Flaky text encoding regression in Node 22

[jarle]

Version

22.7.0

Platform

Debian 12

Subsystem

No response

What steps will reproduce the bug?

• Fetch UTF8 content with special characters from external API (triggers most often for Spanish/Portugese text)
• Try validating the content with isUtf8 or similar, will return false at times

How often does it reproduce? Is there a required condition?

The required condition is that the fetched text should have special characters. It seems to happen sporadically, and is not possible to recreate deterministically.

What is the expected behavior? Why is that the expected behavior?

UTF-8 content should stay encoded as UTF-8.

What do you see instead?

Example when inserting into a Postgres database expecting UTF-8: invalid byte sequence for encoding "UTF8": 0xe8 0x20 0x66

Additional information

I have had this happen in two unrelated systems, both running latest node:22-slim docker images. It also happens locally.

From the byte sequences, it seems there is some mix with latin1 and utf-8, but I'm not sure.

[avivkeller]

This issue will be resolved when v22.8.0 releases (hopefully today)

Duplicate of #54521