started working on a migration for projects that have current import_url issues

1 files changed, 52 insertions, 0 deletions

diff --git a/db/migrate/20160620110927_fix_no_validatable_import_url.rb b/db/migrate/20160620110927_fix_no_validatable_import_url.rb
new file mode 100644
index 00000000000..e56a8a0c853
--- /dev/null
+++ b/db/migrate/20160620110927_fix_no_validatable_import_url.rb
@@ -0,0 +1,52 @@
+# See http://doc.gitlab.com/ce/development/migration_style_guide.html
+# for more information on how to write migrations for GitLab.
+
+class FixNoValidatableImportUrl < ActiveRecord::Migration
+  include Gitlab::Database::MigrationHelpers
+  class SqlBatches
+
+    attr_reader :results, :query
+
+    def initialize(batch_size: 100, query:)
+      @offset = 0
+      @batch_size = batch_size
+      @query = query
+      @results = []
+    end
+
+    def next
+      @results = ActiveRecord::Base.connection.execute(batched_sql)
+      @offset += @batch_size
+      @results.any?
+    end
+
+    private
+
+    def batched_sql
+      "#{@query} OFFSET #{@offset} LIMIT #{@batch_size}"
+    end
+  end
+
+  def up
+    invalid_import_url_project_ids.each { |project_id| cleanup_import_url(project_id) }
+  end
+
+  def invalid_import_url_project_ids
+    ids = []
+    batches = SqlBatches.new(query: "SELECT id, import_url FROM projects WHERE import_url IS NOT NULL")
+
+    while batches.nexts
+      ids += batches.results.map { |result| invalid_url?(result[:import_url]) ? result[:id] : nil }
+    end
+
+    ids.compact
+  end
+
+  def invalid_url?(url)
+    AddressableUrlValidator.new({ attributes: 1 }).valid_url?(url)
+  end
+
+  def cleanup_import_url(project_id)
+    execute("UPDATE projects SET mirror = false, import_url = NULL WHERE id = #{project_id}")
+  end
+end