Some text with code block is not encoded in UTF-8

[ChoCho66]

A similar question is here.

As shown below (left), some chinese texts with code block is not encoded in UTF-8 in quarto.
But is OK for r-markdown (right).

It's the same in VScode.

The following is my .qmd file.

---
title: "test-qmd-chinese"
format: html
---

```{r}
print(1)
```

中文測試

The warning messages is as follows:

Warning messages:
1: In readLines(con, warn = FALSE) :
  invalid input found on input connection 'test-qmd-chinese.qmd'
2: In xfun::read_utf8(input) :
  The file test-qmd-chinese.qmd is not encoded in UTF-8. These lines contain invalid UTF-8 characters: 11
3: In readLines(con, warn = FALSE) :
  invalid input found on input connection 'test-qmd-chinese.knit.md'
4: In readLines(con, warn = FALSE) :
  invalid input found on input connection './test-qmd-chinese.knit.md'
5: In readLines(con, warn = FALSE) :
  invalid input found on input connection 'test-qmd-chinese.qmd'

---

If we just have some chinese texts without code block, then it is OK.

---

I try a chinese word 文, then it becomes <96><87> if we have a code block.

[ChoCho66]

I don't know why, but when I installed R 4.2.2 it solved the problem.