<!DOCTYPE html>
<html>
<head>
<title>Adding new text encoding to OpenOffice.org</title>
</head>
<body>

<h3>Adding new text encoding to OpenOffice.org</h3>

<p>
  Adding a new text encoding involves a number of modifications, mostly in the <a
  href="http://porting.openoffice.org/source/browse/porting/sal/textenc/">sal/textenc</a> source
  directory (see below).  The main benefits of adding a text encoding typically are that you can add
  it to the drop-down list boxes where you choose a character encoding (e.g., for the "Text Encoded"
  filter for importing/exporting plain text files, or in the
  "<b>Tools</b> - <b>Options...</b> - <b>Load/Save</b> - <b>HTML Compatibility</b>" dialog).
</p>

<p>
  The necessary steps include:

  <ol>
    <li>
Add a new <code>RTL_TEXTENCODING_</code> to <a
href="http://porting.openoffice.org/source/browse/porting/sal/inc/rtl/textenc.h">sal/inc/rtl/textenc.h</a>,
and update <code>rtl_isOctetEncoding</code> in <a
href="http://porting.openoffice.org/source/browse/porting/sal/textenc/tencinfo.c">sal/textenc/tencinfo.c</a>.
    </li>


    <li>
In <a
href="http://porting.openoffice.org/source/browse/porting/sal/textenc/">sal/textenc</a>,
create a <code>static ImplTextEncodingData</code> instance for the new
encoding.  If it is a single-byte encoding based on ASCII, you can easily
reuse the <code>ImplTextConverter</code> framework that is already in
place. You also have to specify corresponding Windows/Unix/MIME character
sets, if such exist.  If your text encoding is single-byte, your
<code>ImplTextEncodingData</code> instance would probably best fit into one of the
already existing tcvt<var>XXX</var>1.tab source files (e.g., if your
encoding is for Tibetan, it would fit into <a
href="http://porting.openoffice.org/source/browse/porting/sal/textenc/tcvteas1.tab">tcvteas1.tab</a>).
Do not forget to add a pointer to your <code>ImplTextEncodingData</code>
structure to the <code>aData</code> array in <a
    href="http://porting.openoffice.org/source/browse/porting/sal/textenc/textenc.cxx">textenc.cxx</a>
<code>Impl_getTextEncodingData()</code>. Note that for older versions (OOo1.x
SRX645 branch and up to SRC680_m44) this was <a
    href="http://porting.openoffice.org/source/browse/porting/sal/textenc/textenc.c">textenc.c</a>
    </li>

    <li>
      You should add tests of your new encoding to <a
      href="http://porting.openoffice.org/source/browse/porting/sal/test/testtextenc.cxx">sal/test/texttestenc.cxx</a>.
    </li>

    <li>
      To make the new encoding appear in the drop-down list boxes mentioned above, you have:
      <ol>
  <li>open an issue for <code>Liz@openoffice.org</code>, write the path of where the
    string is going to go (for example, "<b>Tools</b> - <b>Options...</b> - <b>Load/Save</b> -
    <b>HTML Compatibility</b>" dialog) and suggest the new string you want added (in English)
    and wait for approval for the string as well as the German equivalent from Liz. (If you
    can suggest the string in German as well that would be helpful.</li>
  <li>
    add both the English and German strings to the resources in <a
href="http://graphics.openoffice.org/source/browse/graphics/svx/source/dialog/txenctab.src">svx/source/dialog/txenctab.src</a> file manually; that means directly into the source code.
  </li>
      </ol>
    </li>
  </ol>
</p>

<p>

That said, <em>adding any new text encodings also adds complexity</em> to OpenOffice.org (mostly
source and code size in this case), so it should be done only if really necessary (e.g., if there is
demand to import/export plain text files encoded in a special encoding; <em>adding a new text
encoding would generally not be necessary to just make sure that OpenOffice.org can be localized to
some language, or that texts in that language can be written in OpenOffice.org</em>).
</p>

<p>
  If you have any further questions, just ask on our <a
  href="http://mail-archives.apache.org/mod_mbox/incubator-ooo-l10n/">list</a>.
</p>

</body>
</html>