chinese character don't display appropriate

[xpf7622]

// Hello world java file is utf-8
public class HelloWorld {
public static void main(String[] args) {
System.out.println("Hello World! 汉字！");
}
}

output as follows:

Hello World! 姹夊瓧锛�

but java file have the default code utf-8. and Debugger for java default also utf8, why the output is abnormal character!

Operating system : simplified Chinese win10 ; version. : 1709(16299 .334)

launch.json:
"configurations": [
{
"type": "java",
"name": "Java Debug",
"request": "launch",
"cwd": "${workspaceFolder}",
"console": "internalConsole",
"stopOnEntry": false,
"mainClass": "HelloWorld",
"encoding": "UTF-8",
"args": ""
},
……

[akaroml]

@testforstephen please take a look.

[andxu]

@xpf7622 could you please upload your HelloWorld.class to this issue?

[xpf7622]

@andxu Debug doesn't produce HelloWorld.class file。

[xpf7622]

this is my java file. debug process doesn't produce HelloWorld.class file.
HelloWorld.zip

[HaiUIT-zz]

I have the same problem with you, after I set my java file encode to UTF-8 with BOM it work and print the right text. But it is not a solution, because javac doesn't work with UTF-8 with BOM. If i run javac or use other build tool like gradle, maven to build my code, it will throw this error :

> javac -encoding UTF8 Test.java
Test.java:1: error: illegal character: '\ufeff'
?class Test {
^
It seem like there is some problems in compile phase, here is my source code, the class file build with javac and the class file build with java debug extension. I have check both class files with hex editor and confirmed that they have different length and content.

App.zip

[yiailake]

me too.I found the class file in
/Users/xxx/Library/Application Support/Code/User/workspaceStorage/b72b50a9aa05f002d6a626eedc811d53/redhat.java/jdt_ws/jdt.ls-java-project/bin
Then I use command java Hello, the output can't display normal chinese character!

[testforstephen]

The underlying compiler (jdt complier) used by java debug extension looks like using the OS default encoding to parse the source file, when source file encoding is not same as the system encoding, the generated .class file will contain messy string literal. Need find a way to pass the correct source file encoding to the jdt compiler.

[testforstephen]

The encoding configuration item in launch.json mainly affects the character data parsed from the external input, such as console input or network input. It doesn't affect the print of the hard coded string constants in source file.

[testforstephen]

Tracking the issue at vscode-java side redhat-developer/vscode-java#479