Release notes should mention better newtype-deriving
authorsimonpj@microsoft.com <unknown>
Mon, 4 Sep 2006 12:44:58 +0000 (12:44 +0000)
committersimonpj@microsoft.com <unknown>
Mon, 4 Sep 2006 12:44:58 +0000 (12:44 +0000)
docs/users_guide/6.6-notes.xml
docs/users_guide/ffi-chap.xml
docs/users_guide/glasgow_exts.xml

index 9b215b0..f807a33 100644 (file)
           <literal>MyIO <replaceable>t</replaceable></literal>
           rather than just
           <literal>IO <replaceable>t</replaceable></literal>.
+         See <xref linkend="ffi-newtype-io"/>
         </para>
       </listitem>
+      <listitem> <para> GHC's mechansim for deriving user-defined classes
+         for newtypes has been further generalised, to multi-parameter type
+         classes and higher-kinded types.  See <xref
+         linkend="newtype-deriving"/>.
+         </para></listitem>
       <listitem>
         <para>
           By default, pattern bindings in GHC are now monomorphic.
index 78c23f8..ff8ef62 100644 (file)
@@ -54,7 +54,7 @@ the <option>-fglasgow-exts</option><indexterm><primary><option>-fglasgow-exts</o
       and <literal>ByteArray#</literal>.</para>
     </sect2>
 
-    <sect2>
+    <sect2 id="ffi-newtype-io">
       <title>Newtype wrapping of the IO monad</title>
       <para>The FFI spec requires the IO monad to appear in various  places,
       but it can sometimes be convenient to wrap the IO monad in a
index 588dfef..e801b56 100644 (file)
@@ -120,7 +120,7 @@ documentation</ulink> describes all the libraries that come with GHC.
         </term>
        <listitem>
          <para> These two flags control how generalisation is done in
-           See <xlink linkend="monomorphism"/>.
+           See <xref linkend="monomorphism"/>.
           </para>
        </listitem>
       </varlistentry>