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

879 lines
23 KiB
Ruby
Raw Normal View History

2011-11-28 02:39:43 -05:00
class MergeRequest < ActiveRecord::Base
include InternalId
include Issuable
include Referable
2015-02-05 19:49:41 -05:00
include Sortable
include Taskable
include Importable
belongs_to :target_project, foreign_key: :target_project_id, class_name: "Project"
belongs_to :source_project, foreign_key: :source_project_id, class_name: "Project"
2015-11-02 11:27:38 -05:00
belongs_to :merge_user, class_name: "User"
has_many :merge_request_diffs, dependent: :destroy
has_one :merge_request_diff,
-> { order('merge_request_diffs.id DESC') }
2016-07-01 09:34:10 -04:00
has_many :events, as: :target, dependent: :destroy
has_many :merge_requests_closing_issues, class_name: 'MergeRequestsClosingIssues', dependent: :delete_all
Improve performance of the cycle analytics page. 1. These changes bring down page load time for 100 issues from more than a minute to about 1.5 seconds. 2. This entire commit is composed of these types of performance enhancements: - Cache relevant data in `IssueMetrics` wherever possible. - Cache relevant data in `MergeRequestMetrics` wherever possible. - Preload metrics 3. Given these improvements, we now only need to make 4 SQL calls: - Load all issues - Load all merge requests - Load all metrics for the issues - Load all metrics for the merge requests 4. A list of all the data points that are now being pre-calculated: a. The first time an issue is mentioned in a commit - In `GitPushService`, find all issues mentioned by the given commit using `ReferenceExtractor`. Set the `first_mentioned_in_commit_at` flag for each of them. - There seems to be a (pre-existing) bug here - files (and therefore commits) created using the Web CI don't have cross-references created, and issues are not closed even when the commit title is "Fixes #xx". b. The first time a merge request is deployed to production When a `Deployment` is created, find all merge requests that were merged in before the deployment, and set the `first_deployed_to_production_at` flag for each of them. c. The start / end time for a merge request pipeline Hook into the `Pipeline` state machine. When the `status` moves to `running`, find the merge requests whose tip commit matches the pipeline, and record the `latest_build_started_at` time for each of them. When the `status` moves to `success`, record the `latest_build_finished_at` time. d. The merge requests that close an issue - This was a big cause of the performance problems we were having with Cycle Analytics. We need to use `ReferenceExtractor` to make this calculation, which is slow when we have to run it on a large number of merge requests. - When a merge request is created, updated, or refreshed, find the issues it closes, and create an instance of `MergeRequestsClosingIssues`, which acts as a join model between merge requests and issues. - If a `MergeRequestsClosingIssues` instance links a merge request and an issue, that issue closes that merge request. 5. The `Queries` module was changed into a class, so we can cache the results of `issues` and `merge_requests_closing_issues` across various cycle analytics stages. 6. The code added in this commit is untested. Tests will be added in the next commit.
2016-09-15 04:59:36 -04:00
2015-11-02 11:27:38 -05:00
serialize :merge_params, Hash
after_create :ensure_merge_request_diff, unless: :importing?
after_update :reload_diff_if_branch_changed
delegate :commits, :real_size, to: :merge_request_diff, prefix: nil
Merge Request on forked projects The good: - You can do a merge request for a forked commit and it will merge properly (i.e. it does work). - Push events take into account merge requests on forked projects - Tests around merge_actions now present, spinach, and other rspec tests - Satellites now clean themselves up rather then recreate The questionable: - Events only know about target projects - Project's merge requests only hold on to MR's where they are the target - All operations performed in the satellite The bad: - Duplication between project's repositories and satellites (e.g. commits_between) (for reference: http://feedback.gitlab.com/forums/176466-general/suggestions/3456722-merge-requests-between-projects-repos) Fixes: Make test repos/satellites only create when needed -Spinach/Rspec now only initialize test directory, and setup stubs (things that are relatively cheap) -project_with_code, source_project_with_code, and target_project_with_code now create/destroy their repos individually -fixed remote removal -How to merge renders properly -Update emails to show project/branches -Edit MR doesn't set target branch -Fix some failures on editing/creating merge requests, added a test -Added back a test around merge request observer -Clean up project_transfer_spec, Remove duplicate enable/disable observers -Ensure satellite lock files are cleaned up, Attempted to add some testing around these as well -Signifant speed ups for tests -Update formatting ordering in notes_on_merge_requests -Remove wiki schema update Fixes for search/search results -Search results was using by_project for a list of projects, updated this to use in_projects -updated search results to reference the correct (target) project -udpated search results to print both sides of the merge request Change-Id: I19407990a0950945cc95d62089cbcc6262dab1a8
2013-04-25 10:15:33 -04:00
# When this attribute is true some MR validation is ignored
# It allows us to close or modify broken merge requests
attr_accessor :allow_broken
# Temporary fields to store compare vars
# when creating new merge request
attr_accessor :can_be_created, :compare_commits, :diff_options, :compare
2013-02-18 08:22:18 -05:00
state_machine :state, initial: :opened do
2013-02-18 03:40:56 -05:00
event :close do
transition [:reopened, :opened] => :closed
end
event :mark_as_merged do
transition [:reopened, :opened, :locked] => :merged
2013-02-18 03:40:56 -05:00
end
event :reopen do
2013-02-18 08:22:18 -05:00
transition closed: :reopened
2013-02-18 03:40:56 -05:00
end
event :lock_mr do
transition [:reopened, :opened] => :locked
end
event :unlock_mr do
transition locked: :reopened
end
after_transition any => :locked do |merge_request, transition|
merge_request.locked_at = Time.now
merge_request.save
end
2015-02-02 22:30:09 -05:00
after_transition locked: (any - :locked) do |merge_request, transition|
merge_request.locked_at = nil
merge_request.save
end
2013-02-18 03:40:56 -05:00
state :opened
state :reopened
state :closed
state :merged
state :locked
2013-02-18 03:40:56 -05:00
end
state_machine :merge_status, initial: :unchecked do
event :mark_as_unchecked do
transition [:can_be_merged, :cannot_be_merged] => :unchecked
end
event :mark_as_mergeable do
transition [:unchecked, :cannot_be_merged] => :can_be_merged
end
event :mark_as_unmergeable do
transition [:unchecked, :can_be_merged] => :cannot_be_merged
end
state :unchecked
state :can_be_merged
state :cannot_be_merged
around_transition do |merge_request, transition, block|
Gitlab::Timeless.timeless(merge_request, &block)
end
end
validates :source_project, presence: true, unless: [:allow_broken, :importing?, :closed_without_fork?]
2012-10-08 20:10:04 -04:00
validates :source_branch, presence: true
Merge Request on forked projects The good: - You can do a merge request for a forked commit and it will merge properly (i.e. it does work). - Push events take into account merge requests on forked projects - Tests around merge_actions now present, spinach, and other rspec tests - Satellites now clean themselves up rather then recreate The questionable: - Events only know about target projects - Project's merge requests only hold on to MR's where they are the target - All operations performed in the satellite The bad: - Duplication between project's repositories and satellites (e.g. commits_between) (for reference: http://feedback.gitlab.com/forums/176466-general/suggestions/3456722-merge-requests-between-projects-repos) Fixes: Make test repos/satellites only create when needed -Spinach/Rspec now only initialize test directory, and setup stubs (things that are relatively cheap) -project_with_code, source_project_with_code, and target_project_with_code now create/destroy their repos individually -fixed remote removal -How to merge renders properly -Update emails to show project/branches -Edit MR doesn't set target branch -Fix some failures on editing/creating merge requests, added a test -Added back a test around merge request observer -Clean up project_transfer_spec, Remove duplicate enable/disable observers -Ensure satellite lock files are cleaned up, Attempted to add some testing around these as well -Signifant speed ups for tests -Update formatting ordering in notes_on_merge_requests -Remove wiki schema update Fixes for search/search results -Search results was using by_project for a list of projects, updated this to use in_projects -updated search results to reference the correct (target) project -udpated search results to print both sides of the merge request Change-Id: I19407990a0950945cc95d62089cbcc6262dab1a8
2013-04-25 10:15:33 -04:00
validates :target_project, presence: true
2012-10-08 20:10:04 -04:00
validates :target_branch, presence: true
2015-12-03 04:27:34 -05:00
validates :merge_user, presence: true, if: :merge_when_build_succeeds?
validate :validate_branches, unless: [:allow_broken, :importing?, :closed_without_fork?]
validate :validate_fork, unless: :closed_without_fork?
2011-11-28 02:39:43 -05:00
Merge Request on forked projects The good: - You can do a merge request for a forked commit and it will merge properly (i.e. it does work). - Push events take into account merge requests on forked projects - Tests around merge_actions now present, spinach, and other rspec tests - Satellites now clean themselves up rather then recreate The questionable: - Events only know about target projects - Project's merge requests only hold on to MR's where they are the target - All operations performed in the satellite The bad: - Duplication between project's repositories and satellites (e.g. commits_between) (for reference: http://feedback.gitlab.com/forums/176466-general/suggestions/3456722-merge-requests-between-projects-repos) Fixes: Make test repos/satellites only create when needed -Spinach/Rspec now only initialize test directory, and setup stubs (things that are relatively cheap) -project_with_code, source_project_with_code, and target_project_with_code now create/destroy their repos individually -fixed remote removal -How to merge renders properly -Update emails to show project/branches -Edit MR doesn't set target branch -Fix some failures on editing/creating merge requests, added a test -Added back a test around merge request observer -Clean up project_transfer_spec, Remove duplicate enable/disable observers -Ensure satellite lock files are cleaned up, Attempted to add some testing around these as well -Signifant speed ups for tests -Update formatting ordering in notes_on_merge_requests -Remove wiki schema update Fixes for search/search results -Search results was using by_project for a list of projects, updated this to use in_projects -updated search results to reference the correct (target) project -udpated search results to print both sides of the merge request Change-Id: I19407990a0950945cc95d62089cbcc6262dab1a8
2013-04-25 10:15:33 -04:00
scope :by_branch, ->(branch_name) { where("(source_branch LIKE :branch) OR (target_branch LIKE :branch)", branch: branch_name) }
2013-02-20 08:37:20 -05:00
scope :cared, ->(user) { where('assignee_id = :user OR author_id = :user', user: user.id) }
scope :by_milestone, ->(milestone) { where(milestone_id: milestone) }
scope :of_projects, ->(ids) { where(target_project_id: ids) }
scope :from_project, ->(project) { where(source_project_id: project.id) }
scope :merged, -> { with_state(:merged) }
scope :closed_and_merged, -> { with_states(:closed, :merged) }
scope :from_source_branches, ->(branches) { where(source_branch: branches) }
scope :join_project, -> { joins(:target_project) }
scope :references_project, -> { references(:target_project) }
after_save :keep_around_commit
def self.reference_prefix
'!'
end
# Pattern used to extract `!123` merge request references from text
#
# This pattern supports cross-project references.
def self.reference_pattern
@reference_pattern ||= %r{
(#{Project.reference_pattern})?
#{Regexp.escape(reference_prefix)}(?<merge_request>\d+)
}x
end
def self.link_reference_pattern
@link_reference_pattern ||= super("merge_requests", /(?<merge_request>\d+)/)
end
def self.reference_valid?(reference)
reference.to_i > 0 && reference.to_i <= Gitlab::Database::MAX_INT_VALUE
end
# Returns all the merge requests from an ActiveRecord:Relation.
#
# This method uses a UNION as it usually operates on the result of
# ProjectsFinder#execute. PostgreSQL in particular doesn't always like queries
# using multiple sub-queries especially when combined with an OR statement.
# UNIONs on the other hand perform much better in these cases.
#
# relation - An ActiveRecord::Relation that returns a list of Projects.
#
# Returns an ActiveRecord::Relation.
def self.in_projects(relation)
source = where(source_project_id: relation).select(:id)
target = where(target_project_id: relation).select(:id)
union = Gitlab::SQL::Union.new([source, target])
where("merge_requests.id IN (#{union.to_sql})")
end
WIP_REGEX = /\A\s*(\[WIP\]\s*|WIP:\s*|WIP\s+)+\s*/i.freeze
def self.work_in_progress?(title)
!!(title =~ WIP_REGEX)
end
def self.wipless_title(title)
title.sub(WIP_REGEX, "")
end
def self.wip_title(title)
work_in_progress?(title) ? title : "WIP: #{title}"
end
def to_reference(from_project = nil)
reference = "#{self.class.reference_prefix}#{iid}"
if cross_project_reference?(from_project)
reference = project.to_reference + reference
end
reference
end
def first_commit
merge_request_diff ? merge_request_diff.first_commit : compare_commits.first
end
def raw_diffs(*args)
merge_request_diff ? merge_request_diff.raw_diffs(*args) : compare.raw_diffs(*args)
end
def diffs(diff_options = nil)
if compare
compare.diffs(diff_options)
else
merge_request_diff.diffs(diff_options)
end
end
2016-03-03 12:38:44 -05:00
def diff_size
diffs(diff_options).size
2016-03-03 12:38:44 -05:00
end
def diff_base_commit
if persisted?
merge_request_diff.base_commit
else
branch_merge_base_commit
end
end
# MRs created before 8.4 don't store a MergeRequestDiff#base_commit_sha,
# but we need to get a commit for the "View file @ ..." link by deleted files,
# so we find the likely one if we can't get the actual one.
# This will not be the actual base commit if the target branch was merged into
# the source branch after the merge request was created, but it is good enough
# for the specific purpose of linking to a commit.
2016-07-06 19:28:13 -04:00
# It is not good enough for use in `Gitlab::Git::DiffRefs`, which needs the
# true base commit, so we can't simply have `#diff_base_commit` fall back on
# this method.
def likely_diff_base_commit
first_commit.parent || first_commit
end
def diff_start_commit
if persisted?
merge_request_diff.start_commit
else
target_branch_head
end
end
def diff_head_commit
if persisted?
merge_request_diff.head_commit
else
source_branch_head
end
end
def diff_start_sha
diff_start_commit.try(:sha)
end
def diff_base_sha
diff_base_commit.try(:sha)
end
def diff_head_sha
diff_head_commit.try(:sha)
end
# When importing a pull request from GitHub, the old and new branches may no
# longer actually exist by those names, but we need to recreate the merge
# request diff with the right source and target shas.
# We use these attributes to force these to the intended values.
attr_writer :target_branch_sha, :source_branch_sha
def source_branch_head
source_branch_ref = @source_branch_sha || source_branch
2016-08-25 08:46:59 -04:00
source_project.repository.commit(source_branch_ref) if source_branch_ref
end
def target_branch_head
target_branch_ref = @target_branch_sha || target_branch
2016-08-25 08:46:59 -04:00
target_project.repository.commit(target_branch_ref) if target_branch_ref
end
def branch_merge_base_commit
start_sha = target_branch_sha
head_sha = source_branch_sha
if start_sha && head_sha
target_project.merge_base_commit(start_sha, head_sha)
end
end
def target_branch_sha
@target_branch_sha || target_branch_head.try(:sha)
end
def source_branch_sha
@source_branch_sha || source_branch_head.try(:sha)
end
def diff_refs
return unless diff_start_commit || diff_base_commit
Gitlab::Diff::DiffRefs.new(
base_sha: diff_base_sha,
start_sha: diff_start_sha,
head_sha: diff_head_sha
)
end
# Return diff_refs instance trying to not touch the git repository
def diff_sha_refs
if merge_request_diff && merge_request_diff.diff_refs_by_sha?
merge_request_diff.diff_refs
else
diff_refs
end
end
def branch_merge_base_sha
branch_merge_base_commit.try(:sha)
end
def validate_branches
if target_project == source_project && target_branch == source_branch
Merge Request on forked projects The good: - You can do a merge request for a forked commit and it will merge properly (i.e. it does work). - Push events take into account merge requests on forked projects - Tests around merge_actions now present, spinach, and other rspec tests - Satellites now clean themselves up rather then recreate The questionable: - Events only know about target projects - Project's merge requests only hold on to MR's where they are the target - All operations performed in the satellite The bad: - Duplication between project's repositories and satellites (e.g. commits_between) (for reference: http://feedback.gitlab.com/forums/176466-general/suggestions/3456722-merge-requests-between-projects-repos) Fixes: Make test repos/satellites only create when needed -Spinach/Rspec now only initialize test directory, and setup stubs (things that are relatively cheap) -project_with_code, source_project_with_code, and target_project_with_code now create/destroy their repos individually -fixed remote removal -How to merge renders properly -Update emails to show project/branches -Edit MR doesn't set target branch -Fix some failures on editing/creating merge requests, added a test -Added back a test around merge request observer -Clean up project_transfer_spec, Remove duplicate enable/disable observers -Ensure satellite lock files are cleaned up, Attempted to add some testing around these as well -Signifant speed ups for tests -Update formatting ordering in notes_on_merge_requests -Remove wiki schema update Fixes for search/search results -Search results was using by_project for a list of projects, updated this to use in_projects -updated search results to reference the correct (target) project -udpated search results to print both sides of the merge request Change-Id: I19407990a0950945cc95d62089cbcc6262dab1a8
2013-04-25 10:15:33 -04:00
errors.add :branch_conflict, "You can not use same project/branch for source and target"
end
if opened? || reopened?
similar_mrs = self.target_project.merge_requests.where(source_branch: source_branch, target_branch: target_branch, source_project_id: source_project.try(:id)).opened
similar_mrs = similar_mrs.where('id not in (?)', self.id) if self.id
if similar_mrs.any?
errors.add :validate_branches,
2015-12-14 21:53:52 -05:00
"Cannot Create: This merge request already exists: #{similar_mrs.pluck(:title)}"
end
end
end
def validate_fork
return true unless target_project && source_project
return true if target_project == source_project
2016-08-30 07:31:39 -04:00
return true unless forked_source_project_missing?
errors.add :validate_fork,
2016-08-25 09:08:31 -04:00
'Source project is not a fork of the target project'
end
def closed_without_fork?
2016-08-30 07:31:39 -04:00
closed? && forked_source_project_missing?
end
2016-09-02 09:36:59 -04:00
def closed_without_source_project?
closed? && !source_project
end
2016-08-30 07:31:39 -04:00
def forked_source_project_missing?
return false unless for_fork?
return true unless source_project
!source_project.forked_from?(target_project)
end
2016-09-13 08:40:00 -04:00
def reopenable?
2016-09-08 07:25:16 -04:00
return false if closed_without_fork? || closed_without_source_project? || merged?
2016-09-13 08:40:00 -04:00
closed?
2016-09-06 10:48:59 -04:00
end
def ensure_merge_request_diff
merge_request_diff || create_merge_request_diff
end
def create_merge_request_diff
merge_request_diffs.create
reload_merge_request_diff
end
def reload_merge_request_diff
merge_request_diff(true)
end
def reload_diff_if_branch_changed
if source_branch_changed? || target_branch_changed?
reload_diff
end
end
def reload_diff
return unless open?
old_diff_refs = self.diff_refs
create_merge_request_diff
MergeRequests::MergeRequestDiffCacheService.new.execute(self)
new_diff_refs = self.diff_refs
update_diff_notes_positions(
old_diff_refs: old_diff_refs,
new_diff_refs: new_diff_refs
)
end
def check_if_can_be_merged
return unless unchecked?
can_be_merged =
!broken? && project.repository.can_be_merged?(diff_head_sha, target_branch)
if can_be_merged
mark_as_mergeable
else
mark_as_unmergeable
end
2012-03-29 17:27:42 -04:00
end
2012-03-15 17:32:00 -04:00
def merge_event
@merge_event ||= target_project.events.where(target_id: self.id, target_type: "MergeRequest", action: Event::MERGED).last
2012-03-15 17:32:00 -04:00
end
def closed_event
@closed_event ||= target_project.events.where(target_id: self.id, target_type: "MergeRequest", action: Event::CLOSED).last
end
def work_in_progress?
self.class.work_in_progress?(title)
end
def wipless_title
self.class.wipless_title(self.title)
end
def wip_title
self.class.wip_title(self.title)
end
def mergeable?(skip_ci_check: false)
return false unless mergeable_state?(skip_ci_check: skip_ci_check)
check_if_can_be_merged
can_be_merged?
end
def mergeable_state?(skip_ci_check: false)
return false unless open?
return false if work_in_progress?
return false if broken?
return false unless skip_ci_check || mergeable_ci_state?
true
end
2015-11-24 08:59:02 -05:00
def can_cancel_merge_when_build_succeeds?(current_user)
can_be_merged_by?(current_user) || self.author == current_user
end
2015-11-24 08:59:02 -05:00
def can_remove_source_branch?(current_user)
!source_project.protected_branch?(source_branch) &&
!source_project.root_ref?(source_branch) &&
2016-08-08 14:55:13 -04:00
Ability.allowed?(current_user, :push_code, source_project) &&
diff_head_commit == source_branch_head
2015-11-18 05:17:41 -05:00
end
def should_remove_source_branch?
merge_params['should_remove_source_branch'].present?
end
def force_remove_source_branch?
merge_params['force_remove_source_branch'].present?
end
def remove_source_branch?
should_remove_source_branch? || force_remove_source_branch?
end
def mr_and_commit_notes
# Fetch comments only from last 100 commits
commits_for_notes_limit = 100
commit_ids = commits.last(commits_for_notes_limit).map(&:id)
Note.where(
"(project_id = :target_project_id AND noteable_type = 'MergeRequest' AND noteable_id = :mr_id) OR" +
"((project_id = :source_project_id OR project_id = :target_project_id) AND noteable_type = 'Commit' AND commit_id IN (:commit_ids))",
mr_id: id,
commit_ids: commit_ids,
target_project_id: target_project_id,
source_project_id: source_project_id
)
end
def discussions
@discussions ||= self.mr_and_commit_notes.
2016-08-18 19:27:22 -04:00
inc_relations_for_view.
fresh.
discussions
end
def diff_discussions
@diff_discussions ||= self.notes.diff_notes.discussions
end
def find_diff_discussion(discussion_id)
notes = self.notes.diff_notes.where(discussion_id: discussion_id).fresh.to_a
return if notes.empty?
Discussion.new(notes)
end
def discussions_resolvable?
diff_discussions.any?(&:resolvable?)
end
def discussions_resolved?
discussions_resolvable? && diff_discussions.none?(&:to_be_resolved?)
end
2014-09-15 03:10:35 -04:00
def hook_attrs
attrs = {
source: source_project.try(:hook_attrs),
2014-09-15 03:10:35 -04:00
target: target_project.hook_attrs,
last_commit: nil,
work_in_progress: work_in_progress?
2014-09-15 03:10:35 -04:00
}
if diff_head_commit
attrs.merge!(last_commit: diff_head_commit.hook_attrs)
2014-09-15 03:10:35 -04:00
end
attributes.merge!(attrs)
end
Merge Request on forked projects The good: - You can do a merge request for a forked commit and it will merge properly (i.e. it does work). - Push events take into account merge requests on forked projects - Tests around merge_actions now present, spinach, and other rspec tests - Satellites now clean themselves up rather then recreate The questionable: - Events only know about target projects - Project's merge requests only hold on to MR's where they are the target - All operations performed in the satellite The bad: - Duplication between project's repositories and satellites (e.g. commits_between) (for reference: http://feedback.gitlab.com/forums/176466-general/suggestions/3456722-merge-requests-between-projects-repos) Fixes: Make test repos/satellites only create when needed -Spinach/Rspec now only initialize test directory, and setup stubs (things that are relatively cheap) -project_with_code, source_project_with_code, and target_project_with_code now create/destroy their repos individually -fixed remote removal -How to merge renders properly -Update emails to show project/branches -Edit MR doesn't set target branch -Fix some failures on editing/creating merge requests, added a test -Added back a test around merge request observer -Clean up project_transfer_spec, Remove duplicate enable/disable observers -Ensure satellite lock files are cleaned up, Attempted to add some testing around these as well -Signifant speed ups for tests -Update formatting ordering in notes_on_merge_requests -Remove wiki schema update Fixes for search/search results -Search results was using by_project for a list of projects, updated this to use in_projects -updated search results to reference the correct (target) project -udpated search results to print both sides of the merge request Change-Id: I19407990a0950945cc95d62089cbcc6262dab1a8
2013-04-25 10:15:33 -04:00
def for_fork?
target_project != source_project
end
def project
target_project
end
Improve performance of the cycle analytics page. 1. These changes bring down page load time for 100 issues from more than a minute to about 1.5 seconds. 2. This entire commit is composed of these types of performance enhancements: - Cache relevant data in `IssueMetrics` wherever possible. - Cache relevant data in `MergeRequestMetrics` wherever possible. - Preload metrics 3. Given these improvements, we now only need to make 4 SQL calls: - Load all issues - Load all merge requests - Load all metrics for the issues - Load all metrics for the merge requests 4. A list of all the data points that are now being pre-calculated: a. The first time an issue is mentioned in a commit - In `GitPushService`, find all issues mentioned by the given commit using `ReferenceExtractor`. Set the `first_mentioned_in_commit_at` flag for each of them. - There seems to be a (pre-existing) bug here - files (and therefore commits) created using the Web CI don't have cross-references created, and issues are not closed even when the commit title is "Fixes #xx". b. The first time a merge request is deployed to production When a `Deployment` is created, find all merge requests that were merged in before the deployment, and set the `first_deployed_to_production_at` flag for each of them. c. The start / end time for a merge request pipeline Hook into the `Pipeline` state machine. When the `status` moves to `running`, find the merge requests whose tip commit matches the pipeline, and record the `latest_build_started_at` time for each of them. When the `status` moves to `success`, record the `latest_build_finished_at` time. d. The merge requests that close an issue - This was a big cause of the performance problems we were having with Cycle Analytics. We need to use `ReferenceExtractor` to make this calculation, which is slow when we have to run it on a large number of merge requests. - When a merge request is created, updated, or refreshed, find the issues it closes, and create an instance of `MergeRequestsClosingIssues`, which acts as a join model between merge requests and issues. - If a `MergeRequestsClosingIssues` instance links a merge request and an issue, that issue closes that merge request. 5. The `Queries` module was changed into a class, so we can cache the results of `issues` and `merge_requests_closing_issues` across various cycle analytics stages. 6. The code added in this commit is untested. Tests will be added in the next commit.
2016-09-15 04:59:36 -04:00
# If the merge request closes any issues, save this information in the
# `MergeRequestsClosingIssues` model. This is a performance optimization.
# Calculating this information for a number of merge requests requires
# running `ReferenceExtractor` on each of them separately.
# This optimization does not apply to issues from external sources.
Improve performance of the cycle analytics page. 1. These changes bring down page load time for 100 issues from more than a minute to about 1.5 seconds. 2. This entire commit is composed of these types of performance enhancements: - Cache relevant data in `IssueMetrics` wherever possible. - Cache relevant data in `MergeRequestMetrics` wherever possible. - Preload metrics 3. Given these improvements, we now only need to make 4 SQL calls: - Load all issues - Load all merge requests - Load all metrics for the issues - Load all metrics for the merge requests 4. A list of all the data points that are now being pre-calculated: a. The first time an issue is mentioned in a commit - In `GitPushService`, find all issues mentioned by the given commit using `ReferenceExtractor`. Set the `first_mentioned_in_commit_at` flag for each of them. - There seems to be a (pre-existing) bug here - files (and therefore commits) created using the Web CI don't have cross-references created, and issues are not closed even when the commit title is "Fixes #xx". b. The first time a merge request is deployed to production When a `Deployment` is created, find all merge requests that were merged in before the deployment, and set the `first_deployed_to_production_at` flag for each of them. c. The start / end time for a merge request pipeline Hook into the `Pipeline` state machine. When the `status` moves to `running`, find the merge requests whose tip commit matches the pipeline, and record the `latest_build_started_at` time for each of them. When the `status` moves to `success`, record the `latest_build_finished_at` time. d. The merge requests that close an issue - This was a big cause of the performance problems we were having with Cycle Analytics. We need to use `ReferenceExtractor` to make this calculation, which is slow when we have to run it on a large number of merge requests. - When a merge request is created, updated, or refreshed, find the issues it closes, and create an instance of `MergeRequestsClosingIssues`, which acts as a join model between merge requests and issues. - If a `MergeRequestsClosingIssues` instance links a merge request and an issue, that issue closes that merge request. 5. The `Queries` module was changed into a class, so we can cache the results of `issues` and `merge_requests_closing_issues` across various cycle analytics stages. 6. The code added in this commit is untested. Tests will be added in the next commit.
2016-09-15 04:59:36 -04:00
def cache_merge_request_closes_issues!(current_user = self.author)
return if project.has_external_issue_tracker?
Improve performance of the cycle analytics page. 1. These changes bring down page load time for 100 issues from more than a minute to about 1.5 seconds. 2. This entire commit is composed of these types of performance enhancements: - Cache relevant data in `IssueMetrics` wherever possible. - Cache relevant data in `MergeRequestMetrics` wherever possible. - Preload metrics 3. Given these improvements, we now only need to make 4 SQL calls: - Load all issues - Load all merge requests - Load all metrics for the issues - Load all metrics for the merge requests 4. A list of all the data points that are now being pre-calculated: a. The first time an issue is mentioned in a commit - In `GitPushService`, find all issues mentioned by the given commit using `ReferenceExtractor`. Set the `first_mentioned_in_commit_at` flag for each of them. - There seems to be a (pre-existing) bug here - files (and therefore commits) created using the Web CI don't have cross-references created, and issues are not closed even when the commit title is "Fixes #xx". b. The first time a merge request is deployed to production When a `Deployment` is created, find all merge requests that were merged in before the deployment, and set the `first_deployed_to_production_at` flag for each of them. c. The start / end time for a merge request pipeline Hook into the `Pipeline` state machine. When the `status` moves to `running`, find the merge requests whose tip commit matches the pipeline, and record the `latest_build_started_at` time for each of them. When the `status` moves to `success`, record the `latest_build_finished_at` time. d. The merge requests that close an issue - This was a big cause of the performance problems we were having with Cycle Analytics. We need to use `ReferenceExtractor` to make this calculation, which is slow when we have to run it on a large number of merge requests. - When a merge request is created, updated, or refreshed, find the issues it closes, and create an instance of `MergeRequestsClosingIssues`, which acts as a join model between merge requests and issues. - If a `MergeRequestsClosingIssues` instance links a merge request and an issue, that issue closes that merge request. 5. The `Queries` module was changed into a class, so we can cache the results of `issues` and `merge_requests_closing_issues` across various cycle analytics stages. 6. The code added in this commit is untested. Tests will be added in the next commit.
2016-09-15 04:59:36 -04:00
transaction do
self.merge_requests_closing_issues.delete_all
Improve performance of the cycle analytics page. 1. These changes bring down page load time for 100 issues from more than a minute to about 1.5 seconds. 2. This entire commit is composed of these types of performance enhancements: - Cache relevant data in `IssueMetrics` wherever possible. - Cache relevant data in `MergeRequestMetrics` wherever possible. - Preload metrics 3. Given these improvements, we now only need to make 4 SQL calls: - Load all issues - Load all merge requests - Load all metrics for the issues - Load all metrics for the merge requests 4. A list of all the data points that are now being pre-calculated: a. The first time an issue is mentioned in a commit - In `GitPushService`, find all issues mentioned by the given commit using `ReferenceExtractor`. Set the `first_mentioned_in_commit_at` flag for each of them. - There seems to be a (pre-existing) bug here - files (and therefore commits) created using the Web CI don't have cross-references created, and issues are not closed even when the commit title is "Fixes #xx". b. The first time a merge request is deployed to production When a `Deployment` is created, find all merge requests that were merged in before the deployment, and set the `first_deployed_to_production_at` flag for each of them. c. The start / end time for a merge request pipeline Hook into the `Pipeline` state machine. When the `status` moves to `running`, find the merge requests whose tip commit matches the pipeline, and record the `latest_build_started_at` time for each of them. When the `status` moves to `success`, record the `latest_build_finished_at` time. d. The merge requests that close an issue - This was a big cause of the performance problems we were having with Cycle Analytics. We need to use `ReferenceExtractor` to make this calculation, which is slow when we have to run it on a large number of merge requests. - When a merge request is created, updated, or refreshed, find the issues it closes, and create an instance of `MergeRequestsClosingIssues`, which acts as a join model between merge requests and issues. - If a `MergeRequestsClosingIssues` instance links a merge request and an issue, that issue closes that merge request. 5. The `Queries` module was changed into a class, so we can cache the results of `issues` and `merge_requests_closing_issues` across various cycle analytics stages. 6. The code added in this commit is untested. Tests will be added in the next commit.
2016-09-15 04:59:36 -04:00
closes_issues(current_user).each do |issue|
self.merge_requests_closing_issues.create!(issue: issue)
Improve performance of the cycle analytics page. 1. These changes bring down page load time for 100 issues from more than a minute to about 1.5 seconds. 2. This entire commit is composed of these types of performance enhancements: - Cache relevant data in `IssueMetrics` wherever possible. - Cache relevant data in `MergeRequestMetrics` wherever possible. - Preload metrics 3. Given these improvements, we now only need to make 4 SQL calls: - Load all issues - Load all merge requests - Load all metrics for the issues - Load all metrics for the merge requests 4. A list of all the data points that are now being pre-calculated: a. The first time an issue is mentioned in a commit - In `GitPushService`, find all issues mentioned by the given commit using `ReferenceExtractor`. Set the `first_mentioned_in_commit_at` flag for each of them. - There seems to be a (pre-existing) bug here - files (and therefore commits) created using the Web CI don't have cross-references created, and issues are not closed even when the commit title is "Fixes #xx". b. The first time a merge request is deployed to production When a `Deployment` is created, find all merge requests that were merged in before the deployment, and set the `first_deployed_to_production_at` flag for each of them. c. The start / end time for a merge request pipeline Hook into the `Pipeline` state machine. When the `status` moves to `running`, find the merge requests whose tip commit matches the pipeline, and record the `latest_build_started_at` time for each of them. When the `status` moves to `success`, record the `latest_build_finished_at` time. d. The merge requests that close an issue - This was a big cause of the performance problems we were having with Cycle Analytics. We need to use `ReferenceExtractor` to make this calculation, which is slow when we have to run it on a large number of merge requests. - When a merge request is created, updated, or refreshed, find the issues it closes, and create an instance of `MergeRequestsClosingIssues`, which acts as a join model between merge requests and issues. - If a `MergeRequestsClosingIssues` instance links a merge request and an issue, that issue closes that merge request. 5. The `Queries` module was changed into a class, so we can cache the results of `issues` and `merge_requests_closing_issues` across various cycle analytics stages. 6. The code added in this commit is untested. Tests will be added in the next commit.
2016-09-15 04:59:36 -04:00
end
end
end
def closes_issue?(issue)
closes_issues.include?(issue)
end
# Return the set of issues that will be closed if this merge request is accepted.
def closes_issues(current_user = self.author)
if target_branch == project.default_branch
messages = [description]
messages.concat(commits.map(&:safe_message)) if merge_request_diff
Gitlab::ClosingIssueExtractor.new(project, current_user).
closed_by_message(messages.join("\n"))
else
[]
end
end
def target_project_path
if target_project
target_project.path_with_namespace
else
"(removed)"
end
end
def source_project_path
if source_project
source_project.path_with_namespace
else
"(removed)"
end
end
def source_project_namespace
if source_project && source_project.namespace
source_project.namespace.path
else
"(removed)"
end
end
def target_project_namespace
if target_project && target_project.namespace
target_project.namespace.path
else
"(removed)"
end
end
def source_branch_exists?
return false unless self.source_project
self.source_project.repository.branch_names.include?(self.source_branch)
end
def target_branch_exists?
return false unless self.target_project
self.target_project.repository.branch_names.include?(self.target_branch)
end
2013-12-13 14:40:45 -05:00
# Reset merge request events cache
#
# Since we do cache @event we need to reset cache in special cases:
# * when a merge request is updated
# Events cache stored like events/23-20130109142513.
# The cache key includes updated_at timestamp.
# Thus it will automatically generate a new fragment
# when the event is updated because the key changes.
def reset_events_cache
Event.reset_event_cache_for(self)
2013-12-13 14:40:45 -05:00
end
def merge_commit_message
message = "Merge branch '#{source_branch}' into '#{target_branch}'\n\n"
message << "#{title}\n\n"
message << "#{description}\n\n" if description.present?
message << "See merge request #{to_reference}"
message
end
2015-11-02 11:27:38 -05:00
def reset_merge_when_build_succeeds
return unless merge_when_build_succeeds?
2015-11-18 05:17:41 -05:00
2015-11-02 11:27:38 -05:00
self.merge_when_build_succeeds = false
self.merge_user = nil
if merge_params
merge_params.delete('should_remove_source_branch')
merge_params.delete('commit_message')
end
2015-11-02 11:27:38 -05:00
self.save
end
# Return array of possible target branches
# depends on target project of MR
def target_branches
if target_project.nil?
[]
else
target_project.repository.branch_names
end
end
# Return array of possible source branches
# depends on source project of MR
def source_branches
if source_project.nil?
[]
else
source_project.repository.branch_names
end
end
def locked_long_ago?
return false unless locked?
locked_at.nil? || locked_at < (Time.now - 1.day)
end
def has_ci?
source_project.ci_service && commits.any?
end
def branch_missing?
!source_branch_exists? || !target_branch_exists?
end
def broken?
self.commits.blank? || branch_missing? || cannot_be_merged?
end
def can_be_merged_by?(user)
access = ::Gitlab::UserAccess.new(user, project: project)
access.can_push_to_branch?(target_branch) || access.can_merge_to_branch?(target_branch)
end
def can_be_merged_via_command_line_by?(user)
access = ::Gitlab::UserAccess.new(user, project: project)
access.can_push_to_branch?(target_branch)
end
def mergeable_ci_state?
return true unless project.only_allow_merge_if_build_succeeds?
!pipeline || pipeline.success?
end
2016-08-02 08:01:22 -04:00
def environments
return [] unless diff_head_commit
@environments ||=
begin
environments = source_project.environments_for(
source_branch, diff_head_commit)
environments += target_project.environments_for(
target_branch, diff_head_commit, with_tags: true)
environments.uniq
end
2016-08-02 08:01:22 -04:00
end
def state_human_name
if merged?
"Merged"
elsif closed?
"Closed"
else
"Open"
end
end
def state_icon_name
if merged?
"check"
elsif closed?
"times"
else
"circle-o"
end
end
def fetch_ref
target_project.repository.fetch_ref(
source_project.repository.path_to_repo,
"refs/heads/#{source_branch}",
ref_path
)
end
def ref_path
"refs/merge-requests/#{iid}/head"
end
def ref_fetched?
project.repository.ref_exists?(ref_path)
end
def ensure_ref_fetched
fetch_ref unless ref_fetched?
end
def in_locked_state
begin
lock_mr
yield
ensure
unlock_mr if locked?
end
end
def diverged_commits_count
cache = Rails.cache.read(:"merge_request_#{id}_diverged_commits")
if cache.blank? || cache[:source_sha] != source_branch_sha || cache[:target_sha] != target_branch_sha
cache = {
source_sha: source_branch_sha,
target_sha: target_branch_sha,
diverged_commits_count: compute_diverged_commits_count
}
Rails.cache.write(:"merge_request_#{id}_diverged_commits", cache)
end
cache[:diverged_commits_count]
end
def compute_diverged_commits_count
return 0 unless source_branch_sha && target_branch_sha
Gitlab::Git::Commit.between(target_project.repository.raw_repository, source_branch_sha, target_branch_sha).size
end
private :compute_diverged_commits_count
def diverged_from_target_branch?
diverged_commits_count > 0
end
def commits_sha
commits.map(&:sha)
end
def pipeline
return unless diff_head_sha && source_project
@pipeline ||= source_project.pipeline_for(source_branch, diff_head_sha)
end
def all_pipelines
return unless source_project
@all_pipelines ||= begin
sha = if persisted?
all_commits_sha
else
diff_head_sha
end
source_project.pipelines.order(id: :desc).
where(sha: sha, ref: source_branch)
end
end
# Note that this could also return SHA from now dangling commits
def all_commits_sha
merge_request_diffs.flat_map(&:commits_sha).uniq
end
def merge_commit
@merge_commit ||= project.commit(merge_commit_sha) if merge_commit_sha
end
def can_be_reverted?(current_user = nil)
merge_commit && !merge_commit.has_been_reverted?(current_user, self)
end
def can_be_cherry_picked?
merge_commit
end
def has_complete_diff_refs?
diff_sha_refs && diff_sha_refs.complete?
end
def update_diff_notes_positions(old_diff_refs:, new_diff_refs:)
return unless has_complete_diff_refs?
return if new_diff_refs == old_diff_refs
active_diff_notes = self.notes.diff_notes.select do |note|
note.new_diff_note? && note.active?(old_diff_refs)
end
return if active_diff_notes.empty?
paths = active_diff_notes.flat_map { |n| n.diff_file.paths }.uniq
service = Notes::DiffPositionUpdateService.new(
self.project,
nil,
old_diff_refs: old_diff_refs,
new_diff_refs: new_diff_refs,
paths: paths
)
active_diff_notes.each do |note|
service.execute(note)
Gitlab::Timeless.timeless(note, &:save)
end
end
def keep_around_commit
project.repository.keep_around(self.merge_commit_sha)
end
def conflicts
@conflicts ||= Gitlab::Conflict::FileCollection.new(self)
end
def conflicts_can_be_resolved_by?(user)
access = ::Gitlab::UserAccess.new(user, project: source_project)
access.can_push_to_branch?(source_branch)
end
def conflicts_can_be_resolved_in_ui?
return @conflicts_can_be_resolved_in_ui if defined?(@conflicts_can_be_resolved_in_ui)
return @conflicts_can_be_resolved_in_ui = false unless cannot_be_merged?
return @conflicts_can_be_resolved_in_ui = false unless has_complete_diff_refs?
begin
# Try to parse each conflict. If the MR's mergeable status hasn't been updated,
# ensure that we don't say there are conflicts to resolve when there are no conflict
# files.
conflicts.files.each(&:lines)
@conflicts_can_be_resolved_in_ui = conflicts.files.length > 0
rescue Rugged::OdbError, Gitlab::Conflict::Parser::UnresolvableError, Gitlab::Conflict::FileCollection::ConflictSideMissing
@conflicts_can_be_resolved_in_ui = false
end
end
2011-11-28 02:39:43 -05:00
end