is removed with this commit. This simplifies the code, and ensures any queries used for authorizations are not mutated because they are executed in a Rails scope (e.g. Project.visible_to_user). This commit includes a migration to schedule refreshing authorizations for all users, ensuring all of them have their authorizations in place. Said migration schedules users in batches of 5000, with 5 minutes between every batch to smear the load around a bit. == Spec Changes This commit also introduces some changes to various specs. For example, some specs for ProjectTeam assumed that creating a personal project would _not_ lead to the owner having access, which is incorrect. Because we also no longer refresh authorizations on the fly for new users some code had to be added to the "empty_project" factory. This chunk of code ensures that the owner's permissions are refreshed after creating the project, something that is normally done in Projects::CreateService.
2017-04-24 15:19:22 +00:00
let(:project) do
create(:project, :public, :access_requestable)
Use CTEs for nested groups and authorizations This commit introduces the usage of Common Table Expressions (CTEs) to efficiently retrieve nested group hierarchies, without having to rely on the "routes" table (which is an _incredibly_ inefficient way of getting the data). This requires a patch to ActiveRecord (found in the added initializer) to work properly as ActiveRecord doesn't support WITH statements properly out of the box. Unfortunately MySQL provides no efficient way of getting nested groups. For example, the old routes setup could easily take 5-10 seconds depending on the amount of "routes" in a database. Providing vastly different logic for both MySQL and PostgreSQL will negatively impact the development process. Because of this the various nested groups related methods return empty relations when used in combination with MySQL. For project authorizations the logic is split up into two classes: * Gitlab::ProjectAuthorizations::WithNestedGroups * Gitlab::ProjectAuthorizations::WithoutNestedGroups Both classes get the fresh project authorizations (= as they should be in the "project_authorizations" table), including nested groups if PostgreSQL is used. The logic of these two classes is quite different apart from their public interface. This complicates development a bit, but unfortunately there is no way around this. This commit also introduces Gitlab::GroupHierarchy. This class can be used to get the ancestors and descendants of a base relation, or both by using a UNION. This in turn is used by methods such as: * Namespace#ancestors * Namespace#descendants * User#all_expanded_groups Again this class relies on CTEs and thus only works on PostgreSQL. The Namespace methods will return an empty relation when MySQL is used, while User#all_expanded_groups will return only the groups a user is a direct member of. Performance wise the impact is quite large. For example, on GitLab.com Namespace#descendants used to take around 580 ms to retrieve data for a particular user. Using CTEs we are able to reduce this down to roughly 1 millisecond, returning the exact same data. == On The Fly Refreshing Refreshing of authorizations on the fly (= when users.authorized_projects_populated was not set) is removed with this commit. This simplifies the code, and ensures any queries used for authorizations are not mutated because they are executed in a Rails scope (e.g. Project.visible_to_user). This commit includes a migration to schedule refreshing authorizations for all users, ensuring all of them have their authorizations in place. Said migration schedules users in batches of 5000, with 5 minutes between every batch to smear the load around a bit. == Spec Changes This commit also introduces some changes to various specs. For example, some specs for ProjectTeam assumed that creating a personal project would _not_ lead to the owner having access, which is incorrect. Because we also no longer refresh authorizations on the fly for new users some code had to be added to the "empty_project" factory. This chunk of code ensures that the owner's permissions are refreshed after creating the project, something that is normally done in Projects::CreateService.
2017-04-24 15:19:22 +00:00
end
context 'when project is not shared with group' do
before do
project.add_maintainer(maintainer)
project.add_reporter(reporter)
project.add_guest(guest)
project.request_access(requester)
end
it { expect(project.team.max_member_access(maintainer.id)).to eq(Gitlab::Access::MAINTAINER) }
it { expect(project.team.max_member_access(reporter.id)).to eq(Gitlab::Access::REPORTER) }
it { expect(project.team.max_member_access(guest.id)).to eq(Gitlab::Access::GUEST) }
it { expect(project.team.max_member_access(nonmember.id)).to eq(Gitlab::Access::NO_ACCESS) }
it { expect(project.team.max_member_access(requester.id)).to eq(Gitlab::Access::NO_ACCESS) }
end
context 'when project is shared with group' do
before do
group = create(:group)
project.project_group_links.create(
group: group,
group_access: Gitlab::Access::DEVELOPER)
group.add_maintainer(maintainer)
group.add_reporter(reporter)
end
it { expect(project.team.max_member_access(maintainer.id)).to eq(Gitlab::Access::DEVELOPER) }
it { expect(project.team.max_member_access(reporter.id)).to eq(Gitlab::Access::REPORTER) }
it { expect(project.team.max_member_access(nonmember.id)).to eq(Gitlab::Access::NO_ACCESS) }
it { expect(project.team.max_member_access(requester.id)).to eq(Gitlab::Access::NO_ACCESS) }
context 'but share_with_group_lock is true' do
before do
project.namespace.update(share_with_group_lock: true)
end
it { expect(project.team.max_member_access(maintainer.id)).to eq(Gitlab::Access::NO_ACCESS) }
it { expect(project.team.max_member_access(reporter.id)).to eq(Gitlab::Access::NO_ACCESS) }
end
end
end
context 'group project' do
let(:group) { create(:group, :access_requestable) }
Use CTEs for nested groups and authorizations This commit introduces the usage of Common Table Expressions (CTEs) to efficiently retrieve nested group hierarchies, without having to rely on the "routes" table (which is an _incredibly_ inefficient way of getting the data). This requires a patch to ActiveRecord (found in the added initializer) to work properly as ActiveRecord doesn't support WITH statements properly out of the box. Unfortunately MySQL provides no efficient way of getting nested groups. For example, the old routes setup could easily take 5-10 seconds depending on the amount of "routes" in a database. Providing vastly different logic for both MySQL and PostgreSQL will negatively impact the development process. Because of this the various nested groups related methods return empty relations when used in combination with MySQL. For project authorizations the logic is split up into two classes: * Gitlab::ProjectAuthorizations::WithNestedGroups * Gitlab::ProjectAuthorizations::WithoutNestedGroups Both classes get the fresh project authorizations (= as they should be in the "project_authorizations" table), including nested groups if PostgreSQL is used. The logic of these two classes is quite different apart from their public interface. This complicates development a bit, but unfortunately there is no way around this. This commit also introduces Gitlab::GroupHierarchy. This class can be used to get the ancestors and descendants of a base relation, or both by using a UNION. This in turn is used by methods such as: * Namespace#ancestors * Namespace#descendants * User#all_expanded_groups Again this class relies on CTEs and thus only works on PostgreSQL. The Namespace methods will return an empty relation when MySQL is used, while User#all_expanded_groups will return only the groups a user is a direct member of. Performance wise the impact is quite large. For example, on GitLab.com Namespace#descendants used to take around 580 ms to retrieve data for a particular user. Using CTEs we are able to reduce this down to roughly 1 millisecond, returning the exact same data. == On The Fly Refreshing Refreshing of authorizations on the fly (= when users.authorized_projects_populated was not set) is removed with this commit. This simplifies the code, and ensures any queries used for authorizations are not mutated because they are executed in a Rails scope (e.g. Project.visible_to_user). This commit includes a migration to schedule refreshing authorizations for all users, ensuring all of them have their authorizations in place. Said migration schedules users in batches of 5000, with 5 minutes between every batch to smear the load around a bit. == Spec Changes This commit also introduces some changes to various specs. For example, some specs for ProjectTeam assumed that creating a personal project would _not_ lead to the owner having access, which is incorrect. Because we also no longer refresh authorizations on the fly for new users some code had to be added to the "empty_project" factory. This chunk of code ensures that the owner's permissions are refreshed after creating the project, something that is normally done in Projects::CreateService.
2017-04-24 15:19:22 +00:00
let!(:project) do
create(:project, group: group)
Use CTEs for nested groups and authorizations This commit introduces the usage of Common Table Expressions (CTEs) to efficiently retrieve nested group hierarchies, without having to rely on the "routes" table (which is an _incredibly_ inefficient way of getting the data). This requires a patch to ActiveRecord (found in the added initializer) to work properly as ActiveRecord doesn't support WITH statements properly out of the box. Unfortunately MySQL provides no efficient way of getting nested groups. For example, the old routes setup could easily take 5-10 seconds depending on the amount of "routes" in a database. Providing vastly different logic for both MySQL and PostgreSQL will negatively impact the development process. Because of this the various nested groups related methods return empty relations when used in combination with MySQL. For project authorizations the logic is split up into two classes: * Gitlab::ProjectAuthorizations::WithNestedGroups * Gitlab::ProjectAuthorizations::WithoutNestedGroups Both classes get the fresh project authorizations (= as they should be in the "project_authorizations" table), including nested groups if PostgreSQL is used. The logic of these two classes is quite different apart from their public interface. This complicates development a bit, but unfortunately there is no way around this. This commit also introduces Gitlab::GroupHierarchy. This class can be used to get the ancestors and descendants of a base relation, or both by using a UNION. This in turn is used by methods such as: * Namespace#ancestors * Namespace#descendants * User#all_expanded_groups Again this class relies on CTEs and thus only works on PostgreSQL. The Namespace methods will return an empty relation when MySQL is used, while User#all_expanded_groups will return only the groups a user is a direct member of. Performance wise the impact is quite large. For example, on GitLab.com Namespace#descendants used to take around 580 ms to retrieve data for a particular user. Using CTEs we are able to reduce this down to roughly 1 millisecond, returning the exact same data. == On The Fly Refreshing Refreshing of authorizations on the fly (= when users.authorized_projects_populated was not set) is removed with this commit. This simplifies the code, and ensures any queries used for authorizations are not mutated because they are executed in a Rails scope (e.g. Project.visible_to_user). This commit includes a migration to schedule refreshing authorizations for all users, ensuring all of them have their authorizations in place. Said migration schedules users in batches of 5000, with 5 minutes between every batch to smear the load around a bit. == Spec Changes This commit also introduces some changes to various specs. For example, some specs for ProjectTeam assumed that creating a personal project would _not_ lead to the owner having access, which is incorrect. Because we also no longer refresh authorizations on the fly for new users some code had to be added to the "empty_project" factory. This chunk of code ensures that the owner's permissions are refreshed after creating the project, something that is normally done in Projects::CreateService.
2017-04-24 15:19:22 +00:00
end
before do
group.add_maintainer(maintainer)
group.add_reporter(reporter)
group.add_guest(guest)
group.request_access(requester)
end
it { expect(project.team.max_member_access(maintainer.id)).to eq(Gitlab::Access::MAINTAINER) }
it { expect(project.team.max_member_access(reporter.id)).to eq(Gitlab::Access::REPORTER) }
it { expect(project.team.max_member_access(guest.id)).to eq(Gitlab::Access::GUEST) }
it { expect(project.team.max_member_access(nonmember.id)).to eq(Gitlab::Access::NO_ACCESS) }
it { expect(project.team.max_member_access(requester.id)).to eq(Gitlab::Access::NO_ACCESS) }
end
end
describe '#member?' do
let(:group) { create(:group) }
let(:developer) { create(:user) }
let(:maintainer) { create(:user) }
Use CTEs for nested groups and authorizations This commit introduces the usage of Common Table Expressions (CTEs) to efficiently retrieve nested group hierarchies, without having to rely on the "routes" table (which is an _incredibly_ inefficient way of getting the data). This requires a patch to ActiveRecord (found in the added initializer) to work properly as ActiveRecord doesn't support WITH statements properly out of the box. Unfortunately MySQL provides no efficient way of getting nested groups. For example, the old routes setup could easily take 5-10 seconds depending on the amount of "routes" in a database. Providing vastly different logic for both MySQL and PostgreSQL will negatively impact the development process. Because of this the various nested groups related methods return empty relations when used in combination with MySQL. For project authorizations the logic is split up into two classes: * Gitlab::ProjectAuthorizations::WithNestedGroups * Gitlab::ProjectAuthorizations::WithoutNestedGroups Both classes get the fresh project authorizations (= as they should be in the "project_authorizations" table), including nested groups if PostgreSQL is used. The logic of these two classes is quite different apart from their public interface. This complicates development a bit, but unfortunately there is no way around this. This commit also introduces Gitlab::GroupHierarchy. This class can be used to get the ancestors and descendants of a base relation, or both by using a UNION. This in turn is used by methods such as: * Namespace#ancestors * Namespace#descendants * User#all_expanded_groups Again this class relies on CTEs and thus only works on PostgreSQL. The Namespace methods will return an empty relation when MySQL is used, while User#all_expanded_groups will return only the groups a user is a direct member of. Performance wise the impact is quite large. For example, on GitLab.com Namespace#descendants used to take around 580 ms to retrieve data for a particular user. Using CTEs we are able to reduce this down to roughly 1 millisecond, returning the exact same data. == On The Fly Refreshing Refreshing of authorizations on the fly (= when users.authorized_projects_populated was not set) is removed with this commit. This simplifies the code, and ensures any queries used for authorizations are not mutated because they are executed in a Rails scope (e.g. Project.visible_to_user). This commit includes a migration to schedule refreshing authorizations for all users, ensuring all of them have their authorizations in place. Said migration schedules users in batches of 5000, with 5 minutes between every batch to smear the load around a bit. == Spec Changes This commit also introduces some changes to various specs. For example, some specs for ProjectTeam assumed that creating a personal project would _not_ lead to the owner having access, which is incorrect. Because we also no longer refresh authorizations on the fly for new users some code had to be added to the "empty_project" factory. This chunk of code ensures that the owner's permissions are refreshed after creating the project, something that is normally done in Projects::CreateService.
2017-04-24 15:19:22 +00:00
let(:personal_project) do
create(:project, namespace: developer.namespace)
Use CTEs for nested groups and authorizations This commit introduces the usage of Common Table Expressions (CTEs) to efficiently retrieve nested group hierarchies, without having to rely on the "routes" table (which is an _incredibly_ inefficient way of getting the data). This requires a patch to ActiveRecord (found in the added initializer) to work properly as ActiveRecord doesn't support WITH statements properly out of the box. Unfortunately MySQL provides no efficient way of getting nested groups. For example, the old routes setup could easily take 5-10 seconds depending on the amount of "routes" in a database. Providing vastly different logic for both MySQL and PostgreSQL will negatively impact the development process. Because of this the various nested groups related methods return empty relations when used in combination with MySQL. For project authorizations the logic is split up into two classes: * Gitlab::ProjectAuthorizations::WithNestedGroups * Gitlab::ProjectAuthorizations::WithoutNestedGroups Both classes get the fresh project authorizations (= as they should be in the "project_authorizations" table), including nested groups if PostgreSQL is used. The logic of these two classes is quite different apart from their public interface. This complicates development a bit, but unfortunately there is no way around this. This commit also introduces Gitlab::GroupHierarchy. This class can be used to get the ancestors and descendants of a base relation, or both by using a UNION. This in turn is used by methods such as: * Namespace#ancestors * Namespace#descendants * User#all_expanded_groups Again this class relies on CTEs and thus only works on PostgreSQL. The Namespace methods will return an empty relation when MySQL is used, while User#all_expanded_groups will return only the groups a user is a direct member of. Performance wise the impact is quite large. For example, on GitLab.com Namespace#descendants used to take around 580 ms to retrieve data for a particular user. Using CTEs we are able to reduce this down to roughly 1 millisecond, returning the exact same data. == On The Fly Refreshing Refreshing of authorizations on the fly (= when users.authorized_projects_populated was not set) is removed with this commit. This simplifies the code, and ensures any queries used for authorizations are not mutated because they are executed in a Rails scope (e.g. Project.visible_to_user). This commit includes a migration to schedule refreshing authorizations for all users, ensuring all of them have their authorizations in place. Said migration schedules users in batches of 5000, with 5 minutes between every batch to smear the load around a bit. == Spec Changes This commit also introduces some changes to various specs. For example, some specs for ProjectTeam assumed that creating a personal project would _not_ lead to the owner having access, which is incorrect. Because we also no longer refresh authorizations on the fly for new users some code had to be added to the "empty_project" factory. This chunk of code ensures that the owner's permissions are refreshed after creating the project, something that is normally done in Projects::CreateService.
2017-04-24 15:19:22 +00:00
end
let(:group_project) do
create(:project, namespace: group)
Use CTEs for nested groups and authorizations This commit introduces the usage of Common Table Expressions (CTEs) to efficiently retrieve nested group hierarchies, without having to rely on the "routes" table (which is an _incredibly_ inefficient way of getting the data). This requires a patch to ActiveRecord (found in the added initializer) to work properly as ActiveRecord doesn't support WITH statements properly out of the box. Unfortunately MySQL provides no efficient way of getting nested groups. For example, the old routes setup could easily take 5-10 seconds depending on the amount of "routes" in a database. Providing vastly different logic for both MySQL and PostgreSQL will negatively impact the development process. Because of this the various nested groups related methods return empty relations when used in combination with MySQL. For project authorizations the logic is split up into two classes: * Gitlab::ProjectAuthorizations::WithNestedGroups * Gitlab::ProjectAuthorizations::WithoutNestedGroups Both classes get the fresh project authorizations (= as they should be in the "project_authorizations" table), including nested groups if PostgreSQL is used. The logic of these two classes is quite different apart from their public interface. This complicates development a bit, but unfortunately there is no way around this. This commit also introduces Gitlab::GroupHierarchy. This class can be used to get the ancestors and descendants of a base relation, or both by using a UNION. This in turn is used by methods such as: * Namespace#ancestors * Namespace#descendants * User#all_expanded_groups Again this class relies on CTEs and thus only works on PostgreSQL. The Namespace methods will return an empty relation when MySQL is used, while User#all_expanded_groups will return only the groups a user is a direct member of. Performance wise the impact is quite large. For example, on GitLab.com Namespace#descendants used to take around 580 ms to retrieve data for a particular user. Using CTEs we are able to reduce this down to roughly 1 millisecond, returning the exact same data. == On The Fly Refreshing Refreshing of authorizations on the fly (= when users.authorized_projects_populated was not set) is removed with this commit. This simplifies the code, and ensures any queries used for authorizations are not mutated because they are executed in a Rails scope (e.g. Project.visible_to_user). This commit includes a migration to schedule refreshing authorizations for all users, ensuring all of them have their authorizations in place. Said migration schedules users in batches of 5000, with 5 minutes between every batch to smear the load around a bit. == Spec Changes This commit also introduces some changes to various specs. For example, some specs for ProjectTeam assumed that creating a personal project would _not_ lead to the owner having access, which is incorrect. Because we also no longer refresh authorizations on the fly for new users some code had to be added to the "empty_project" factory. This chunk of code ensures that the owner's permissions are refreshed after creating the project, something that is normally done in Projects::CreateService.
2017-04-24 15:19:22 +00:00
end
let(:members_project) { create(:project) }
let(:shared_project) { create(:project) }
before do
group.add_maintainer(maintainer)
group.add_developer(developer)
members_project.add_developer(developer)
members_project.add_maintainer(maintainer)
create(:project_group_link, project: shared_project, group: group)
end
it 'returns false for no user' do
expect(personal_project.team.member?(nil)).to be(false)
end
it 'returns true for personal projects of the user' do
expect(personal_project.team.member?(developer)).to be(true)
end
it 'returns true for projects of groups the user is a member of' do
expect(group_project.team.member?(developer)).to be(true)
end
it 'returns true for projects for which the user is a member of' do
expect(members_project.team.member?(developer)).to be(true)
end
it 'returns true for projects shared on a group the user is a member of' do
expect(shared_project.team.member?(developer)).to be(true)
end
it 'checks for the correct minimum level access' do
expect(group_project.team.member?(developer, Gitlab::Access::MAINTAINER)).to be(false)
expect(group_project.team.member?(maintainer, Gitlab::Access::MAINTAINER)).to be(true)
expect(members_project.team.member?(developer, Gitlab::Access::MAINTAINER)).to be(false)
expect(members_project.team.member?(maintainer, Gitlab::Access::MAINTAINER)).to be(true)
expect(shared_project.team.member?(developer, Gitlab::Access::MAINTAINER)).to be(false)
expect(shared_project.team.member?(maintainer, Gitlab::Access::MAINTAINER)).to be(false)
expect(shared_project.team.member?(developer, Gitlab::Access::DEVELOPER)).to be(true)
expect(shared_project.team.member?(maintainer, Gitlab::Access::DEVELOPER)).to be(true)
end
end
shared_examples 'max member access for users' do
let(:project) { create(:project) }
let(:group) { create(:group) }
let(:second_group) { create(:group) }
let(:maintainer) { create(:user) }
let(:reporter) { create(:user) }
let(:guest) { create(:user) }
let(:promoted_guest) { create(:user) }
let(:group_developer) { create(:user) }
let(:second_developer) { create(:user) }
let(:user_without_access) { create(:user) }
let(:second_user_without_access) { create(:user) }
let(:users) do
[maintainer, reporter, promoted_guest, guest, group_developer, second_developer, user_without_access].map(&:id)
end
let(:expected) do
{
maintainer.id => Gitlab::Access::MAINTAINER,
reporter.id => Gitlab::Access::REPORTER,
promoted_guest.id => Gitlab::Access::DEVELOPER,
guest.id => Gitlab::Access::GUEST,
group_developer.id => Gitlab::Access::DEVELOPER,
second_developer.id => Gitlab::Access::MAINTAINER,
user_without_access.id => Gitlab::Access::NO_ACCESS
}
end
before do
project.add_maintainer(maintainer)
project.add_reporter(reporter)
project.add_guest(promoted_guest)
project.add_guest(guest)
project.project_group_links.create(
group: group,
group_access: Gitlab::Access::DEVELOPER
)
group.add_maintainer(promoted_guest)
group.add_developer(group_developer)
group.add_developer(second_developer)
project.project_group_links.create(
group: second_group,
group_access: Gitlab::Access::MAINTAINER
)
second_group.add_maintainer(second_developer)
end
it 'returns correct roles for different users' do
expect(project.team.max_member_access_for_user_ids(users)).to eq(expected)
end
end
describe '#max_member_access_for_user_ids' do
2017-06-09 17:12:51 +00:00
context 'with RequestStore enabled', :request_store do
include_examples 'max member access for users'
def access_levels(users)
project.team.max_member_access_for_user_ids(users)
end
it 'does not perform extra queries when asked for users who have already been found' do
access_levels(users)
expect { access_levels(users) }.not_to exceed_query_limit(0)
expect(access_levels(users)).to eq(expected)
end
it 'only requests the extra users when uncached users are passed' do
new_user = create(:user)
second_new_user = create(:user)
all_users = users + [new_user.id, second_new_user.id]
expected_all = expected.merge(new_user.id => Gitlab::Access::NO_ACCESS,
second_new_user.id => Gitlab::Access::NO_ACCESS)
access_levels(users)
queries = ActiveRecord::QueryRecorder.new { access_levels(all_users) }
expect(queries.count).to eq(1)
expect(queries.log_message).to match(/\W#{new_user.id}\W/)
expect(queries.log_message).to match(/\W#{second_new_user.id}\W/)
expect(queries.log_message).not_to match(/\W#{promoted_guest.id}\W/)
expect(access_levels(all_users)).to eq(expected_all)
end
end
context 'with RequestStore disabled' do
include_examples 'max member access for users'
end
end
end