A Javadoc comment with accented characters mixes blue and black color in its text on a Pandoc generated pdf. One without accented characters doesn't

Why do I have mixed colors in Javadoc comments when using accented characters?

This markdown:

/** Première manière **/
@GetMapping
public String index() {
   return "Hello Spring Boot";
}
   
/** Deuxième manière */
@RequestMapping("/")
public String greetings() {
   return "Greetings";
}

produces this on a pdf, with Pandoc:

Blue and black colors are mixed in the middle of the text comment. Each time it encounters an accented character, here: è.

But without accented characters, it produces that:

Here, it's more correct, but I wonder why the whole Javadoc comment isn't blue?

my header-includes are these ones, but I have the same behavior whatever:

• I put some one header-includes or not,
• the Java code is under a citation or not,
• in a box,
• in a group

header-includes:
- \usepackage{tcolorbox}
- \usepackage{fvextra}
- \DefineVerbatimEnvironment{Highlighting}{Verbatim}{breaklines,breakanywhere,breaksymbol=,breakanywheresymbolpre=,commandchars=\\\{\}}

Is there a way to correct the behavior with accented characters? It doesn't have a nice effect.

@abra If I change è with è, é with é and all the characters here it produces this: