Add TOML syntax highlighting

src/Highlighter.php

@@ -36,6 +36,7 @@
 use Tempest\Highlight\Languages\Terminal\TerminalLanguage;
 use Tempest\Highlight\Languages\Terraform\TerraformLanguage;
 use Tempest\Highlight\Languages\Text\TextLanguage;
+use Tempest\Highlight\Languages\Toml\TomlLanguage;
 use Tempest\Highlight\Languages\Tsx\TsxLanguage;
 use Tempest\Highlight\Languages\Twig\TwigLanguage;
 use Tempest\Highlight\Languages\TypeScript\TypeScriptLanguage;
@@ -94,6 +95,7 @@ public function __construct(
             ->addLanguage(new TerminalLanguage())
             ->addLanguage(new TextLanguage())
             ->addLanguage(new TerraformLanguage())
+            ->addLanguage(new TomlLanguage())
             ->addLanguage(new TypeScriptLanguage())
             ->addLanguage(new XmlLanguage())
             ->addLanguage(new YamlLanguage())

src/Languages/Toml/Patterns/TomlBooleanPattern.php

@@ -0,0 +1,28 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Toml\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenType;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: 'enabled = true', output: 'true')]
+#[PatternTest(input: 'enabled = false', output: 'false')]
+final class TomlBooleanPattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return '/(?<![A-Za-z0-9_-])(?<match>true|false)(?![A-Za-z0-9_-])/';
+    }
+
+    public function getTokenType(): TokenType
+    {
+        return TokenTypeEnum::LITERAL;
+    }
+}

src/Languages/Toml/Patterns/TomlCommentPattern.php

@@ -0,0 +1,27 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Toml\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenType;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: '# Server settings', output: '# Server settings')]
+final class TomlCommentPattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return '/(?<match>#[^\r\n]*)/';
+    }
+
+    public function getTokenType(): TokenType
+    {
+        return TokenTypeEnum::COMMENT;
+    }
+}

src/Languages/Toml/Patterns/TomlDateTimePattern.php

@@ -0,0 +1,28 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Toml\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenType;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: 'created = 1979-05-27T07:32:00Z', output: '1979-05-27T07:32:00Z')]
+#[PatternTest(input: 'alarm = 07:32:00', output: '07:32:00')]
+final class TomlDateTimePattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return '/(?<![A-Za-z0-9_-])(?<match>(?:\d{4}-\d{2}-\d{2}(?:[Tt ]\d{2}:\d{2}:\d{2}(?:\.\d+)?(?:[Zz]|[+-]\d{2}:\d{2})?)?|\d{2}:\d{2}:\d{2}(?:\.\d+)?))(?![A-Za-z0-9_-])/';
+    }
+
+    public function getTokenType(): TokenType
+    {
+        return TokenTypeEnum::VALUE;
+    }
+}

src/Languages/Toml/Patterns/TomlKeyPattern.php

@@ -0,0 +1,30 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Toml\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenType;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: 'title = "TOML"', output: 'title')]
+#[PatternTest(input: 'fruit.physical.color = "orange"', output: 'fruit.physical.color')]
+final class TomlKeyPattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return <<<'REGEX'
+/(?:^[\t ]*|(?<=[{,])[\t ]*)(?<match>(?:"(?:\\(?:["\\bfnrt]|u[0-9A-Fa-f]{4}|U[0-9A-Fa-f]{8})|[^"\\\r\n])*"|'[^'\r\n]*'|[A-Za-z0-9_-]+)(?:[\t ]*\.[\t ]*(?:"(?:\\(?:["\\bfnrt]|u[0-9A-Fa-f]{4}|U[0-9A-Fa-f]{8})|[^"\\\r\n])*"|'[^'\r\n]*'|[A-Za-z0-9_-]+))*)(?=[\t ]*=)/m
+REGEX;
+    }
+
+    public function getTokenType(): TokenType
+    {
+        return TokenTypeEnum::PROPERTY;
+    }
+}

src/Languages/Toml/Patterns/TomlNumberPattern.php

@@ -0,0 +1,31 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Toml\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenType;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: 'answer = 42', output: '42')]
+#[PatternTest(input: 'color = 0xDEAD_BEEF', output: '0xDEAD_BEEF')]
+#[PatternTest(input: 'limit = +inf', output: '+inf')]
+final class TomlNumberPattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return <<<'REGEX'
+/(?<![A-Za-z0-9_.-])(?<match>[+-]?(?:inf|nan)|0x[0-9A-Fa-f](?:_?[0-9A-Fa-f])*|0o[0-7](?:_?[0-7])*|0b[01](?:_?[01])*|[+-]?(?:\d(?:_?\d)*\.\d(?:_?\d)*(?:[eE][+-]?\d(?:_?\d)*)?|\d(?:_?\d)*[eE][+-]?\d(?:_?\d)*|\d(?:_?\d)*))(?![A-Za-z0-9_.-])/
+REGEX;
+    }
+
+    public function getTokenType(): TokenType
+    {
+        return TokenTypeEnum::NUMBER;
+    }
+}

src/Languages/Toml/Patterns/TomlOperatorPattern.php

@@ -0,0 +1,28 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Toml\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenType;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: 'ports = [8000, 8001]', output: '=')]
+#[PatternTest(input: '[[products]]', output: '[[')]
+final class TomlOperatorPattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return '/(?<match>\[\[|\]\]|\[|\]|\{|\}|=|,)/';
+    }
+
+    public function getTokenType(): TokenType
+    {
+        return TokenTypeEnum::OPERATOR;
+    }
+}

src/Languages/Toml/Patterns/TomlStringPattern.php

@@ -0,0 +1,30 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Toml\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenType;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: 'name = "TOML"', output: '"TOML"')]
+#[PatternTest(input: 'path = \'C:\\Users\'', output: '\'C:\\Users\'')]
+final class TomlStringPattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return <<<'REGEX'
+/(?<match>"""(?:\\[\s\S]|"{1,2}(?!")|[^"\\])*"""|'''(?:'{1,2}(?!')|[^'])*'''|"(?:\\(?:["\\bfnrt]|u[0-9A-Fa-f]{4}|U[0-9A-Fa-f]{8})|[^"\\\r\n])*"|'[^'\r\n]*')/
+REGEX;
+    }
+
+    public function getTokenType(): TokenType
+    {
+        return TokenTypeEnum::VALUE;
+    }
+}

src/Languages/Toml/Patterns/TomlTablePattern.php

@@ -0,0 +1,30 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Toml\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenType;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: '[database]', output: 'database')]
+#[PatternTest(input: '[[products]]', output: 'products')]
+final class TomlTablePattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return <<<'REGEX'
+/^[\t ]*\[{1,2}[\t ]*(?<match>(?:"(?:\\(?:["\\bfnrt]|u[0-9A-Fa-f]{4}|U[0-9A-Fa-f]{8})|[^"\\\r\n])*"|'[^'\r\n]*'|[A-Za-z0-9_-]+)(?:[\t ]*\.[\t ]*(?:"(?:\\(?:["\\bfnrt]|u[0-9A-Fa-f]{4}|U[0-9A-Fa-f]{8})|[^"\\\r\n])*"|'[^'\r\n]*'|[A-Za-z0-9_-]+))*)(?=[\t ]*\]{1,2})/m
+REGEX;
+    }
+
+    public function getTokenType(): TokenType
+    {
+        return TokenTypeEnum::TYPE;
+    }
+}

src/Languages/Toml/TomlLanguage.php

@@ -0,0 +1,40 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Toml;
+
+use Override;
+use Tempest\Highlight\Languages\Base\BaseLanguage;
+use Tempest\Highlight\Languages\Toml\Patterns\TomlBooleanPattern;
+use Tempest\Highlight\Languages\Toml\Patterns\TomlCommentPattern;
+use Tempest\Highlight\Languages\Toml\Patterns\TomlDateTimePattern;
+use Tempest\Highlight\Languages\Toml\Patterns\TomlKeyPattern;
+use Tempest\Highlight\Languages\Toml\Patterns\TomlNumberPattern;
+use Tempest\Highlight\Languages\Toml\Patterns\TomlOperatorPattern;
+use Tempest\Highlight\Languages\Toml\Patterns\TomlStringPattern;
+use Tempest\Highlight\Languages\Toml\Patterns\TomlTablePattern;
+
+final class TomlLanguage extends BaseLanguage
+{
+    public function getName(): string
+    {
+        return 'toml';
+    }
+
+    #[Override]
+    public function getPatterns(): array
+    {
+        return [
+            ...parent::getPatterns(),
+            new TomlKeyPattern(),
+            new TomlTablePattern(),
+            new TomlStringPattern(),
+            new TomlDateTimePattern(),
+            new TomlNumberPattern(),
+            new TomlBooleanPattern(),
+            new TomlCommentPattern(),
+            new TomlOperatorPattern(),
+        ];
+    }
+}

tests/Bench/Fixtures/toml.txt

@@ -0,0 +1,23 @@
+# Application settings
+title = "Nightly Service"
+"site owner" = 'Operations'
+enabled = true
+ports = [8000, 8001, 8002]
+ratio = 6.626e-34
+mask = 0xFF_EC_DE_5E
+released = 2026-10-19T03:45:00Z
+alarm = 07:30:00
+description = """A multiline
+service description."""
+
+[database.primary]
+host = "db.internal"
+credentials = { user = "service", retries = 3 }
+
+[[servers]]
+name = "api-01"
+roles = ["web", "worker"]
+
+[[servers]]
+name = "api-02"
+roles = ["web"]

tests/Bench/HighlighterBench.php

@@ -44,6 +44,7 @@ final class HighlighterBench
         'swift' => 'swift.txt',
         'terminal' => 'terminal.txt',
         'terraform' => 'terraform.txt',
+        'toml' => 'toml.txt',
         'typescript' => 'typescript.txt',
         'txt' => 'text.txt',
         'twig' => 'twig.txt',