Investigate our handling of the UTF BOM (byte order mark) #4712
Description
The following tests in fetch/api/basic/text-utf8.any.js complain that we don't handle the BOM when calling .text().
- UTF-8 with BOM with Request.text()
- UTF-8 with BOM with Response.text()
- UTF-8 with BOM with fetched data (UTF-8 charset)
- UTF-8 with BOM with fetched data (UTF-16 charset)
We could probably just patch TextDecoder but we need to know some things:
- How many requests/responses do we even have that have BOMs? We should have some logging.
- Do we need a compat flag to fix this?
This is very low priority.