1 files changed, 135 insertions, 0 deletions
diff --git a/lib/mime/types/_columnar.rb b/lib/mime/types/_columnar.rb
new file mode 100644
index 0000000..acafbb9
--- /dev/null
+++ b/lib/mime/types/_columnar.rb
@@ -0,0 +1,135 @@
+require 'mime/type/columnar'
+
+# MIME::Types::Columnar is used to extend a MIME::Types container to load data
+# by columns instead of from JSON or YAML. Column loads of MIME types loaded
+# through the columnar store are synchronized with a Mutex.
+#
+# MIME::Types::Columnar is not intended to be used directly, but will be added
+# to an instance of MIME::Types when it is loaded with
+# MIME::Types::Loader#load_columnar.
+module MIME::Types::Columnar
+  LOAD_MUTEX = Mutex.new # :nodoc:
+
+  def self.extended(obj) # :nodoc:
+    super
+    obj.instance_variable_set(:@__mime_data__, [])
+    obj.instance_variable_set(:@__files__, Set.new)
+  end
+
+  # Load the first column data file (type and extensions).
+  def load_base_data(path) #:nodoc:
+    @__root__ = path
+
+    each_file_line('content_type', false) do |line|
+      line = line.split
+      content_type = line.shift
+      extensions = line
+      # content_type, *extensions = line.split
+
+      type = MIME::Type::Columnar.new(self, content_type, extensions)
+      @__mime_data__ << type
+      add(type)
+    end
+
+    self
+  end
+
+  private
+
+  def each_file_line(name, lookup = true)
+    LOAD_MUTEX.synchronize do
+      next if @__files__.include?(name)
+
+      i = -1
+      column = File.join(@__root__, "mime.#{name}.column")
+
+      IO.readlines(column, encoding: 'UTF-8'.freeze).each do |line|
+        line.chomp!
+
+        if lookup
+          type = @__mime_data__[i += 1] or next
+          yield type, line
+        else
+          yield line
+        end
+      end
+
+      @__files__ << name
+    end
+  end
+
+  def load_encoding
+    each_file_line('encoding') do |type, line|
+      pool ||= {}
+      line.freeze
+      type.instance_variable_set(:@encoding, (pool[line] ||= line))
+    end
+  end
+
+  def load_docs
+    each_file_line('docs') do |type, line|
+      type.instance_variable_set(:@docs, opt(line))
+    end
+  end
+
+  def load_preferred_extension
+    each_file_line('pext') do |type, line|
+      type.instance_variable_set(:@preferred_extension, opt(line))
+    end
+  end
+
+  def load_flags
+    each_file_line('flags') do |type, line|
+      line = line.split
+      type.instance_variable_set(:@obsolete, flag(line.shift))
+      type.instance_variable_set(:@registered, flag(line.shift))
+      type.instance_variable_set(:@signature, flag(line.shift))
+    end
+  end
+
+  def load_xrefs
+    each_file_line('xrefs') { |type, line|
+      type.instance_variable_set(:@xrefs, dict(line, array: true))
+    }
+  end
+
+  def load_friendly
+    each_file_line('friendly') { |type, line|
+      type.instance_variable_set(:@friendly, dict(line))
+    }
+  end
+
+  def load_use_instead
+    each_file_line('use_instead') do |type, line|
+      type.instance_variable_set(:@use_instead, opt(line))
+    end
+  end
+
+  def dict(line, array: false)
+    if line == '-'.freeze
+      {}
+    else
+      line.split('|'.freeze).each_with_object({}) { |l, h|
+        k, v = l.split('^'.freeze)
+        v = nil if v.empty?
+        h[k] = array ? Array(v) : v
+      }
+    end
+  end
+
+  def arr(line)
+    if line == '-'.freeze
+      []
+    else
+      line.split('|'.freeze).flatten.compact.uniq
+    end
+  end
+
+  def opt(line)
+    line unless line == '-'.freeze
+  end
+
+  def flag(line)
+    line == '1'.freeze ? true : false
+  end
+end