JiBX does not generate Java from XSD files with correct encoding
-----------------------------------------------------------------
Key: JIBX-434
URL: http://jira.codehaus.org/browse/JIBX-434
Project: JiBX
Issue Type: Bug
Components: CodeGen
Affects Versions: JiBX 1.2.2
Environment: Windows Server 2008 R2
Reporter: Christian Callsen
Per email conversation (see bottom for request to file bug here):
---
Hello there,
I have several XSD files with danish national characters in it that I include
in my own XSD file (here's one of them:
http://rep.oio.dk/ebxml/xml/schemas/dkcc/2003/02/13/DKCC_CountryIdentificationCode.xsd).
When the JiBX CodeGen tool v1.2.2 runs on it (from maven via maven-jibx-plugin
v1.2.2) the output file is not UTF-8 even though I've set
project.build.outputEncoding to UTF-8. I've checked the source, and the
SourceBuilder.java line 325 seems suspicious:
FileWriter fwrit = new FileWriter(file);
According to the maven plugin guide, one needs to be careful around FileWriter
instantiation and file encodings. Would it be possible to respect the
maven-requested encoding or by adding a flag to control the encoding written to
the file?
I've tried setting file.encoding and project.build.sourceEncoding in pom.xml,
and supplied -Dfile.encoding=UTF-8. No luck. I've tried turning off javadoc and
annotations via:
<show-schema>false</show-schema>
<delete-annotations>false</delete-annotations>
in pom.xml. Still no luck.
Any pointers/workarounds?
---
Hi Christian,
Don is handling all the Maven issue for JiBX, so hopefully he can comment on
this.
Sorry for the delay in responding to this. In general, it's best to ask this
type of question on the JiBX users list (or enter a Jira bug report).
- Dennis
---
Christian,
I tried generating code for this schema and all the special characters look
fine. I have attached the generated java source files.
You may want to check your default java encoding. Here is a great article:
http://stackoverflow.com/questions/361975/setting-the-default-java-character-encoding
If you can supply a schema definition that is not generating code correctly,
that would be a great help.
Don
---
Hey Don,
My problem could be that I'm running this under Windows Server 2008, in CP1252
locale. Which locale and OS did you try out the test on? I'm attaching a ZIP of
the generated sources I get. Notice the generated class
_CountryIdentificationSchemeType.java - there are comments that contain danish
national characters, but the file's not in UTF-8 encoding :-( according to
Notepad++ (it says ANSI).
Below's a snippet from pom.xml, showing you how I invoke the maven-jibx-plugin.
<plugin>
<groupId>org.jibx</groupId>
<artifactId>maven-jibx-plugin</artifactId>
<version>${jibx.plugin.version}</version>
<configuration>
<directory>target/generated-sources/src/main/java/</directory>
<includes>
<include>binding.xml</include>
</includes>
<load>true</load>
<verbose>false</verbose>
</configuration>
<executions>
<execution>
<id>generate-java-code-from-schema</id>
<phase>generate-sources</phase>
<goals>
<goal>schema-codegen</goal>
</goals>
<configuration>
<targetDirectory>${generated.sources.directory}</targetDirectory>
<directory>${webapp.directory}/WEB-INF/xsd</directory>
<verbose>true</verbose>
<includes>
<include>file.xsd</include>
</includes>
<options>
<package>test</package>
<show-schema>false</show-schema>
<delete-annotations>false</delete-annotations>
</options>
</configuration>
</execution>
<execution>
<id>compile-binding</id>
<phase>process-classes</phase>
<goals>
<goal>bind</goal>
</goals>
</execution>
</executions>
</plugin>
The "file.xsd" file is a file that includes these files (can't show you the
entire file, but here are the includes):
<xsd:import
namespace="http://rep.oio.dk/cpr.dk/xml/schemas/core/2005/03/18/";
schemaLocation="http://rep.oio.dk/cpr.dk/xml/schemas/core/2005/03/18/CPR_PersonCivilRegistrationIdentifier.xsd"/>
<xsd:import
namespace="http://rep.oio.dk/itst.dk/xml/schemas/2006/01/17/";
schemaLocation="http://rep.oio.dk/itst.dk/xml/schemas/2006/01/17/ITST_PersonNameStructure.xsd"/>
<xsd:import
namespace="http://rep.oio.dk/xkom.dk/xml/schemas/2006/01/06/";
schemaLocation="http://rep.oio.dk/xkom.dk/xml/schemas/2006/01/06/XKOM_AddressPostal.xsd"/>
<xsd:import
namespace="http://rep.oio.dk/ebxml/xml/schemas/dkcc/2003/02/13/";
schemaLocation="http://rep.oio.dk/ebxml/xml/schemas/dkcc/2003/02/13/DKCC_BirthDate.xsd"/>
<xsd:import
namespace="http://rep.oio.dk/ebxml/xml/schemas/dkcc/2003/02/13/";
schemaLocation="http://rep.oio.dk/ebxml/xml/schemas/dkcc/2003/02/13/DKCC_CountryIdentificationCode.xsd"/>
I suppose we could use the fix in the stackoverflow article (great website
btw), but I'd rather be able to say in my pom.xml that I want output encoding
UTF8 for generated files. For some odd reason - possibly the idea I pointed out
below - that does not seem to work.
I've done a different workaround - using native2ascii, a copy step and a
cleanup step in pom.xml - which turns the entire thing into \uabcd notation. It
works, but makes my pom quite a bit longer and unwieldy. The maven folks
suggest not using a FileWriter here:
http://maven.apache.org/plugin-developers/common-bugs.html and point to issues
for review
http://docs.codehaus.org/display/MAVENUSER/POM+Element+for+Source+File+Encoding
There's a guy suggesting a better way here:
http://www.malcolmhardie.com/weblogs/angus/2004/10/23/java-filewriter-xml-and-utf-8/
Notice in his example that the encoding is supplied, so that the maven plugin
easily could respect the projects source file encoding (if set).
Christian
---
Christian,
I use SUSE linux. UTF-8 must be the default.
You are correct, there should be an option to output the code in UTF-8.
You need to file a enhancement request for JiBX. Unfortunately, I'm not the one
responsible for the SourceBuilder.java file, I handle the maven-jibx-plugin.
File a bug at: http://jira.codehaus.org/secure/BrowseProject.jspa?id=10410.
File it under 'CodeGen' and assign it to Dennis. Say you need a way to specify
UTF-8 encoding on the source code output. Dennis will have to add a parameter
to cause UTF-8 output, I can pass this automatically using the
<project.build.sourceEncoding>UTF-8</project.build.sourceEncoding> property
from the maven plugin.
Please include the links you supplied in this email. Especially the one that
shows how to create an OutputStream with UTF encoding. As you probably know, in
open source, you will usually get a faster response if you write the correct
code, test it, and submit it with your request as a patch.
Thanks,
Don
---
So is the issue just with the encoding of the generated source code? That would
be a problem in the actual generation, rather than anything to do with maven.
Right now there's no way to pass in the character encoding for the generated
code. It looks like you can tell javac the character encoding on the command
line, as "javac -encoding utf8 ...". If I set the code generation to always
output UTF-8, will that work for you?
- Dennis
---
Hey Dennis,
Yes, always using UTF-8 will work for us, as we're using source encoding =
UTF-8.
Btw: JiBX is a serious time saver. I've had to make a few adjustments to the
WSDL/XSD I'm using, and every time it turned out all I had to do was adjust the
bindings file. Slightly.
Excellent tool!
Best Regards
Christian
---
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
------------------------------------------------------------------------------
Create and publish websites with WebMatrix
Use the most popular FREE web apps or write code yourself;
WebMatrix provides all the features you need to develop and
publish your website. http://p.sf.net/sfu/ms-webmatrix-sf
_______________________________________________
jibx-devs mailing list
jibx-devs@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/jibx-devs
