From 8d83609a120d27b89e80828b4972408984ba56e2 Mon Sep 17 00:00:00 2001
From: DorZuberi <admin@dorzki.io>
Date: Mon, 30 Mar 2026 14:56:28 +0300
Subject: [PATCH] Add Bash language support with patterns and tests

---
 src/Highlighter.php                           |  4 +-
 src/Languages/Bash/BashLanguage.php           | 72 +++++++++++++++++++
 .../Bash/Patterns/BashBuiltinPattern.php      | 33 +++++++++
 .../Bash/Patterns/BashCommentPattern.php      | 26 +++++++
 .../Bash/Patterns/BashFlagPattern.php         | 27 +++++++
 .../Bash/Patterns/BashKeywordPattern.php      | 35 +++++++++
 .../Bash/Patterns/BashNumberPattern.php       | 26 +++++++
 .../Bash/Patterns/BashOperatorPattern.php     | 27 +++++++
 .../Bash/Patterns/BashShebangPattern.php      | 27 +++++++
 .../Bash/Patterns/BashVariablePattern.php     | 27 +++++++
 tests/Bench/Fixtures/bash.txt                 | 60 ++++++++++++++++
 tests/Bench/HighlighterBench.php              |  1 +
 tests/Languages/Bash/BashLanguageTest.php     | 71 ++++++++++++++++++
 13 files changed, 435 insertions(+), 1 deletion(-)
 create mode 100644 src/Languages/Bash/BashLanguage.php
 create mode 100644 src/Languages/Bash/Patterns/BashBuiltinPattern.php
 create mode 100644 src/Languages/Bash/Patterns/BashCommentPattern.php
 create mode 100644 src/Languages/Bash/Patterns/BashFlagPattern.php
 create mode 100644 src/Languages/Bash/Patterns/BashKeywordPattern.php
 create mode 100644 src/Languages/Bash/Patterns/BashNumberPattern.php
 create mode 100644 src/Languages/Bash/Patterns/BashOperatorPattern.php
 create mode 100644 src/Languages/Bash/Patterns/BashShebangPattern.php
 create mode 100644 src/Languages/Bash/Patterns/BashVariablePattern.php
 create mode 100644 tests/Bench/Fixtures/bash.txt
 create mode 100644 tests/Languages/Bash/BashLanguageTest.php

diff --git a/src/Highlighter.php b/src/Highlighter.php
index 1374bd4..b253643 100644
--- a/src/Highlighter.php
+++ b/src/Highlighter.php
@@ -6,6 +6,7 @@
 
 use ReflectionClass;
 use Tempest\Highlight\Languages\Base\Injections\GutterInjection;
+use Tempest\Highlight\Languages\Bash\BashLanguage;
 use Tempest\Highlight\Languages\Blade\BladeLanguage;
 use Tempest\Highlight\Languages\Css\CssLanguage;
 use Tempest\Highlight\Languages\Diff\DiffLanguage;
@@ -51,7 +52,8 @@ final class Highlighter
 
     public function __construct(private readonly Theme $theme = new CssTheme())
     {
-        $this->addLanguage(new BladeLanguage())
+        $this->addLanguage(new BashLanguage())
+            ->addLanguage(new BladeLanguage())
             ->addLanguage(new CssLanguage())
             ->addLanguage(new DiffLanguage())
             ->addLanguage(new DocCommentLanguage())
diff --git a/src/Languages/Bash/BashLanguage.php b/src/Languages/Bash/BashLanguage.php
new file mode 100644
index 0000000..830ba7b
--- /dev/null
+++ b/src/Languages/Bash/BashLanguage.php
@@ -0,0 +1,72 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Bash;
+
+use Override;
+use Tempest\Highlight\Languages\Base\BaseLanguage;
+use Tempest\Highlight\Languages\Bash\Patterns\BashBuiltinPattern;
+use Tempest\Highlight\Languages\Bash\Patterns\BashCommentPattern;
+use Tempest\Highlight\Languages\Bash\Patterns\BashFlagPattern;
+use Tempest\Highlight\Languages\Bash\Patterns\BashKeywordPattern;
+use Tempest\Highlight\Languages\Bash\Patterns\BashNumberPattern;
+use Tempest\Highlight\Languages\Bash\Patterns\BashOperatorPattern;
+use Tempest\Highlight\Languages\Bash\Patterns\BashShebangPattern;
+use Tempest\Highlight\Languages\Bash\Patterns\BashVariablePattern;
+use Tempest\Highlight\Languages\Php\Patterns\DoubleQuoteValuePattern;
+use Tempest\Highlight\Languages\Php\Patterns\SingleQuoteValuePattern;
+
+class BashLanguage extends BaseLanguage
+{
+    public function getName(): string
+    {
+        return 'bash';
+    }
+
+    #[Override]
+    public function getAliases(): array
+    {
+        return ['sh', 'shell'];
+    }
+
+    #[Override]
+    public function getInjections(): array
+    {
+        return [
+            ...parent::getInjections(),
+        ];
+    }
+
+    #[Override]
+    public function getPatterns(): array
+    {
+        return [
+            ...parent::getPatterns(),
+
+            // COMMENTS
+            new BashShebangPattern(),
+            new BashCommentPattern(),
+
+            // VALUES
+            new SingleQuoteValuePattern(),
+            new DoubleQuoteValuePattern(),
+
+            // KEYWORDS
+            new BashKeywordPattern(),
+            new BashBuiltinPattern(),
+
+            // VARIABLES
+            new BashVariablePattern(),
+
+            // OPERATORS
+            new BashOperatorPattern(),
+
+            // NUMBERS
+            new BashNumberPattern(),
+
+            // FLAGS
+            new BashFlagPattern(),
+        ];
+    }
+}
\ No newline at end of file
diff --git a/src/Languages/Bash/Patterns/BashBuiltinPattern.php b/src/Languages/Bash/Patterns/BashBuiltinPattern.php
new file mode 100644
index 0000000..d0ec35e
--- /dev/null
+++ b/src/Languages/Bash/Patterns/BashBuiltinPattern.php
@@ -0,0 +1,33 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Bash\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+final readonly class BashBuiltinPattern implements Pattern
+{
+    use IsPattern;
+
+    public function __construct(private array $builtins = [
+        'echo', 'printf', 'cd', 'pwd', 'test', 'eval', 'exec',
+        'set', 'read', 'exit',
+    ])
+    {
+    }
+
+    public function getPattern(): string
+    {
+        $builtins = implode('|', $this->builtins);
+
+        return "\b(?<match>(?:{$builtins}))\b";
+    }
+
+    public function getTokenType(): TokenTypeEnum
+    {
+        return TokenTypeEnum::TYPE;
+    }
+}
\ No newline at end of file
diff --git a/src/Languages/Bash/Patterns/BashCommentPattern.php b/src/Languages/Bash/Patterns/BashCommentPattern.php
new file mode 100644
index 0000000..742b807
--- /dev/null
+++ b/src/Languages/Bash/Patterns/BashCommentPattern.php
@@ -0,0 +1,26 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Bash\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: '# this is a comment', output: '# this is a comment')]
+final readonly class BashCommentPattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return '(^|[^\\\\])(?<match>#.*)';
+    }
+
+    public function getTokenType(): TokenTypeEnum
+    {
+        return TokenTypeEnum::COMMENT;
+    }
+}
\ No newline at end of file
diff --git a/src/Languages/Bash/Patterns/BashFlagPattern.php b/src/Languages/Bash/Patterns/BashFlagPattern.php
new file mode 100644
index 0000000..814c389
--- /dev/null
+++ b/src/Languages/Bash/Patterns/BashFlagPattern.php
@@ -0,0 +1,27 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Bash\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: 'ls -la', output: '-la')]
+#[PatternTest(input: 'grep --color=auto', output: '--color')]
+final readonly class BashFlagPattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return '(?<match>--?[a-zA-Z][\w-]*)';
+    }
+
+    public function getTokenType(): TokenTypeEnum
+    {
+        return TokenTypeEnum::GENERIC;
+    }
+}
\ No newline at end of file
diff --git a/src/Languages/Bash/Patterns/BashKeywordPattern.php b/src/Languages/Bash/Patterns/BashKeywordPattern.php
new file mode 100644
index 0000000..9679c87
--- /dev/null
+++ b/src/Languages/Bash/Patterns/BashKeywordPattern.php
@@ -0,0 +1,35 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Bash\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+final readonly class BashKeywordPattern implements Pattern
+{
+    use IsPattern;
+
+    public function __construct(private array $keywords = [
+        'if', 'then', 'else', 'elif', 'fi', 'for', 'while', 'do', 'done',
+        'case', 'esac', 'function', 'return', 'in', 'select', 'until',
+        'break', 'continue', 'declare', 'local', 'export', 'readonly',
+        'unset', 'shift', 'trap', 'source',
+    ])
+    {
+    }
+
+    public function getPattern(): string
+    {
+        $keywords = implode('|', $this->keywords);
+
+        return "\b(?<match>(?:{$keywords}))\b";
+    }
+
+    public function getTokenType(): TokenTypeEnum
+    {
+        return TokenTypeEnum::KEYWORD;
+    }
+}
\ No newline at end of file
diff --git a/src/Languages/Bash/Patterns/BashNumberPattern.php b/src/Languages/Bash/Patterns/BashNumberPattern.php
new file mode 100644
index 0000000..2accaa8
--- /dev/null
+++ b/src/Languages/Bash/Patterns/BashNumberPattern.php
@@ -0,0 +1,26 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Bash\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: 'exit 0', output: '0')]
+final readonly class BashNumberPattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return '\b(?<match>\d+)\b';
+    }
+
+    public function getTokenType(): TokenTypeEnum
+    {
+        return TokenTypeEnum::NUMBER;
+    }
+}
\ No newline at end of file
diff --git a/src/Languages/Bash/Patterns/BashOperatorPattern.php b/src/Languages/Bash/Patterns/BashOperatorPattern.php
new file mode 100644
index 0000000..82e60db
--- /dev/null
+++ b/src/Languages/Bash/Patterns/BashOperatorPattern.php
@@ -0,0 +1,27 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Bash\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: 'cmd1 | cmd2', output: '|')]
+#[PatternTest(input: 'cmd1 && cmd2', output: '&&')]
+final readonly class BashOperatorPattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return '(?<match>\|\||&&|>>|<<|;;|[|><;])';
+    }
+
+    public function getTokenType(): TokenTypeEnum
+    {
+        return TokenTypeEnum::OPERATOR;
+    }
+}
\ No newline at end of file
diff --git a/src/Languages/Bash/Patterns/BashShebangPattern.php b/src/Languages/Bash/Patterns/BashShebangPattern.php
new file mode 100644
index 0000000..2b08c82
--- /dev/null
+++ b/src/Languages/Bash/Patterns/BashShebangPattern.php
@@ -0,0 +1,27 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Bash\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: '#!/bin/bash', output: '#!/bin/bash')]
+#[PatternTest(input: '#!/usr/bin/env bash', output: '#!/usr/bin/env bash')]
+final readonly class BashShebangPattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return '/(?<match>\#!\/.*)/';
+    }
+
+    public function getTokenType(): TokenTypeEnum
+    {
+        return TokenTypeEnum::COMMENT;
+    }
+}
\ No newline at end of file
diff --git a/src/Languages/Bash/Patterns/BashVariablePattern.php b/src/Languages/Bash/Patterns/BashVariablePattern.php
new file mode 100644
index 0000000..c1076da
--- /dev/null
+++ b/src/Languages/Bash/Patterns/BashVariablePattern.php
@@ -0,0 +1,27 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Bash\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: 'echo $HOME', output: '$HOME')]
+#[PatternTest(input: 'echo ${PATH}', output: '${PATH}')]
+final readonly class BashVariablePattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return '(?<match>\$\{[^}]+\}|\$[a-zA-Z_]\w*|\$[\d@?#!\$\-\*])';
+    }
+
+    public function getTokenType(): TokenTypeEnum
+    {
+        return TokenTypeEnum::VARIABLE;
+    }
+}
\ No newline at end of file
diff --git a/tests/Bench/Fixtures/bash.txt b/tests/Bench/Fixtures/bash.txt
new file mode 100644
index 0000000..a6a4381
--- /dev/null
+++ b/tests/Bench/Fixtures/bash.txt
@@ -0,0 +1,60 @@
+#!/bin/bash
+
+# Configuration
+MAX_RETRIES=3
+LOG_FILE="/var/log/deploy.log"
+
+# Function to check service health
+check_health() {
+    local url=$1
+    local timeout=${2:-5}
+
+    if curl --silent --fail --max-time "$timeout" "$url" > /dev/null 2>&1; then
+        echo "Service is healthy"
+        return 0
+    else
+        echo "Service is not responding" >> "$LOG_FILE"
+        return 1
+    fi
+}
+
+# Deploy with retries
+deploy() {
+    local attempt=1
+
+    while [ $attempt -le $MAX_RETRIES ]; do
+        echo "Deployment attempt $attempt of $MAX_RETRIES"
+
+        if docker-compose up -d --build 2>&1 | tee -a "$LOG_FILE"; then
+            echo "Deployment successful"
+            break
+        fi
+
+        attempt=$((attempt + 1))
+        sleep 5
+    done
+
+    if [ $attempt -gt $MAX_RETRIES ]; then
+        echo "Deployment failed after $MAX_RETRIES attempts" >&2
+        exit 1
+    fi
+}
+
+# Parse arguments
+case "$1" in
+    start)
+        deploy
+        ;;
+    status)
+        check_health "http://localhost:8080/health"
+        ;;
+    *)
+        echo "Usage: $0 {start|status}"
+        exit 1
+        ;;
+esac
+
+for service in web api worker; do
+    printf "Checking %s...\n" "$service"
+    systemctl is-active --quiet "$service" && echo "OK" || echo "FAILED"
+done
\ No newline at end of file
diff --git a/tests/Bench/HighlighterBench.php b/tests/Bench/HighlighterBench.php
index 0516f96..2979dc3 100644
--- a/tests/Bench/HighlighterBench.php
+++ b/tests/Bench/HighlighterBench.php
@@ -17,6 +17,7 @@ final class HighlighterBench
     private const string FIXTURES_DIR = __DIR__ . '/Fixtures';
 
     public const array LANGUAGES = [
+        'bash' => 'bash.txt',
         'blade' => 'blade.txt',
         'css' => 'css.txt',
         'diff' => 'diff.txt',
diff --git a/tests/Languages/Bash/BashLanguageTest.php b/tests/Languages/Bash/BashLanguageTest.php
new file mode 100644
index 0000000..a025c89
--- /dev/null
+++ b/tests/Languages/Bash/BashLanguageTest.php
@@ -0,0 +1,71 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Tests\Languages\Bash;
+
+use PHPUnit\Framework\Attributes\DataProvider;
+use PHPUnit\Framework\TestCase;
+use Tempest\Highlight\Highlighter;
+
+class BashLanguageTest extends TestCase
+{
+    #[DataProvider('provide_highlight_cases')]
+    public function test_highlight(string $content, string $expected): void
+    {
+        $highlighter = new Highlighter();
+
+        $this->assertSame(
+            $expected,
+            $highlighter->parse($content, 'bash'),
+        );
+
+        $this->assertSame(
+            $expected,
+            $highlighter->parse($content, 'sh'),
+        );
+
+        $this->assertSame(
+            $expected,
+            $highlighter->parse($content, 'shell'),
+        );
+    }
+
+    public static function provide_highlight_cases(): iterable
+    {
+        return [
+            [<<<'TXT'
+#!/bin/bash
+
+# Check if file exists
+if [ -f "$1" ]; then
+    echo "File exists"
+    cat "$1" | wc -l
+else
+    echo "File not found" >> /tmp/log
+    exit 1
+fi
+
+for i in 1 2 3; do
+    echo $i
+done
+TXT,
+            <<<'TXT'
+<span class="hl-comment">#!/bin/bash</span>
+
+<span class="hl-comment"># Check if file exists</span>
+<span class="hl-keyword">if</span> [ <span class="hl-generic">-f</span> <span class="hl-value">&quot;$1&quot;</span> ]<span class="hl-operator">;</span> <span class="hl-keyword">then</span>
+    <span class="hl-type">echo</span> <span class="hl-value">&quot;File exists&quot;</span>
+    cat <span class="hl-value">&quot;$1&quot;</span> <span class="hl-operator">|</span> wc <span class="hl-generic">-l</span>
+<span class="hl-keyword">else</span>
+    <span class="hl-type">echo</span> <span class="hl-value">&quot;File not found&quot;</span> <span class="hl-operator">&gt;&gt;</span> /tmp/log
+    <span class="hl-type">exit</span> <span class="hl-number">1</span>
+<span class="hl-keyword">fi</span>
+
+<span class="hl-keyword">for</span> i <span class="hl-keyword">in</span> <span class="hl-number">1</span> <span class="hl-number">2</span> <span class="hl-number">3</span><span class="hl-operator">;</span> <span class="hl-keyword">do</span>
+    <span class="hl-type">echo</span> <span class="hl-variable">$i</span>
+<span class="hl-keyword">done</span>
+TXT],
+        ];
+    }
+}
\ No newline at end of file