debian-mirror-gitlab/lib/gitlab/bitbucket_server_import/importer.rb
2023-04-23 21:23:45 +05:30

476 lines
17 KiB
Ruby

# frozen_string_literal: true
module Gitlab
module BitbucketServerImport
class Importer
attr_reader :recover_missing_commits
attr_reader :project, :project_key, :repository_slug, :client, :errors, :users, :already_imported_cache_key
attr_accessor :logger
BATCH_SIZE = 100
# The base cache key to use for tracking already imported objects.
ALREADY_IMPORTED_CACHE_KEY =
'bitbucket_server-importer/already-imported/%{project}/%{collection}'
TempBranch = Struct.new(:name, :sha)
def self.imports_repository?
true
end
def self.refmap
# We omit :heads and :tags since these are fetched in the import_repository
['+refs/pull-requests/*/to:refs/merge-requests/*/head']
end
# Unlike GitHub, you can't grab the commit SHAs for pull requests that
# have been closed but not merged even though Bitbucket has these
# commits internally. We can recover these pull requests by creating a
# branch with the Bitbucket REST API, but by default we turn this
# behavior off.
def initialize(project, recover_missing_commits: false)
@project = project
@recover_missing_commits = recover_missing_commits
@project_key = project.import_data.data['project_key']
@repository_slug = project.import_data.data['repo_slug']
@client = BitbucketServer::Client.new(project.import_data.credentials)
@formatter = Gitlab::ImportFormatter.new
@errors = []
@users = {}
@temp_branches = []
@logger = Gitlab::Import::Logger.build
@already_imported_cache_key = ALREADY_IMPORTED_CACHE_KEY %
{ project: project.id, collection: collection_method }
end
def collection_method
:pull_requests
end
def execute
import_repository
import_pull_requests
download_lfs_objects
delete_temp_branches
handle_errors
metrics.track_finished_import
log_info(import_stage: "complete")
Gitlab::Cache::Import::Caching.expire(already_imported_cache_key, Gitlab::Cache::Import::Caching::SHORTER_TIMEOUT)
true
end
private
def handle_errors
return unless errors.any?
project.import_state.update_column(:last_error, {
message: 'The remote data could not be fully imported.',
errors: errors
}.to_json)
end
def find_user_id(by:, value:)
return unless value
return users[value] if users.key?(value)
user = if by == :email
User.find_by_any_email(value, confirmed: true)
else
User.find_by_username(value)
end
users[value] = user&.id
user&.id
end
def repo
@repo ||= client.repo(project_key, repository_slug)
end
def sha_exists?(sha)
project.repository.commit(sha)
end
def temp_branch_name(pull_request, suffix)
"gitlab/import/pull-request/#{pull_request.iid}/#{suffix}"
end
# This method restores required SHAs that GitLab needs to create diffs
# into branch names as the following:
#
# gitlab/import/pull-request/N/{to,from}
def restore_branches(pull_requests)
shas_to_restore = []
pull_requests.each do |pull_request|
shas_to_restore << TempBranch.new(temp_branch_name(pull_request, :from),
pull_request.source_branch_sha)
shas_to_restore << TempBranch.new(temp_branch_name(pull_request, :to),
pull_request.target_branch_sha)
end
# Create the branches on the Bitbucket Server first
created_branches = restore_branch_shas(shas_to_restore)
@temp_branches += created_branches
# Now sync the repository so we get the new branches
import_repository unless created_branches.empty?
end
def restore_branch_shas(shas_to_restore)
shas_to_restore.each_with_object([]) do |temp_branch, branches_created|
branch_name = temp_branch.name
sha = temp_branch.sha
next if sha_exists?(sha)
begin
client.create_branch(project_key, repository_slug, branch_name, sha)
branches_created << temp_branch
rescue BitbucketServer::Connection::ConnectionError => e
log_warn(message: "Unable to recreate branch", sha: sha, error: e.message)
end
end
end
def import_repository
log_info(import_stage: 'import_repository', message: 'starting import')
project.repository.import_repository(project.import_url)
project.repository.fetch_as_mirror(project.import_url, refmap: self.class.refmap)
log_info(import_stage: 'import_repository', message: 'finished import')
rescue ::Gitlab::Git::CommandError => e
Gitlab::ErrorTracking.log_exception(
e,
import_stage: 'import_repository', message: 'failed import', error: e.message
)
# Expire cache to prevent scenarios such as:
# 1. First import failed, but the repo was imported successfully, so +exists?+ returns true
# 2. Retried import, repo is broken or not imported but +exists?+ still returns true
project.repository.expire_content_cache if project.repository_exists?
raise
end
def download_lfs_objects
result = Projects::LfsPointers::LfsImportService.new(project).execute
if result[:status] == :error
errors << { type: :lfs_objects, errors: "The Lfs import process failed. #{result[:message]}" }
end
end
# Bitbucket Server keeps tracks of references for open pull requests in
# refs/heads/pull-requests, but closed and merged requests get moved
# into hidden internal refs under stash-refs/pull-requests. Unless the
# SHAs involved are at the tip of a branch or tag, there is no way to
# retrieve the server for those commits.
#
# To avoid losing history, we use the Bitbucket API to re-create the branch
# on the remote server. Then we have to issue a `git fetch` to download these
# branches.
def import_pull_requests
page = 0
log_info(import_stage: 'import_pull_requests', message: "starting")
loop do
log_debug(import_stage: 'import_pull_requests', message: "importing page #{page} and batch-size #{BATCH_SIZE} from #{page * BATCH_SIZE} to #{(page + 1) * BATCH_SIZE}")
pull_requests = client.pull_requests(project_key, repository_slug, page_offset: page, limit: BATCH_SIZE).to_a
break if pull_requests.empty?
# Creating branches on the server and fetching the newly-created branches
# may take a number of network round-trips. This used to be done in batches to
# avoid doing a git fetch for every new branch, as the whole process is now
# batched, we do not need to separately do this in batches.
restore_branches(pull_requests) if recover_missing_commits
pull_requests.each do |pull_request|
if already_imported?(pull_request)
log_info(import_stage: 'import_pull_requests', message: 'already imported', iid: pull_request.iid)
else
import_bitbucket_pull_request(pull_request)
end
rescue StandardError => e
Gitlab::ErrorTracking.log_exception(
e,
import_stage: 'import_pull_requests', iid: pull_request.iid, error: e.message
)
backtrace = Gitlab::BacktraceCleaner.clean_backtrace(e.backtrace)
errors << { type: :pull_request, iid: pull_request.iid, errors: e.message, backtrace: backtrace.join("\n"), raw_response: pull_request.raw }
end
log_debug(import_stage: 'import_pull_requests', message: "finished page #{page} and batch-size #{BATCH_SIZE}")
page += 1
end
end
# Returns true if the given object has already been imported, false
# otherwise.
#
# object - The object to check.
def already_imported?(pull_request)
Gitlab::Cache::Import::Caching.set_includes?(already_imported_cache_key, pull_request.iid)
end
# Marks the given object as "already imported".
def mark_as_imported(pull_request)
Gitlab::Cache::Import::Caching.set_add(already_imported_cache_key, pull_request.iid)
end
def delete_temp_branches
@temp_branches.each do |branch|
client.delete_branch(project_key, repository_slug, branch.name, branch.sha)
project.repository.delete_branch(branch.name)
rescue BitbucketServer::Connection::ConnectionError => e
Gitlab::ErrorTracking.log_exception(
e,
import_stage: 'delete_temp_branches', branch: branch.name, error: e.message
)
@errors << { type: :delete_temp_branches, branch_name: branch.name, errors: e.message }
end
end
def import_bitbucket_pull_request(pull_request)
log_info(import_stage: 'import_bitbucket_pull_requests', message: 'starting', iid: pull_request.iid)
description = ''
description += author_line(pull_request)
description += pull_request.description if pull_request.description
attributes = {
iid: pull_request.iid,
title: pull_request.title,
description: description,
source_project_id: project.id,
source_branch: Gitlab::Git.ref_name(pull_request.source_branch_name),
source_branch_sha: pull_request.source_branch_sha,
target_project_id: project.id,
target_branch: Gitlab::Git.ref_name(pull_request.target_branch_name),
target_branch_sha: pull_request.target_branch_sha,
state_id: MergeRequest.available_states[pull_request.state],
author_id: author_id(pull_request),
created_at: pull_request.created_at,
updated_at: pull_request.updated_at
}
creator = Gitlab::Import::MergeRequestCreator.new(project)
merge_request = creator.execute(attributes)
if merge_request.persisted?
import_pull_request_comments(pull_request, merge_request)
metrics.merge_requests_counter.increment
end
log_info(import_stage: 'import_bitbucket_pull_requests', message: 'finished', iid: pull_request.iid)
mark_as_imported(pull_request)
end
def import_pull_request_comments(pull_request, merge_request)
log_info(import_stage: 'import_pull_request_comments', message: 'starting', iid: merge_request.iid)
comments, other_activities = client.activities(project_key, repository_slug, pull_request.iid).partition(&:comment?)
merge_event = other_activities.find(&:merge_event?)
import_merge_event(merge_request, merge_event) if merge_event
inline_comments, pr_comments = comments.partition(&:inline_comment?)
import_inline_comments(inline_comments.map(&:comment), merge_request)
import_standalone_pr_comments(pr_comments.map(&:comment), merge_request)
log_info(import_stage: 'import_pull_request_comments', message: 'finished', iid: merge_request.iid,
merge_event_found: merge_event.present?,
inline_comments_count: inline_comments.count,
standalone_pr_comments: pr_comments.count)
end
# rubocop: disable CodeReuse/ActiveRecord
def import_merge_event(merge_request, merge_event)
log_info(import_stage: 'import_merge_event', message: 'starting', iid: merge_request.iid)
committer = merge_event.committer_email
user_id = find_user_id(by: :email, value: committer) || project.creator_id
timestamp = merge_event.merge_timestamp
merge_request.update({ merge_commit_sha: merge_event.merge_commit })
metric = MergeRequest::Metrics.find_or_initialize_by(merge_request: merge_request)
metric.update(merged_by_id: user_id, merged_at: timestamp)
log_info(import_stage: 'import_merge_event', message: 'finished', iid: merge_request.iid)
end
# rubocop: enable CodeReuse/ActiveRecord
def import_inline_comments(inline_comments, merge_request)
log_info(import_stage: 'import_inline_comments', message: 'starting', iid: merge_request.iid)
inline_comments.each do |comment|
position = build_position(merge_request, comment)
parent = create_diff_note(merge_request, comment, position)
next unless parent&.persisted?
discussion_id = parent.discussion_id
comment.comments.each do |reply|
create_diff_note(merge_request, reply, position, discussion_id)
end
end
log_info(import_stage: 'import_inline_comments', message: 'finished', iid: merge_request.iid)
end
def create_diff_note(merge_request, comment, position, discussion_id = nil)
attributes = pull_request_comment_attributes(comment)
attributes.merge!(position: position, type: 'DiffNote')
attributes[:discussion_id] = discussion_id if discussion_id
note = merge_request.notes.build(attributes)
if note.valid?
note.save
return note
end
log_info(import_stage: 'create_diff_note', message: 'creating fallback DiffNote', iid: merge_request.iid)
# Bitbucket Server supports the ability to comment on any line, not just the
# line in the diff. If we can't add the note as a DiffNote, fallback to creating
# a regular note.
create_fallback_diff_note(merge_request, comment, position)
rescue StandardError => e
Gitlab::ErrorTracking.log_exception(
e,
import_stage: 'create_diff_note', comment_id: comment.id, error: e.message
)
errors << { type: :pull_request, id: comment.id, errors: e.message }
nil
end
def create_fallback_diff_note(merge_request, comment, position)
attributes = pull_request_comment_attributes(comment)
note = "*Comment on"
note += " #{position.old_path}:#{position.old_line} -->" if position.old_line
note += " #{position.new_path}:#{position.new_line}" if position.new_line
note += "*\n\n#{comment.note}"
attributes[:note] = note
merge_request.notes.create!(attributes)
end
def build_position(merge_request, pr_comment)
params = {
diff_refs: merge_request.diff_refs,
old_path: pr_comment.file_path,
new_path: pr_comment.file_path,
old_line: pr_comment.old_pos,
new_line: pr_comment.new_pos
}
Gitlab::Diff::Position.new(params)
end
def import_standalone_pr_comments(pr_comments, merge_request)
pr_comments.each do |comment|
merge_request.notes.create!(pull_request_comment_attributes(comment))
comment.comments.each do |replies|
merge_request.notes.create!(pull_request_comment_attributes(replies))
end
rescue StandardError => e
Gitlab::ErrorTracking.log_exception(
e,
import_stage: 'import_standalone_pr_comments', merge_request_id: merge_request.id, comment_id: comment.id, error: e.message
)
errors << { type: :pull_request, comment_id: comment.id, errors: e.message }
end
end
def pull_request_comment_attributes(comment)
author = uid(comment)
note = ''
unless author
author = project.creator_id
note = "*By #{comment.author_username} (#{comment.author_email})*\n\n"
end
note +=
# Provide some context for replying
if comment.parent_comment
"> #{comment.parent_comment.note.truncate(80)}\n\n#{comment.note}"
else
comment.note
end
{
project: project,
note: note,
author_id: author,
created_at: comment.created_at,
updated_at: comment.updated_at
}
end
def log_debug(details)
logger.debug(log_base_data.merge(details))
end
def log_info(details)
logger.info(log_base_data.merge(details))
end
def log_warn(details)
logger.warn(log_base_data.merge(details))
end
def log_base_data
{
class: self.class.name,
project_id: project.id,
project_path: project.full_path
}
end
def metrics
@metrics ||= Gitlab::Import::Metrics.new(:bitbucket_server_importer, @project)
end
def author_line(rep_object)
return '' if uid(rep_object)
@formatter.author_line(rep_object.author)
end
def author_id(rep_object)
uid(rep_object) || project.creator_id
end
def uid(rep_object)
# We want this explicit to only be username on the FF
# Otherwise, match email.
# There should be no default fall-through on username. Fall-through to import user
if Feature.enabled?(:bitbucket_server_user_mapping_by_username)
find_user_id(by: :username, value: rep_object.author_username)
else
find_user_id(by: :email, value: rep_object.author_email)
end
end
end
end
end