Don't allow null in filenames

1 files changed, 77 insertions, 1 deletions

diff --git a/lib/stdlib/doc/src/filename.xml b/lib/stdlib/doc/src/filename.xml
index 14fd5ef787..b6028fc066 100644
--- a/lib/stdlib/doc/src/filename.xml
+++ b/lib/stdlib/doc/src/filename.xml
@@ -46,7 +46,10 @@
       filename by removing redundant directory separators, use
       <seealso marker="#join/1"><c>join/1</c></seealso>.</p>
 
-    <p>The module supports raw filenames in the way that if a binary is
+    <p>
+      The module supports
+      <seealso marker="unicode_usage#notes-about-raw-filenames">raw
+      filenames</seealso> in the way that if a binary is
       present, or the filename cannot be interpreted according to the return
       value of <seealso marker="kernel:file#native_name_encoding/0">
       <c>file:native_name_encoding/0</c></seealso>, a raw filename is also
@@ -56,6 +59,30 @@
       (the join operation is performed of course). For more information
       about raw filenames, see the
       <seealso marker="kernel:file"><c>file</c></seealso> module.</p>
+
+    <note>
+      <p>
+	Functionality in this module generally assumes valid input and
+	does not necessarily fail on input that does not use a valid
+	encoding. You can validate the encoding of a filename using
+	<seealso marker="#validate/1">filename:validate/1</seealso>.
+      </p>
+      <p>
+	File operations used to accept filenames containing
+	null characters (integer value zero). This caused
+	the name to be truncated at the first null character.
+	Filenames containing null characters inside the filename
+	are now <em>rejected</em> and will cause primitive
+	file operations fail.
+      </p>
+    </note>
+    <warning><p>
+      Currently null characters at the end of the filename
+      will be accepted by primitive file operations. Such
+      filenames are however still documented as invalid. The
+      implementation will also change in the future and
+      reject such filenames.
+    </p></warning>
   </description>
   <datatypes>
       <datatype>
@@ -555,6 +582,55 @@ unsafe</pre>
 ["a:/","msdev","include"]</pre>
       </desc>
     </func>
+
+    <func>
+      <name name="validate" arity="1"/>
+      <fsummary>Validate encoding of filename</fsummary>
+      <desc>
+	<p>
+	  Validates filename encoding. Returns <c>true</c> if
+	  <c><anno>FileName</anno></c> has a valid encoding;
+	  otherwise, returns <c>false</c>.
+	</p>
+	<taglist>
+	  <tag>Ordinary Filename</tag>
+	  <item>
+	    <p>
+	      Type: <c><anno>FileName</anno> = </c><seealso marker="kernel:file#type-name"><c>file:name()</c></seealso>
+	    </p>
+	    <p>
+	      Validates encoding against the
+	      <seealso marker="kernel:file#native_name_encoding/0">native file
+	      name encoding</seealso>, and the
+	      capabilities of the operating system used.
+	      Regardless of configuration and OS, null
+	      characters (integer value zero) will be
+	      rejected by the validation  (even when only
+	      present at the end of the filename).
+	    </p>
+	  </item>
+	  <tag><seealso marker="unicode_usage#notes-about-raw-filenames">Raw
+	  Filename</seealso></tag>
+	  <item>
+	    <p>
+	      Type: <c><anno>FileName</anno> = binary()</c>
+	    </p>
+	    <p>
+	      The encoding will not be interpreted, but
+	      null bytes (integer value zero) will be
+	      rejected by the validation (even when only
+	      present at the end of the filename).
+	    </p>
+	  </item>
+	</taglist>
+	<p>
+	  For information on filename encoding see the documentation
+	  of unicode filenames in
+	  <seealso marker="stdlib:unicode_usage#unicode_file_names">STDLIB
+	  Users Guide ➜ Using Unicode in Erlang ➜ Unicode Filenames</seealso>.
+	</p>
+      </desc>
+    </func>
   </funcs>
 </erlref>