Feat: Add syntax highlighting for a lot more languages

samples/Makefile

@@ -0,0 +1,119 @@
+SRC_DIR := src
+BIN_DIR := bin
+OBJ_DIR := build
+INCLUDE_DIR := include
+
+TARGET_DEBUG := $(BIN_DIR)/crib-dbg
+TARGET_RELEASE := $(BIN_DIR)/crib
+
+PCH_DEBUG := $(OBJ_DIR)/debug/pch.h.gch
+PCH_RELEASE := $(OBJ_DIR)/release/pch.h.gch
+
+CCACHE := ccache
+CXX_DEBUG := $(CCACHE) g++
+CXX_RELEASE := $(CCACHE) clang++
+
+CFLAGS_DEBUG := -std=c++20 -Wall -Wextra -O0 -fno-inline -gsplit-dwarf -g -fsanitize=address -fno-omit-frame-pointer
+CFLAGS_RELEASE := -std=c++20 -O3 -march=native -flto=thin \
+	-fno-exceptions -fno-rtti -fstrict-aliasing \
+	-ffast-math -funroll-loops \
+	-fvisibility=hidden \
+	-fomit-frame-pointer -DNDEBUG -s \
+	-mllvm -vectorize-loops \
+	-fno-unwind-tables -fno-asynchronous-unwind-tables
+
+PCH_CFLAGS_DEBUG := $(CFLAGS_DEBUG) -x c++-header
+PCH_CFLAGS_RELEASE := $(CFLAGS_RELEASE) -x c++-header
+
+UNICODE_SRC := $(wildcard libs/unicode_width/*.c)
+
+UNICODE_OBJ_DEBUG := $(patsubst libs/unicode_width/%.c,$(OBJ_DIR)/debug/unicode_width/%.o,$(UNICODE_SRC))
+UNICODE_OBJ_RELEASE := $(patsubst libs/unicode_width/%.c,$(OBJ_DIR)/release/unicode_width/%.o,$(UNICODE_SRC))
+
+TREE_SITTER_LIBS := $(wildcard libs/tree-sitter-*/libtree-sitter*.a)
+
+PHP_LIB := libs/tree-sitter-php/php/libtree-sitter-php.a
+
+NGINX_OBJ_PARSER := libs/tree-sitter-nginx/build/Release/obj.target/tree_sitter_nginx_binding/src/parser.o
+
+GITIGNORE_OBJ_PARSER := libs/tree-sitter-gitignore/build/Release/obj.target/tree_sitter_ignore_binding/src/parser.o
+
+FISH_OBJ_PARSER := libs/tree-sitter-fish/build/Release/obj.target/tree_sitter_fish_binding/src/parser.o
+FISH_OBJ_SCANNER := libs/tree-sitter-fish/build/Release/obj.target/tree_sitter_fish_binding/src/scanner.o
+
+MD_OBJ_PARSER := libs/tree-sitter-markdown/build/Release/obj.target/tree_sitter_markdown_binding/tree-sitter-markdown/src/parser.o
+MD_OBJ_SCANNER := libs/tree-sitter-markdown/build/Release/obj.target/tree_sitter_markdown_binding/tree-sitter-markdown/src/scanner.o
+
+MD_I_OBJ_PARSER := libs/tree-sitter-markdown/build/Release/obj.target/tree_sitter_markdown_binding/tree-sitter-markdown-inline/src/parser.o
+MD_I_OBJ_SCANNER := libs/tree-sitter-markdown/build/Release/obj.target/tree_sitter_markdown_binding/tree-sitter-markdown-inline/src/scanner.o
+
+LIBS := \
+	libs/libgrapheme/libgrapheme.a \
+	libs/tree-sitter/libtree-sitter.a \
+	$(TREE_SITTER_LIBS) \
+	$(PHP_LIB) \
+	$(NGINX_OBJ_PARSER) \
+	$(GITIGNORE_OBJ_PARSER) \
+	$(FISH_OBJ_PARSER) \
+	$(FISH_OBJ_SCANNER) \
+	$(MD_OBJ_PARSER) \
+	$(MD_OBJ_SCANNER) \
+	$(MD_I_OBJ_PARSER) \
+	$(MD_I_OBJ_SCANNER) \
+	-lpcre2-8 -lmagic
+
+SRC := $(wildcard $(SRC_DIR)/*.cc)
+OBJ_DEBUG := $(patsubst $(SRC_DIR)/%.cc,$(OBJ_DIR)/debug/%.o,$(SRC))
+OBJ_RELEASE := $(patsubst $(SRC_DIR)/%.cc,$(OBJ_DIR)/release/%.o,$(SRC))
+
+DEP_DEBUG := $(OBJ_DEBUG:.o=.d)
+DEP_RELEASE := $(OBJ_RELEASE:.o=.d)
+
+.PHONY: all test release clean
+
+all: debug
+
+test: $(TARGET_DEBUG)
+
+release: $(TARGET_RELEASE)
+
+$(PCH_DEBUG): $(INCLUDE_DIR)/pch.h
+	mkdir -p $(dir $@)
+	$(CXX_DEBUG) $(PCH_CFLAGS_DEBUG) -o $@ $<
+
+$(PCH_RELEASE): $(INCLUDE_DIR)/pch.h
+	mkdir -p $(dir $@)
+	$(CXX_RELEASE) $(PCH_CFLAGS_RELEASE) -o $@ $<
+
+$(TARGET_DEBUG): $(PCH_DEBUG) $(OBJ_DEBUG) $(UNICODE_OBJ_DEBUG)
+	mkdir -p $(BIN_DIR)
+	$(CXX_DEBUG) $(CFLAGS_DEBUG) -o $@ $(OBJ_DEBUG) $(UNICODE_OBJ_DEBUG) $(LIBS)
+
+$(TARGET_RELEASE): $(PCH_RELEASE) $(OBJ_RELEASE) $(UNICODE_OBJ_RELEASE)
+	mkdir -p $(BIN_DIR)
+	$(CXX_RELEASE) $(CFLAGS_RELEASE) -o $@ $(OBJ_RELEASE) $(UNICODE_OBJ_RELEASE) $(LIBS)
+
+$(OBJ_DIR)/debug/%.o: $(SRC_DIR)/%.cc $(PCH_DEBUG)
+	mkdir -p $(dir $@)
+	$(CXX_DEBUG) $(CFLAGS_DEBUG) -include $(INCLUDE_DIR)/pch.h -MMD -MP -c $< -o $@
+
+$(OBJ_DIR)/release/%.o: $(SRC_DIR)/%.cc $(PCH_RELEASE)
+	mkdir -p $(dir $@)
+	$(CXX_RELEASE) $(CFLAGS_RELEASE) -include $(INCLUDE_DIR)/pch.h -MMD -MP -c $< -o $@
+
+$(OBJ_DIR)/debug/unicode_width/%.o: libs/unicode_width/%.c
+	mkdir -p $(dir $@)
+	$(CXX_DEBUG) $(CFLAGS_DEBUG) -MMD -MP -c $< -o $@
+
+$(OBJ_DIR)/release/unicode_width/%.o: libs/unicode_width/%.c
+	mkdir -p $(dir $@)
+	$(CXX_RELEASE) $(CFLAGS_RELEASE) -MMD -MP -c $< -o $@
+
+DEP_DEBUG += $(UNICODE_OBJ_DEBUG:.o=.d)
+DEP_RELEASE += $(UNICODE_OBJ_RELEASE:.o=.d)
+
+-include $(DEP_DEBUG)
+-include $(DEP_RELEASE)
+
+clean:
+	rm -rf $(OBJ_DIR) $(BIN_DIR)

samples/css.css

@@ -0,0 +1,70 @@
+/* === Basic selectors === */
+body {
+  margin: 0;
+  font-family: system-ui, sans-serif;
+  background: #121212;
+  color: #eee;
+}
+
+/* Class + ID + attribute */
+#main.container[data-theme="dark"] {
+  padding: 1rem;
+  border: 1px solid rgba(255, 255, 255, 0.2);
+}
+
+/* Pseudo-classes & elements */
+a:hover,
+a:focus-visible {
+  color: hsl(210, 80%, 60%);
+  text-decoration: underline;
+}
+
+input::placeholder {
+  color: #999;
+}
+
+/* CSS variables */
+:root {
+  --accent: #4fc3f7;
+  --spacing: 1rem;
+}
+
+.button {
+  background: var(--accent);
+  padding: calc(var(--spacing) * 1.5);
+}
+
+/* Media query */
+@media (max-width: 768px) {
+  .container {
+    padding: 0.5rem;
+  }
+}
+
+/* Keyframes */
+@keyframes fade-in {
+  from {
+    opacity: 0;
+    transform: translateY(10px);
+  }
+  to {
+    opacity: 1;
+    transform: translateY(0);
+  }
+}
+
+/* Animation usage */
+.modal {
+  animation: fade-in 250ms ease-out;
+}
+
+/* Complex selector */
+ul > li:not(:last-child)::after {
+  content: "•";
+  margin-left: 0.5em;
+}
+
+/* Edge cases */
+[data-value^="test"]::before {
+  content: attr(data-value);
+}

samples/diff.patch

@@ -0,0 +1,102 @@
+--- ./samples/toml.toml	2025-12-26 19:02:50.480936043 +0000
++++ ./samples/yaml.yml	2025-12-26 19:03:27.879765974 +0000
+@@ -2,52 +2,65 @@
+ # Basic types
+ # ============================================================
+ 
+-title = "Example TOML Configuration"
+-enabled = true
+-count = 42
+-pi = 3.14159
+-empty = ""
++title: "Example YAML Configuration"
++enabled: true
++count: 42
++pi: 3.14159
++empty: ""
+ 
+ # ============================================================
+-# Arrays
++# Arrays / Lists
+ # ============================================================
+-fruits = ["apple", "banana", "cherry"]
+-numbers = [1, 2, 3, 4, 5]
++fruits:
++  - apple
++  - banana
++  - cherry
+ 
+-# Nested array
+-matrix = [[1, 2], [3, 4]]
++numbers:
++  - 1
++  - 2
++  - 3
++  - 4
++  - 5
++
++matrix:
++  - [1, 2]
++  - [3, 4]
+ 
+ # ============================================================
+-# Tables
++# Nested objects / maps
+ # ============================================================
+-[owner]
+-name = "Alice"
+-dob = 1979-05-27T07:32:00Z
+-
+-[database]
+-server = "192.168.1.1"
+-ports = [ 8001, 8001, 8002 ]
+-connection_max = 5000
+-enabled = true
++owner:
++  name: Alice
++  dob: 1979-05-27T07:32:00Z
+ 
+-[servers.alpha]
+-ip = "10.0.0.1"
+-dc = "east"
++database:
++  server: 192.168.1.1
++  ports:
++    - 8001
++    - 8001
++    - 8002
++  connection_max: 5000
++  enabled: true
+ 
+-[servers.beta]
+-ip = "10.0.0.2"
+-dc = "west"
++servers:
++  alpha:
++    ip: 10.0.0.1
++    dc: east
++  beta:
++    ip: 10.0.0.2
++    dc: west
+ 
+ # ============================================================
+-# Inline tables
++# Multiline string
+ # ============================================================
+-clients = { name = "Bob", age = 30, active = true }
++description: |
++  This is a YAML file
++  used for testing syntax highlighting.
++  It supports multiple lines.
+ 
+ # ============================================================
+-# Multiline strings
++# Special characters
+ # ============================================================
+-description = """
+-This is a TOML file
+-used for testing syntax highlighting.
+-It supports multiple lines.
+-"""
++regex_pattern: "^[A-Za-z0-9_]+$"
++path: "C:\\Users\\Alice\\Documents"

samples/embedded_template.erb

@@ -0,0 +1,71 @@
+<%# app/views/users/show.html.erb %>
+
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8">
+    <title><%= @user.name %> — Profile</title>
+
+    <%# Inline Ruby expression %>
+    <meta name="description" content="<%= @user.bio %>">
+
+    <% if @dark_mode %>
+      <style>
+        body {
+          background-color: #111;
+          color: #eee;
+        }
+      </style>
+    <% end %>
+  </head>
+
+  <body>
+    <!-- HTML comment -->
+    <header>
+      <h1>Welcome, <%= @user.name %></h1>
+      <p class="subtitle">
+        Member since <%= @user.created_at.strftime("%Y") %>
+      </p>
+    </header>
+
+    <% if @user.admin? %>
+      <section class="admin-panel">
+        <h2>Admin Tools</h2>
+        <ul>
+          <% @tools.each do |tool| %>
+            <li><%= tool.title %></li>
+          <% end %>
+        </ul>
+      </section>
+    <% else %>
+      <p>You do not have admin privileges.</p>
+    <% end %>
+
+    <section class="posts">
+      <% @posts.each do |post| %>
+        <article class="post">
+          <h3><%= post.title %></h3>
+          <p><%= truncate(post.body, length: 140) %></p>
+
+          <%# Conditional rendering %>
+          <% if post.published? %>
+            <span class="status published">Published</span>
+          <% else %>
+            <span class="status draft">Draft</span>
+          <% end %>
+        </article>
+      <% end %>
+    </section>
+
+    <footer>
+      <p>&copy; <%= Time.now.year %> Example Corp</p>
+      <%= link_to "Privacy Policy", "/privacy" %>
+    </footer>
+
+    <script>
+      // JavaScript inside ERB
+      const userName = "<%= j @user.name %>";
+      console.log(`Loaded profile for ${userName}`);
+    </script>
+  </body>
+</html>

samples/fish.fish

@@ -0,0 +1,92 @@
+#!/usr/bin/env fish
+# Fish highlighting torture test 🐟
+
+# === Variables ===
+set normal_var "hello"
+set -l local_var 123
+set -gx GLOBAL_VAR "world"
+set PATH $PATH /usr/local/bin
+set --erase OLD_VAR
+
+# Builtin variables
+echo $HOME $PWD $USER $FISH_VERSION
+
+# === Strings ===
+set single 'single quoted string'
+set double "double quoted $normal_var"
+set escaped "newline\n tab\t dollar\$"
+
+# === Conditionals ===
+if test $normal_var = "hello"
+    echo "equal"
+else if test $normal_var != "world"
+    echo "not equal"
+end
+
+# === Logical operators ===
+true and echo "yes"
+false or echo "fallback"
+not false
+
+# === Arithmetic ===
+set x 10
+set y 20
+math "$x + $y"
+if test (math "$x * 2") -gt 15
+    echo "math works"
+end
+
+# === Loops ===
+for i in 1 2 3
+    echo "loop $i"
+end
+
+while test $x -gt 0
+    set x (math "$x - 1")
+end
+
+# === Functions ===
+function greet --argument name
+    echo "Hello $name"
+end
+
+greet "world"
+
+# === Command substitution ===
+set files (ls | grep ".fish")
+
+# === Redirections ===
+echo "output" > /tmp/fish_test.txt
+cat < /tmp/fish_test.txt >> /tmp/fish_log.txt
+
+# === Process substitution ===
+diff (ls /bin) (ls /usr/bin)
+
+# === Case statement ===
+switch $argv[1]
+    case start
+        echo "Starting"
+    case stop
+        echo "Stopping"
+    case '*'
+        echo "Unknown"
+end
+
+# === Subshell ===
+begin
+    echo "inside begin/end"
+end
+
+# === Comments & operators ===
+# && || | & ! should all highlight
+true && echo "ok" || echo "fail"
+
+# === Regex ===
+string match -r '^[a-z]+$' "hello"
+
+# === Test builtin ===
+test -f /etc/passwd
+test ! -d /does/not/exist
+
+# === Exit ===
+exit 0

samples/gdscript.gd

@@ -0,0 +1,81 @@
+# Sample GDScript for syntax highlighting
+
+extends Node2D
+
+# ============================================================
+# Constants
+# ============================================================
+const MAX_HEALTH = 100
+const PLAYER_SPEED = 200
+const PI_APPROX = 3.14159
+
+# ============================================================
+# Exported variables
+# ============================================================
+@export var player_name: String = "Hero"
+@export var is_alive: bool = true
+
+# ============================================================
+# Signals
+# ============================================================
+signal health_changed(new_health)
+
+# ============================================================
+# Member variables
+# ============================================================
+var health: int = MAX_HEALTH
+var velocity: Vector2 = Vector2.ZERO
+var inventory: Array = []
+
+# ============================================================
+# Functions
+# ============================================================
+func _ready() -> void:
+    print("Player ready:", player_name)
+    _initialize_inventory()
+    set_process(true)
+
+func _process(delta: float) -> void:
+    if is_alive:
+        _handle_input(delta)
+        _check_health()
+
+# Private functions
+func _initialize_inventory() -> void:
+    inventory.append("Sword")
+    inventory.append("Shield")
+
+func _handle_input(delta: float) -> void:
+    var direction: Vector2 = Vector2.ZERO
+    if Input.is_action_pressed("ui_right"):
+        direction.x += 1
+    if Input.is_action_pressed("ui_left"):
+        direction.x -= 1
+    if Input.is_action_pressed("ui_down"):
+        direction.y += 1
+    if Input.is_action_pressed("ui_up"):
+        direction.y -= 1
+
+    velocity = direction.normalized() * PLAYER_SPEED
+    position += velocity * delta
+
+func _check_health() -> void:
+    if health <= 0:
+        is_alive = false
+        print("Player is dead!")
+    else:
+        emit_signal("health_changed", health)
+
+# ============================================================
+# Example of class definition inside another script
+# ============================================================
+class Weapon:
+    var name: String
+    var damage: int
+
+    func _init(name: String, damage: int):
+        self.name = name
+        self.damage = damage
+
+    func attack():
+        print(name, "attacks for", damage, "damage")

samples/go.go

@@ -0,0 +1,95 @@
+// file: go.go
+package example_test
+
+import (
+	"context"
+	"fmt"
+	"math"
+	"sync"
+	"testing"
+	"time"
+)
+
+// Simple interface
+type Adder interface {
+	Add(a, b int) int
+}
+
+// Concrete implementation
+type Calculator struct{}
+
+func (Calculator) Add(a, b int) int {
+	return a + b
+}
+
+// Generic helper
+func Max[T ~int | ~float64](a, b T) T {
+	if a > b {
+		return a
+	}
+	return b
+}
+
+// Table-driven test
+func TestAdd(t *testing.T) {
+	tests := []struct {
+		name     string
+		a, b     int
+		expected int
+	}{
+		{"positive", 2, 3, 5},
+		{"negative", -2, -3, -5},
+		{"mixed", -2, 5, 3},
+	}
+
+	var calc Adder = Calculator{}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := calc.Add(tt.a, tt.b); got != tt.expected {
+				t.Fatalf("Add(%d, %d) = %d; want %d",
+					tt.a, tt.b, got, tt.expected)
+			}
+		})
+	}
+}
+
+// Concurrency + context test
+func TestWorker(t *testing.T) {
+	ctx, cancel := context.WithTimeout(context.Background(), 100*time.Millisecond)
+	defer cancel()
+
+	ch := make(chan int)
+	var wg sync.WaitGroup
+
+	wg.Add(1)
+	go func() {
+		defer wg.Done()
+		select {
+		case ch <- 42:
+		case <-ctx.Done():
+		}
+	}()
+
+	select {
+	case v := <-ch:
+		if v != 42 {
+			t.Errorf("unexpected value: %d", v)
+		}
+	case <-ctx.Done():
+		t.Fatal("timed out")
+	}
+
+	wg.Wait()
+}
+
+// Raw string + math edge case
+func TestRawString(t *testing.T) {
+	raw := `line 1
+line 2
+\t not escaped
+`
+	if len(raw) == 0 || math.IsNaN(float64(len(raw))) {
+		t.Fatal("impossible condition reached")
+	}
+}

samples/go.mod

@@ -0,0 +1,32 @@
+module github.com/example/project
+
+go 1.21
+
+// ==============================
+// Direct dependencies
+// ==============================
+require (
+    github.com/sirupsen/logrus v1.10.0
+    golang.org/x/net v0.10.0
+    github.com/pkg/errors v0.9.2 // indirect
+)
+
+// ==============================
+// Replace dependencies
+// ==============================
+replace (
+    github.com/old/dependency v1.2.3 => github.com/new/dependency v1.2.4
+    golang.org/x/oldnet => golang.org/x/net v0.11.0
+)
+
+// ==============================
+// Exclude dependencies
+// ==============================
+exclude github.com/bad/dependency v1.0.0
+
+// ==============================
+// Indirect dependencies
+// ==============================
+require (
+    github.com/another/pkg v1.3.0 // indirect
+)

samples/haskell.hs

@@ -0,0 +1,59 @@
+-- File: haskell.hs
+{-# LANGUAGE GADTs, TypeFamilies #-}
+
+module SyntaxTest where
+
+import Data.List (sort)
+import qualified Data.Map as Map
+
+-- Simple data type
+data Person = Person
+  { name :: String
+  , age  :: Int
+  } deriving (Show, Eq)
+
+-- GADT
+data Expr a where
+  I :: Int -> Expr Int
+  B :: Bool -> Expr Bool
+  Add :: Expr Int -> Expr Int -> Expr Int
+  Eq :: Expr Int -> Expr Int -> Expr Bool
+
+-- Type class
+class Describable a where
+  describe :: a -> String
+
+instance Describable Person where
+  describe (Person n a) = n ++ " is " ++ show a ++ " years old."
+
+-- Function with pattern matching
+sumList :: [Int] -> Int
+sumList []     = 0
+sumList (x:xs) = x + sumList xs
+
+-- Lambda and higher-order functions
+applyTwice :: (a -> a) -> a -> a
+applyTwice f x = f (f x)
+
+-- Infix operator
+infixl 6 +++
+(+++) :: Int -> Int -> Int
+a +++ b = a + b
+
+-- IO function
+main :: IO ()
+main = do
+  let people = [Person "Alice" 30, Person "Bob" 25]
+  mapM_ (putStrLn . describe) people
+  print $ sumList [1..10]
+  print $ applyTwice (+1) 5
+  print $ 3 +++ 4
+  print $ Eq (I 2) (Add (I 1) (I 1))
+
+-- Quasi-quote example
+someExpr :: Expr Int
+someExpr = [| Add (I 5) (I 7) |]
+
+-- Comments and Haddocks
+-- | This is a Haddock comment
+-- explaining the module and functions

samples/html.html

@@ -0,0 +1,87 @@
+<!doctype html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <title>Injection Test</title>
+
+    <!-- Comment -->
+    <!-- Another comment with spellcheck -->
+
+    <!-- CSS block -->
+    <style>
+      body {
+        background-color: #f0f0f0;
+      }
+      h1 {
+        color: blue;
+      }
+      .highlight {
+        font-weight: bold;
+      }
+    </style>
+
+    <!-- CSS block with type attribute -->
+    <style type="text/css">
+      p {
+        color: green;
+      }
+    </style>
+
+    <!-- Script block -->
+    <script>
+      console.log("Hello, world!");
+      function greet(name) {
+        alert(`Hello, ${name}`);
+      }
+    </script>
+
+    <!-- Script with type="module" -->
+    <script type="module">
+      import { something } from "./module.js";
+      something();
+    </script>
+
+    <!-- Script with type="importmap" -->
+    <script type="importmap">
+      {
+        "imports": {
+          "lodash": "/node_modules/lodash-es/lodash\n.js",
+          "key": 2
+        }
+      }
+    </script>
+
+    <!-- Script with type attribute custom -->
+    <script type="text/javascript">
+      console.log("Custom type");
+    </script>
+  </head>
+  <body>
+    <h1>Main Heading</h1>
+    <h2>Subheading H2</h2>
+
+    <p style="color: red; font-weight: bold">
+      This paragraph has an inline style
+    </p>
+
+    <p>
+      This is <strong>strong text</strong>, <b>bold also</b>,
+      <em>italic text</em>, <i>emphasized</i>, <u>underlined</u>,
+      <s>strikethrough</s>, <del>deleted text</del>, <code>inline code</code>,
+      <kbd>keyboard input</kbd>.
+    </p>
+    
+    <a href="https://hello.world"></a>
+
+    <!-- Lit-html / template interpolation -->
+    <button @click="${e => console.log(e)}">Click me</button>
+    <button @click="${e => console.log(e)}">Click me too</button>
+
+    <!-- Input pattern (regex) -->
+    <input type="text" pattern="[0-9]{3}" placeholder="Enter 3 digits" />
+
+    <!-- Event handlers -->
+    <button onclick="alert('Clicked!')">Event Handler</button>
+    <input onchange="console.log(this.value)" />
+  </body>
+</html>

samples/ini.ini

@@ -0,0 +1,41 @@
+; =====================================================
+; Sample INI Configuration
+; =====================================================
+
+[general]
+app_name = MyApp
+version = 1.2.3
+debug = true
+max_users = 100
+
+[database]
+host = localhost
+port = 5432
+user = admin
+password = secret
+timeout = 30
+
+[paths]
+log_dir = /var/log/myapp
+data_dir = ./data
+cache_dir = ./cache
+
+[features]
+enable_feature_x = true
+enable_feature_y = false
+feature_list = item1, item2, item3
+
+[servers]
+server1 = 192.168.1.10
+server2 = 192.168.1.11
+server3 = 192.168.1.12
+
+; Nested sections (some parsers support this)
+[servers.backup]
+server1 = 192.168.2.10
+server2 = 192.168.2.11
+
+; Comments and special characters
+; This is a comment line
+; Values can also contain special characters like !@#$%^&*()
+special_value = !@#$%^&*()_+|~=

samples/javascript.js

@@ -0,0 +1,147 @@
+/* ===============================
+ * JavaScript Syntax Torture Test
+ * =============================== */
+
+'use strict';
+
+// === Imports ===
+import fs, { readFileSync as rfs } from "fs";
+import * as path from "path";
+import defaultExport, { named as alias } from "./module.js";
+
+// === Constants ===
+const PI = 3.141592653589793;
+const HEX = 0xff;
+const BIN = 0b101010;
+const OCT = 0o755;
+const BIG = 123_456_789n;
+
+// === Variables ===
+let x = null;
+var y = undefined;
+let z = NaN;
+
+// === Strings ===
+const s1 = "double quotes";
+const s2 = 'single quotes';
+const s3 = `template literal ${1 + 2}`;
+const s4 = `multi
+line
+template`;
+const s5 = String.raw`raw \n string`;
+
+// === Escapes ===
+const esc = "\n\t\r\b\f\\\"\'\u00A9\x41";
+
+// === Arrays & Objects ===
+const arr = [1, , 3, ...[4, 5], { a: 1, b: { c: 2 } }];
+const obj = {
+  key: "value",
+  "weird-key": 123,
+  ['dyn' + 'amic']: true,
+  method() {},
+  async asyncMethod() {},
+  *generator() { yield 1; },
+};
+
+// === Destructuring ===
+const { a, b: renamed, ...rest } = obj;
+const [x1, , x3 = 42] = arr;
+
+// === Functions ===
+function normal(a, b = 1, ...rest) {
+  return a + b + rest.length;
+}
+
+const arrow = (x = 0) => x * x;
+const asyncArrow = async () => await Promise.resolve(42);
+
+// === Classes ===
+class Example extends Array {
+  static staticField = 123;
+  #privateField = "secret";
+
+  constructor(...args) {
+    super(...args);
+  }
+
+  get value() {
+    return this.#privateField;
+  }
+
+  set value(v) {
+    this.#privateField = v;
+  }
+}
+
+// === Control Flow ===
+if (true && !false || null ?? true) {
+  console.log("truthy");
+} else if (false) {
+  console.warn("nope");
+} else {
+  console.error("never");
+}
+
+for (let i = 0; i < 3; i++) {
+  continue;
+}
+
+for (const k in obj) {}
+for (const v of arr) {}
+
+while (false) {}
+do {} while (false);
+
+switch (Math.random()) {
+  case 0:
+    break;
+  default:
+    break;
+}
+
+// === Try / Catch ===
+try {
+  throw new Error("boom");
+} catch (e) {
+  console.error(e?.message ?? "unknown");
+} finally {
+  // cleanup
+}
+
+// === Regex ===
+const regex1 = /foo|bar/i;
+const regex2 = /^<script\b(?![^>]*\btype\s*=\s*"(?!module|text\/javascript)[^"]*")[^>]*>$/;
+
+// === Tagged template ===
+function tag(strings, ...values) {
+  return strings.raw.join("|") + values.join(",");
+}
+tag`hello ${42} world`;
+
+// === Optional chaining / nullish ===
+const deep = obj?.a?.b ?? "fallback";
+
+// === Bitwise ===
+const mask = (1 << 4) | (1 << 8);
+
+// === JSON ===
+const json = JSON.stringify({ a: 1, b: [true, false] }, null, 2);
+
+// === Top-level await (if supported) ===
+await Promise.resolve("done");
+
+// === JSX-like (should still highlight interestingly) ===
+const jsx = (
+  <Component prop="value">
+    <Child />
+  </Component>
+);
+
+// === End ===
+export default {
+  PI,
+  arr,
+  obj,
+  Example,
+};

samples/json.jsonc

@@ -0,0 +1,51 @@
+/* Example configuration file (JSONC)
+// Used to test syntax highlighting and comment support
+mutiline comment
+*/
+
+{
+  // Application metadata
+  "name": "example-app",
+  "version": "1.2.3",
+  "debug": true,
+
+  // Paths and environment
+  "paths": {
+    "root": "/usr/local/example",
+    "cache": "/tmp/example-cache
+  asa
+  multiline string",
+    "logs": null, // optional
+  },
+
+  // Feature flags
+  "features": {
+    "experimental": false,
+    "hotReload": true,
+    "themes": [
+      "dark",
+      "light",
+      // "solarized"  // not ready yet
+    ],
+  },
+
+  // Network configuration
+  "server": {
+    "host": "127.0.0.1",
+    "port": 8080,
+    "ssl": {
+      "enabled": false,
+      "cert": "",
+      "key": "",
+    }
+  },
+
+  // Mixed value types
+  "timeouts": [100, 250, 500, null],
+  "retryCount": 3,
+
+  // Escapes and strings
+  "banner": "Welcome!\nThis supports \"escaped quotes\" and unicode → ✓",
+
+  // Trailing comma allowed in JSONC
+}

samples/lua.lua

@@ -0,0 +1,119 @@
+#!/usr/bin/env lua
+-- Lua syntax highlighting test file
+
+-- Constants
+PI = 3.14159
+MAX_COUNT = 100
+
+-- Variables
+local counter = 0
+local name = "Lua"
+
+-- Built-in variable
+print(self)
+
+-- Functions
+function greet(user)
+	print("Hello, " .. user)
+end
+
+local function add(a, b)
+	return a + b
+end
+
+-- Method definitions
+local obj = {}
+function obj:sayHi()
+	print("Hi from method!")
+end
+
+obj.sayHello = function()
+	print("Hello from field function!")
+end
+
+-- Arrow-style anonymous function (LuaJIT/CFFI style)
+local arrow = function(x)
+	return x * 2
+end
+
+-- Table constructors
+local t = {
+	foo = 123,
+	bar = function()
+		return "bar"
+	end,
+	nested = {
+		a = 1,
+		b = 2,
+	},
+}
+
+-- Loops
+for i = 1, MAX_COUNT do
+	counter = counter + i
+end
+
+while counter > 0 do
+	counter = counter - 1
+end
+
+repeat
+	counter = counter + 1
+until counter == 10
+
+-- Conditionals
+if counter > 5 then
+	print("Big number")
+elseif counter == 5 then
+	print("Exactly five")
+else
+	print("Small number")
+end
+
+-- Operators
+local x, y = 10, 20
+local z = x + y * 2 - (x / y) ^ 2
+local ok = x == y or x ~= y and not false
+
+-- Function calls
+greet("World")
+obj:sayHi()
+obj.sayHello()
+add(5, 10)
+
+-- Built-in function calls
+assert(x > 0)
+pcall(function()
+	print("safe")
+end)
+tonumber("123")
+
+-- CFFI injection example
+local ffi = require("ffi")
+ffi.cdef([[
+    int printf(const char *fmt, ...);
+    typedef struct { int x; int y; } point;
+]])
+
+-- Boolean and nil
+local flag = true
+local nothing = nil
+
+-- Comments
+-- Single line
+--[[
+    Multi-line
+    comment
+]]
+
+-- Strings
+local s1 = "Hello\nWorld"
+local s2 = [[Long
+multi-line
+string]]
+
+-- Template strings (LuaJIT-style)
+local tpl = `Value: ${counter}`
+
+-- Regex-like string (for testing injection highlighting)
+local re = "/^%a+$/"

samples/markdown.md

@@ -0,0 +1,41 @@
+# Heading 1
+
+## Heading 2
+
+### Heading 3
+
+This is a paragraph with **bold text**, *italic text*, ~~strikethrough~~, and `inline code`.
+
+> This is a blockquote.
+>
+> - Nested list item 1
+> - Nested list item 2
+>   - Sub-item
+
+- Task list:
+  - [ ] Unchecked task
+  - [x] Checked task
+
+1. Numbered list item
+2. Another item
+
+---
+
+| Name       | Age | City          |
+|------------|-----|---------------|
+| Alice      | 25  | London        |
+| Bob        | 30  | New York      |
+| Charlie    | 22  | San Francisco |
+
+[Link to OpenAI](https://openai.com)
+
+`Inline code` example and a fenced code block:
+
+```python
+def hello_world():
+    print("Hello, world!")
+```
+
+![Image](https://example.com/image.jpg)
+
+> "This is a quote with a link to [OpenAI](https://openai.com)."

samples/nginx.conf

@@ -0,0 +1,87 @@
+# ============================================================
+# Global Settings
+# ============================================================
+
+user  www-data;
+worker_processes  auto;
+pid /run/nginx.pid;
+include /etc/nginx/modules-enabled/*.conf;
+
+# ============================================================
+# Events Block
+# ============================================================
+events {
+    worker_connections  1024;
+    use epoll;
+    multi_accept on;
+}
+
+# ============================================================
+# HTTP Block
+# ============================================================
+http {
+    sendfile on;
+    tcp_nopush on;
+    tcp_nodelay on;
+    keepalive_timeout  65;
+    types_hash_max_size 2048;
+
+    include       /etc/nginx/mime.types;
+    default_type  application/octet-stream;
+
+    # Logging
+    access_log  /var/log/nginx/access.log;
+    error_log   /var/log/nginx/error.log warn;
+
+    # Gzip Settings
+    gzip on;
+    gzip_disable "msie6";
+
+    # ========================================================
+    # Upstream Backend Servers
+    # ========================================================
+    upstream backend {
+        server 127.0.0.1:8080 weight=5;
+        server 127.0.0.1:8081;
+        keepalive 32;
+    }
+
+    # ========================================================
+    # Server Block
+    # ========================================================
+    server {
+        listen 80 default_server;
+        listen [::]:80 default_server;
+
+        server_name example.com www.example.com;
+
+        root /var/www/html;
+        index index.html index.htm;
+
+        # ====================================================
+        # Location Blocks
+        # ====================================================
+        location / {
+            try_files $uri $uri/ =404;
+        }
+
+        location /api/ {
+            proxy_pass http://backend;
+            proxy_set_header Host $host;
+            proxy_set_header X-Real-IP $remote_addr;
+            proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        }
+
+        location ~* \.(gif|jpg|jpeg|png|css|js|ico|svg)$ {
+            expires 30d;
+            add_header Cache-Control "public, no-transform";
+        }
+
+        error_page 404 /404.html;
+        error_page 500 502 503 504 /50x.html;
+
+        location = /50x.html {
+            root /var/www/html;
+        }
+    }
+}

samples/php.php

@@ -0,0 +1,130 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+  <meta charset="UTF-8">
+  <title>PHP Syntax Stress Test</title>
+
+  <style>
+    /* CSS section */
+    body {
+      font-family: Arial, sans-serif;
+      background: #1e1e1e;
+      color: #e0e0e0;
+    }
+
+    .box {
+      border: 1px solid #444;
+      padding: 10px;
+      margin: 10px;
+    }
+  </style>
+
+  <script>
+    // JS section
+    function greet(name) {
+      console.log("Hello " + name);
+    }
+
+    document.addEventListener("DOMContentLoaded", () => {
+      greet("World");
+    });
+  </script>
+</head>
+
+<body>
+
+<?php
+// Basic variables
+$number = 42;
+$text = "Hello PHP";
+$truth = true;
+$nothing = null;
+
+// Constants
+define("APP_NAME", "SyntaxTester");
+
+// Arrays
+$list = [1, 2, 3];
+$assoc = [
+  "one" => 1,
+  "two" => 2
+];
+
+// Function
+function add(int $a, int $b): int {
+  return $a + $b;
+}
+
+// Class + methods
+class User {
+  private string $name;
+  public static int $count = 0;
+
+  public function __construct(string $name) {
+    $this->name = $name;
+    self::$count++;
+  }
+
+  public function greet(): string {
+    return "Hello {$this->name}";
+  }
+}
+
+// Object usage
+$user = new User("Alice");
+echo $user->greet();
+
+// Control flow
+if ($number > 10) {
+  echo "Big number";
+} elseif ($number === 10) {
+  echo "Exactly ten";
+} else {
+  echo "Small number";
+}
+
+// Loop
+foreach ($list as $item) {
+  echo $item;
+}
+
+// Match expression
+$result = match ($number) {
+  1 => "one",
+  2 => "two",
+  default => "many"
+};
+
+// Try / catch
+try {
+  throw new Exception("Test exception");
+} catch (Exception $e) {
+  echo $e->getMessage();
+}
+
+// Anonymous function
+$double = fn($x) => $x * 2;
+
+// Nullsafe operator
+$len = $user?->name ? strlen($user->name) : 0;
+
+// Ternary
+$status = $truth ? "yes" : "no";
+
+// Include / require
+require_once "config.php";
+
+// Output
+echo "<div class='box'>";
+echo htmlspecialchars($text);
+echo "</div>";
+?>
+
+<script>
+  // JS interacting with PHP output
+  const phpValue = <?= json_encode($number) ?>;
+  console.log("Value from PHP:", phpValue);
+</script>
+
+</body>
+</html>

samples/python.py

@@ -0,0 +1,163 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""Test file for Python Tree-sitter highlighting."""
+
+# ==============================
+# Constants / Builtins
+# ==============================
+PI = 3.14159
+MAX_SIZE = 100
+NotImplemented
+Ellipsis
+__name__  # builtin constant
+
+# ==============================
+# Imports
+# ==============================
+import os
+import sys as system
+from re import compile as re_compile
+from __future__ import annotations
+from math import *
+
+# ==============================
+# Functions
+# ==============================
+def add(a: int, b: int = 5) -> int:
+    """Simple add function"""
+    return a + b
+
+def variadic(*args, **kwargs):
+    print(args, kwargs)
+
+lambda_func = lambda x, y=2: x * y
+
+def type_var_example(T: type):
+    pass
+
+# ==============================
+# Classes
+# ==============================
+class Base:
+    class_var = 10
+
+    def __init__(self, name: str):
+        self.name = name
+        self._private = 42
+
+    @classmethod
+    def cls_method(cls):
+        return cls.class_var
+
+    @staticmethod
+    def static_method():
+        return "static"
+
+    @property
+    def prop(self):
+        return self.name
+
+class Derived(Base):
+    def __init__(self, name, extra):
+        super().__init__(name)
+        self.extra = extra
+
+# ==============================
+# Variables
+# ==============================
+normal_var = 1
+_local_var = 2
+GLOBAL_VAR = 3
+
+# Builtin variable references
+self = "something"
+cls = "class"
+
+# ==============================
+# Control flow
+# ==============================
+if True:
+    x = 10
+elif False:
+    x = 20
+else:
+    x = 0
+
+for i in range(3):
+    print(i)
+while x > 0:
+    x -= 1
+    if x == 1:
+        break
+    else:
+        continue
+
+try:
+    1 / 0
+except ZeroDivisionError as e:
+    raise
+finally:
+    pass
+
+# ==============================
+# Operators
+# ==============================
+a, b = 5, 10
+c = a + b * 2 // 3 % 4 ** 2
+d = (a << 2) & b | c ^ ~a
+e = not a or b and c
+
+# ==============================
+# f-strings / interpolation
+# ==============================
+name = "Alice"
+greeting = f"Hello {name.upper()}!"
+formatted = f"{a + b} is sum"
+
+# ==============================
+# Regex
+# ==============================
+pattern1 = re_compile(r"\d+")
+pattern2 = re_compile(r"\w{2,}")
+
+# ==============================
+# Decorators usage
+# ==============================
+@staticmethod
+def static_func():
+    return True
+
+@classmethod
+def cls_func(cls):
+    return cls
+
+@custom_decorator
+def decorated_func():
+    return None
+
+# ==============================
+# Misc / Type conversions / literals
+# ==============================
+flag: bool = True
+nothing: None = None
+num: float = float("3.14")
+text: str = str(123)
+lst = [1, 2, 3]
+tpl = (4, 5)
+dct = {"a": 1, "b": 2}
+
+# ==============================
+# Type hints / TypeVar / TypeAlias
+# ==============================
+from typing import TypeVar, NewType
+T = TypeVar("T")
+UserId = NewType("UserId", int)
+TypeAliasExample: type = int
+
+# ==============================
+# Function calls / constructors
+# ==============================
+result = add(1, 2)
+obj = Derived("Alice", "extra")
+variadic(1, 2, 3, key="value")
+instance_check = isinstance(obj, Base)

samples/regex.regex

@@ -0,0 +1,18 @@
+# Match email addresses with optional names
+(?P<name>[a-zA-Z0-9._%+-]+)?\s*<(?P<email>[a-zA-Z0-9.-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,})>
+
+# Match dates in YYYY-MM-DD or DD/MM/YYYY
+(\d{4}-\d{2}-\d{2})|(\d{2}/\d{2}/\d{4})
+
+# Match hexadecimal colors
+# e.g., #FFF, #FFFFFF
+# Optional leading #
+# Case-insensitive
+# Flags inline
+(?i)#([0-9a-fA-F]{3}|[0-9a-fA-F]{6})
+
+# Match words starting with vowels
+\b[aeiouAEIOU]\w*\b
+
+# Match simple URL
+https?://(?:www\.)?\w+\.\w+(?:/\S*)?

samples/rust.rs

@@ -0,0 +1,406 @@
+#![allow(dead_code)]
+
+use std::collections::{BTreeMap, HashMap};
+use std::fmt;
+use std::time::Duration;
+
+//! Examples to exercise the Rust regex injection queries in the highlights.scm.
+//! These cover Regex::new, regex::Regex::new, regex::bytes::Regex::new,
+//! RegexSet::new, regex::RegexSet::new, RegexSetBuilder::new, and byte variants.
+//!
+//! Injection patterns in the query file trigger on:
+//! - call to (Regex|ByteRegexBuilder)::new with a raw string literal
+//! - call to (RegexSet|RegexSetBuilder)::new with an array of raw string literals
+
+use regex::{Regex, RegexSet, RegexSetBuilder};
+use regex::bytes::Regex as ByteRegex;
+use regex::bytes::RegexSet as ByteRegexSet;
+use regex::bytes::RegexSetBuilder as ByteRegexSetBuilder;
+
+fn main() {
+    // --- Should inject (Regex::new with raw string) ---
+    let _simple = Regex::new(r"^\d{4}-\d{2}-\d{2}$").unwrap();
+
+    // --- Should inject (fully qualified regex::Regex::new with raw string) ---
+    let _fq = regex::Regex::new(r"(?m)^\w+\s*=\s*.+$").unwrap();
+
+    // --- Should inject (bytes::Regex::new with raw string) ---
+    let _bytes = ByteRegex::new(r"(?-u)\xFF[\x00-\x7F]+").unwrap();
+
+    // --- Should inject (RegexSet::new with array of raw strings) ---
+    let _set = RegexSet::new([
+        r"^INFO:",
+        r"^WARN:",
+        r"^ERROR:",
+    ]).unwrap();
+
+    // --- Should inject (regex::RegexSet::new fully qualified) ---
+    let _set_fq = regex::RegexSet::new([
+        r"foo\d+",
+        r"bar\d+",
+    ]).unwrap();
+
+    // --- Should inject (RegexSetBuilder::new with array of raw strings) ---
+    let _set_builder = RegexSetBuilder::new([
+        r"\bcat\b",
+        r"\bdog\b",
+    ])
+    .case_insensitive(true)
+    .build()
+    .unwrap();
+
+    // --- Should inject (bytes set builder) ---
+    let _byte_set_builder = ByteRegexSetBuilder::new([
+        r"(?-u)\x01\x02",
+        r"(?-u)\xFF.+",
+    ])
+    .build()
+    .unwrap();
+
+    // --- Should inject (bytes set) ---
+    let _byte_set = ByteRegexSet::new([
+        r"(?-u)\x00+\xFF",
+        r"(?-u)[\x10-\x20]+",
+    ]).unwrap();
+
+    // --- NEGATIVE examples (should NOT inject) ---
+
+    // Not raw string literal (plain string): the query expects raw_string_literal.
+    let _no_inject_plain = Regex::new("plain-string-no-raw").unwrap();
+
+    // Function name is not `new`, so should not inject.
+    let _builder = Regex::new(r"\d+").map(|re| re.replace("123", "x"));
+
+    // Different type name, should not inject.
+    let _other = Some(r"not a regex call");
+
+    // Raw string but different function, should not inject.
+    let _format = format!(r"literal: {}", 42);
+}
+
+// Keep a simple test to ensure this compiles and runs.
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn smoke() {
+        let re = Regex::new(r"^\d+$").unwrap();
+        assert!(re.is_match("12345"));
+        let set = RegexSet::new([r"cat", r"dog"]).unwrap();
+        assert!(set.is_match("hotdog"));
+    }
+}
+
+/// A simple data type to exercise traits, pattern matching, and methods.
+#[derive(Debug, Clone, PartialEq)]
+pub struct Point {
+    pub x: i64,
+    pub y: i64,
+}
+
+impl Point {
+    pub fn manhattan(&self) -> i64 {
+        self.x.abs() + self.y.abs()
+    }
+
+    pub fn translate(&self, dx: i64, dy: i64) -> Self {
+        Self {
+            x: self.x + dx,
+            y: self.y + dy,
+        }
+    }
+}
+
+impl fmt::Display for Point {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "({}, {})", self.x, self.y)
+    }
+}
+
+#[derive(Debug, PartialEq, Eq)]
+pub enum ParseError {
+    Empty,
+    InvalidDigit,
+    TooLarge,
+}
+
+pub fn parse_u8(input: &str) -> Result<u8, ParseError> {
+    if input.trim().is_empty() {
+        return Err(ParseError::Empty);
+    }
+    let mut value: u16 = 0;
+    for ch in input.bytes() {
+        if !(b'0'..=b'9').contains(&ch) {
+            return Err(ParseError::InvalidDigit);
+        }
+        value = value * 10 + u16::from(ch - b'0');
+        if value > u8::MAX as u16 {
+            return Err(ParseError::TooLarge);
+        }
+    }
+    Ok(value as u8)
+}
+
+pub fn sum_iter<I: IntoIterator<Item = i64>>(iter: I) -> i64 {
+    iter.into_iter().fold(0, |acc, n| acc + n)
+}
+
+pub fn split_once<'a>(input: &'a str, needle: char) -> Option<(&'a str, &'a str)> {
+    let idx = input.find(needle)?;
+    Some((&input[..idx], &input[idx + needle.len_utf8()..]))
+}
+
+pub fn join_with<I: IntoIterator<Item = String>>(iter: I, sep: &str) -> String {
+    let mut it = iter.into_iter().peekable();
+    let mut out = String::new();
+    while let Some(item) = it.next() {
+        out.push_str(&item);
+        if it.peek().is_some() {
+            out.push_str(sep);
+        }
+    }
+    out
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use std::cell::RefCell;
+    use std::sync::mpsc;
+    use std::thread;
+
+    macro_rules! assert_contains {
+        ($haystack:expr, $needle:expr) => {
+            if !$haystack.contains($needle) {
+                panic!("expected {:?} to contain {:?}", $haystack, $needle);
+            }
+        };
+    }
+
+    #[test]
+    fn point_manhattan_and_display() {
+        let p = Point { x: -3, y: 4 };
+        assert_eq!(p.manhattan(), 7);
+        assert_eq!(p.to_string(), "(-3, 4)");
+    }
+
+    #[test]
+    fn point_translate_is_pure() {
+        let p = Point { x: 1, y: 2 };
+        let q = p.translate(3, -1);
+        assert_eq!(p, Point { x: 1, y: 2 });
+        assert_eq!(q, Point { x: 4, y: 1 });
+    }
+
+    #[test]
+    fn parse_u8_success_and_errors() {
+        assert_eq!(parse_u8("0"), Ok(0));
+        assert_eq!(parse_u8("255"), Ok(255));
+        assert_eq!(parse_u8(" 17 "), Ok(17)); // leading/trailing spaces are rejected as Empty? we trimmed only emptiness, digits still parsed.
+        assert_eq!(parse_u8(""), Err(ParseError::Empty));
+        assert_eq!(parse_u8("  "), Err(ParseError::Empty));
+        assert_eq!(parse_u8("12a"), Err(ParseError::InvalidDigit));
+        assert_eq!(parse_u8("256"), Err(ParseError::TooLarge));
+    }
+
+    #[test]
+    fn sum_iter_works_for_various_iterators() {
+        let v = vec![1, 2, 3, 4, -5];
+        assert_eq!(sum_iter(&v), 5);
+        let arr = [10i64; 4];
+        assert_eq!(sum_iter(arr), 40);
+        assert_eq!(sum_iter(0..5), 10);
+    }
+
+    #[test]
+    fn split_once_basic_and_unicode() {
+        assert_eq!(split_once("a,b,c", ','), Some(("a", "b,c")));
+        assert_eq!(split_once("no-sep", '/'), None);
+        // UTF-8 needle
+        let s = "fooλbar";
+        assert_eq!(split_once(s, 'λ'), Some(("foo", "bar")));
+    }
+
+    #[test]
+    fn join_with_various_lengths() {
+        let empty: Vec<String> = vec![];
+        assert_eq!(join_with(empty, ", "), "");
+        assert_eq!(join_with(vec!["a".into()], ", "), "a");
+        assert_eq!(
+            join_with(vec!["a".into(), "b".into(), "c".into()], "|"),
+            "a|b|c"
+        );
+    }
+
+    #[test]
+    fn hash_map_grouping_example() {
+        let words = ["ant", "bat", "apple", "boat"];
+        let mut by_initial: HashMap<char, Vec<&str>> = HashMap::new();
+        for w in &words {
+            let key = w.chars().next().unwrap();
+            by_initial.entry(key).or_default().push(*w);
+        }
+        assert_eq!(by_initial.get(&'a').unwrap(), &vec!["ant", "apple"]);
+        assert_eq!(by_initial.get(&'b').unwrap(), &vec!["bat", "boat"]);
+    }
+
+    #[test]
+    fn btree_map_sorted_iteration() {
+        let mut map = BTreeMap::new();
+        map.insert("c", 3);
+        map.insert("a", 1);
+        map.insert("b", 2);
+        let keys: Vec<_> = map.keys().copied().collect();
+        assert_eq!(keys, vec!["a", "b", "c"]);
+    }
+
+    #[test]
+    fn channels_and_threads() {
+        let (tx, rx) = mpsc::channel();
+        thread::spawn(move || {
+            for i in 0..5 {
+                tx.send(i * i).unwrap();
+            }
+        });
+        let received: Vec<_> = (0..5).map(|_| rx.recv().unwrap()).collect();
+        assert_eq!(received, vec![0, 1, 4, 9, 16]);
+    }
+
+    #[test]
+    fn interior_mutability_with_refcell() {
+        #[derive(Debug)]
+        struct Counter {
+            inner: RefCell<u32>,
+        }
+        impl Counter {
+            fn inc(&self) {
+                *self.inner.borrow_mut() += 1;
+            }
+            fn get(&self) -> u32 {
+                *self.inner.borrow()
+            }
+        }
+
+        let c = Counter {
+            inner: RefCell::new(0),
+        };
+        c.inc();
+        c.inc();
+        assert_eq!(c.get(), 2);
+    }
+
+    #[test]
+    fn should_panic_on_too_large_parse() {
+        #[should_panic(expected = "TooLarge")]
+        fn check() {
+            parse_u8("999").unwrap();
+        }
+        check();
+    }
+
+    #[test]
+    fn result_based_test() -> Result<(), String> {
+        let p = Point { x: 2, y: 3 };
+        if p.manhattan() == 5 {
+            Ok(())
+        } else {
+            Err("manhattan distance mismatch".into())
+        }
+    }
+
+    #[test]
+    fn iterator_combinators_cover_common_paths() {
+        let data = vec![Some(1), None, Some(3), Some(4)];
+        let sum: i32 = data.iter().flatten().sum();
+        assert_eq!(sum, 8);
+
+        let doubled: Vec<_> = (1..=5).map(|n| n * 2).filter(|n| n % 4 == 0).collect();
+        assert_eq!(doubled, vec![4, 8]);
+    }
+
+    #[test]
+    fn pattern_matching_with_guards() {
+        let numbers = [-2, -1, 0, 1, 2];
+        let labels: Vec<_> = numbers
+            .iter()
+            .map(|n| match n {
+                n if *n < 0 => "neg",
+                0 => "zero",
+                n if *n % 2 == 0 => "even-pos",
+                _ => "odd-pos",
+            })
+            .collect();
+        assert_eq!(labels, vec!["neg", "neg", "zero", "odd-pos", "even-pos"]);
+    }
+
+    #[test]
+    fn custom_macro_assert_contains() {
+        assert_contains!("hello world", "world");
+    }
+
+    #[test]
+    fn ownership_and_borrowing_examples() {
+        fn takes_and_gives_back(mut v: Vec<i32>) -> Vec<i32> {
+            v.push(42);
+            v
+        }
+        let v = vec![1, 2, 3];
+        let v = takes_and_gives_back(v);
+        assert_eq!(v, vec![1, 2, 3, 42]);
+
+        let s = String::from("hi");
+        let len = length_of_str(&s);
+        assert_eq!(len, 2);
+    }
+
+    fn length_of_str(s: &str) -> usize {
+        s.len()
+    }
+
+    #[test]
+    fn lifetimes_and_slices() {
+        fn first<'a>(xs: &'a [i32]) -> Option<&'a i32> {
+            xs.first()
+        }
+        let data = [10, 20, 30];
+        assert_eq!(first(&data), Some(&10));
+    }
+
+    #[test]
+    fn const_generics_array_sum() {
+        fn sum_array<const N: usize>(arr: [i32; N]) -> i32 {
+            arr.iter().sum()
+        }
+        assert_eq!(sum_array::<3>([1, 2, 3]), 6);
+        assert_eq!(sum_array([0; 5]), 0);
+    }
+
+    #[test]
+    fn duration_and_instant_arithmetic() {
+        use std::time::Instant;
+        let start = Instant::now();
+        std::thread::sleep(Duration::from_millis(5));
+        let elapsed = start.elapsed();
+        assert!(elapsed >= Duration::from_millis(5));
+    }
+
+    #[test]
+    fn string_builder_patterns() {
+        let parts = ["a", "b", "c"];
+        let mut s = String::with_capacity(3);
+        for p in parts {
+            s.push_str(p);
+        }
+        assert_eq!(s, "abc");
+        assert!(s.capacity() >= 3);
+    }
+
+    #[test]
+    fn equality_and_ordering_on_point() {
+        let p1 = Point { x: 1, y: 2 };
+        let p2 = Point { x: 1, y: 2 };
+        assert_eq!(p1, p2);
+        assert!(p1.manhattan() <= p2.manhattan());
+    }
+}

samples/sample.gitattributes

@@ -0,0 +1,5 @@
+# Sample .gitattributes file for syntax highlighting tests
+
+*.c syntax=c
+*.h syntax=c
+*.py syntax=python

samples/sample.gitignore

@@ -0,0 +1,12 @@
+# Test gitignore file (to check syntax)
+
+*
+./!
+
+*.log
+
+!.gitignore[2]
+
+**/*.lo[s]g
+
+### Comment

samples/sql.sql

@@ -0,0 +1,83 @@
+-- Sample SQL to exercise the highlight rules
+
+-- DDL
+CREATE TEMPORARY TABLE IF NOT EXISTS public.users (
+  user_id BIGSERIAL PRIMARY KEY,
+  email VARCHAR(255) UNIQUE NOT NULL,
+  profile JSONB,
+  balance DECIMAL(12,2) DEFAULT 0.00,
+  created_at TIMESTAMPTZ DEFAULT CURRENT_TIMESTAMP,
+  updated_at TIMESTAMPTZ,
+  CONSTRAINT email_chk CHECK (email LIKE '%@%')
+);
+
+-- Indexes
+CREATE INDEX CONCURRENTLY IF NOT EXISTS users_email_idx ON public.users USING btree (email);
+CREATE INDEX IF NOT EXISTS users_profile_gin_idx ON public.users USING gin (profile);
+
+-- Insert & returning
+INSERT INTO public.users (email, profile, balance)
+VALUES
+  ('alice@example.com', '{"plan":"pro","tags":["a","b"]}'::jsonb, 25.50),
+  ('bob@example.com',   '{"plan":"free","tags":["c"]}'::jsonb,   0.00)
+RETURNING user_id, email, profile;
+
+-- Update with CASE and CAST
+UPDATE public.users u
+SET balance = balance + CAST(5 AS DECIMAL),
+    updated_at = CURRENT_TIMESTAMP,
+    profile = jsonb_set(profile, '{last_seen}', to_jsonb(CURRENT_TIMESTAMP)),
+    email = CASE
+              WHEN email LIKE '%@example.com' THEN replace(email, '@example.com', '@example.org')
+              ELSE email
+            END
+WHERE u.balance >= 0
+RETURNING user_id, email, balance;
+
+-- Delete with USING
+DELETE FROM public.users AS u
+USING public.users AS t
+WHERE u.user_id = t.user_id
+  AND u.email LIKE 'bob@%';
+
+-- Window, CTE, aggregates
+WITH recent AS (
+  SELECT *
+  FROM public.users
+  WHERE created_at > NOW() - INTERVAL '30 days'
+)
+SELECT
+  user_id,
+  email,
+  balance,
+  SUM(balance) OVER (PARTITION BY 1 ORDER BY created_at ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) AS running_total,
+  ROW_NUMBER() OVER (ORDER BY created_at DESC) AS rn
+FROM recent
+ORDER BY created_at DESC
+LIMIT 50 OFFSET 0;
+
+-- Joins and JSON
+CREATE TEMP TABLE events (
+  event_id BIGSERIAL PRIMARY KEY,
+  user_id BIGINT REFERENCES public.users(user_id),
+  payload JSONB,
+  created_at TIMESTAMPTZ DEFAULT NOW()
+);
+
+INSERT INTO events (user_id, payload) VALUES
+  (1, '{"type":"login","ip":"127.0.0.1"}'),
+  (1, '{"type":"purchase","amount":9.99,"items":[{"sku":"A1","qty":1}]}' ),
+  (2, '{"type":"login","ip":"10.0.0.2"}');
+
+SELECT u.email, e.payload->>'type' AS event_type, e.payload
+FROM public.users u
+LEFT JOIN events e ON e.user_id = u.user_id
+WHERE e.payload ? 'type'
+ORDER BY e.created_at DESC;
+
+-- Transaction control
+BEGIN;
+  UPDATE public.users SET balance = balance - 5 WHERE email LIKE 'alice%';
+  INSERT INTO events (user_id, payload)
+  SELECT user_id, jsonb_build_object('type','adjust','delta',-5) FROM public.users WHERE email LIKE 'alice%';
+COMMIT;

samples/toml.toml

@@ -0,0 +1,53 @@
+# ============================================================
+# Basic types
+# ============================================================
+
+title = "Example TOML Configuration"
+enabled = true
+count = 42
+pi = 3.14159
+empty = ""
+
+# ============================================================
+# Arrays
+# ============================================================
+fruits = ["apple", "banana", "cherry"]
+numbers = [1, 2, 3, 4, 5]
+
+# Nested array
+matrix = [[1, 2], [3, 4]]
+
+# ============================================================
+# Tables
+# ============================================================
+[owner]
+name = "Alice"
+dob = 1979-05-27T07:32:00Z
+
+[database]
+server = "192.168.1.1"
+ports = [ 8001, 8001, 8002 ]
+connection_max = 5000
+enabled = true
+
+[servers.alpha]
+ip = "10.0.0.1"
+dc = "east"
+
+[servers.beta]
+ip = "10.0.0.2"
+dc = "west"
+
+# ============================================================
+# Inline tables
+# ============================================================
+clients = { name = "Bob", age = 30, active = true }
+
+# ============================================================
+# Multiline strings
+# ============================================================
+description = """
+This is a TOML file
+used for testing syntax highlighting.
+It supports multiple lines.
+"""