path: root/lib/kernel/doc/src/file.xml

<?xml version="1.0" encoding="latin1" ?>
<!DOCTYPE erlref SYSTEM "erlref.dtd">

<erlref>
  <header>
    <copyright>
      <year>1996</year><year>2013</year>
      <holder>Ericsson AB. All Rights Reserved.</holder>
    </copyright>
    <legalnotice>
      The contents of this file are subject to the Erlang Public License,
      Version 1.1, (the "License"); you may not use this file except in
      compliance with the License. You should have received a copy of the
      Erlang Public License along with this software. If not, it can be
      retrieved online at http://www.erlang.org/.

      Software distributed under the License is distributed on an "AS IS"
      basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See
      the License for the specific language governing rights and limitations
      under the License.

    </legalnotice>

    <title>file</title>
    <prepared></prepared>
    <docno></docno>
    <date></date>
    <rev></rev>
  </header>
  <module>file</module>
  <modulesummary>File Interface Module</modulesummary>
  <description>
    <p>The module <c>file</c> provides an interface to the file system.</p>
    <p>On operating systems with thread support, it is possible to let
      file operations be performed in threads of their own, allowing
      other Erlang processes to continue executing in parallel with
      the file operations. See the command line flag
      <c>+A</c> in <seealso marker="erts:erl">erl(1)</seealso>.</p>

    <p>The Erlang VM supports file names in Unicode to a limited
    extent.  Depending on how the VM is started (with the parameter
    <c>+fnu</c> or <c>+fnl</c>), file names given can contain
    characters > 255 and the VM system will convert file names
    back and forth to the native file name encoding.</p>

    <p>The default behavior for Unicode character translation depends
    on to what extent the underlying OS/filesystem enforces consistent
    naming. On OSes where all file names are ensured to be in one or
    another encoding, Unicode is the default (currently this holds for
    Windows and MacOSX). On OSes with completely transparent file
    naming (i.e. all Unixes except MacOSX), ISO-latin-1 file naming is
    the default. The reason for the ISO-latin-1 default is that
    file names are not guaranteed to be possible to interpret according to
    the Unicode encoding expected (i.e. UTF-8), and file names that
    cannot be decoded will only be accessible by using &quot;raw
    file names&quot;, in other word file names given as binaries.</p>
 
    <p>As file names are traditionally not binaries in Erlang,
    applications that need to handle raw file names need to be
    converted, why the Unicode mode for file names is not default on
    systems having completely transparent file naming.</p>

    <note>
      <p>As of R14B01, the most basic file handling modules
      (<c>file</c>, <c>prim_file</c>, <c>filelib</c> and
      <c>filename</c>) accept raw file names, but the rest of OTP is not
      guaranteed to handle them, why Unicode file naming on systems
      where it is not default is still considered experimental.</p>
    </note>

    <p>Raw file names is a new feature in OTP R14B01, which allows the
    user to supply completely uninterpreted file names to the
    underlying OS/filesystem. They are supplied as binaries, where it
    is up to the user to supply a correct encoding for the
    environment. The function <c>file:native_name_encoding()</c> can
    be used to check what encoding the VM is working in. If the
    function returns <c>latin1</c> file names are not in any way
    converted to Unicode, if it is <c>utf8</c>, raw file names should
    be encoded as UTF-8 if they are to follow the convention of the VM
    (and usually the convention of the OS as well). Using raw
    file names is useful if you have a filesystem with inconsistent
    file naming, where some files are named in UTF-8 encoding while
    others are not. A file:list_dir on such mixed file name systems
    when the VM is in Unicode file name mode might return file names as
    raw binaries as they cannot be interpreted as Unicode
    file names. Raw file names can also be used to give UTF-8 encoded
    file names even though the VM is not started in Unicode file name
    translation mode.</p>

    <p>Note that on Windows, <c>file:native_name_encoding()</c>
    returns <c>utf8</c> per default, which is the format for raw
    file names even on Windows, although the underlying OS specific
    code works in a limited version of little endian UTF16. As far as
    the Erlang programmer is concerned, Windows native Unicode format
    is UTF-8...</p>
  </description>

  <datatypes>
    <datatype>
      <name name="deep_list"/>
    </datatype>
    <datatype>
      <name><marker id="type-fd">fd()</marker></name>
      <desc>
        <p>A file descriptor representing a file opened in <seealso
          marker="#raw">raw</seealso> mode.</p>
      </desc>
    </datatype>
    <datatype>
      <name name="filename"/>
    </datatype>
    <datatype>
      <name name="filename_all"/>
    </datatype>
    <datatype>
      <name name="io_device"/>
      <desc>
        <p>As returned by
          <seealso marker="#open/2">file:open/2</seealso>;
          <c>pid()</c> is a process handling I/O-protocols.</p>
      </desc>
    </datatype>
    <datatype>
      <name name="name"/>
      <desc>
        <p>If VM is in Unicode filename mode, <c>string()</c> and <c>char()</c>
          are allowed to be > 255.
          <c><anno>RawFilename</anno></c> is a filename not subject to
          Unicode translation,
          meaning that it can contain characters not conforming to
          the Unicode encoding expected from the filesystem
          (i.e. non-UTF-8 characters although the VM is started
          in Unicode filename mode).
        </p>
      </desc>
    </datatype>
    <datatype>
      <name name="posix"/>
      <desc>
        <p>An atom which is named from the POSIX error codes used in
          Unix, and in the runtime libraries of most C compilers.</p>
      </desc>
    </datatype>
    <datatype>
      <name name="date_time"/>
      <desc>
        <p>Must denote a valid date and time.</p>
      </desc>
    </datatype>
    <datatype>
      <name name="file_info"/>
    </datatype>
    <datatype>
      <name name="location"/>
    </datatype>
    <datatype>
      <name name="mode"/>
    </datatype>
    <datatype>
      <name name="file_info_option"/>
    </datatype>
   </datatypes>

  <funcs>
    <func>
      <name name="advise" arity="4"/>
      <fsummary>Predeclare an access pattern for file data</fsummary>
      <type name="posix_file_advise"/>
      <desc>
        <p><c>advise/4</c> can be used to announce an intention to access file
          data in a specific pattern in the future, thus allowing the
          operating system to perform appropriate optimizations.</p>
        <p>On some platforms, this function might have no effect.</p>
      </desc>
    </func>
    <func>
      <name name="allocate" arity="3"/>
      <fsummary>Allocate file space</fsummary>
      <desc>
        <p><c>allocate/3</c> can be used to preallocate space for a file.</p>
        <p>This function only succeeds in platforms that implement this
          feature. When it succeeds, space is preallocated for the file but
          the file size might not be updated. This behaviour depends on the
          preallocation implementation. To guarantee the file size is updated
          one must truncate the file to the new size.</p>
      </desc>
    </func>
    <func>
      <name name="change_group" arity="2"/>
      <fsummary>Change group of a file</fsummary>
      <desc>
        <p>Changes group of a file. See
          <seealso marker="#write_file_info/2">write_file_info/2</seealso>.</p>
      </desc>
    </func>
    <func>
      <name name="change_mode" arity="2"/>
      <fsummary>Change permissions of a file</fsummary>
      <desc>
        <p>Changes permissions of a file. See
          <seealso marker="#write_file_info/2">write_file_info/2</seealso>.</p>
      </desc>
    </func>
    <func>
      <name name="change_owner" arity="2"/>
      <fsummary>Change owner of a file</fsummary>
      <desc>
        <p>Changes owner of a file. See
          <seealso marker="#write_file_info/2">write_file_info/2</seealso>.</p>
      </desc>
    </func>
    <func>
      <name name="change_owner" arity="3"/>
      <fsummary>Change owner and group of a file</fsummary>
      <desc>
        <p>Changes owner and group of a file. See
          <seealso marker="#write_file_info/2">write_file_info/2</seealso>.</p>
      </desc>
    </func>
    <func>
      <name name="change_time" arity="2"/>
      <fsummary>Change the modification time of a file</fsummary>
      <desc>
        <p>Changes the modification and access times of a file. See
          <seealso marker="#write_file_info/2">write_file_info/2</seealso>.</p>
      </desc>
    </func>
    <func>
      <name name="change_time" arity="3"/>
      <fsummary>Change the modification and last access time of a file</fsummary>
      <desc>
        <p>Changes the modification and last access times of a file. See
          <seealso marker="#write_file_info/2">write_file_info/2</seealso>.</p>
      </desc>
    </func>
    <func>
      <name name="close" arity="1"/>
      <fsummary>Close a file</fsummary>
      <desc>
        <p>Closes the file referenced by <c><anno>IoDevice</anno></c>. It mostly
          returns <c>ok</c>, expect for some severe errors such as out
          of memory.</p>
        <p>Note that if the option <c>delayed_write</c> was
          used when opening the file, <c>close/1</c> might return an
          old write error and not even try to close the file. See
          <seealso marker="#open/2">open/2</seealso>.</p>
      </desc>
    </func>
    <func>
      <name name="consult" arity="1"/>
      <fsummary>Read Erlang terms from a file</fsummary>
      <desc>
        <p>Reads Erlang terms, separated by '.', from
           <c><anno>Filename</anno></c>. Returns one of the following:</p>
        <taglist>
          <tag><c>{ok, <anno>Terms</anno>}</c></tag>
          <item>
            <p>The file was successfully read.</p>
          </item>
          <tag><c>{error, atom()}</c></tag>
          <item>
            <p>An error occurred when opening the file or reading it.
              See <seealso marker="#open/2">open/2</seealso> for a list
              of typical error codes.</p>
          </item>
          <tag><c>{error, {<anno>Line</anno>, <anno>Mod</anno>,
                           <anno>Term</anno>}}</c></tag>
          <item>
            <p>An error occurred when interpreting the Erlang terms in
              the file. Use <c>format_error/1</c> to convert
              the three-element tuple to an English description of
              the error.</p>
          </item>
        </taglist>
        <p>Example:</p>
<code type="none">f.txt:  {person, "kalle", 25}.
        {person, "pelle", 30}.</code>
<pre>1> <input>file:consult("f.txt").</input>
{ok,[{person,"kalle",25},{person,"pelle",30}]}</pre>
        <p>The encoding of of <c><anno>Filename</anno></c> can be set
          by a comment as described in <seealso
          marker="stdlib:epp#encoding">epp(3)</seealso>.</p>
      </desc>
    </func>
    <func>
      <name name="copy" arity="2"/>
      <name name="copy" arity="3"/>
      <fsummary>Copy file contents</fsummary>
      <desc>
        <p>Copies <c><anno>ByteCount</anno></c> bytes from
          <c><anno>Source</anno></c> to <c><anno>Destination</anno></c>.
          <c><anno>Source</anno></c> and <c><anno>Destination</anno></c> refer
          to either filenames or IO devices from e.g. <c>open/2</c>.
          <c><anno>ByteCount</anno></c> defaults to <c>infinity</c>, denoting an
          infinite number of bytes.</p>
        <p>The argument <c><anno>Modes</anno></c> is a list of possible modes,
          see <seealso marker="#open/2">open/2</seealso>, and defaults to
          [].</p>
        <p>If both <c><anno>Source</anno></c> and
          <c><anno>Destination</anno></c> refer to
          filenames, the files are opened with <c>[read, binary]</c>
          and <c>[write, binary]</c> prepended to their mode lists,
          respectively, to optimize the copy.</p>
        <p>If <c><anno>Source</anno></c> refers to a filename, it is opened with
          <c>read</c> mode prepended to the mode list before the copy,
          and closed when done.</p>
        <p>If <c><anno>Destination</anno></c> refers to a filename, it is opened
          with <c>write</c> mode prepended to the mode list before
          the copy, and closed when done.</p>
        <p>Returns <c>{ok, <anno>BytesCopied</anno>}</c> where
          <c><anno>BytesCopied</anno></c> is
          the number of bytes that actually was copied, which may be
          less than <c><anno>ByteCount</anno></c> if end of file was
          encountered on the source. If the operation fails,
          <c>{error, <anno>Reason</anno>}</c> is returned.</p>
        <p>Typical error reasons: As for <c>open/2</c> if a file had to
          be opened, and as for <c>read/2</c> and <c>write/2</c>.</p>
      </desc>
    </func>
    <func>
      <name name="del_dir" arity="1"/>
      <fsummary>Delete a directory</fsummary>
      <desc>
        <p>Tries to delete the directory <c><anno>Dir</anno></c>.
          The directory must
          be empty before it can be deleted. Returns <c>ok</c> if
          successful.</p>
        <p>Typical error reasons are:</p>
        <taglist>
          <tag><c>eacces</c></tag>
          <item>
            <p>Missing search or write permissions for the parent
              directories of <c><anno>Dir</anno></c>.</p>
          </item>
          <tag><c>eexist</c></tag>
          <item>
            <p>The directory is not empty.</p>
          </item>
          <tag><c>enoent</c></tag>
          <item>
            <p>The directory does not exist.</p>
          </item>
          <tag><c>enotdir</c></tag>
          <item>
            <p>A component of <c><anno>Dir</anno></c> is not a directory.
              On some platforms, <c>enoent</c> is returned instead.</p>
          </item>
          <tag><c>einval</c></tag>
          <item>
            <p>Attempt to delete the current directory. On some
              platforms, <c>eacces</c> is returned instead.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="delete" arity="1"/>
      <fsummary>Delete a file</fsummary>
      <desc>
        <p>Tries to delete the file <c><anno>Filename</anno></c>.
          Returns <c>ok</c> if successful.</p>
        <p>Typical error reasons are:</p>
        <taglist>
          <tag><c>enoent</c></tag>
          <item>
            <p>The file does not exist.</p>
          </item>
          <tag><c>eacces</c></tag>
          <item>
            <p>Missing permission for the file or one of its parents.</p>
          </item>
          <tag><c>eperm</c></tag>
          <item>
            <p>The file is a directory and the user is not super-user.</p>
          </item>
          <tag><c>enotdir</c></tag>
          <item>
            <p>A component of the file name is not a directory. On some
              platforms, <c>enoent</c> is returned instead.</p>
          </item>
          <tag><c>einval</c></tag>
          <item>
            <p><c><anno>Filename</anno></c> had an improper type, such as tuple.</p>
          </item>
        </taglist>
        <warning>
          <p>In a future release, a bad type for the
            <c><anno>Filename</anno></c> argument will probably generate
            an exception.</p>
        </warning>
      </desc>
    </func>
    <func>
      <name name="eval" arity="1"/>
      <fsummary>Evaluate Erlang expressions in a file</fsummary>
      <desc>
        <p>Reads and evaluates Erlang expressions, separated by '.' (or
          ',', a sequence of expressions is also an expression), from
          <c><anno>Filename</anno></c>. The actual result of the evaluation
          is not returned; any expression sequence in the file must be there
          for its side effect. Returns one of the following:</p>
        <taglist>
          <tag><c>ok</c></tag>
          <item>
            <p>The file was read and evaluated.</p>
          </item>
          <tag><c>{error, atom()}</c></tag>
          <item>
            <p>An error occurred when opening the file or reading it.
              See <c>open/2</c> for a list of typical error codes.</p>
          </item>
          <tag><c>{error, {<anno>Line</anno>, <anno>Mod</anno>,
                           <anno>Term</anno>}}</c></tag>
          <item>
            <p>An error occurred when interpreting the Erlang
              expressions in the file. Use <c>format_error/1</c> to
              convert the three-element tuple to an English description
              of the error.</p>
          </item>
        </taglist>
        <p>The encoding of of <c><anno>Filename</anno></c> can be set
          by a comment as described in <seealso
          marker="stdlib:epp#encoding">epp(3)</seealso>.</p>
      </desc>
    </func>
    <func>
      <name name="eval" arity="2"/>
      <fsummary>Evaluate Erlang expressions in a file</fsummary>
      <desc>
        <p>The same as <c>eval/1</c> but the variable bindings
          <c><anno>Bindings</anno></c> are used in the evaluation. See
          <seealso marker="stdlib:erl_eval">erl_eval(3)</seealso> about
          variable bindings.</p>
      </desc>
    </func>
    <func>
      <name name="file_info" arity="1"/>
      <fsummary>Get information about a file (deprecated)</fsummary>
      <desc>
        <p>This function is obsolete. Use <c>read_file_info/1,2</c>
          instead.</p>
      </desc>
    </func>
    <func>
      <name name="format_error" arity="1"/>
      <fsummary>Return a descriptive string for an error reason</fsummary>
      <desc>
        <p>Given the error reason returned by any function in this
          module, returns a descriptive string of the error in English.</p>
      </desc>
    </func>
    <func>
      <name name="get_cwd" arity="0"/>
      <fsummary>Get the current working directory</fsummary>
      <desc>
        <p>Returns <c>{ok, <anno>Dir</anno>}</c>, where <c><anno>Dir</anno></c>
          is the current
          working directory of the file server.</p>
        <note>
          <p>In rare circumstances, this function can fail on Unix.
            It may happen if read permission does not exist for
            the parent directories of the current directory.</p>
        </note>
        <p>Typical error reasons are:</p>
        <taglist>
          <tag><c>eacces</c></tag>
          <item>
            <p>Missing read permission for one of the parents of
              the current directory.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="get_cwd" arity="1"/>
      <fsummary>Get the current working directory for the drive specified</fsummary>
      <desc>
        <p><c><anno>Drive</anno></c> should be of the form
          "<c>Letter</c><c>:</c>",
          for example "c:". Returns <c>{ok, <anno>Dir</anno>}</c> or
          <c>{error, <anno>Reason</anno>}</c>, where <c><anno>Dir</anno></c>
          is the current
          working directory of the drive specified.</p>
        <p>This function returns <c>{error, enotsup}</c> on platforms
          which have no concept of current drive (Unix, for example).</p>
        <p>Typical error reasons are:</p>
        <taglist>
          <tag><c>enotsup</c></tag>
          <item>
            <p>The operating system has no concept of drives.</p>
          </item>
          <tag><c>eacces</c></tag>
          <item>
            <p>The drive does not exist.</p>
          </item>
          <tag><c>einval</c></tag>
          <item>
            <p>The format of <c><anno>Drive</anno></c> is invalid.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="list_dir" arity="1"/>
      <fsummary>List files in a directory</fsummary>
      <desc>
        <p>Lists all the files in a directory. Returns
          <c>{ok, <anno>Filenames</anno>}</c> if successful.
          Otherwise, it returns <c>{error, <anno>Reason</anno>}</c>.
          <c><anno>Filenames</anno></c> is a list of
          the names of all the files in the directory. The names are
          not sorted.</p>
        <p>Typical error reasons are:</p>
        <taglist>
          <tag><c>eacces</c></tag>
          <item>
            <p>Missing search or write permissions for <c><anno>Dir</anno></c>
              or one of its parent directories.</p>
          </item>
          <tag><c>enoent</c></tag>
          <item>
            <p>The directory does not exist.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="make_dir" arity="1"/>
      <fsummary>Make a directory</fsummary>
      <desc>
        <p>Tries to create the directory <c><anno>Dir</anno></c>. Missing parent
          directories are <em>not</em> created. Returns <c>ok</c> if
          successful.</p>
        <p>Typical error reasons are:</p>
        <taglist>
          <tag><c>eacces</c></tag>
          <item>
            <p>Missing search or write permissions for the parent
              directories of <c><anno>Dir</anno></c>.</p>
          </item>
          <tag><c>eexist</c></tag>
          <item>
            <p>There is already a file or directory named <c><anno>Dir</anno></c>.</p>
          </item>
          <tag><c>enoent</c></tag>
          <item>
            <p>A component of <c><anno>Dir</anno></c> does not exist.</p>
          </item>
          <tag><c>enospc</c></tag>
          <item>
            <p>There is a no space left on the device.</p>
          </item>
          <tag><c>enotdir</c></tag>
          <item>
            <p>A component of <c><anno>Dir</anno></c> is not a directory.
              On some platforms, <c>enoent</c> is returned instead.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="make_link" arity="2"/>
      <fsummary>Make a hard link to a file</fsummary>
      <desc>
        <p>Makes a hard link from <c><anno>Existing</anno></c> to
          <c><anno>New</anno></c>, on
          platforms that support links (Unix and Windows). This function returns
          <c>ok</c> if the link was successfully created, or
          <c>{error, <anno>Reason</anno>}</c>. On platforms that do not support
          links, <c>{error,enotsup}</c> is returned.</p>
        <p>Typical error reasons:</p>
        <taglist>
          <tag><c>eacces</c></tag>
          <item>
            <p>Missing read or write permissions for the parent
              directories of <c><anno>Existing</anno></c> or
              <c><anno>New</anno></c>.</p>
          </item>
          <tag><c>eexist</c></tag>
          <item>
            <p><c><anno>New</anno></c> already exists.</p>
          </item>
          <tag><c>enotsup</c></tag>
          <item>
            <p>Hard links are not supported on this platform.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="make_symlink" arity="2"/>
      <fsummary>Make a symbolic link to a file or directory</fsummary>
      <desc>
        <p>This function creates a symbolic link <c><anno>New</anno></c> to
          the file or directory <c><anno>Existing</anno></c>, on platforms that
          support symbolic links (most Unix systems and Windows beginning with
	  Vista).
	  <c><anno>Existing</anno></c> need not exist.
          This function returns <c>ok</c> if the link was
          successfully created, or <c>{error, <anno>Reason</anno>}</c>.
          On platforms
          that do not support symbolic links, <c>{error, enotsup}</c>
          is returned.</p>
        <p>Typical error reasons:</p>
        <taglist>
          <tag><c>eacces</c></tag>
          <item>
            <p>Missing read or write permissions for the parent directories
            of <c><anno>Existing</anno></c> or <c><anno>New</anno></c>.</p>
          </item>
          <tag><c>eexist</c></tag>
          <item>
            <p><c><anno>New</anno></c> already exists.</p>
          </item>
          <tag><c>enotsup</c></tag>
          <item>
            <p>Symbolic links are not supported on this platform.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="native_name_encoding" arity="0"/>
      <fsummary>Return the VM's configured filename encoding.</fsummary>
      <desc>
        <p>This function returns the configured default file name encoding to use for raw file names. Generally an application supplying file names raw (as binaries), should obey the character encoding returned by this function.</p>
	<p>By default, the VM uses ISO-latin-1 file name encoding on filesystems and/or OSes that use completely transparent file naming. This includes all Unix versions except MacOSX, where the vfs layer enforces UTF-8 file naming. By giving the experimental option <c>+fnu</c> when starting Erlang, UTF-8 translation of file names can be turned on even for those systems. If Unicode file name translation is in effect, the system behaves as usual as long as file names conform to the encoding, but will return file names that are not properly encoded in UTF-8 as raw file names (i.e. binaries).</p> 
	<p>On Windows, this function also returns <c>utf8</c> by default. The OS uses a pure Unicode naming scheme and file names are always possible to interpret as valid Unicode. The fact that the underlying Windows OS actually encodes file names using little endian UTF-16 can be ignored by the Erlang programmer. Windows and MacOSX are the only operating systems where the VM operates in Unicode file name mode by default.</p>
      </desc>
    </func>
    <func>
      <name name="open" arity="2"/>
      <fsummary>Open a file</fsummary>
      <desc>
        <p>Opens the file <c><anno>File</anno></c> in the mode determined
          by <c><anno>Modes</anno></c>, which may contain one or more of the
          following items:</p>
        <taglist>
          <tag><c>read</c></tag>
          <item>
            <p>The file, which must exist, is opened for reading.</p>
          </item>
          <tag><c>write</c></tag>
          <item>
            <p>The file is opened for writing. It is created if it does
              not exist. If the file exists, and if <c>write</c> is not
              combined with <c>read</c>, the file will be truncated.</p>
          </item>
          <tag><c>append</c></tag>
          <item>
            <p>The file will be opened for writing, and it will be
              created if it does not exist. Every write operation to a
              file opened with <c>append</c> will take place at
              the end of the file.</p>
          </item>
          <tag><c>exclusive</c></tag>
          <item>
            <p>The file, when opened for writing, is created if it
              does not exist. If the file exists, open will return
              <c>{error, eexist}</c>.</p>
	    <warning><p>This option does not guarantee exclusiveness on
	    file systems that do not support O_EXCL properly,
	    such as NFS. Do not depend on this option unless you
	    know that the file system supports it (in general, local
	    file systems should be safe).</p></warning>
          </item>
          <tag><c>raw</c></tag>
          <item>
            <p><marker id="raw"/>
              The <c>raw</c> option allows faster access to a file,
              because no Erlang process is needed to handle the file.
              However, a file opened in this way has the following
              limitations:</p>
            <list type="bulleted">
              <item>The functions in the <c>io</c> module cannot be used,
               because they can only talk to an Erlang process.
               Instead, use the <c>read/2</c>, <c>read_line/1</c> and 
	       <c>write/2</c>
               functions.</item>
	      <item>Especially if <c>read_line/1</c> is to be used on a <c>raw</c> file, it is recommended to combine this option with the <c>{read_ahead, Size}</c> option as line oriented I/O is inefficient without buffering.</item>
              <item>Only the Erlang process which opened the file can use
               it.</item>
              <item>A remote Erlang file server cannot be used;
               the computer on which the Erlang node is running must
               have access to the file system (directly or through
               NFS).</item>
            </list>
          </item>
          <tag><c>binary</c></tag>
          <item>
            <p>When this option has been given, read operations on the file
	       will return binaries rather than lists.</p>
          </item>
          <tag><c>{delayed_write, Size, Delay}</c></tag>
          <item>
            <p>If this option is used, the data in subsequent
              <c>write/2</c> calls is buffered until there are at least
              <c>Size</c> bytes buffered, or until the oldest buffered
              data is <c>Delay</c> milliseconds old. Then all buffered
              data is written in one operating system call.
              The buffered data is also flushed before some other file
              operation than <c>write/2</c> is executed.</p>
            <p>The purpose of this option is to increase performance
              by reducing the number of operating system calls, so the
              <c>write/2</c> calls should be for sizes significantly
              less than <c>Size</c>, and not interspersed by to many
              other file operations, for this to happen.</p>
            <p>When this option is used, the result of <c>write/2</c>
              calls may prematurely be reported as successful, and if
              a write error should actually occur the error is
              reported as the result of the next file operation, which
              is not executed.</p>
            <p>For example, when <c>delayed_write</c> is used, after a
              number of <c>write/2</c> calls, <c>close/1</c> might
              return <c>{error, enospc}</c> because there was not enough
              space on the disc for previously written data, and
              <c>close/1</c> should probably be called again since the
              file is still open.</p>
          </item>
          <tag><c>delayed_write</c></tag>
          <item>
            <p>The same as <c>{delayed_write, Size, Delay}</c> with
              reasonable default values for <c>Size</c> and
              <c>Delay</c>. (Roughly some 64 KBytes, 2 seconds)</p>
          </item>
          <tag><c>{read_ahead, Size}</c></tag>
          <item>
            <p>This option activates read data buffering. If
              <c>read/2</c> calls are for significantly less than
              <c>Size</c> bytes, read operations towards the operating
              system are still performed for blocks of <c>Size</c>
              bytes. The extra data is buffered and returned in
              subsequent <c>read/2</c> calls, giving a performance gain
              since the number of operating system calls is reduced.</p>
	    <p>The <c>read_ahead</c> buffer is also highly utilized
	      by the <c>read_line/1</c> function in <c>raw</c> mode,
	      why this option is recommended (for performance reasons)
	      when accessing raw files using that function.</p>
            <p>If <c>read/2</c> calls are for sizes not significantly
              less than, or even greater than <c>Size</c> bytes, no
              performance gain can be expected.</p>
          </item>
          <tag><c>read_ahead</c></tag>
          <item>
            <p>The same as <c>{read_ahead, Size}</c> with a reasonable
              default value for <c>Size</c>. (Roughly some 64 KBytes)</p>
          </item>
          <tag><c>compressed</c></tag>
          <item>
            <p>Makes it possible to read or write gzip compressed
              files. The <c>compressed</c> option must be combined
	      with either <c>read</c> or <c>write</c>, but not both.
	      Note that the file size obtained with
              <c>read_file_info/1</c> will most probably not match the
              number of bytes that can be read from a compressed file.</p>
          </item>
          <tag><c>{encoding, Encoding}</c></tag>
          <item>
            <p>Makes the file perform automatic translation of characters to and from a specific (Unicode) encoding. Note that the data supplied to file:write or returned by file:read still is byte oriented, this option only denotes how data is actually stored in the disk file.</p>
	    <p>Depending on the encoding, different methods of reading and writing data is preferred. The default encoding of <c>latin1</c> implies using this (the file) module for reading and writing data, as the interfaces provided here work with byte-oriented data, while using other (Unicode) encodings makes the  <seealso marker="stdlib:io">io(3)</seealso> module's <c>get_chars</c>, <c>get_line</c> and <c>put_chars</c> functions more suitable, as they can work with the full Unicode range.</p>
	    <p>If data is sent to an <c>io_device()</c> in a format that cannot be converted to the specified encoding, or if data is read by a function that returns data in a format that cannot cope with the character range of the data, an error occurs and the file will be closed.</p> 
	    <p>The allowed values for <c>Encoding</c> are:</p>
	    <taglist>
	    <tag><c>latin1</c></tag>
	    <item>
	    <p>The default encoding. Bytes supplied to i.e. file:write are written as is on the file, likewise bytes read from the file are returned to i.e. file:read as is. If the <seealso marker="stdlib:io">io(3)</seealso> module is used for writing, the file can only cope with Unicode characters up to codepoint 255 (the ISO-latin-1 range).</p>
	    </item>
	    <tag><c>unicode</c> or <c>utf8</c></tag>
	    <item>
	    <p>Characters are translated to and from the UTF-8 encoding before being written to or read from the file. A file opened in this way might be readable using the file:read function, as long as no data stored on the file lies beyond the ISO-latin-1 range (0..255), but failure will occur if the data contains Unicode codepoints beyond that range. The file is best read with the functions in the Unicode aware   <seealso marker="stdlib:io">io(3)</seealso> module.</p>
	    <p>Bytes written to the file by any means are translated to UTF-8 encoding before actually being stored on the disk file.</p>
	    </item>
	    <tag><c>utf16</c> or <c>{utf16,big}</c></tag>
	    <item>
	    <p>Works like <c>unicode</c>, but translation is done to and from big endian UTF-16 instead of UTF-8.</p>
	    </item>
	    <tag><c>{utf16,little}</c></tag>
	    <item>
	    <p>Works like <c>unicode</c>, but translation is done to and from little endian UTF-16 instead of UTF-8.</p>
	    </item>
	    <tag><c>utf32</c> or <c>{utf32,big}</c></tag>
	    <item>
	    <p>Works like <c>unicode</c>, but translation is done to and from big endian UTF-32 instead of UTF-8.</p>
	    </item>
	    <tag><c>{utf32,little}</c></tag>
	    <item>
	    <p>Works like <c>unicode</c>, but translation is done to and from little endian UTF-32 instead of UTF-8.</p>
	    </item>
	    </taglist>
	    <p>The Encoding can be changed for a file "on the fly" by using the <seealso marker="stdlib:io#setopts/2">io:setopts/2</seealso> function, why a file can be analyzed in latin1 encoding for i.e. a BOM, positioned beyond the BOM and then be set for the right encoding before further reading.See the <seealso marker="stdlib:unicode">unicode(3)</seealso> module for functions identifying BOM's.</p>
	    <p>This option is not allowed on <c>raw</c> files.</p>
          </item>
          <tag><c>ram</c></tag>
          <item>
            <p><c>File</c> must be <c>iodata()</c>. Returns an <c>fd()</c> which lets the <c>file</c> module operate on the data in-memory as if it is a file.</p>
          </item>
        </taglist>
        <p>Returns:</p>
        <taglist>
          <tag><c>{ok, <anno>IoDevice</anno>}</c></tag>
          <item>
            <p>The file has been opened in the requested mode.
              <c><anno>IoDevice</anno></c> is a reference to the file.</p>
          </item>
          <tag><c>{error, <anno>Reason</anno>}</c></tag>
          <item>
            <p>The file could not be opened.</p>
          </item>
        </taglist>
        <p><c><anno>IoDevice</anno></c> is really the pid of the process which
          handles the file. This process is linked to the process
          which originally opened the file. If any process to which
          the <c><anno>IoDevice</anno></c> is linked terminates, the file will
          be closed and the process itself will be terminated.
          An <c><anno>IoDevice</anno></c> returned from this call can be used
          as an argument to the IO functions (see
          <seealso marker="stdlib:io">io(3)</seealso>).</p>
        <note>
          <p>In previous versions of <c>file</c>, modes were given
            as one of the atoms <c>read</c>, <c>write</c>, or
            <c>read_write</c> instead of a list. This is still allowed
            for reasons of backwards compatibility, but should not be
            used for new code. Also note that <c>read_write</c> is not
            allowed in a mode list.</p>
        </note>
        <p>Typical error reasons:</p>
        <taglist>
          <tag><c>enoent</c></tag>
          <item>
            <p>The file does not exist.</p>
          </item>
          <tag><c>eacces</c></tag>
          <item>
            <p>Missing permission for reading the file or searching one
              of the parent directories.</p>
          </item>
          <tag><c>eisdir</c></tag>
          <item>
            <p>The named file is not a regular file. It may be a
              directory, a fifo, or a device.</p>
          </item>
          <tag><c>enotdir</c></tag>
          <item>
            <p>A component of the file name is not a directory. On some
              platforms, <c>enoent</c> is returned instead.</p>
          </item>
          <tag><c>enospc</c></tag>
          <item>
            <p>There is a no space left on the device (if <c>write</c>
              access was specified).</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="path_consult" arity="2"/>
      <fsummary>Read Erlang terms from a file</fsummary>
      <desc>
        <p>Searches the path <c><anno>Path</anno></c> (a list of directory
          names) until the file <c><anno>Filename</anno></c> is found.
          If <c><anno>Filename</anno></c>
          is an absolute filename, <c><anno>Path</anno></c> is ignored.
          Then reads Erlang terms, separated by '.', from the file.
          Returns one of the following:</p>
        <taglist>
          <tag><c>{ok, <anno>Terms</anno>, <anno>FullName</anno>}</c></tag>
          <item>
            <p>The file was successfully read. <c><anno>FullName</anno></c> is
              the full name of the file.</p>
          </item>
          <tag><c>{error, enoent}</c></tag>
          <item>
            <p>The file could not be found in any of the directories in
              <c><anno>Path</anno></c>.</p>
          </item>
          <tag><c>{error, atom()}</c></tag>
          <item>
            <p>An error occurred when opening the file or reading it.
              See <seealso marker="#open/2">open/2</seealso> for a list
              of typical error codes.</p>
          </item>
          <tag><c>{error, {<anno>Line</anno>, <anno>Mod</anno>,
                           <anno>Term</anno>}}</c></tag>
          <item>
            <p>An error occurred when interpreting the Erlang terms in
              the file. Use <c>format_error/1</c> to convert
              the three-element tuple to an English description of
              the error.</p>
          </item>
        </taglist>
        <p>The encoding of of <c><anno>Filename</anno></c> can be set
          by a comment as described in <seealso
          marker="stdlib:epp#encoding">epp(3)</seealso>.</p>
      </desc>
    </func>
    <func>
      <name name="path_eval" arity="2"/>
      <fsummary>Evaluate Erlang expressions in a file</fsummary>
      <desc>
        <p>Searches the path <c><anno>Path</anno></c> (a list of directory
          names) until the file <c><anno>Filename</anno></c> is found.
          If <c><anno>Filename</anno></c> is an absolute file name,
          <c><anno>Path</anno></c> is ignored. Then reads
          and evaluates Erlang expressions, separated by '.' (or ',', a
          sequence of expressions is also an expression), from the file.
          The actual result of evaluation is not returned; any
          expression sequence in the file must be there for its side
          effect. Returns one of the following:</p>
        <taglist>
          <tag><c>{ok, <anno>FullName</anno>}</c></tag>
          <item>
            <p>The file was read and evaluated. <c><anno>FullName</anno></c> is
              the full name of the file.</p>
          </item>
          <tag><c>{error, enoent}</c></tag>
          <item>
            <p>The file could not be found in any of the directories in
              <c><anno>Path</anno></c>.</p>
          </item>
          <tag><c>{error, atom()}</c></tag>
          <item>
            <p>An error occurred when opening the file or reading it.
              See <seealso marker="#open/2">open/2</seealso> for a list
              of typical error codes.</p>
          </item>
          <tag><c>{error, {<anno>Line</anno>, <anno>Mod</anno>,
                           <anno>Term</anno>}}</c></tag>
          <item>
            <p>An error occurred when interpreting the Erlang
              expressions in the file. Use <c>format_error/1</c> to
              convert the three-element tuple to an English description
              of the error.</p>
          </item>
        </taglist>
        <p>The encoding of of <c><anno>Filename</anno></c> can be set
          by a comment as described in <seealso
          marker="stdlib:epp#encoding">epp(3)</seealso>.</p>
      </desc>
    </func>
    <func>
      <name name="path_open" arity="3"/>
      <fsummary>Open a file</fsummary>
      <desc>
        <p>Searches the path <c><anno>Path</anno></c> (a list of directory
          names) until the file <c><anno>Filename</anno></c> is found.
          If <c><anno>Filename</anno></c>
          is an absolute file name, <c><anno>Path</anno></c> is ignored.
          Then opens the file in the mode determined by <c><anno>Modes</anno></c>.
          Returns one of the following:</p>
        <taglist>
          <tag><c>{ok, <anno>IoDevice</anno>, <anno>FullName</anno>}</c></tag>
          <item>
            <p>The file has been opened in the requested mode.
              <c><anno>IoDevice</anno></c> is a reference to the file and
              <c><anno>FullName</anno></c> is the full name of the file.</p>
          </item>
          <tag><c>{error, enoent}</c></tag>
          <item>
            <p>The file could not be found in any of the directories in
              <c><anno>Path</anno></c>.</p>
          </item>
          <tag><c>{error, atom()}</c></tag>
          <item>
            <p>The file could not be opened.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="path_script" arity="2"/>
      <fsummary>Evaluate and return the value of Erlang expressions in a file</fsummary>
      <desc>
        <p>Searches the path <c><anno>Path</anno></c> (a list of directory
          names) until the file <c><anno>Filename</anno></c> is found.
          If <c><anno>Filename</anno></c> is an absolute file name,
          <c><anno>Path</anno></c> is ignored. Then reads
          and evaluates Erlang expressions, separated by '.' (or ',', a
          sequence of expressions is also an expression), from the file.
          Returns one of the following:</p>
        <taglist>
          <tag><c>{ok, <anno>Value</anno>, <anno>FullName</anno>}</c></tag>
          <item>
            <p>The file was read and evaluated. <c><anno>FullName</anno></c> is
              the full name of the file and <c><anno>Value</anno></c> the value of
              the last expression.</p>
          </item>
          <tag><c>{error, enoent}</c></tag>
          <item>
            <p>The file could not be found in any of the directories in
              <c><anno>Path</anno></c>.</p>
          </item>
          <tag><c>{error, atom()}</c></tag>
          <item>
            <p>An error occurred when opening the file or reading it.
              See <seealso marker="#open/2">open/2</seealso> for a list
              of typical error codes.</p>
          </item>
          <tag><c>{error, {<anno>Line</anno>, <anno>Mod</anno>,
                           <anno>Term</anno>}}</c></tag>
          <item>
            <p>An error occurred when interpreting the Erlang
              expressions in the file. Use <c>format_error/1</c> to
              convert the three-element tuple to an English description
              of the error.</p>
          </item>
        </taglist>
        <p>The encoding of of <c><anno>Filename</anno></c> can be set
          by a comment as described in <seealso
          marker="stdlib:epp#encoding">epp(3)</seealso>.</p>
   </desc>
    </func>
    <func>
      <name name="path_script" arity="3"/>
      <fsummary>Evaluate and return the value of Erlang expressions in a file</fsummary>
      <desc>
        <p>The same as <c>path_script/2</c> but the variable bindings
          <c><anno>Bindings</anno></c> are used in the evaluation. See
          <seealso marker="stdlib:erl_eval">erl_eval(3)</seealso> about
          variable bindings.</p>
      </desc>
    </func>
    <func>
      <name name="pid2name" arity="1"/>
      <fsummary>Return the name of the file handled by a pid</fsummary>
      <desc>
        <p>If <c><anno>Pid</anno></c> is an IO device, that is, a pid returned from
          <c>open/2</c>, this function returns the filename, or rather:</p>
        <taglist>
          <tag><c>{ok, <anno>Filename</anno>}</c></tag>
          <item>
            <p>If this node's file server is not a slave, the file was
              opened by this node's file server, (this implies that
              <c><anno>Pid</anno></c> must be a local pid) and the file is not
              closed. <c><anno>Filename</anno></c> is the filename in flat string
              format.</p>
          </item>
          <tag><c>undefined</c></tag>
          <item>
            <p>In all other cases.</p>
          </item>
        </taglist>
        <warning>
          <p>This function is intended for debugging only.</p>
        </warning>
      </desc>
    </func>
    <func>
      <name name="position" arity="2"/>
      <fsummary>Set position in a file</fsummary>
      <desc>
        <p>Sets the position of the file referenced by <c><anno>IoDevice</anno></c>
          to <c><anno>Location</anno></c>. Returns
          <c>{ok, <anno>NewPosition</anno>}</c> (as
          absolute offset) if successful, otherwise
          <c>{error, <anno>Reason</anno>}</c>. <c><anno>Location</anno></c> is
          one of the following:</p>
        <taglist>
          <tag><c>Offset</c></tag>
          <item>
            <p>The same as <c>{bof, Offset}</c>.</p>
          </item>
          <tag><c>{bof, Offset}</c></tag>
          <item>
            <p>Absolute offset.</p>
          </item>
          <tag><c>{cur, Offset}</c></tag>
          <item>
            <p>Offset from the current position.</p>
          </item>
          <tag><c>{eof, Offset}</c></tag>
          <item>
            <p>Offset from the end of file.</p>
          </item>
          <tag><c>bof | cur | eof</c></tag>
          <item>
            <p>The same as above with <c>Offset</c> 0.</p>
          </item>
        </taglist>
	<p>Note that offsets are counted in bytes, not in characters. If the file is opened using some other <c>encoding</c> than <c>latin1</c>, one byte does not correspond to one character. Positioning in such a file can only be done to known character boundaries, i.e. to a position earlier retrieved by getting a current position, to the beginning/end of the file or to some other position <em>known</em> to be on a correct character boundary by some other means (typically beyond a byte order mark in the file, which has a known byte-size).</p>
        <p>Typical error reasons are:</p>
        <taglist>
          <tag><c>einval</c></tag>
          <item>
            <p>Either <c><anno>Location</anno></c> was illegal, or it
              evaluated to a
              negative offset in the file. Note that if the resulting
              position is a negative value, the result is an error, and
              after the call the file position is undefined.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="pread" arity="2"/>
      <fsummary>Read from a file at certain positions</fsummary>
      <desc>
        <p>Performs a sequence of <c>pread/3</c> in one operation, 
          which is more efficient than calling them one at a time.
          Returns <c>{ok, [<anno>Data</anno>, ...]}</c> or
          <c>{error, <anno>Reason</anno>}</c>,
          where each <c><anno>Data</anno></c>, the result of the corresponding
          <c>pread</c>, is either a list or a binary depending on
          the mode of the file, or <c>eof</c> if the requested position
          was beyond end of file.</p>
	  <p>As the position is given as a byte-offset, special caution has to be taken when working with files where <c>encoding</c> is set to something else than <c>latin1</c>, as not every byte position will be a valid character boundary on such a file.</p>
      </desc>
    </func>
    <func>
      <name name="pread" arity="3"/>
      <fsummary>Read from a file at a certain position</fsummary>
      <desc>
        <p>Combines <c>position/2</c> and <c>read/2</c> in one
          operation, which is more efficient than calling them one at a
          time. If <c><anno>IoDevice</anno></c> has been opened in raw mode,
          some restrictions apply: <c><anno>Location</anno></c> is only allowed
          to be an
          integer; and the current position of the file is undefined
          after the operation.</p>
	  <p>As the position is given as a byte-offset, special caution has to be taken when working with files where <c>encoding</c> is set to something else than <c>latin1</c>, as not every byte position will be a valid character boundary on such a file.</p>
        </desc>
    </func>
    <func>
      <name name="pwrite" arity="2"/>
      <fsummary>Write to a file at certain positions</fsummary>
      <desc>
        <p>Performs a sequence of <c>pwrite/3</c> in one operation, 
          which is more efficient than calling them one at a time.
          Returns <c>ok</c> or <c>{error, {<anno>N</anno>,
          <anno>Reason</anno>}}</c>, where
          <c><anno>N</anno></c> is the number of successful writes that was done
          before the failure.</p>
	<p>When positioning in a file with other <c>encoding</c> than <c>latin1</c>, caution must be taken to set the position on a correct character boundary, see <seealso marker="#position/2">position/2</seealso> for details.</p>
      </desc>
    </func>
    <func>
      <name name="pwrite" arity="3"/>
      <fsummary>Write to a file at a certain position</fsummary>
      <desc>
        <p>Combines <c>position/2</c> and <c>write/2</c> in one
          operation, which is more efficient than calling them one at a
          time. If <c><anno>IoDevice</anno></c> has been opened in raw mode,
          some restrictions apply: <c><anno>Location</anno></c> is only allowed
          to be an
          integer; and the current position of the file is undefined
          after the operation.</p>
	<p>When positioning in a file with other <c>encoding</c> than <c>latin1</c>, caution must be taken to set the position on a correct character boundary, see <seealso marker="#position/2">position/2</seealso> for details.</p>
      </desc>
    </func>
    <func>
      <name name="read" arity="2"/>
      <fsummary>Read from a file</fsummary>
      <desc>
        <p>Reads <c><anno>Number</anno></c> bytes/characters from the file
          referenced by <c><anno>IoDevice</anno></c>. The functions
          <c>read/2</c>, <c>pread/3</c>
          and <c>read_line/1</c> are the only ways to read from a file
          opened in raw mode (although they work for normally opened
          files, too).</p>
	  <p>For files where <c>encoding</c> is set to something else than <c>latin1</c>, one character might be represented by more than one byte on the file. The parameter <c>Number</c> always denotes the number of <em>characters</em> read from the file, why the position in the file might be moved a lot more than this number when reading a Unicode file.</p>
	  <p>Also if <c>encoding</c> is set to something else than <c>latin1</c>, the <c>read/3</c> call will fail if the data contains characters larger than 255, why the <seealso marker="stdlib:io">io(3)</seealso> module is to be preferred when reading such a file.</p> 
	  <p>The function returns:</p>
        <taglist>
          <tag><c>{ok, <anno>Data</anno>}</c></tag>
          <item>
            <p>If the file was opened in binary mode, the read bytes are
              returned in a binary, otherwise in a list. The list or
              binary will be shorter than the number of bytes requested
              if end of file was reached.</p>
          </item>
          <tag><c>eof</c></tag>
          <item>
            <p>Returned if <c><anno>Number</anno>>0</c> and end of file was
              reached before anything at all could be read.</p>
          </item>
          <tag><c>{error, <anno>Reason</anno>}</c></tag>
          <item>
            <p>An error occurred.</p>
          </item>
        </taglist>
        <p>Typical error reasons:</p>
        <taglist>
          <tag><c>ebadf</c></tag>
          <item>
            <p>The file is not opened for reading.</p>
          </item>
          <tag><c>{no_translation, unicode, latin1}</c></tag>
          <item>
            <p>The file was opened with another <c>encoding</c> than <c>latin1</c> and the data in the file can not be translated to the byte-oriented data that this function returns.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="read_file" arity="1"/>
      <fsummary>Read a file</fsummary>
      <desc>
        <p>Returns <c>{ok, <anno>Binary</anno>}</c>, where
          <c><anno>Binary</anno></c> is a binary
          data object that contains the contents of
          <c><anno>Filename</anno></c>, or
          <c>{error, <anno>Reason</anno>}</c> if an error occurs.</p>
        <p>Typical error reasons:</p>
        <taglist>
          <tag><c>enoent</c></tag>
          <item>
            <p>The file does not exist.</p>
          </item>
          <tag><c>eacces</c></tag>
          <item>
            <p>Missing permission for reading the file, or for
              searching one of the parent directories.</p>
          </item>
          <tag><c>eisdir</c></tag>
          <item>
            <p>The named file is a directory.</p>
          </item>
          <tag><c>enotdir</c></tag>
          <item>
            <p>A component of the file name is not a directory. On some
              platforms, <c>enoent</c> is returned instead.</p>
          </item>
          <tag><c>enomem</c></tag>
          <item>
            <p>There is not enough memory for the contents of the file.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="read_file_info" arity="1"/>
      <name name="read_file_info" arity="2"/>
      <fsummary>Get information about a file</fsummary>
      <desc>
        <p>Retrieves information about a file. Returns
          <c>{ok, <anno>FileInfo</anno>}</c> if successful, otherwise
          <c>{error, <anno>Reason</anno>}</c>. <c><anno>FileInfo</anno></c>
          is a record
          <c>file_info</c>, defined in the Kernel include file
          <c>file.hrl</c>. Include the following directive in the module
          from which the function is called:</p>
        <code type="none">
-include_lib("kernel/include/file.hrl").</code>
		<p>The time type returned in <c>atime</c>, <c>mtime</c> and <c>ctime</c>
			is dependent on the time type set in <c>Opts :: {time, Type}</c>.
			Type <c>local</c> will return local time, <c>universal</c> will
			return universal time and <c>posix</c> will return seconds since
			or before unix time epoch which is 1970-01-01 00:00 UTC. 
			Default is <c>{time, local}</c>.
		</p> 
		<note>
          <p>
			  Since file times is stored in posix time on most OS it is
			  faster to query file information with the <c>posix</c> option.
		  </p> 
        </note>
        
        <p>The record <c>file_info</c> contains the following fields.</p>
        <taglist>
          <tag><c>size = integer() >= 0</c></tag>
          <item>
            <p>Size of file in bytes.</p>
          </item>
          <tag><c>type = device | directory | other | regular | symlink</c></tag>
          <item>
            <p>The type of the file.</p>
          </item>
          <tag><c>access = read | write | read_write | none</c></tag>
          <item>
            <p>The current system access to the file.</p>
          </item>
          <tag><c>atime = <seealso marker="#type-date_time">date_time()</seealso> | integer() >= 0</c></tag>
          <item>
            <p>The last time the file was read.</p>
          </item>
          <tag><c>mtime = <seealso marker="#type-date_time">date_time()</seealso> | integer() >= 0</c></tag>
          <item>
            <p>The last time the file was written.</p>
          </item>
          <tag><c>ctime = <seealso marker="#type-date_time">date_time()</seealso> | integer() >=0</c></tag>
          <item>
            <p>The interpretation of this time field depends on
              the operating system. On Unix, it is the last time
              the file or the inode was changed. In Windows, it is
              the create time.</p>
          </item>
          <tag><c>mode = integer() >= 0</c></tag>
          <item>
            <p>The file permissions as the sum of the following bit
              values:</p>
            <taglist>
              <tag>8#00400</tag>
              <item>read permission: owner</item>
              <tag>8#00200</tag>
              <item>write permission: owner</item>
              <tag>8#00100</tag>
              <item>execute permission: owner</item>
              <tag>8#00040</tag>
              <item>read permission: group</item>
              <tag>8#00020</tag>
              <item>write permission: group</item>
              <tag>8#00010</tag>
              <item>execute permission: group</item>
              <tag>8#00004</tag>
              <item>read permission: other</item>
              <tag>8#00002</tag>
              <item>write permission: other</item>
              <tag>8#00001</tag>
              <item>execute permission: other</item>
              <tag>16#800</tag>
              <item>set user id on execution</item>
              <tag>16#400</tag>
              <item>set group id on execution</item>
            </taglist>
            <p>On Unix platforms, other bits than those listed above
              may be set.</p>
          </item>
          <tag><c>links = integer() >= 0</c></tag>
          <item>
            <p>Number of links to the file (this will always be 1 for
              file systems which have no concept of links).</p>
          </item>
          <tag><c>major_device = integer() >= 0</c></tag>
          <item>
            <p>Identifies the file system where the file is located.
              In Windows, the number indicates a drive as follows:
              0 means A:, 1 means B:, and so on.</p>
          </item>
          <tag><c>minor_device = integer() >= 0</c></tag>
          <item>
            <p>Only valid for character devices on Unix. In all other
              cases, this field is zero.</p>
          </item>
          <tag><c>inode = integer() >= 0</c></tag>
          <item>
            <p>Gives the <c>inode</c> number. On non-Unix file systems,
              this field will be zero.</p>
          </item>
          <tag><c>uid = integer() >= 0</c></tag>
          <item>
            <p>Indicates the owner of the file. Will be zero for
              non-Unix file systems.</p>
          </item>
          <tag><c>gid = integer() >= 0</c></tag>
          <item>
            <p>Gives the group that the owner of the file belongs to.
              Will be zero for non-Unix file systems.</p>
          </item>
        </taglist>
        <p>Typical error reasons:</p>
        <taglist>
          <tag><c>eacces</c></tag>
          <item>
            <p>Missing search permission for one of the parent
              directories of the file.</p>
          </item>
          <tag><c>enoent</c></tag>
          <item>
            <p>The file does not exist.</p>
          </item>
          <tag><c>enotdir</c></tag>
          <item>
            <p>A component of the file name is not a directory. On some
              platforms, <c>enoent</c> is returned instead.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="read_line" arity="1"/>
      <fsummary>Read a line from a file</fsummary>
      <desc>
        <p>Reads a line of bytes/characters from the file referenced by
          <c><anno>IoDevice</anno></c>. Lines are defined to be delimited by the linefeed (LF, <c>\n</c>) character, but any carriage return (CR, <c>\r</c>) followed by a newline is also treated as a single LF character (the carriage return is silently ignored). The line is returned <em>including</em> the LF, but excluding any CR immediately followed by a LF. This behaviour is consistent with the behaviour of <seealso marker="stdlib:io#get_line/2">io:get_line/2</seealso>. If end of file is reached without any LF ending the last line, a line with no trailing LF is returned.</p>
	  <p>The function can be used on files opened in <c>raw</c> mode. It is however inefficient to use it on <c>raw</c> files if the file is not opened with the option <c>{read_ahead, Size}</c> specified, why combining <c>raw</c> and <c>{read_ahead, Size}</c> is highly recommended when opening a text file for raw line oriented reading.</p> 
	  <p>If <c>encoding</c> is set to something else than <c>latin1</c>, the <c>read_line/1</c> call will fail if the data contains characters larger than 255, why the <seealso marker="stdlib:io">io(3)</seealso> module is to be preferred when reading such a file.</p> 
	  <p>The function returns:</p>
        <taglist>
          <tag><c>{ok, <anno>Data</anno>}</c></tag>
          <item>
	    <p>One line from the file is returned, including the trailing LF, but with CRLF sequences replaced by a single LF (see above).</p>
            <p>If the file was opened in binary mode, the read bytes are
              returned in a binary, otherwise in a list.</p>
          </item>
          <tag><c>eof</c></tag>
          <item>
            <p>Returned if end of file was reached
              before anything at all could be read.</p>
          </item>
          <tag><c>{error, <anno>Reason</anno>}</c></tag>
          <item>
            <p>An error occurred.</p>
          </item>
        </taglist>
        <p>Typical error reasons:</p>
        <taglist>
          <tag><c>ebadf</c></tag>
          <item>
            <p>The file is not opened for reading.</p>
          </item>
          <tag><c>{no_translation, unicode, latin1}</c></tag>
          <item>
            <p>The file is was opened with another <c>encoding</c> than <c>latin1</c> and the data on the file can not be translated to the byte-oriented data that this function returns.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="read_link" arity="1"/>
      <fsummary>See what a link is pointing to</fsummary>
      <desc>
        <p>This function returns <c>{ok, <anno>Filename</anno>}</c> if
          <c><anno>Name</anno></c> refers to a symbolic link or
          <c>{error, <anno>Reason</anno>}</c> otherwise.
          On platforms that do not support symbolic links, the return
          value will be <c>{error,enotsup}</c>.</p>
        <p>Typical error reasons:</p>
        <taglist>
          <tag><c>einval</c></tag>
          <item>
            <p><c><anno>Name</anno></c> does not refer to a symbolic link.</p>
          </item>
          <tag><c>enoent</c></tag>
          <item>
            <p>The file does not exist.</p>
          </item>
          <tag><c>enotsup</c></tag>
          <item>
            <p>Symbolic links are not supported on this platform.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="read_link_info" arity="1"/>
      <name name="read_link_info" arity="2"/>
      <fsummary>Get information about a link or file</fsummary>
      <desc>
		  <p>This function works like 
			 <seealso marker="#read_file_info/2">read_file_info/1,2</seealso> except that
          if <c><anno>Name</anno></c> is a symbolic link, information about
          the link will be returned in the <c>file_info</c> record and
          the <c>type</c> field of the record will be set to
          <c>symlink</c>.</p>
        <p>If <c><anno>Name</anno></c> is not a symbolic link, this function returns
          exactly the same result as <c>read_file_info/1</c>.
          On platforms that do not support symbolic links, this function
          is always equivalent to <c>read_file_info/1</c>.</p>
      </desc>
    </func>
    <func>
      <name name="rename" arity="2"/>
      <fsummary>Rename a file</fsummary>
      <desc>
        <p>Tries to rename the file <c><anno>Source</anno></c> to
          <c><anno>Destination</anno></c>.
          It can be used to move files (and directories) between
          directories, but it is not sufficient to specify
          the destination only. The destination file name must also be
          specified. For example, if <c>bar</c> is a normal file and
          <c>foo</c> and <c>baz</c> are directories,
          <c>rename("foo/bar", "baz")</c> returns an error, but
          <c>rename("foo/bar", "baz/bar")</c> succeeds. Returns
          <c>ok</c> if it is successful.</p>
        <note>
          <p>Renaming of open files is not allowed on most platforms
            (see <c>eacces</c> below).</p>
        </note>
        <p>Typical error reasons:</p>
        <taglist>
          <tag><c>eacces</c></tag>
          <item>
            <p>Missing read or write permissions for the parent
              directories of <c><anno>Source</anno></c> or
              <c><anno>Destination</anno></c>. On
              some platforms, this error is given if either
              <c><anno>Source</anno></c> or <c><anno>Destination</anno></c>
              is open.</p>
          </item>
          <tag><c>eexist</c></tag>
          <item>
            <p><c><anno>Destination</anno></c> is not an empty directory.
              On some platforms, also given when <c><anno>Source</anno></c> and
              <c><anno>Destination</anno></c> are not of the same type.</p>
          </item>
          <tag><c>einval</c></tag>
          <item>
            <p><c><anno>Source</anno></c> is a root directory, or
              <c><anno>Destination</anno></c>
              is a sub-directory of <c><anno>Source</anno></c>.</p>
          </item>
          <tag><c>eisdir</c></tag>
          <item>
            <p><c><anno>Destination</anno></c> is a directory, but
              <c><anno>Source</anno></c> is not.</p>
          </item>
          <tag><c>enoent</c></tag>
          <item>
            <p><c>Source</c> does not exist.</p>
          </item>
          <tag><c>enotdir</c></tag>
          <item>
            <p><c><anno>Source</anno></c> is a directory, but
              <c><anno>Destination</anno></c> is not.</p>
          </item>
          <tag><c>exdev</c></tag>
          <item>
            <p><c><anno>Source</anno></c> and <c><anno>Destination</anno></c>
              are on different file systems.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="script" arity="1"/>
      <fsummary>Evaluate and return the value of Erlang expressions in a file</fsummary>
      <desc>
        <p>Reads and evaluates Erlang expressions, separated by '.' (or
          ',', a sequence of expressions is also an expression), from
          the file. Returns one of the following:</p>
        <taglist>
          <tag><c>{ok, <anno>Value</anno>}</c></tag>
          <item>
            <p>The file was read and evaluated. <c><anno>Value</anno></c> is
              the value of the last expression.</p>
          </item>
          <tag><c>{error, atom()}</c></tag>
          <item>
            <p>An error occurred when opening the file or reading it.
              See <seealso marker="#open/2">open/2</seealso> for a list
              of typical error codes.</p>
          </item>
          <tag><c>{error, {<anno>Line</anno>, <anno>Mod</anno>,
                           <anno>Term</anno>}}</c></tag>
          <item>
            <p>An error occurred when interpreting the Erlang
              expressions in the file. Use <c>format_error/1</c> to
              convert the three-element tuple to an English description
              of the error.</p>
          </item>
        </taglist>
        <p>The encoding of of <c><anno>Filename</anno></c> can be set
          by a comment as described in <seealso
          marker="stdlib:epp#encoding">epp(3)</seealso>.</p>
      </desc>
    </func>
    <func>
      <name name="script" arity="2"/>
      <fsummary>Evaluate and return the value of Erlang expressions in a file</fsummary>
      <desc>
        <p>The same as <c>script/1</c> but the variable bindings
          <c><anno>Bindings</anno></c> are used in the evaluation. See
          <seealso marker="stdlib:erl_eval">erl_eval(3)</seealso> about
          variable bindings.</p>
      </desc>
    </func>
    <func>
      <name name="set_cwd" arity="1"/>
      <fsummary>Set the current working directory</fsummary>
      <desc>
        <p>Sets the current working directory of the file server to
          <c><anno>Dir</anno></c>. Returns <c>ok</c> if successful.</p>
        <p>Typical error reasons are:</p>
        <taglist>
          <tag><c>enoent</c></tag>
          <item>
            <p>The directory does not exist.</p>
          </item>
          <tag><c>enotdir</c></tag>
          <item>
            <p>A component of <c><anno>Dir</anno></c> is not a directory.
              On some platforms, <c>enoent</c> is returned.</p>
          </item>
          <tag><c>eacces</c></tag>
          <item>
            <p>Missing permission for the directory or one of its
              parents.</p>
          </item>
          <tag><c>badarg</c></tag>
          <item>
            <p><c><anno>Dir</anno></c> had an improper type,
              such as tuple.</p>
          </item>
        </taglist>
        <warning>
          <p>In a future release, a bad type for the
            <c><anno>Dir</anno></c>
            argument will probably generate an exception.</p>
        </warning>
      </desc>
    </func>
    <func>
      <name name="sync" arity="1"/>
      <fsummary>Synchronizes the in-memory state of a file with that on the physical medium</fsummary>
      <desc>
        <p>Makes sure that any buffers kept by the operating system
          (not by the Erlang runtime system) are written to disk. On
          some platforms, this function might have no effect.</p>
        <p>Typical error reasons are:</p>
        <taglist>
          <tag><c>enospc</c></tag>
          <item>
            <p>Not enough space left to write the file.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="datasync" arity="1"/>
      <fsummary>Synchronizes the in-memory data of a file, ignoring most of its metadata, with that on the physical medium</fsummary>
      <desc>
        <p>Makes sure that any buffers kept by the operating system
          (not by the Erlang runtime system) are written to disk. In
          many ways it's resembles fsync but it not requires to update
          some of file's metadata such as the access time. On
          some platforms, this function might have no effect.</p>
        <p>Applications that access databases or log files often write
          a tiny data fragment (e.g., one line in a log file) and then
          call fsync() immediately in order to ensure that the written
          data is physically stored on the harddisk. Unfortunately, fsync()
          will always initiate two write operations: one for the newly
          written data and another one in order to update the modification
          time stored in the inode. If the modification time is not a part
          of the transaction concept fdatasync() can be used to avoid
          unnecessary inode disk write operations.</p>
        <p>Available only in some POSIX systems. This call results in a
          call to fsync(), or has no effect, in systems not implementing
          the fdatasync syscall.</p>
      </desc>
    </func>
    <func>
      <name name="truncate" arity="1"/>
      <fsummary>Truncate a file</fsummary>
      <desc>
        <p>Truncates the file referenced by <c><anno>IoDevice</anno></c> at
          the current position.  Returns <c>ok</c> if successful,
          otherwise <c>{error, <anno>Reason</anno>}</c>.</p>
      </desc>
    </func>
    <func>
      <name name="sendfile" arity="2"/>
      <fsummary>send a file to a socket</fsummary>
      <desc>
        <p>Sends the file <c>Filename</c> to <c>Socket</c>.
        Returns <c>{ok, BytesSent}</c> if successful,
        otherwise <c>{error, Reason}</c>.</p>
      </desc>
    </func>
    <func>
      <name name="sendfile" arity="5"/>
      <fsummary>send a file to a socket</fsummary>
      <type name="sendfile_option"/>
      <desc>
        <p>Sends <c>Bytes</c> from the file
        referenced by <c>RawFile</c> beginning at <c>Offset</c> to
        <c>Socket</c>.
        Returns <c>{ok, BytesSent}</c> if successful,
        otherwise <c>{error, Reason}</c>. If <c>Bytes</c> is set to
	0 all data after the given <c>Offset</c> is sent.</p>
	<p>The file used must be opened using the raw flag, and the process
	calling sendfile must be the controlling process of the socket.
	See <seealso marker="gen_tcp#controlling_process-2">gen_tcp:controlling_process/2</seealso></p>
	<p>If the OS used does not support sendfile, an Erlang fallback
	using file:read and gen_tcp:send is used.</p>
	<p>The option list can contain the following options:
	<taglist>
          <tag><c>chunk_size</c></tag>
          <item>The chunk size used by the erlang fallback to send
	  data. If using the fallback, this should be set to a value
	  which comfortably fits in the systems memory. Default is 20 MB.</item>
	</taglist>
	</p>
	<p>On operating systems with thread support, it is recommended to use
	async threads. See the command line flag
	<c>+A</c> in <seealso marker="erts:erl">erl(1)</seealso>. If it is not
	possible to use async threads for sendfile, it is recommended to use
	a relatively small value for the send buffer on the socket. Otherwise
	the Erlang VM might loose some of its soft realtime guarantees.
	Which size to use depends on the OS/hardware and the requirements
	of the application.</p>
      </desc>
    </func>
    <func>
      <name name="write" arity="2"/>
      <fsummary>Write to a file</fsummary>
      <desc>
        <p>Writes <c><anno>Bytes</anno></c> to the file referenced by
          <c><anno>IoDevice</anno></c>. This function is the only way to write to a
          file opened in raw mode (although it works for normally
          opened files, too). Returns <c>ok</c> if successful, and
          <c>{error, <anno>Reason</anno>}</c> otherwise.</p>
	  <p>If the file is opened with <c>encoding</c> set to something else than <c>latin1</c>, each byte written might result in several bytes actually being written to the file, as the byte range 0..255 might represent anything between one and four bytes depending on value and UTF encoding type.</p>
        <p>Typical error reasons are:</p>
        <taglist>
          <tag><c>ebadf</c></tag>
          <item>
            <p>The file is not opened for writing.</p>
          </item>
          <tag><c>enospc</c></tag>
          <item>
            <p>There is a no space left on the device.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="write_file" arity="2"/>
      <fsummary>Write a file</fsummary>
      <desc>
        <p>Writes the contents of the iodata term <c><anno>Bytes</anno></c>
          to the file <c><anno>Filename</anno></c>.
          The file is created if it does not
          exist. If it exists, the previous contents are
          overwritten. Returns <c>ok</c>, or <c>{error, <anno>Reason</anno>}</c>.</p>
        <p>Typical error reasons are:</p>
        <taglist>
          <tag><c>enoent</c></tag>
          <item>
            <p>A component of the file name does not exist.</p>
          </item>
          <tag><c>enotdir</c></tag>
          <item>
            <p>A component of the file name is not a directory. On some
              platforms, <c>enoent</c> is returned instead.</p>
          </item>
          <tag><c>enospc</c></tag>
          <item>
            <p>There is a no space left on the device.</p>
          </item>
          <tag><c>eacces</c></tag>
          <item>
            <p>Missing permission for writing the file or searching one
              of the parent directories.</p>
          </item>
          <tag><c>eisdir</c></tag>
          <item>
            <p>The named file is a directory.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="write_file" arity="3"/>
      <fsummary>Write a file</fsummary>
      <desc>
        <p>Same as <c>write_file/2</c>, but takes a third argument
          <c><anno>Modes</anno></c>, a list of possible modes, see
          <seealso marker="#open/2">open/2</seealso>. The mode flags
          <c>binary</c> and <c>write</c> are implicit, so they should
          not be used.</p>
      </desc>
    </func>
    <func>
      <name name="write_file_info" arity="2"/>
      <name name="write_file_info" arity="3"/>
      <fsummary>Change information about a file</fsummary>
      <desc>
        <p>Change file information. Returns <c>ok</c> if successful,
          otherwise <c>{error, <anno>Reason</anno>}</c>.
          <c><anno>FileInfo</anno></c> is a record
          <c>file_info</c>, defined in the Kernel include file
          <c>file.hrl</c>. Include the following directive in the module
          from which the function is called:</p>
        <code type="none">
-include_lib("kernel/include/file.hrl").</code>
		<p>The time type set in <c>atime</c>, <c>mtime</c> and <c>ctime</c>
			is dependent on the time type set in <c>Opts :: {time, Type}</c>.
			Type <c>local</c> will interpret the time set as local, <c>universal</c> will
			interpret it as universal time and <c>posix</c> must be seconds since
			or before unix time epoch which is 1970-01-01 00:00 UTC.
			Default is <c>{time, local}</c>.
		</p>
        <p>The following fields are used from the record, if they are
          given.</p>
        <taglist>
          <tag><c>atime = <seealso marker="#type-date_time">date_time()</seealso> | integer() >= 0</c></tag>
          <item>
            <p>The last time the file was read.</p>
          </item>
          <tag><c>mtime = <seealso marker="#type-date_time">date_time()</seealso> | integer() >= 0</c></tag>
          <item>
            <p>The last time the file was written.</p>
          </item>
          <tag><c>ctime = <seealso marker="#type-date_time">date_time()</seealso> | integer() >= 0</c></tag>
          <item>
            <p>On Unix, any value give for this field will be ignored
              (the "ctime" for the file will be set to the current
              time). On Windows, this field is the new creation time to
              set for the file.</p>
          </item>
          <tag><c>mode = integer() >= 0</c></tag>
          <item>
            <p>The file permissions as the sum of the following bit
              values:</p>
            <taglist>
              <tag>8#00400</tag>
              <item>read permission: owner</item>
              <tag>8#00200</tag>
              <item>write permission: owner</item>
              <tag>8#00100</tag>
              <item>execute permission: owner</item>
              <tag>8#00040</tag>
              <item>read permission: group</item>
              <tag>8#00020</tag>
              <item>write permission: group</item>
              <tag>8#00010</tag>
              <item>execute permission: group</item>
              <tag>8#00004</tag>
              <item>read permission: other</item>
              <tag>8#00002</tag>
              <item>write permission: other</item>
              <tag>8#00001</tag>
              <item>execute permission: other</item>
              <tag>16#800</tag>
              <item>set user id on execution</item>
              <tag>16#400</tag>
              <item>set group id on execution</item>
            </taglist>
            <p>On Unix platforms, other bits than those listed above
              may be set.</p>
          </item>
          <tag><c>uid = integer() >= 0</c></tag>
          <item>
            <p>Indicates the owner of the file. Ignored for non-Unix
              file systems.</p>
          </item>
          <tag><c>gid = integer() >= 0</c></tag>
          <item>
            <p>Gives the group that the owner of the file belongs to.
              Ignored for non-Unix file systems.</p>
          </item>
        </taglist>
        <p>Typical error reasons:</p>
        <taglist>
          <tag><c>eacces</c></tag>
          <item>
            <p>Missing search permission for one of the parent
              directories of the file.</p>
          </item>
          <tag><c>enoent</c></tag>
          <item>
            <p>The file does not exist.</p>
          </item>
          <tag><c>enotdir</c></tag>
          <item>
            <p>A component of the file name is not a directory. On some
              platforms, <c>enoent</c> is returned instead.</p>
          </item>
        </taglist>
      </desc>
    </func>
  </funcs>

  <section>
    <title>POSIX Error Codes</title>
    <list type="bulleted">
      <item><c>eacces</c> - permission denied</item>
      <item><c>eagain</c> - resource temporarily unavailable</item>
      <item><c>ebadf</c> - bad file number</item>
      <item><c>ebusy</c> - file busy</item>
      <item><c>edquot</c> - disk quota exceeded</item>
      <item><c>eexist</c> - file already exists</item>
      <item><c>efault</c> - bad address in system call argument</item>
      <item><c>efbig</c> - file too large</item>
      <item><c>eintr</c> - interrupted system call</item>
      <item><c>einval</c> - invalid argument</item>
      <item><c>eio</c> - IO error</item>
      <item><c>eisdir</c> - illegal operation on a directory</item>
      <item><c>eloop</c> - too many levels of symbolic links</item>
      <item><c>emfile</c> - too many open files</item>
      <item><c>emlink</c> - too many links</item>
      <item><c>enametoolong</c> - file name too long</item>
      <item><c>enfile</c> - file table overflow</item>
      <item><c>enodev</c> - no such device</item>
      <item><c>enoent</c> - no such file or directory</item>
      <item><c>enomem</c> - not enough memory</item>
      <item><c>enospc</c> - no space left on device</item>
      <item><c>enotblk</c> - block device required</item>
      <item><c>enotdir</c> - not a directory</item>
      <item><c>enotsup</c> - operation not supported</item>
      <item><c>enxio</c> - no such device or address</item>
      <item><c>eperm</c> - not owner</item>
      <item><c>epipe</c> - broken pipe</item>
      <item><c>erofs</c> - read-only file system</item>
      <item><c>espipe</c> - invalid seek</item>
      <item><c>esrch</c> - no such process</item>
      <item><c>estale</c> - stale remote file handle</item>
      <item><c>exdev</c> - cross-domain link</item>
    </list>
  </section>

  <section>
    <title>Performance</title>
    <p>Some operating system file operations, for example a
      <c>sync/1</c> or <c>close/1</c> on a huge file, may block their 
      calling thread for seconds. If this befalls the emulator main
      thread, the response time is no longer in the order of
      milliseconds, depending on the definition of "soft" in soft
      real-time system.</p>
    <p>If the device driver thread pool is active, file operations are
      done through those threads instead, so the emulator can go on
      executing Erlang processes. Unfortunately, the time for serving a
      file operation increases due to the extra scheduling required
      from the operating system.</p>
    <p>If the device driver thread pool is disabled or of size 0, large
      file reads and writes are segmented into several smaller, which
      enables the emulator so server other processes during the file
      operation. This gives the same effect as when using the thread
      pool, but with larger overhead. Other file operations, for
      example <c>sync/1</c> or <c>close/1</c> on a huge file, still are
      a problem.</p>
    <p>For increased performance, raw files are recommended. Raw files
      uses the file system of the node's host machine. For normal files
      (non-raw), the file server is used to find the files, and if
      the node is running its file server as slave to another node's,
      and the other node runs on some other host machine, they may have
      different file systems. This is seldom a problem, but you have
      now been warned.</p>
    <p>A normal file is really a process so it can be used as an IO
      device (see <c>io</c>). Therefore when data is written to a
      normal file, the sending of the data to the file process, copies 
      all data that are not binaries. Opening the file in binary mode
      and writing binaries is therefore recommended. If the file is
      opened on another node, or if the file server runs as slave to
      another node's, also binaries are copied.</p>
    <p>Caching data to reduce the number of file operations, or rather
      the number of calls to the file driver, will generally increase
      performance. The following function writes 4 MBytes in 23
      seconds when tested:</p>
    <code type="none"><![CDATA[
create_file_slow(Name, N) when integer(N), N >= 0 ->
    {ok, FD} = file:open(Name, [raw, write, delayed_write, binary]),
    ok = create_file_slow(FD, 0, N),
    ok = ?FILE_MODULE:close(FD),
    ok.
      
create_file_slow(FD, M, M) ->
    ok;
create_file_slow(FD, M, N) ->
    ok = file:write(FD, <<M:32/unsigned>>),
    create_file_slow(FD, M+1, N).]]></code>
    <p>The following, functionally equivalent, function collects 1024
      entries into a list of 128 32-byte binaries before each call to
      <c>file:write/2</c> and so does the same work in 0.52 seconds,
      which is 44 times faster.</p>
    <code type="none"><![CDATA[
create_file(Name, N) when integer(N), N >= 0 ->
    {ok, FD} = file:open(Name, [raw, write, delayed_write, binary]),
    ok = create_file(FD, 0, N),
    ok = ?FILE_MODULE:close(FD),
    ok.
      
create_file(FD, M, M) ->
    ok;
create_file(FD, M, N) when M + 1024 =&lt; N ->
    create_file(FD, M, M + 1024, []),
    create_file(FD, M + 1024, N);
create_file(FD, M, N) ->
    create_file(FD, M, N, []).
      
create_file(FD, M, M, R) ->
    ok = file:write(FD, R);
create_file(FD, M, N0, R) when M + 8 =&lt; N0 ->
    N1  = N0-1,  N2  = N0-2,  N3  = N0-3,  N4  = N0-4, 
    N5  = N0-5,  N6  = N0-6,  N7  = N0-7,  N8  = N0-8, 
    create_file(FD, M, N8, 
                [<<N8:32/unsigned,  N7:32/unsigned, 
                   N6:32/unsigned,  N5:32/unsigned, 
                   N4:32/unsigned,  N3:32/unsigned, 
                   N2:32/unsigned,  N1:32/unsigned>> | R]);
create_file(FD, M, N0, R) ->
    N1 = N0-1,
    create_file(FD, M, N1, [<<N1:32/unsigned>> | R]).]]></code>
    <note>
      <p>Trust only your own benchmarks. If the list length in 
        <c>create_file/2</c> above is increased, it will run slightly
        faster, but consume more memory and cause more memory
        fragmentation. How much this affects your application is
        something that this simple benchmark can not predict.</p>
      <p>If the size of each binary is increased to 64 bytes, it will
        also run slightly faster, but the code will be twice as clumsy.
        In the current implementation are binaries larger than 64 bytes
        stored in memory common to all processes and not copied when
        sent between processes, while these smaller binaries are stored
        on the process heap and copied when sent like any other term.</p>
      <p>So, with a binary size of 68 bytes <c>create_file/2</c> runs
        30 percent slower then with 64 bytes, and will cause much more
        memory fragmentation. Note that if the binaries were to be sent
        between processes (for example a non-raw file) the results
        would probably be completely different.</p>
    </note>
    <p>A raw file is really a port. When writing data to a port, it is
      efficient to write a list of binaries. There is no need to
      flatten a deep list before writing. On Unix hosts, scatter output,
      which writes a set of buffers in one operation, is used when
      possible. In this way <c>file:write(FD, [Bin1, Bin2 | Bin3])</c>
      will write the contents of the binaries without copying the data
      at all except for perhaps deep down in the operating system
      kernel.</p>
    <p>For raw files, <c>pwrite/2</c> and <c>pread/2</c> are
      efficiently implemented. The file driver is called only once for
      the whole operation, and the list iteration is done in the file
      driver.</p>
    <p>The options <c>delayed_write</c> and <c>read_ahead</c> to 
      <c>file:open/2</c> makes the file driver cache data to reduce
      the number of operating system calls. The function
      <c>create_file/2</c> in the example above takes 60 seconds
      seconds without the <c>delayed_write</c> option, which is 2.6
      times slower.</p>
    <p>And, as a really bad example, <c>create_file_slow/2</c> above
      without the <c>raw</c>, <c>binary</c> and <c>delayed_write</c>
      options, that is it calls <c>file:open(Name, [write])</c>, needs
      1 min 20 seconds for the job, which is 3.5 times slower than
      the first example, and 150 times slower than the optimized
      <c>create_file/2</c>. </p>
  </section>

  <section>
    <title>Warnings</title>
    <p>If an error occurs when accessing an open file with the <c>io</c>
      module, the process which handles the file will exit. The dead
      file process might hang if a process tries to access it later.
      This will be fixed in a future release.</p>
  </section>

  <section>
    <title>SEE ALSO</title>
    <p><seealso marker="stdlib:filename">filename(3)</seealso></p>
  </section>
</erlref>