Can I set file encoding to GBK?

[zjulzq]

In my case, the header files are encoded with GBK. As a result, the java files contain messy code.

/////////////////////////////////////////////////////////////////////////
/**\system ��һ������ϵͳ
 * \company �Ϻ��ڻ���Ϣ�������޹�˾
 * \file ThostFtdcMdApi.h
 * \brief �����˿ͻ��˽ӿ�
 * \history 
 * 20060106 �Ժ��     �������ļ�
///////////////////////////////////////////////////////////////////////// */

I have tried -Dfile.encoding=GBK, but it does not work.

What shall I do?

[saudet]

-Dfile.encoding=GBK should work. Did you confirm that your version of Java supports GBK?

[zjulzq]

I print the available Charsets, it contains GBK.

Big5, Big5
Big5-HKSCS, Big5-HKSCS
EUC-JP, EUC-JP
EUC-KR, EUC-KR
GB18030, GB18030
GB2312, GB2312
GBK, GBK
IBM-Thai, IBM-Thai
IBM00858, IBM00858
IBM01140, IBM01140
......

[saudet]

I'm more familiar with Japanese, so I tried SHIFT_JIS this way:

$ java -Dfile.encoding=SHIFT_JIS -jar javacpp.jar LibraryName

And it works just fine!

[zjulzq]

My command is

mvn clean install -Dfile.encoding=GBK -Djavacpp.platform=linux-x86_64 -Djavacpp.platform.dependency=false -Djavacpp.cppbuild.skip=true --projects .,ctp

Will your javacpp plugin pick the -Dfile.encoding=UTF-8?

[saudet]

Not sure, we might have to set the MAVEN_OPTS or _JAVA_OPTIONS environment variable for that...

[zjulzq]

I solve it by this:

export JAVA_TOOL_OPTIONS="-Dfile.encoding=GBK"

[saudet]

Cool! Thanks for testing :)

[zjulzq]

There is another problem.

After I set file.encoding to GBK, another problem occurs. I meet this problem after you have fixed thetypedef char Thost*Type[32] bug. #192

[ERROR] /home/lzq/workspace/javacpp-presets/ctp/src/main/java/org/bytedeco/javacpp/ThostFtdcMdApi.java:16660: error: unmappable character for encoding UTF-8
[ERROR] /**??????? */
[ERROR] ^
[ERROR] /home/lzq/workspace/javacpp-presets/ctp/src/main/java/org/bytedeco/javacpp/ThostFtdcMdApi.java:16660: error: unmappable character for encoding UTF-8
[ERROR] /**??????? */
[ERROR] ^
[ERROR] /home/lzq/workspace/javacpp-presets/ctp/src/main/java/org/bytedeco/javacpp/ThostFtdcMdApi.java:16660: error: unmappable character for encoding UTF-8
[ERROR] /**??????? */
[ERROR] ^
[ERROR] /home/lzq/workspace/javacpp-presets/ctp/src/main/java/org/bytedeco/javacpp/ThostFtdcMdApi.java:16660: error: unmappable character for encoding UTF-8
[ERROR] /**??????? */

/**??????? */ are comments. These comments are written in Chinese.

[saudet]

That sounds like an issue with the Java compiler? Make sure you don't force it to UTF-8 in your pom.xml file.

[zjulzq]

I set project.build.sourceEncoding to GBK in parent pom.xml to solve it.

<project.build.sourceEncoding>GBK</project.build.sourceEncoding>

[zjulzq]

There are some warnings, but not fatal. @param followed by param name and chinese comment can not be parsed.

They are Javadoc warnings.

[WARNING] /home/lzq/workspace/javacpp-presets/ctp/src/main/java/org/bytedeco/javacpp/ThostFtdcMdApi.java:135: warning - @param argument "pszFrontAddress��ǰ�û�������ַ��" is not a parameter name.
[WARNING] /home/lzq/workspace/javacpp-presets/ctp/src/main/java/org/bytedeco/javacpp/ThostFtdcMdApi.java:144: warning - @param argument "pszNsAddress�����ַ�����������ַ��" is not a parameter name.
[WARNING] /home/lzq/workspace/javacpp-presets/ctp/src/main/java/org/bytedeco/javacpp/ThostFtdcMdApi.java:150: warning - @param argument "pFensUserInfo���û���Ϣ��" is not a parameter name.
[WARNING] /home/lzq/workspace/javacpp-presets/ctp/src/main/java/org/bytedeco/javacpp/ThostFtdcTraderApi.java:416: warning - @param argument "pszFrontAddress��ǰ�û�������ַ��" is not a parameter name.
[WARNING] /home/lzq/workspace/javacpp-presets/ctp/src/main/java/org/bytedeco/javacpp/ThostFtdcTraderApi.java:425: warning - @param argument "pszNsAddress�����ַ�����������ַ��" is not a parameter name.
[WARNING] /home/lzq/workspace/javacpp-presets/ctp/src/main/java/org/bytedeco/javacpp/ThostFtdcTraderApi.java:431: warning - @param argument "pFensUserInfo���û���Ϣ��" is not a parameter name.

[saudet]

That's Javadoc. Don't use UTF-8 with Javadoc either.

[zjulzq]

I don't think it is a good solution to set encoding along the tool chains. So, I convert the original header files from GBK to UTF-8, in order to avoid the encoding problem.