Logging unknown values when have special characters about quarkus

[quarkus-bot]

/cc @dmlloyd (logging)

[geoand]

I can't reproduce this with Greek characters

[Willian199]

Maybe it only happens with Latin characters.

Reproducer: https://github.com/Willian199/quarkus_log_chars

[geoand]

That one works fine for me as well

[dmlloyd]

By "special" characters I assume you mean "non-ASCII" (aka Unicode code points whose value is greater than 127), which are only "special" outside of the USA (the "A" in "ASCII" stands for "American"). The problems you are having likely indicate that your system's default character encoding does not match what Java is expecting for whatever reason.

You did not indicate anything about your system environment (uname -a or ver along with java -version at least is needed) so that's all I can say for certain. You also did not indicate where you are seeing this issue: is it on the console, or in a file, or somewhere else?

What changed in Quarkus 3.6.5 is that the console charset is now detected more actively, and if no charset is given, we are falling back to the system default charset (which normally defaults to UTF-8 if all else fails). So I'm guessing that your problem is occurring on the console, in which case setting quarkus.log.file.encoding will do nothing, since that applies only to file logging. But, you should specify, otherwise I'm just guessing.

[Willian199]

I'm using Windows 11 22H2 and this happens in console. If I do System.out.println("àáéíóúç"); works as I expect

ver

Microsoft Windows [versão 10.0.22621.3007]

Java -version

java version "21" 2023-09-19
Java(TM) SE Runtime Environment Oracle GraalVM 21+35.1 (build 21+35-jvmci-23.1-b15)
Java HotSpot(TM) 64-Bit Server VM Oracle GraalVM 21+35.1 (build 21+35-jvmci-23.1-b15, mixed mode, sharing)

[Willian199]

The Charset.defaultCharset() returns UTF-8.
But the StreamEncoder use IBM850.

Here is the last place where I found the correct value.

Here is wrong

[Willian199]

Well, following this tutorial solved the problem.

https://itsontreinamentos.com/2022/05/05/how-to-change-character-encoding-windows-11/

[dmlloyd]

To add more information: System.out does not use the console encoding, which is determined by java.io.Console separately (see Console.charset() for more information). The log manager uses this console encoding for console logging.