Game logs are now encoded in UTF-8

[marrrrrrme6]

fix #2419

[TayouVR]

question is, if that breaks any other codepoints? if someone has their OS set to e.g. that asian one, would that break the logs then if we decode it as UTF-8?
PolyMC/PolyMC#968
3c4b45c

[marrrrrrme6]

For now, I've added -Dfile.encoding=UTF-8 and tested Japanese.
There were no problems with Japanese.
I don't understand other languages ​​such as Chinese, so could you please test them for me?

[marrrrrrme6]

The following tests were done on minecraft 1.12.2 and 1.21.1. In 1.12.2, -Dfile.encoding=UTF-8 was added to the arguments.

The issue of garbled characters in sections (§) has been resolved.
1.21.1:

1.12.2:

MultiMC/Launcher#5435

Chinese, accent aigu and Russian are also no problem.
1.21.1:

1.12.2:

PolyMC/PolyMC#968 (comment)
MultiMC/Launcher#2998

Special characters are no problem.
1.21.1:

1.12.2:

PolyMC/PolyMC#806

Use utf-8 is not checked. PolyMC/PolyMC#968 (comment)

[Ryex]

Wasn't there an issue with using UTF-8 in 1.7 and earlier?
we do want to support those early versions too.

[marrrrrrme6]

Everything was fine.
I couldn't enter §.
I also tried 1.2.5 and b1.7.3 but the chat wouldn't open.

1.7.10:

1.5.2:

1.3.2:

[Ryex]

huh. I could have sworn this was brought up before and it broke things...
let's ask @Scrumplex. He has a better memory of the project.

[Scrumplex]

He has a better memory of the project.

I sadly dont remember

[Ryex]

alright well, if we're going to add this, -Dfile.encoding=UTF-8 needs to be added as a default java argument or this actually will break things.

[TayouVR]

forgot to send this comment yesterday as I was thinking of asking for some feedback on discord...

then it sounds like the launcher should automatically add that flag.
But I'm not sure if that could cause any issues for anyone

from what you are describing it seems that this should work even with older versions without any issues, so I suppose its fine to blanket apply the flag to every version.
But I'm still skeptical of it working in some really early versions.

I was thinking if we could have a toggle in the settings "Use UTF-8 for log files", which both adds the launch flag and configures the log reader to use UTF-8 instead of the system charset
But I would need other peoples opinions on this.
I'm gonna ask for some input on discord, see if anyone maybe has more to say about this topic

[marrrrrrme6]

alright well, if we're going to add this, -Dfile.encoding=UTF-8 needs to be added as a default java argument or this actually will break things.

In https://github.com/marrrrrrme6/PrismLauncher/tree/test, I add -Dfile.encoding=UTF-8 to the default arguments, but I'm not sure if this is ok.

[TheKodeToad]

alright well, if we're going to add this, -Dfile.encoding=UTF-8 needs to be added as a default java argument or this actually will break things.

This will potentially break things too. Since this also controls the encoding for loading files (at least I think it does and the name suggests that) it could mean loading files encoded differently as utf-8...

[marrrrrrme6]

I am not originally a C++ engineer, and I modified it based on #2419 (comment), but I don't know any more about it than that, so I would like someone to take over.

[TheKodeToad]

Alright... closing this then