TextDecoder should not output BOM

[joyeecheung]

Found when porting encoding WPT into core.

Spec: https://encoding.spec.whatwg.org/#concept-td-serialize

If encoding is UTF-8, UTF-16BE, or UTF-16LE, and ignore BOM flag and BOM seen flag are unset, then:

• If token is U+FEFF, then set BOM seen flag.
• Otherwise, if token is not end-of-stream, then set BOM seen flag and append token to output.
• Otherwise, return output.

Failing tests:

• wpt/encoding/textdecoder-byte-order-marks.any.js
• wpt/encoding/textdecoder-copy.any.js

[joyeecheung]

cc @nodejs/i18n

[joyeecheung]

Sorry, pinged the wrong team @nodejs/intl

[Xstoudi]

PR merged, should be close.

[targos]

PR was merged, but the bug still exists. Both failing tests are skipped:

node/test/wpt/status/encoding.json

Lines 24 to 29 in 7c9ee6d

	"textdecoder-byte-order-marks.any.js": {
	"fail": "Mismatching BOM should not be ignored"
	},
	"textdecoder-copy.any.js": {
	"fail": "Should not have output BOM: https://encoding.spec.whatwg.org/#concept-td-serialize"
	},