NetUI APT does not create UTF-8 encoded output in the generated XML documents
for other multibyte encoded source files.
-----------------------------------------------------------------------------------------------------------------------
Key: BEEHIVE-1033
URL: http://issues.apache.org/jira/browse/BEEHIVE-1033
Project: Beehive
Type: Bug
Components: NetUI
Reporter: Carlin Rogers
Assigned to: Carlin Rogers
The struts module config files and validator rules generated by NetUI APT are
XML files with the encoding attribute of the declaration set to "UTF-8" (<?xml
version="1.0" encoding="UTF-8"?>). However, if a different multibyte encoding
is used in the source files, NetUI APT does not create UTF-8 encoded output in
the generated XML documents. Instead, the encoding of the input source file is
used for the output, even though we set the declaration attrbute as UTF-8.
Then, when this XML file is parsed there will be errors such as
SAXParseException about an incorrect byte sequence for UTF-8.
The problem occurs with the difference in the file encoding of the input Java
file and the charset used by our APT processor to write out the struts config
file. We just delegate to APT to create the file and have it use the encoding
of the source file. We get a PrintWriter for the generated XML output by
calling com.sun.mirror.apt.Filer.createTextFile() with null as the charset
name. The null charset name tells APT to use the charset used to encode source
files when we wite the file.
For example, on a Japanese Windows box the JVM file.encoding property may be
SJIS. When a source file is edited and multybite characters are used for the
names of actions, the encoding is SJIS. The beehive ant task for compiling a
Controller file with annotations will use the JVM file.encoding of the system
(SJIS) and correctly compile the source file. However, during the annotation
processing, when we get the PrintWriter to output the generated XML, it will
also use SJIS, regardless of the fact that we write a declaration that says the
output file encoding is UTF-8.
Note, you can set the encoding of the APT process with an option. So, one
possible workaround is to write your Controller.java file using the UTF-8
encoding (or use the native2ascii converter in the JDK). Then explicitly set
the encoding to UTF-8 for the APT processor (and javac) in our ant macrodef for
"build-pageflows". To try this out, please follow these steps...
- Before building your sample app, edit the beehive-tools.xml file in your
Beehive distribution. It's in the ant subdirectory of your beehive home.
apache-beehive-svn-snapshot/ant/beehive-tools.xml
- Add the encoding option to the call to the <apt> task in the <macrodef
name="build-pageflows">. The line to add should be encoding="UTF-8". In the
macrodef, the <apt> task should look like this...
<apt srcdir="@{srcdir}"
destdir="@{destdir}"
gendir="@{tempdir}"
classpathref="_pageflow.build.classpath"
sourcepathref="@{sourcepathref}"
srcExtensions="*.jpf,*.jpfs,*.app,*.jsfb,*.java"
encoding="UTF-8"
processorOptions="[EMAIL PROTECTED]"
debug="true"/>
- Now try to build your modified sample app.
The output for the struts config file will be in UTF-8
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira
