<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML//EN">
<html>
<head>
<meta http-equiv="Content-Type"
content="text/html; charset=iso-8859-1">
<meta name="GENERATOR" content="Microsoft FrontPage 2.0">
<title>ReadMe: Character Set Conversion Tool for Unicode</title>
</head>
<body bgcolor="#FFFFFF">
<p><font size="5"><b>ReadMe: The Character Set Conversion Tool
for Unicode</b></font> </p>
<p>Version: <!--webbot bot="Timestamp" startspan
s-type="EDITED" s-format="%m/%d/%y" -->10/01/98<!--webbot
bot="Timestamp" i-checksum="13821" endspan --> <br>
</p>
<hr>
<p>COPYRIGHT: <br>
(C) Copyright International Business Machines Corporation, 1998 <br>
Licensed Material - Program-Property of IBM - All Rights
Reserved. <br>
US Government Users Restricted Rights - Use, duplication, or
disclosure restricted by GSA ADP Schedule Contract with IBM Corp.
<br>
</p>
<hr>
<p><font size="4"><b>Introduction</b></font> </p>
<p><b>makeconv</b> is a tool that converts character set
conversion data into binary files for use by the UnicodeConverter
C code. The UnicodeConverter code will not function with text
format data files. The user of charset conversion library will
need to convert all the data files in text format into binary
format. </p>
<p>For a IBM registered code page to/from Unicode conversion data
file in the specified data path, it generates .cnv files, which
it puts in the same directory. The tool takes a list of file
names as argument, with their absolute or relative path, and will
generate a corresponding ".cnv" file with the same
path. On Unix shells one can easily convert all the
".cnv" files from the text files in the data directory
by using wildcards (e.g. <b>makeconv ../../data/locales/*.ucm</b>).
On Win32 we have provided a batch file "<b>mkcnvfle.bat</b>"
in the same directory you will find <b>makeconv.c</b>. <b>mkcnvfle</b>
converts all the ".ucm" files in intlwork/data/locales
to ".cnv" files. It requires one argument : Debug or
Release so it knows where to look for the makeconv.exe. <br>
(Win32 usage:<b> \intlwork\tools\makeconv\mkcnvfle.bat Debug</b>
for debug build, <b>\intlwork\tools\makeconv\mkcnvfle.bat Release</b>,
otherwise<b>.</b>) </p>
<ul>
<li><b>Create a binary file for codepage IBM-949 in the data
path "\intlwork\data\locales\"</b>: type in at
the command prompt, "<b>makeconv</b><font
face="Courier"> \intlwork\data\locales\ibm-949.ucm".
</font>The generated binary file will be located in the
same directory.</li>
<li><b>Look for the codepage IBM-949 data file in the current
working directory</b>: type in at the command prompt,
"<b>makeconv</b><font face="Courier"> ibm-949.ucm</font>".
The tool will try to locate the file "<font
face="Courier">ibm-949.ucm</font>" in the current
directory. The generated binary file, "<font
face="Courier">ibm-949.cnv</font>" will be located
in current working directory.</li>
</ul>
<p><font size="4"><b>Building the Tool</b></font> </p>
<p><font size="3">To build the tool on Win32 environment is very
easy, simply open the "makeconv</font><font size="3"
face="Courier">.dsw</font><font size="3">" workspace project
file in MSVC++ 5.0 compiler.</font> <br>
On Unix the process of building the tool and the ".cnv"
files is included in the build process of the library. The
provided makefiles will <br>
attempt to build <b>makeconv</b> as it builds other parts of the
library. It will also convert all the initial ".ucm"
files in data/locales to ".cnv" files. </p>
<p><font size="4"><b>Limitations</b></font> </p>
<p>The tools uses stdio.h functions to communicate success or
failure to the user. Platforms that don't support stdio will have
to adapt <br>
that portion of the code. <br>
</p>
</body>
</html>
