String.capitalize now considers Unicode's titlecase property

Author: José Valim

lib/elixir/lib/string.ex

@@ -179,23 +179,25 @@ defmodule String do
   defdelegate downcase(binary), to: String.Unicode
 
   @doc """
-  Convert the first character on the given string to uppercase
-  and the remaining to downcase.
+  Converts the first character in the given string to
+  titlecase and the remaining to downcase.
+
+  This relies on the titlecase information provided
+  by the Unicode Standard. Note this function makes
+  no attempt in capitalizing all words in the string
+  (usually known as titlecase).
 
   ## Examples
 
       String.capitalize("abcd") #=> "Abcd"
-      String.capitalize("ab 123 xpto") #=> "Ab 123 xpto"
+      String.capitalize("fin")   #=> "Fin"
       String.capitalize("josé") #=> "José"
 
   """
   @spec capitalize(t) :: t
-
   def capitalize(string) when is_binary(string) do
-    case next_grapheme(string) do
-      { char, rest } -> upcase(char) <> downcase(rest)
-      :no_grapheme -> ""
-    end
+    { char, rest } = String.Unicode.titlecase_once(string)
+    char <> downcase(rest)
   end
 
   @doc """

lib/elixir/priv/unicode.ex

@@ -23,11 +23,13 @@ defmodule String.Unicode do
       _class, bidi, _decomposition,
       _numeric_1, _numeric_2, _numeric_3,
       _bidi_mirror, _unicode_1, _iso,
-      upper, lower, _title ] = :binary.split(line, ";", [:global])
+      upper, lower, title ] = :binary.split(line, ";", [:global])
+
+    title = :binary.part(title, 0, size(title) - 1)
 
     cond do
-      upper != "" or lower != "" ->
-        { [{ to_binary.(codepoint), to_binary.(upper), to_binary.(lower) } | cacc], wacc }
+      upper != "" or lower != "" or title != "" ->
+        { [{ to_binary.(codepoint), to_binary.(upper), to_binary.(lower), to_binary.(title) } | cacc], wacc }
       bidi in ["B", "S", "WS"] ->
         { cacc, [to_binary.(codepoint) | wacc] }
       true ->
@@ -38,9 +40,9 @@ defmodule String.Unicode do
   special_path = Path.expand("../SpecialCasing.txt", __FILE__)
 
   codes = Enum.reduce File.iterator!(special_path), codes, fn(line, acc) ->
-    [ codepoint, lower, _title, upper, _comment ] = :binary.split(line, "; ", [:global])
+    [ codepoint, lower, title, upper, _comment ] = :binary.split(line, "; ", [:global])
     key = to_binary.(codepoint)
-    :lists.keystore(key, 1, acc, { key, to_binary.(upper), to_binary.(lower) })
+    :lists.keystore(key, 1, acc, { key, to_binary.(upper), to_binary.(lower), to_binary.(title) })
   end
 
   seqs_path = Path.expand("../NamedSequences.txt", __FILE__)
@@ -55,7 +57,7 @@ defmodule String.Unicode do
 
   # Downcase
 
-  lc { codepoint, _upper, lower } inlist codes, lower && lower != codepoint do
+  lc { codepoint, _upper, lower, _title } inlist codes, lower && lower != codepoint do
     args = quote do: [unquote(codepoint) <> t]
     code = quote do: unquote(lower) <> downcase(t)
     def :downcase, args, [], do: code
@@ -71,7 +73,7 @@ defmodule String.Unicode do
 
   # Upcase
 
-  lc { codepoint, upper, _lower } inlist codes, upper && upper != codepoint do
+  lc { codepoint, upper, _lower, _title } inlist codes, upper && upper != codepoint do
     args = quote do: [unquote(codepoint) <> t]
     code = quote do: unquote(upper) <> upcase(t)
     def :upcase, args, [], do: code
@@ -85,6 +87,22 @@ defmodule String.Unicode do
     << >>
   end
 
+  # Titlecase once
+
+  lc { codepoint, _upper, _lower, title } inlist codes, title && title != codepoint do
+    args = quote do: [unquote(codepoint) <> t]
+    code = quote do: { unquote(title), t }
+    def :titlecase_once, args, [], do: code
+  end
+
+  def titlecase_once(<< h, t :: binary >>) do
+    { <<h>>, t }
+  end
+
+  def titlecase_once(<< >>) do
+    { <<>>, <<>> }
+  end
+
   # Strip
 
   def lstrip(""), do: ""

lib/elixir/test/elixir/string_test.exs

@@ -74,6 +74,7 @@ defmodule StringTest do
     assert String.capitalize("ÂÁÀ") == "Âáà"
     assert String.capitalize("òóôõö") == "Òóôõö"
     assert String.capitalize("ÒÓÔÕÖ") == "Òóôõö"
+    assert String.capitalize("fin") == "Fin"
   end
 
   test :rstrip do