unicode-line-breaks¶

What this rule does¶

Reports raw occurrences of three Unicode characters that YAML 1.1 treated as line breaks but YAML 1.2 does not:

Character	Name	YAML escape
`U+0085`	next line (NEL)	`\N`
`U+2028`	line separator	`\L`
`U+2029`	paragraph separator	`\P`

The rule scans the decoded source text and flags every raw occurrence, wherever it appears — inside a scalar, a key, or a comment. It is off by default.

Why this matters¶

YAML 1.1 treated a broad Unicode set as line breaks, including NEL, LS, and PS. YAML 1.2 narrowed line breaks to just line feed (\n) and carriage return (\r), and the 1.2.2 changes page records that these three characters "are no longer considered line-break characters." ryl targets YAML 1.2.

This makes a raw occurrence a portability trap: a YAML 1.1 parser splits the line where a 1.2 parser keeps the character as ordinary scalar content, so the same bytes can produce a different parsed structure depending on the tool. The characters are also invisible in virtually every editor, so a stray one — pasted in from a word processor, a PDF, or a web page — is almost impossible to spot by eye.

If you genuinely need one of these characters, YAML 1.2 gives each a dedicated escape (\N, \L, \P) that includes it visibly and portably inside a double-quoted scalar.

Sources: YAML 1.2.2 changes page; YAML 1.2.2 spec §5.1 (character set), §5.4 (line-break characters), §5.7 (escaped characters).

Configuration¶

unicode-line-breaks is a ryl-only rule (yamllint has no equivalent), so it is configured only in TOML — [rules.unicode-line-breaks] in .ryl.toml/ryl.toml or [tool.ryl.rules.unicode-line-breaks] in pyproject.toml. It is rejected in yamllint-compatible YAML config (including -d data) so the YAML namespace stays reserved for any future yamllint rule.

[rules.unicode-line-breaks]
level = "error"

The rule has no options: when enabled it flags all three characters everywhere.

Examples¶

Reported¶

A double-quoted scalar containing a raw U+2028 (shown here as <LS>; in a real file the character is invisible):

title: "first line<LS>second line"

1:19  error  forbidden raw line separator U+2028; escape as "\L" in a double-quoted scalar  (unicode-line-breaks)

The rule also fires on a raw U+0085/U+2029 in a plain scalar or a comment.

Allowed¶

Use the dedicated escape inside a double-quoted scalar:

title: "first line\Lsecond line"

Automatic fixing¶

This rule does not auto-fix. The escape (\N/\L/\P) is only valid inside a double-quoted scalar, so rewriting a plain or single-quoted scalar, a comment, or a block scalar would require changing the quoting style or guessing intent; no single rewrite is universally safe.

new-lines — enforces a consistent line ending (LF vs CRLF) for the real line breaks.
quoted-strings — governs the quoting style you need in order to write a \L/\N/\P escape.