Linter Rule: Disallow a byte order mark at the start of a template

Rule: erb-no-byte-order-mark

Description

Reports a UTF-8 byte order mark (U+FEFF, the bytes EF BB BF) at the very start of a template.

Rationale

A byte order mark exists to record the endianness of UTF-16 and UTF-32 text. UTF-8 is a sequence of single bytes and has no endianness, so the mark carries no information there. Windows editors write it anyway as an encoding signature, and it survives in templates for years because it is zero width and nothing displays it.

In a template it is not metadata, it is content. ERB copies it into the response ahead of everything else, so the document starts with an invisible character before the doctype. That can push a browser into quirks mode, break code that checks whether a response begins with <, and show up as unexplained whitespace at the top of a page.

It also confuses tooling that pairs byte offsets with source text, because a leading U+FEFF is stripped by default in most UTF-8 decoders while the bytes are still counted everywhere else.

The formatter removes a leading byte order mark, so herb format fixes this too.

Examples

✅ Good

erb

<!DOCTYPE html>
<html></html>

A byte order mark that is genuinely part of the content is left alone:

erb

<div>ab</div>

🚫 Bad

erb

<!DOCTYPE html>Remove the byte order mark from the start of the file. It is an encoding signature, not template content, so it renders as an invisible character ahead of everything else and can push browsers into quirks mode. Save the file as UTF-8 without a BOM. (erb-no-byte-order-mark)
<html></html>

Linter Rule: Disallow a byte order mark at the start of a template ​

Description ​

Rationale ​

Examples ​

✅ Good ​

🚫 Bad ​

References ​