mbolli/tempest-highlight-ron

RON (Readable Object Notation) language support for tempest/highlight — parser-backed, role-aware highlighting of keys, values, numbers, literals, and structure.

v1.0.1 2026-06-15 11:17 UTC

Requires

• php: ^8.4
• mbolli/php-ron: ^0.1.1
• tempest/highlight: ^2.17

Requires (Dev)

• friendsofphp/php-cs-fixer: ^3.0
• phpstan/phpstan: ^2.0
• phpunit/phpunit: ^11.0

MIT

• Michael Bolli

• ron
• syntax-highlighting
• tempest-highlight
• readable-object-notation

README

RON (Readable Object Notation) language support for tempest/highlight.

Unlike a regex grammar, this plugin tokenizes with the real RON parser (mbolli/php-ron), so highlighting is byte-exact and context-aware: keys are distinguished from values even though RON has no :/= separator, and repeated-quote strings (''''' , ""a "quoted" phrase"") are spanned correctly.

Installation

composer require mbolli/tempest-highlight-ron

Usage

use Tempest\Highlight\Highlighter;
use Mbolli\TempestHighlightRon\RonLanguage;

$highlighter = new Highlighter();
$highlighter->addLanguage(new RonLanguage());

echo $highlighter->parse($ron, 'ron');

Registering the language as ron also highlights fenced ```ron blocks in markdown rendered through tempest/highlight.

What gets highlighted

For example RON like:

users [
  {id 100 name Ada roles [admin writer] active true}
]
settings {retry {max 3} tags [llm json]}

RON construct	Role	tempest token	CSS class
{ } [ ]	punctuation	PROPERTY	hl-property
object keys (bare/quoted, elided or braced)	key	KEYWORD	hl-keyword
string values (bare, quoted, repeated-quote)	string	VALUE	hl-value
numbers (100, -12.5e+2)	number	NUMBER	hl-number
true / false / null	literal	LITERAL	hl-literal

Because classification comes from the parser, 'true' (quoted) is a string, a key named true is a key, and a bare 123 in key position stays a key — none are mis-coloured.

How it works

RonLanguage adds five Patterns on top of tempest's base language. Each delegates to Mbolli\Ron\Ron::tokenize(), which returns role-aware source spans; the spans are handed to tempest in PREG_OFFSET_CAPTURE shape. Tokenization is memoized per content string, so the source is lexed once per parse regardless of how many patterns run. Ron::tokenize() is lenient — malformed RON never throws, it just highlights as much as it can classify.

License

MIT