<i18n dev> RFR: 8274544: Langtools command's usage were garbled on Japanese Windows

[Ichiroh Takiguchi]

On Thu, 14 Oct 2021 23:43:47 GMT, Naoto Sato <naoto at openjdk.org> wrote:

>> @takiguc - if JShell is still an issue, is there a chance you could try this:
>> https://github.com/lahodaj/jdk/commit/cfa6b3eebbc22c5a48d31cfd692ff98690653686
>> 
>> Not sure if it will help, but it might (this won't change the default charset, but could send the output in a sensible encoding for the console.
>> 
>> Thanks!
>
> I believe both @lahodaj and my proposed changes are needed, which I provided here: https://github.com/openjdk/jdk/commit/82a9033953655042480c90d388fcbc8053fc5ff5
> Can you check this works?

Hello @naotoj .
[82a9033](https://github.com/openjdk/jdk/commit/82a9033953655042480c90d388fcbc8053fc5ff5) worked on Windows platform, then I got your point.
To minimize changes, I think I should create new method and call from JShellToolBuilder.java and JShellToolProvider.java.

    static PrintStream derivePrintStream(PrintStream ps) {
        if (Charset.defaultCharset().equals(NATIVE_CHARSET)) {
            return ps;
        } else {
            return new PrintStream(new WriterOutputStream(
                new OutputStreamWriter(ps, NATIVE_CHARSET), Charset.defaultCharset()));
        }
    }


Additionally, I tested this issue on non-UTF-8 platform like RHEL7/CentOS7 EUC locale (ja_JP.eucjp).
It seemed JLine's terminal's encoding was still UTF-8.
I could not input Japanese character as expected, also Cut & Paste with Japanese string did not work.
In my investigation, Charset.defaultCharset() was used on AbstractTerminal.java
https://github.com/openjdk/jdk/blob/master/src/jdk.internal.le/share/classes/jdk/internal/org/jline/terminal/impl/AbstractTerminal.java#L55-L62
I should be like,

        this.encoding = encoding != null ? encoding : nativeCharset;

(JLine was upgraded by b19, I need to rebase if it's required)

One more thing,
For Inter-Process Communication between jshell front-end and agent, I think default charset should be same for non UTF-8 environment.
I think JdiInitiator.java should be modified.

Please give me your suggestion.

-------------

PR: https://git.openjdk.java.net/jdk/pull/5771