6.0.0

Reverted back to the 4.0.0 behavior of returning encoding names instead of labels.

Added the xml option.

5.0.0

Raised the minimum required Node.js version to ^20.19.0 || ^22.12.0 || >=24.0.0.

The package now returns encoding labels, not encoding names. (In practice, it returns the lowercase versioned of what it previously returned.)

4.0.0

Raised the minimum required Node.js version to v18.

<meta charset="x-user-defined"> is now handled correctly.

3.0.0

Raised the minimum required Node.js version to v12.

Although it worked before by accident, as of this version any Uint8Array input is officially supported, not just Node.js Buffer objects.

2.0.1

• Fixed cases involving trailing spaces inside attribute values, e.g. content="text/html; charset=iso8859-2 ".
• Fixed cases involving two appearances of the string charset, e.g. content="charsetcharset=iso-8859-2".
• Fixed cases involving multiple attributes with the same name, e.g. <meta http-equiv="refresh" http-equiv="Content-Type" content="text/html; charset=iso8859-2">.

2.0.0

• (Breaking) updated minimum required Node.js version to v10.
• Fixed cases involving the character sequence ><meta.
• Fixed cases involving "short comments", i.e. <!-->.

1.0.2

Relicensed as MIT

1.0.1

Fixed an off-by-one error in the <meta> sniffing code that led to it failing to detect the encoding correctly when attributes were unquoted.

1.0.0

Extracted from jsdom, with some changes to do manual per-spec content-type parsing.