Encoding problem with ./player/(PLAYER_NAME)/config.json file about beatoraja HOT 2 CLOSED

I believe this can be resolved by adding this environment variable to your launch script:

set JAVA_TOOL_OPTIONS=-Dfile.encoding=UTF-8 - Windows
or
export JAVA_TOOL_OPTIONS=-Dfile.encoding=UTF-8 - Linux/OS X

Ideally this would've been included initially when beatoraja came out, but instead we're too late in development ... it would break a lot of installs when the file is ANSI encoded rather than UTF-8. note, if your config.json is not UTF-8 there might be strange bugs (i.e: save/load of certain features don't work when ANSI!)

from beatoraja.

Okay, so after I studied Java by myself, I've figured out that a usage of FileWriter and FileReader causes this problem because that classes use the system encoding when its OS is Windows (which is obviously not UTF-8 in general condition.)

Of course, you can use such a 'UTF-8 fix', editing a batch file with an addition of an environment variable JAVA_TOOL_OPTIONS=-Dfile.encoding=UTF-8. However, when it comes to executable binaries like the one in Mocha-Repository, we can't do that black magic.
Luckily, there are standard classes that can read/write files in specific locale, OutputStreamWriter and InputStreamReader. If these classes are used for this section, users won't need that 'UTF-8 fix' anymore.

try (OutputStreamWriter writer =
    new OutputStreamWriter(new FileOutputStream(Paths.get(playerpath + "/" + player.getId() + "/config.json"), StandardCharsets.UTF_8))
) {

The problem is, if users who haven't used a 'UTF-8 fix' (typically users who are using Windows in Shift-JIS encoding), their configuration file will be deleted because it has not been in UTF-8 encoding so they'll loose all their configurations.

This can be solved by using an external library such as icu4j, but I'm not sure it is proper solution for this case. Any ideas?

(Sorry for my poor English)

from beatoraja.