Process commits in a separate worker

This moves the code used for processing commits from GitPushService to
its own Sidekiq worker: ProcessCommitWorker.

Using a Sidekiq worker allows us to process multiple commits in
parallel. This in turn will lead to issues being closed faster and cross
references being created faster. Furthermore by isolating this code into
a separate class it's easier to test and maintain the code.

The new worker also ensures it can efficiently check which issues can be
closed, without having to run numerous SQL queries for every issue.

1 files changed, 67 insertions, 0 deletions

diff --git a/app/workers/process_commit_worker.rb b/app/workers/process_commit_worker.rb
new file mode 100644
index 00000000000..071741fbacd
--- /dev/null
+++ b/app/workers/process_commit_worker.rb
@@ -0,0 +1,67 @@
+# Worker for processing individiual commit messages pushed to a repository.
+#
+# Jobs for this worker are scheduled for every commit that is being pushed. As a
+# result of this the workload of this worker should be kept to a bare minimum.
+# Consider using an extra worker if you need to add any extra (and potentially
+# slow) processing of commits.
+class ProcessCommitWorker
+  include Sidekiq::Worker
+  include DedicatedSidekiqQueue
+
+  # project_id - The ID of the project this commit belongs to.
+  # user_id - The ID of the user that pushed the commit.
+  # commit_sha - The SHA1 of the commit to process.
+  # default - The data was pushed to the default branch.
+  def perform(project_id, user_id, commit_sha, default = false)
+    project = Project.find_by(id: project_id)
+
+    return unless project
+
+    user = User.find_by(id: user_id)
+
+    return unless user
+
+    commit = find_commit(project, commit_sha)
+
+    return unless commit
+
+    author = commit.author || user
+
+    process_commit_message(project, commit, user, author, default)
+
+    update_issue_metrics(commit, author)
+  end
+
+  def process_commit_message(project, commit, user, author, default = false)
+    closed_issues = default ? commit.closes_issues(user) : []
+
+    unless closed_issues.empty?
+      close_issues(project, user, author, commit, closed_issues)
+    end
+
+    commit.create_cross_references!(author, closed_issues)
+  end
+
+  def close_issues(project, user, author, commit, issues)
+    # We don't want to run permission related queries for every single issue,
+    # therefor we use IssueCollection here and skip the authorization check in
+    # Issues::CloseService#execute.
+    IssueCollection.new(issues).updatable_by_user(user).each do |issue|
+      Issues::CloseService.new(project, author).
+        close_issue(issue, commit: commit)
+    end
+  end
+
+  def update_issue_metrics(commit, author)
+    mentioned_issues = commit.all_references(author).issues
+
+    Issue::Metrics.where(issue_id: mentioned_issues.map(&:id), first_mentioned_in_commit_at: nil).
+      update_all(first_mentioned_in_commit_at: commit.committed_date)
+  end
+
+  private
+
+  def find_commit(project, sha)
+    project.commit(sha)
+  end
+end