@mlightcad/mtext-parser 1.5.2 → 1.5.3

package/src/parser.ts

@@ -1,2129 +1,2167 @@
-/**
- * Token types used in MText parsing
- */
-export enum TokenType {
-  /** No token */
-  NONE = 0,
-  /** Word token with string data */
-  WORD = 1,
-  /** Stack token with [numerator, denominator, type] data */
-  STACK = 2,
-  /** Space token with no data */
-  SPACE = 3,
-  /** Non-breaking space token with no data */
-  NBSP = 4,
-  /** Tab token with no data */
-  TABULATOR = 5,
-  /** New paragraph token with no data */
-  NEW_PARAGRAPH = 6,
-  /** New column token with no data */
-  NEW_COLUMN = 7,
-  /** Wrap at dimension line token with no data */
-  WRAP_AT_DIMLINE = 8,
-  /** Properties changed token with string data (full command) */
-  PROPERTIES_CHANGED = 9,
-  /** AutoCAD percent-sign symbol code (`%%c`, `%%d`, `%%p`, `%%nnn`, `%%%`) */
-  PERCENT_SYMBOL = 10,
-}
-
-/**
- * AutoCAD percent-sign symbol emitted when {@link MTextParserOptions.yieldPercentSymbols}
- * is enabled.
- */
-export type PercentSymbolData =
-  | {
-      kind: 'named';
-      /** Percent code letter: `%%c`, `%%d`, or `%%p`. */
-      code: 'c' | 'd' | 'p';
-      /** Unicode expansion used for display (e.g. `%%c` → `Ø`). */
-      char: string;
-    }
-  | {
-      kind: 'numeric';
-      /**
-       * Decimal code from `%%nnn`, where `nnn` is 1–3 digits (Unicode/ASCII
-       * code point, e.g. `%%34` → `"`, `%%176` → `°`).
-       */
-      charCode: number;
-      /** `String.fromCharCode(charCode)`. */
-      char: string;
-    }
-  | {
-      kind: 'literal';
-      /** Literal percent from `%%%`. */
-      char: '%';
-    };
-
-/**
- * Represents a factor value that can be either absolute or relative.
- * Used for properties like height, width, and character tracking in MText formatting.
- */
-export interface FactorValue {
-  /** The numeric value of the factor */
-  value: number;
-  /** Whether the value is relative (true) or absolute (false) */
-  isRelative: boolean;
-}
-
-/**
- * Format properties of MText word tokens.
- * This interface defines all the formatting properties that can be applied to MText content,
- * including text styling, colors, alignment, font properties, and paragraph formatting.
- */
-export interface Properties {
-  /** Whether text is underlined */
-  underline?: boolean;
-  /** Whether text has an overline */
-  overline?: boolean;
-  /** Whether text has strike-through */
-  strikeThrough?: boolean;
-  /** AutoCAD Color Index (ACI) color value (0-256), or null if not set */
-  aci?: number | null;
-  /** RGB color tuple [r, g, b], or null if not set */
-  rgb?: RGB | null;
-  /** Line alignment for the text */
-  align?: MTextLineAlignment;
-  /** Font face properties including family, style, and weight */
-  fontFace?: FontFace;
-  /** Capital letter height factor (can be relative or absolute) */
-  capHeight?: FactorValue;
-  /** Character width factor (can be relative or absolute) */
-  widthFactor?: FactorValue;
-  /** Character tracking factor for spacing between characters (can be relative or absolute) */
-  charTrackingFactor?: FactorValue;
-  /** Oblique angle in degrees for text slant */
-  oblique?: number;
-  /** Paragraph formatting properties (partial to allow selective updates) */
-  paragraph?: Partial<ParagraphProperties>;
-}
-
-/**
- * Represents a change in MText properties, including the command, the changed properties, and the context depth.
- */
-export interface ChangedProperties {
-  /**
-   * The property command that triggered the change (e.g., 'L', 'C', 'f').
-   * The command will be undefined if it is to restore context.
-   */
-  command: string | undefined;
-  /**
-   * The set of properties that have changed as a result of the command.
-   */
-  changes: Properties;
-  /**
-   * The current context stack depth when the property change occurs.
-   * - 0: The change is global (applies outside of any `{}` block).
-   * - >0: The change is local (applies within one or more nested `{}` blocks).
-   */
-  depth: number; // 0 = global, >0 = local
-}
-
-/**
- * Type for token data based on token type
- */
-export type TokenData = {
-  [TokenType.NONE]: null;
-  [TokenType.WORD]: string;
-  [TokenType.STACK]: [string, string, string];
-  [TokenType.SPACE]: null;
-  [TokenType.NBSP]: null;
-  [TokenType.TABULATOR]: null;
-  [TokenType.NEW_PARAGRAPH]: null;
-  [TokenType.NEW_COLUMN]: null;
-  [TokenType.WRAP_AT_DIMLINE]: null;
-  [TokenType.PROPERTIES_CHANGED]: ChangedProperties;
-  [TokenType.PERCENT_SYMBOL]: PercentSymbolData;
-};
-
-/**
- * Line alignment options for MText
- */
-export enum MTextLineAlignment {
-  /** Align text to bottom */
-  BOTTOM = 0,
-  /** Align text to middle */
-  MIDDLE = 1,
-  /** Align text to top */
-  TOP = 2,
-}
-
-/**
- * Paragraph alignment options for MText
- */
-export enum MTextParagraphAlignment {
-  /** Default alignment */
-  DEFAULT = 0,
-  /** Left alignment */
-  LEFT = 1,
-  /** Right alignment */
-  RIGHT = 2,
-  /** Center alignment */
-  CENTER = 3,
-  /** Justified alignment */
-  JUSTIFIED = 4,
-  /** Distributed alignment */
-  DISTRIBUTED = 5,
-}
-
-/**
- * Text stroke options for MText
- */
-export enum MTextStroke {
-  /** No stroke */
-  NONE = 0,
-  /** Underline stroke */
-  UNDERLINE = 1,
-  /** Overline stroke */
-  OVERLINE = 2,
-  /** Strike-through stroke */
-  STRIKE_THROUGH = 4,
-}
-
-/**
- * RGB color tuple
- */
-export type RGB = [number, number, number];
-
-/**
- * Font style type
- */
-export type FontStyle = 'Regular' | 'Italic';
-
-/**
- * Font face properties
- */
-export interface FontFace {
-  /** Font family name */
-  family: string;
-  /** Font style (e.g., 'Regular', 'Italic') */
-  style: FontStyle;
-  /** Font weight (e.g., 400 for normal, 700 for bold) */
-  weight: number;
-}
-
-/**
- * Paragraph properties
- */
-export interface ParagraphProperties {
-  /** Indentation value */
-  indent: number;
-  /** Left margin value */
-  left: number;
-  /** Right margin value */
-  right: number;
-  /** Paragraph alignment */
-  align: MTextParagraphAlignment;
-  /** Tab stop positions and types */
-  tabs: (number | string)[];
-}
-
-/**
- * Special character encoding mapping
- */
-const SPECIAL_CHAR_ENCODING: Record<string, string> = {
-  c: 'Ø',
-  d: '°',
-  p: '±',
-  '%': '%',
-};
-
-/**
- * Character to paragraph alignment mapping
- */
-const CHAR_TO_ALIGN: Record<string, MTextParagraphAlignment> = {
-  l: MTextParagraphAlignment.LEFT,
-  r: MTextParagraphAlignment.RIGHT,
-  c: MTextParagraphAlignment.CENTER,
-  j: MTextParagraphAlignment.JUSTIFIED,
-  d: MTextParagraphAlignment.DISTRIBUTED,
-};
-
-/**
- * Convert RGB tuple to integer color value
- * @param rgb - RGB color tuple
- * @returns Integer color value
- */
-export function rgb2int(rgb: RGB): number {
-  const [r, g, b] = rgb;
-  return (r << 16) | (g << 8) | b;
-}
-
-/**
- * Convert integer color value to RGB tuple
- * @param value - Integer color value
- * @returns RGB color tuple
- */
-export function int2rgb(value: number): RGB {
-  const r = (value >> 16) & 0xff;
-  const g = (value >> 8) & 0xff;
-  const b = value & 0xff;
-  return [r, g, b];
-}
-
-function clampColorChannel(value: number): number {
-  return Math.max(0, Math.min(255, Math.round(value)));
-}
-
-function normalizeColorNumber(color: number): number {
-  return Math.max(0, Math.min(0xffffff, Math.round(color)));
-}
-
-function colorNumberToHex(color: number | null): string | null {
-  if (color === null) return null;
-  return `#${normalizeColorNumber(color).toString(16).padStart(6, '0')}`;
-}
-
-function normalizeHexColor(value: string | null | undefined): string | null {
-  if (!value) return null;
-  const normalized = value.trim().toLowerCase();
-  if (/^#[0-9a-f]{6}$/.test(normalized)) return normalized;
-  if (/^[0-9a-f]{6}$/.test(normalized)) return `#${normalized}`;
-  if (/^#[0-9a-f]{3}$/.test(normalized)) {
-    const r = normalized[1];
-    const g = normalized[2];
-    const b = normalized[3];
-    return `#${r}${r}${g}${g}${b}${b}`;
-  }
-  if (/^[0-9a-f]{3}$/.test(normalized)) {
-    const r = normalized[0];
-    const g = normalized[1];
-    const b = normalized[2];
-    return `#${r}${r}${g}${g}${b}${b}`;
-  }
-  return null;
-}
-
-function cssColorToRgbValue(value: string | null | undefined): number | null {
-  if (!value) return null;
-  const raw = value.trim().toLowerCase();
-  if (raw === 'transparent') return null;
-
-  const hex = normalizeHexColor(raw);
-  if (hex) {
-    return normalizeColorNumber(Number.parseInt(hex.slice(1), 16));
-  }
-
-  const fnMatch = raw.match(/^rgba?\((.*)\)$/);
-  if (!fnMatch) return null;
-
-  const parts = fnMatch[1]
-    .replace(/\s*\/\s*/g, ' ')
-    .split(/[,\s]+/)
-    .map(p => p.trim())
-    .filter(Boolean);
-
-  if (parts.length < 3) return null;
-
-  const toChannel = (token: string): number => {
-    if (token.endsWith('%')) {
-      const percent = Number.parseFloat(token.slice(0, -1));
-      return clampColorChannel((percent / 100) * 255);
-    }
-    const num = Number.parseFloat(token);
-    return clampColorChannel(num);
-  };
-
-  const r = toChannel(parts[0]);
-  const g = toChannel(parts[1]);
-  const b = toChannel(parts[2]);
-  return rgb2int([r, g, b]);
-}
-
-/**
- * Escape DXF line endings
- * @param text - Text to escape
- * @returns Escaped text
- */
-export function escapeDxfLineEndings(text: string): string {
-  return text.replace(/\r\n|\r|\n/g, '\\P');
-}
-
-/**
- * Check if text contains inline formatting codes
- * @param text - Text to check
- * @returns True if text contains formatting codes
- */
-export function hasInlineFormattingCodes(text: string): boolean {
-  return text.replace(/\\P/g, '').replace(/\\~/g, '').includes('\\');
-}
-
-/**
- * Extracts all unique font names used in an MText string.
- * This function searches for font commands in the format \f{fontname}| or \f{fontname}; and returns a set of unique font names.
- * Font names are converted to lowercase to ensure case-insensitive uniqueness.
- *
- * @param mtext - The MText string to analyze for font names
- * @param removeExtension - Whether to remove font file extensions (e.g., .ttf, .shx) from font names. Defaults to false.
- * @returns A Set containing all unique font names found in the MText string, converted to lowercase
- * @example
- * ```ts
- * const mtext = "\\fArial.ttf|Hello\\fTimes New Roman.otf|World";
- * const fonts = getFonts(mtext, true);
- * // Returns: Set(2) { "arial", "times new roman" }
- * ```
- */
-export function getFonts(mtext: string, removeExtension: boolean = false) {
-  const fonts: Set<string> = new Set();
-  const regex = /\\[fF](.*?)[;|]/g;
-
-  [...mtext.matchAll(regex)].forEach(match => {
-    let fontName = match[1].toLowerCase();
-    if (removeExtension) {
-      fontName = fontName.replace(/\.(ttf|otf|woff|shx)$/, '');
-    }
-    fonts.add(fontName);
-  });
-
-  return fonts;
-}
-
-/**
- * ContextStack manages a stack of MTextContext objects for character-level formatting.
- *
- * - Character-level formatting (underline, color, font, etc.) is scoped to `{}` blocks and managed by the stack.
- * - Paragraph-level formatting (\p) is not scoped, but when a block ends, any paragraph property changes are merged into the parent context.
- * - On pop, paragraph properties from the popped context are always merged into the new top context.
- */
-class ContextStack {
-  private stack: MTextContext[] = [];
-
-  /**
-   * Creates a new ContextStack with an initial context.
-   * @param initial The initial MTextContext to use as the base of the stack.
-   */
-  constructor(initial: MTextContext) {
-    this.stack.push(initial);
-  }
-
-  /**
-   * Pushes a copy of the given context onto the stack.
-   * @param ctx The MTextContext to push (copied).
-   */
-  push(ctx: MTextContext) {
-    this.stack.push(ctx);
-  }
-
-  /**
-   * Pops the top context from the stack and merges its paragraph properties into the new top context.
-   * If only one context remains, nothing is popped.
-   * @returns The popped MTextContext, or undefined if the stack has only one context.
-   */
-  pop(): MTextContext | undefined {
-    if (this.stack.length <= 1) return undefined;
-    const popped = this.stack.pop()!;
-    // Merge paragraph properties into the new top context
-    const top = this.stack[this.stack.length - 1];
-    if (JSON.stringify(top.paragraph) !== JSON.stringify(popped.paragraph)) {
-      top.paragraph = { ...popped.paragraph };
-    }
-    return popped;
-  }
-
-  /**
-   * Returns the current (top) context on the stack.
-   */
-  get current(): MTextContext {
-    return this.stack[this.stack.length - 1];
-  }
-
-  /**
-   * Returns the current stack depth (number of nested blocks), not counting the root context.
-   */
-  get depth(): number {
-    return this.stack.length - 1;
-  }
-
-  /**
-   * Returns the root (bottom) context, which represents the global formatting state.
-   * Used for paragraph property application.
-   */
-  get root(): MTextContext {
-    return this.stack[0];
-  }
-
-  /**
-   * Replaces the current (top) context with the given context.
-   * @param ctx The new context to set as the current context.
-   */
-  setCurrent(ctx: MTextContext) {
-    this.stack[this.stack.length - 1] = ctx;
-  }
-}
-
-/**
- * Configuration options for the MText parser.
- * These options control how the parser behaves during tokenization and property handling.
- */
-export interface MTextParserOptions {
-  /**
-   * Whether to yield PROPERTIES_CHANGED tokens when formatting properties change.
-   * When true, the parser will emit tokens whenever properties like color, font, or alignment change.
-   * When false, property changes are applied silently to the context without generating tokens.
-   * @default false
-   */
-  yieldPropertyCommands?: boolean;
-  /**
-   * Whether to reset paragraph parameters when encountering a new paragraph token.
-   * When true, paragraph properties (indent, margins, alignment, tab stops) are reset to defaults
-   * at the start of each new paragraph.
-   * @default false
-   */
-  resetParagraphParameters?: boolean;
-  /**
-   * Whether to emit {@link TokenType.PERCENT_SYMBOL} tokens for AutoCAD `%%` symbol
-   * codes instead of expanding them into {@link TokenType.WORD} characters.
-   * @default false
-   */
-  yieldPercentSymbols?: boolean;
-  /**
-   * Custom decoder function for MIF (Multibyte Interchange Format) codes.
-   * If provided, this function will be used instead of the default decodeMultiByteChar.
-   * The function receives the hex code string and should return the decoded character.
-   * @param hex - Hex code string (e.g., "C4E3" or "1A2B3")
-   * @returns Decoded character or empty square (▯) if invalid
-   * @default undefined (uses default decoder)
-   */
-  mifDecoder?: (hex: string) => string;
-  /**
-   * The length of MIF hex codes to parse. MIF codes in AutoCAD can vary in length
-   * depending on the specific SHX big font used (typically 4 or 5 digits).
-   * If not specified, the parser will try to auto-detect the length by attempting
-   * to match 4 digits first, then 5 digits if needed.
-   * @default undefined (auto-detect)
-   */
-  mifCodeLength?: 4 | 5 | 'auto';
-}
-
-/**
- * Main parser class for MText content
- */
-export class MTextParser {
-  private scanner: TextScanner;
-  private ctxStack: ContextStack;
-  private continueStroke: boolean = false;
-  private yieldPropertyCommands: boolean;
-  private resetParagraphParameters: boolean;
-  private yieldPercentSymbols: boolean;
-  private inStackContext: boolean = false;
-  private mifDecoder: (hex: string) => string;
-  private mifCodeLength: 4 | 5 | 'auto';
-
-  /**
-   * Creates a new MTextParser instance
-   * @param content - The MText content to parse
-   * @param ctx - Optional initial MText context
-   * @param options - Parser options
-   */
-  constructor(content: string, ctx?: MTextContext, options: MTextParserOptions = {}) {
-    this.scanner = new TextScanner(content);
-    const initialCtx = ctx ?? new MTextContext();
-    this.ctxStack = new ContextStack(initialCtx);
-    this.yieldPropertyCommands = options.yieldPropertyCommands ?? false;
-    this.resetParagraphParameters = options.resetParagraphParameters ?? false;
-    this.yieldPercentSymbols = options.yieldPercentSymbols ?? false;
-    this.mifDecoder = options.mifDecoder ?? this.decodeMultiByteChar.bind(this);
-    this.mifCodeLength = options.mifCodeLength ?? 'auto';
-  }
-
-  /**
-   * Decode multi-byte character from hex code
-   * @param hex - Hex code string (e.g. "C4E3" or "1A2B3")
-   * @returns Decoded character or empty square if invalid
-   */
-  private decodeMultiByteChar(hex: string): string {
-    try {
-      // For 5-digit codes, return placeholder directly
-      if (hex.length === 5) {
-        const prefix = hex[0];
-
-        // Notes:
-        // I know AutoCAD uses prefix 1 for Shift-JIS, 2 for big5, and 5 for gbk.
-        // But I don't know whether there are other prefixes and their meanings.
-        let encoding = 'gbk';
-        if (prefix === '1') {
-          encoding = 'shift-jis';
-        } else if (prefix === '2') {
-          encoding = 'big5';
-        }
-        const bytes = new Uint8Array([
-          parseInt(hex.substr(1, 2), 16),
-          parseInt(hex.substr(3, 2), 16),
-        ]);
-        const decoder = new TextDecoder(encoding);
-        const result = decoder.decode(bytes);
-        return result;
-      } else if (hex.length === 4) {
-        // For 4-digit hex codes, decode as 2-byte character
-        const bytes = new Uint8Array([
-          parseInt(hex.substr(0, 2), 16),
-          parseInt(hex.substr(2, 2), 16),
-        ]);
-
-        // Try GBK first
-        const gbkDecoder = new TextDecoder('gbk');
-        const gbkResult = gbkDecoder.decode(bytes);
-        if (gbkResult !== '▯') {
-          return gbkResult;
-        }
-
-        // Try BIG5 if GBK fails
-        const big5Decoder = new TextDecoder('big5');
-        const big5Result = big5Decoder.decode(bytes);
-        if (big5Result !== '▯') {
-          return big5Result;
-        }
-      }
-
-      return '▯';
-    } catch {
-      return '▯';
-    }
-  }
-
-  /**
-   * Extract MIF hex code from scanner
-   * @param length - The length of the hex code to extract (4 or 5), or 'auto' to detect
-   * @returns The extracted hex code, or null if not found
-   */
-  private extractMifCode(length: 4 | 5 | 'auto'): string | null {
-    if (length === 'auto') {
-      // Try 5 digits first if available, then fall back to 4 digits
-      const code5 = this.scanner.tail.match(/^[0-9A-Fa-f]{5}/)?.[0];
-      if (code5) {
-        return code5;
-      }
-      const code4 = this.scanner.tail.match(/^[0-9A-Fa-f]{4}/)?.[0];
-      if (code4) {
-        return code4;
-      }
-      return null;
-    } else {
-      const code = this.scanner.tail.match(new RegExp(`^[0-9A-Fa-f]{${length}}`))?.[0];
-      return code ?? null;
-    }
-  }
-
-  /**
-   * Parse an AutoCAD `\U+nnnn` escape after the leading `U` has been consumed.
-   *
-   * @remarks
-   * Exactly four hexadecimal digits are read. A high surrogate followed by
-   * another `\U+nnnn` low surrogate is combined into one supplementary-plane
-   * character (AutoCAD's encoding for code points above U+FFFF).
-   *
-   * @returns The decoded character, or `null` if the escape is incomplete
-   */
-  private parseUnicodeEscape(): string | null {
-    if (this.scanner.peek() !== '+') {
-      return null;
-    }
-
-    this.scanner.consume(1); // Consume the '+'
-    const hexMatch = this.scanner.tail.match(/^[0-9A-Fa-f]{4}/);
-    if (!hexMatch) {
-      this.scanner.consume(-1); // Rewind '+'
-      return null;
-    }
-
-    const codeUnit = parseInt(hexMatch[0], 16);
-    this.scanner.consume(hexMatch[0].length);
-
-    // High surrogate: combine with a following \U+ low-surrogate escape
-    if (codeUnit >= 0xd800 && codeUnit <= 0xdbff) {
-      const pairMatch = this.scanner.tail.match(/^\\U\+([0-9A-Fa-f]{4})/);
-      if (pairMatch) {
-        const lowUnit = parseInt(pairMatch[1], 16);
-        if (lowUnit >= 0xdc00 && lowUnit <= 0xdfff) {
-          this.scanner.consume(pairMatch[0].length);
-          const codePoint =
-            ((codeUnit - 0xd800) << 10) + (lowUnit - 0xdc00) + 0x10000;
-          try {
-            return String.fromCodePoint(codePoint);
-          } catch {
-            return '▯';
-          }
-        }
-      }
-    }
-
-    // BMP code unit, or an unpaired surrogate kept as a UTF-16 unit
-    return String.fromCharCode(codeUnit);
-  }
-
-  /**
-   * Push current context onto the stack
-   */
-  private pushCtx(): void {
-    this.ctxStack.push(this.ctxStack.current);
-  }
-
-  /**
-   * Pop context from the stack
-   */
-  private popCtx(): void {
-    this.ctxStack.pop();
-  }
-
-  /**
-   * Parse stacking expression (numerator/denominator)
-   * @returns Tuple of [TokenType.STACK, [numerator, denominator, type]]
-   */
-  private parseStacking(): [TokenType, [string, string, string]] {
-    const scanner = new TextScanner(this.extractExpression(true));
-    let numerator = '';
-    let denominator = '';
-    let stackingType = '';
-
-    const getNextChar = (): [string, boolean] => {
-      let c = scanner.peek();
-      let escape = false;
-      if (c.charCodeAt(0) < 32) {
-        c = ' ';
-      }
-      if (c === '\\') {
-        escape = true;
-        scanner.consume(1);
-        c = scanner.peek();
-      }
-      scanner.consume(1);
-      return [c, escape];
-    };
-
-    const parseNumerator = (): [string, string] => {
-      let word = '';
-      while (scanner.hasData) {
-        const [c, escape] = getNextChar();
-        // Check for stacking operators first
-        if (!escape && (c === '/' || c === '#' || c === '^')) {
-          return [word, c];
-        }
-        word += c;
-      }
-      return [word, ''];
-    };
-
-    const parseDenominator = (skipLeadingSpace: boolean): string => {
-      let word = '';
-      let skipping = skipLeadingSpace;
-      while (scanner.hasData) {
-        const [c, escape] = getNextChar();
-        if (skipping && c === ' ') {
-          continue;
-        }
-        skipping = false;
-        // Stop at terminator unless escaped
-        if (!escape && c === ';') {
-          break;
-        }
-        word += c;
-      }
-      return word;
-    };
-
-    [numerator, stackingType] = parseNumerator();
-    if (stackingType) {
-      // Only skip leading space for caret divider
-      denominator = parseDenominator(stackingType === '^');
-    }
-
-    // Special case for \S^!/^?;
-    if (numerator === '' && denominator.includes('I/')) {
-      return [TokenType.STACK, [' ', ' ', '/']];
-    }
-
-    // Handle caret as a stacking operator
-    if (stackingType === '^') {
-      return [TokenType.STACK, [numerator, denominator, '^']];
-    }
-
-    return [TokenType.STACK, [numerator, denominator, stackingType]];
-  }
-
-  /**
-   * Parse MText properties
-   * @param cmd - The property command to parse
-   * @returns Property changes if yieldPropertyCommands is true and changes occurred
-   */
-  private parseProperties(cmd: string): TokenData[TokenType.PROPERTIES_CHANGED] | void {
-    const prevCtx = this.ctxStack.current.copy();
-    const newCtx = this.ctxStack.current.copy();
-    switch (cmd) {
-      case 'L':
-        newCtx.underline = true;
-        this.continueStroke = true;
-        break;
-      case 'l':
-        newCtx.underline = false;
-        if (!newCtx.hasAnyStroke) {
-          this.continueStroke = false;
-        }
-        break;
-      case 'O':
-        newCtx.overline = true;
-        this.continueStroke = true;
-        break;
-      case 'o':
-        newCtx.overline = false;
-        if (!newCtx.hasAnyStroke) {
-          this.continueStroke = false;
-        }
-        break;
-      case 'K':
-        newCtx.strikeThrough = true;
-        this.continueStroke = true;
-        break;
-      case 'k':
-        newCtx.strikeThrough = false;
-        if (!newCtx.hasAnyStroke) {
-          this.continueStroke = false;
-        }
-        break;
-      case 'A':
-        this.parseAlign(newCtx);
-        break;
-      case 'C':
-        this.parseAciColor(newCtx);
-        break;
-      case 'c':
-        this.parseRgbColor(newCtx);
-        break;
-      case 'H':
-        this.parseHeight(newCtx);
-        break;
-      case 'W':
-        this.parseWidth(newCtx);
-        break;
-      case 'Q':
-        this.parseOblique(newCtx);
-        break;
-      case 'T':
-        this.parseCharTracking(newCtx);
-        break;
-      case 'p':
-        this.parseParagraphProperties(newCtx);
-        break;
-      case 'f':
-      case 'F':
-        this.parseFontProperties(newCtx);
-        break;
-      default:
-        throw new Error(`Unknown command: ${cmd}`);
-    }
-
-    // Update continueStroke based on current stroke state
-    this.continueStroke = newCtx.hasAnyStroke;
-    newCtx.continueStroke = this.continueStroke;
-    // Use setCurrent to replace the current context
-    this.ctxStack.setCurrent(newCtx);
-
-    if (this.yieldPropertyCommands) {
-      const changes = this.getPropertyChanges(prevCtx, newCtx);
-      if (Object.keys(changes).length > 0) {
-        return {
-          command: cmd,
-          changes,
-          depth: this.ctxStack.depth,
-        };
-      }
-    }
-  }
-
-  /**
-   * Get property changes between two contexts
-   * @param oldCtx - The old context
-   * @param newCtx - The new context
-   * @returns Object containing changed properties
-   */
-  private getPropertyChanges(
-    oldCtx: MTextContext,
-    newCtx: MTextContext
-  ): TokenData[TokenType.PROPERTIES_CHANGED]['changes'] {
-    const changes: TokenData[TokenType.PROPERTIES_CHANGED]['changes'] = {};
-
-    if (oldCtx.underline !== newCtx.underline) {
-      changes.underline = newCtx.underline;
-    }
-    if (oldCtx.overline !== newCtx.overline) {
-      changes.overline = newCtx.overline;
-    }
-    if (oldCtx.strikeThrough !== newCtx.strikeThrough) {
-      changes.strikeThrough = newCtx.strikeThrough;
-    }
-    if (oldCtx.color.aci !== newCtx.color.aci) {
-      changes.aci = newCtx.color.aci;
-    }
-    if (oldCtx.color.rgbValue !== newCtx.color.rgbValue) {
-      changes.rgb = newCtx.color.rgb;
-    }
-    if (oldCtx.align !== newCtx.align) {
-      changes.align = newCtx.align;
-    }
-    if (JSON.stringify(oldCtx.fontFace) !== JSON.stringify(newCtx.fontFace)) {
-      changes.fontFace = newCtx.fontFace;
-    }
-    if (
-      oldCtx.capHeight.value !== newCtx.capHeight.value ||
-      oldCtx.capHeight.isRelative !== newCtx.capHeight.isRelative
-    ) {
-      changes.capHeight = newCtx.capHeight;
-    }
-    if (
-      oldCtx.widthFactor.value !== newCtx.widthFactor.value ||
-      oldCtx.widthFactor.isRelative !== newCtx.widthFactor.isRelative
-    ) {
-      changes.widthFactor = newCtx.widthFactor;
-    }
-    if (
-      oldCtx.charTrackingFactor.value !== newCtx.charTrackingFactor.value ||
-      oldCtx.charTrackingFactor.isRelative !== newCtx.charTrackingFactor.isRelative
-    ) {
-      changes.charTrackingFactor = newCtx.charTrackingFactor;
-    }
-    if (oldCtx.oblique !== newCtx.oblique) {
-      changes.oblique = newCtx.oblique;
-    }
-    if (JSON.stringify(oldCtx.paragraph) !== JSON.stringify(newCtx.paragraph)) {
-      // Only include changed paragraph properties
-      const changedProps: Partial<ParagraphProperties> = {};
-      if (oldCtx.paragraph.indent !== newCtx.paragraph.indent) {
-        changedProps.indent = newCtx.paragraph.indent;
-      }
-      if (oldCtx.paragraph.align !== newCtx.paragraph.align) {
-        changedProps.align = newCtx.paragraph.align;
-      }
-      if (oldCtx.paragraph.left !== newCtx.paragraph.left) {
-        changedProps.left = newCtx.paragraph.left;
-      }
-      if (oldCtx.paragraph.right !== newCtx.paragraph.right) {
-        changedProps.right = newCtx.paragraph.right;
-      }
-      if (JSON.stringify(oldCtx.paragraph.tabs) !== JSON.stringify(newCtx.paragraph.tabs)) {
-        changedProps.tabs = newCtx.paragraph.tabs;
-      }
-      if (Object.keys(changedProps).length > 0) {
-        changes.paragraph = changedProps;
-      }
-    }
-
-    return changes;
-  }
-
-  /**
-   * Parse alignment property
-   * @param ctx - The context to update
-   */
-  private parseAlign(ctx: MTextContext): void {
-    const char = this.scanner.get();
-    if ('012'.includes(char)) {
-      ctx.align = parseInt(char) as MTextLineAlignment;
-    } else {
-      ctx.align = MTextLineAlignment.BOTTOM;
-    }
-    this.consumeOptionalTerminator();
-  }
-
-  /**
-   * Parse height property
-   * @param ctx - The context to update
-   */
-  private parseHeight(ctx: MTextContext): void {
-    const expr = this.extractFloatExpression(true);
-    if (expr) {
-      try {
-        if (expr.endsWith('x')) {
-          // For height command, treat x suffix as relative value
-          ctx.capHeight = {
-            value: parseFloat(expr.slice(0, -1)),
-            isRelative: true,
-          };
-        } else {
-          ctx.capHeight = {
-            value: parseFloat(expr),
-            isRelative: false,
-          };
-        }
-      } catch {
-        // If parsing fails, treat the entire command as literal text
-        this.scanner.consume(-expr.length); // Rewind to before the expression
-        return;
-      }
-    }
-    this.consumeOptionalTerminator();
-  }
-
-  /**
-   * Parse width property
-   * @param ctx - The context to update
-   */
-  private parseWidth(ctx: MTextContext): void {
-    const expr = this.extractFloatExpression(true);
-    if (expr) {
-      try {
-        if (expr.endsWith('x')) {
-          // For width command, treat x suffix as relative value
-          ctx.widthFactor = {
-            value: parseFloat(expr.slice(0, -1)),
-            isRelative: true,
-          };
-        } else {
-          ctx.widthFactor = {
-            value: parseFloat(expr),
-            isRelative: false,
-          };
-        }
-      } catch {
-        // If parsing fails, treat the entire command as literal text
-        this.scanner.consume(-expr.length); // Rewind to before the expression
-        return;
-      }
-    }
-    this.consumeOptionalTerminator();
-  }
-
-  /**
-   * Parse character tracking property
-   * @param ctx - The context to update
-   */
-  private parseCharTracking(ctx: MTextContext): void {
-    const expr = this.extractFloatExpression(true);
-    if (expr) {
-      try {
-        if (expr.endsWith('x')) {
-          // For tracking command, treat x suffix as relative value
-          ctx.charTrackingFactor = {
-            value: Math.abs(parseFloat(expr.slice(0, -1))),
-            isRelative: true,
-          };
-        } else {
-          ctx.charTrackingFactor = {
-            value: Math.abs(parseFloat(expr)),
-            isRelative: false,
-          };
-        }
-      } catch {
-        // If parsing fails, treat the entire command as literal text
-        this.scanner.consume(-expr.length); // Rewind to before the expression
-        return;
-      }
-    }
-    this.consumeOptionalTerminator();
-  }
-
-  /**
-   * Parse float value or factor
-   * @param value - Current value to apply factor to
-   * @returns New value
-   */
-  private parseFloatValueOrFactor(value: number): number {
-    const expr = this.extractFloatExpression(true);
-    if (expr) {
-      if (expr.endsWith('x')) {
-        const factor = parseFloat(expr.slice(0, -1));
-        value *= factor; // Allow negative factors
-      } else {
-        value = parseFloat(expr); // Allow negative values
-      }
-    }
-    return value;
-  }
-
-  /**
-   * Parse oblique angle property
-   * @param ctx - The context to update
-   */
-  private parseOblique(ctx: MTextContext): void {
-    const obliqueExpr = this.extractFloatExpression(false);
-    if (obliqueExpr) {
-      ctx.oblique = parseFloat(obliqueExpr);
-    }
-    this.consumeOptionalTerminator();
-  }
-
-  /**
-   * Parse ACI color property
-   * @param ctx - The context to update
-   */
-  private parseAciColor(ctx: MTextContext): void {
-    const aciExpr = this.extractIntExpression();
-    if (aciExpr) {
-      const aci = parseInt(aciExpr);
-      if (aci < 257) {
-        ctx.color.aci = aci;
-      }
-    }
-    this.consumeOptionalTerminator();
-  }
-
-  /**
-   * Parse RGB color property
-   * @param ctx - The context to update
-   */
-  private parseRgbColor(ctx: MTextContext): void {
-    const rgbExpr = this.extractIntExpression();
-    if (rgbExpr) {
-      const value = parseInt(rgbExpr) & 0xffffff;
-      ctx.color.rgbValue = value;
-    }
-    this.consumeOptionalTerminator();
-  }
-
-  /**
-   * Extract float expression from scanner
-   * @param relative - Whether to allow relative values (ending in 'x')
-   * @returns Extracted expression
-   */
-  private extractFloatExpression(relative: boolean = false): string {
-    const pattern = relative
-      ? /^[+-]?(?:\d+(?:\.\d*)?|\.\d+)(?:[eE][+-]?\d+)?x?/
-      : /^[+-]?(?:\d+(?:\.\d*)?|\.\d+)(?:[eE][+-]?\d+)?/;
-    const match = this.scanner.tail.match(pattern);
-    if (match) {
-      const result = match[0];
-      this.scanner.consume(result.length);
-      return result;
-    }
-    return '';
-  }
-
-  /**
-   * Extract integer expression from scanner
-   * @returns Extracted expression
-   */
-  private extractIntExpression(): string {
-    const match = this.scanner.tail.match(/^\d+/);
-    if (match) {
-      const result = match[0];
-      this.scanner.consume(result.length);
-      return result;
-    }
-    return '';
-  }
-
-  /**
-   * Extract expression until semicolon or end
-   * @param escape - Whether to handle escaped semicolons
-   * @returns Extracted expression
-   */
-  private extractExpression(escape: boolean = false): string {
-    const stop = this.scanner.find(';', escape);
-    if (stop < 0) {
-      const expr = this.scanner.tail;
-      this.scanner.consume(expr.length);
-      return expr;
-    }
-    // Check if the semicolon is escaped by looking at the previous character
-    const prevChar = this.scanner.peek(stop - this.scanner.currentIndex - 1);
-    const isEscaped = prevChar === '\\';
-    const expr = this.scanner.tail.slice(0, stop - this.scanner.currentIndex + (isEscaped ? 1 : 0));
-    this.scanner.consume(expr.length + 1);
-    return expr;
-  }
-
-  /**
-   * Parse font properties
-   * @param ctx - The context to update
-   */
-  private parseFontProperties(ctx: MTextContext): void {
-    const parts = this.extractExpression().split('|');
-    if (parts.length > 0 && parts[0]) {
-      const name = parts[0];
-      let style: FontStyle = 'Regular';
-      let weight = 400;
-
-      for (const part of parts.slice(1)) {
-        if (part.startsWith('b1')) {
-          weight = 700;
-        } else if (part === 'i' || part.startsWith('i1')) {
-          style = 'Italic';
-        } else if (part === 'i0' || part.startsWith('i0')) {
-          style = 'Regular';
-        }
-      }
-
-      ctx.fontFace = {
-        family: name,
-        style,
-        weight,
-      };
-    }
-  }
-
-  /**
-   * Parse paragraph properties from the MText content
-   * Handles properties like indentation, alignment, and tab stops
-   * @param ctx - The context to update
-   */
-  private parseParagraphProperties(ctx: MTextContext): void {
-    const scanner = new TextScanner(this.extractExpression());
-    /** Current indentation value */
-    let indent = ctx.paragraph.indent;
-    /** Left margin value */
-    let left = ctx.paragraph.left;
-    /** Right margin value */
-    let right = ctx.paragraph.right;
-    /** Current paragraph alignment */
-    let align = ctx.paragraph.align;
-    /** Array of tab stop positions and types */
-    let tabStops: (number | string)[] = [];
-
-    /**
-     * Parse a floating point number from the scanner's current position
-     * Handles optional sign, decimal point, and scientific notation
-     * @returns The parsed float value, or 0 if no valid number is found
-     */
-    const parseFloatValue = (): number => {
-      const match = scanner.tail.match(/^[+-]?\d+(?:\.\d*)?(?:[eE][+-]?\d+)?/);
-      if (match) {
-        const value = parseFloat(match[0]);
-        scanner.consume(match[0].length);
-        while (scanner.peek() === ',') {
-          scanner.consume(1);
-        }
-        return value;
-      }
-      return 0;
-    };
-
-    while (scanner.hasData) {
-      const cmd = scanner.get();
-      switch (cmd) {
-        case 'i': // Indentation
-          indent = parseFloatValue();
-          break;
-        case 'l': // Left margin
-          left = parseFloatValue();
-          break;
-        case 'r': // Right margin
-          right = parseFloatValue();
-          break;
-        case 'x': // Skip
-          break;
-        case 'q': {
-          // Alignment
-          const adjustment = scanner.get();
-          align = CHAR_TO_ALIGN[adjustment] || MTextParagraphAlignment.DEFAULT;
-          while (scanner.peek() === ',') {
-            scanner.consume(1);
-          }
-          break;
-        }
-        case 't': // Tab stops
-          tabStops = [];
-          while (scanner.hasData) {
-            const type = scanner.peek();
-            if (type === 'r' || type === 'c') {
-              scanner.consume(1);
-              const value = parseFloatValue();
-              tabStops.push(type + value.toString());
-            } else {
-              const value = parseFloatValue();
-              if (!isNaN(value)) {
-                tabStops.push(value);
-              } else {
-                scanner.consume(1);
-              }
-            }
-          }
-          break;
-      }
-    }
-
-    ctx.paragraph = {
-      indent,
-      left,
-      right,
-      align,
-      tabs: tabStops,
-    };
-  }
-
-  /**
-   * Consume optional terminator (semicolon)
-   */
-  private consumeOptionalTerminator(): void {
-    if (this.scanner.peek() === ';') {
-      this.scanner.consume(1);
-    }
-  }
-
-  /**
-   * Builds {@link PercentSymbolData} for a recognized `%%` code letter.
-   */
-  private buildPercentSymbolData(
-    code: string,
-    specialChar: string
-  ): PercentSymbolData | null {
-    if (code === 'c' || code === 'd' || code === 'p') {
-      return { kind: 'named', code, char: specialChar };
-    }
-    if (code === '%') {
-      return { kind: 'literal', char: '%' };
-    }
-    return null;
-  }
-
-  /**
-   * Parse MText content into tokens
-   * @yields MTextToken objects
-   */
-  *parse(): Generator<MTextToken> {
-    const wordToken = TokenType.WORD;
-    const spaceToken = TokenType.SPACE;
-    let followupToken: TokenType | null = null;
-    let followupData: TokenData[TokenType] | undefined;
-
-    function resetParagraph(ctx: MTextContext): Partial<ParagraphProperties> {
-      const prev = { ...ctx.paragraph };
-      ctx.paragraph = {
-        indent: 0,
-        left: 0,
-        right: 0,
-        align: MTextParagraphAlignment.DEFAULT,
-        tabs: [],
-      };
-      const changed: Partial<ParagraphProperties> = {};
-      if (prev.indent !== 0) changed.indent = 0;
-      if (prev.left !== 0) changed.left = 0;
-      if (prev.right !== 0) changed.right = 0;
-      if (prev.align !== MTextParagraphAlignment.DEFAULT)
-        changed.align = MTextParagraphAlignment.DEFAULT;
-      if (JSON.stringify(prev.tabs) !== JSON.stringify([])) changed.tabs = [];
-      return changed;
-    }
-
-    const nextToken = (): [TokenType, TokenData[TokenType]] => {
-      let word = '';
-      while (this.scanner.hasData) {
-        let escape = false;
-        let letter = this.scanner.peek();
-        const cmdStartIndex = this.scanner.currentIndex;
-
-        // Handle control characters first
-        if (letter.charCodeAt(0) < 32) {
-          this.scanner.consume(1); // Always consume the control character
-          if (letter === '\t') {
-            return [TokenType.TABULATOR, null];
-          }
-          if (letter === '\n') {
-            return [TokenType.NEW_PARAGRAPH, null];
-          }
-          letter = ' ';
-        }
-
-        if (letter === '\\') {
-          if ('\\{}'.includes(this.scanner.peek(1))) {
-            escape = true;
-            this.scanner.consume(1);
-            letter = this.scanner.peek();
-          } else {
-            if (word) {
-              return [wordToken, word];
-            }
-            this.scanner.consume(1);
-            const cmd = this.scanner.get();
-            switch (cmd) {
-              case '~':
-                return [TokenType.NBSP, null];
-              case 'P':
-                return [TokenType.NEW_PARAGRAPH, null];
-              case 'N':
-                return [TokenType.NEW_COLUMN, null];
-              case 'X':
-                return [TokenType.WRAP_AT_DIMLINE, null];
-              case 'S': {
-                this.inStackContext = true;
-                const result = this.parseStacking();
-                this.inStackContext = false;
-                return result;
-              }
-              case 'm':
-              case 'M':
-                // Handle multi-byte character encoding (MIF)
-                if (this.scanner.peek() === '+') {
-                  this.scanner.consume(1); // Consume the '+'
-                  const hexCode = this.extractMifCode(this.mifCodeLength);
-                  if (hexCode) {
-                    this.scanner.consume(hexCode.length);
-                    const decodedChar = this.mifDecoder(hexCode);
-                    if (word) {
-                      return [wordToken, word];
-                    }
-                    return [wordToken, decodedChar];
-                  }
-                  // If no valid hex code found, rewind the '+' character
-                  this.scanner.consume(-1);
-                }
-                // If not a valid multi-byte code, treat as literal text
-                word += '\\M';
-                continue;
-              case 'U': {
-                // AutoCAD \U+nnnn uses exactly four hex digits. Characters
-                // outside the BMP are encoded as two consecutive \U+XXXX
-                // UTF-16 surrogate escapes.
-                const decodedChar = this.parseUnicodeEscape();
-                if (decodedChar !== null) {
-                  if (word) {
-                    return [wordToken, word];
-                  }
-                  return [wordToken, decodedChar];
-                }
-                // If not a valid Unicode code, treat as literal text
-                word += '\\U';
-                continue;
-              }
-              default:
-                if (cmd) {
-                  try {
-                    const propertyChanges = this.parseProperties(cmd);
-                    if (this.yieldPropertyCommands && propertyChanges) {
-                      return [TokenType.PROPERTIES_CHANGED, propertyChanges];
-                    }
-                    // After processing a property command, continue with normal parsing
-                    continue;
-                  } catch {
-                    const commandText = this.scanner.tail.slice(
-                      cmdStartIndex,
-                      this.scanner.currentIndex
-                    );
-                    word += commandText;
-                  }
-                }
-            }
-            continue;
-          }
-        }
-
-        if (letter === '%' && this.scanner.peek(1) === '%') {
-          const code = this.scanner.peek(2).toLowerCase();
-          const specialChar = SPECIAL_CHAR_ENCODING[code];
-          if (specialChar) {
-            this.scanner.consume(3);
-            if (this.yieldPercentSymbols) {
-              const symbolData = this.buildPercentSymbolData(code, specialChar);
-              if (symbolData) {
-                if (word) {
-                  followupToken = TokenType.PERCENT_SYMBOL;
-                  followupData = symbolData;
-                  return [wordToken, word];
-                }
-                return [TokenType.PERCENT_SYMBOL, symbolData];
-              }
-            }
-            word += specialChar;
-            continue;
-          } else {
-            /**
-             * Supports Control Codes: `%%nnn`, where nnn is 1–3 decimal digits
-             * for the character's Unicode/ASCII code point (e.g. `%%34` → `"`,
-             * `%%176` → `°`). Digit count is not fixed at three.
-             *
-             * Reference: https://help.autodesk.com/view/ACD/2026/ENU/?guid=GUID-968CBC1D-BA99-4519-ABDD-88419EB2BF92
-             */
-            const digitChars: string[] = [];
-            for (let i = 0; i < 3; i++) {
-              const d = this.scanner.peek(2 + i);
-              if (d >= '0' && d <= '9') {
-                digitChars.push(d);
-              } else {
-                break;
-              }
-            }
-
-            if (digitChars.length > 0) {
-              const charCode = Number.parseInt(digitChars.join(''), 10);
-              this.scanner.consume(2 + digitChars.length);
-              if (this.yieldPercentSymbols) {
-                const symbolData: PercentSymbolData = {
-                  kind: 'numeric',
-                  charCode,
-                  char: String.fromCharCode(charCode),
-                };
-                if (word) {
-                  followupToken = TokenType.PERCENT_SYMBOL;
-                  followupData = symbolData;
-                  return [wordToken, word];
-                }
-                return [TokenType.PERCENT_SYMBOL, symbolData];
-              }
-              word += String.fromCharCode(charCode);
-            } else {
-              // Skip invalid special character codes (`%%` + non-digit)
-              this.scanner.consume(3);
-            }
-
-            continue;
-          }
-        }
-
-        if (letter === ' ') {
-          if (word) {
-            this.scanner.consume(1);
-            followupToken = spaceToken;
-            return [wordToken, word];
-          }
-          this.scanner.consume(1);
-          return [spaceToken, null];
-        }
-
-        if (!escape) {
-          if (letter === '{') {
-            if (word) {
-              return [wordToken, word];
-            }
-            this.scanner.consume(1);
-            this.pushCtx();
-            continue;
-          } else if (letter === '}') {
-            if (word) {
-              return [wordToken, word];
-            }
-            this.scanner.consume(1);
-            // Context restoration with yieldPropertyCommands
-            if (this.yieldPropertyCommands) {
-              const prevCtx = this.ctxStack.current;
-              this.popCtx();
-              const changes = this.getPropertyChanges(prevCtx, this.ctxStack.current);
-              if (Object.keys(changes).length > 0) {
-                return [
-                  TokenType.PROPERTIES_CHANGED,
-                  { command: undefined, changes, depth: this.ctxStack.depth },
-                ];
-              }
-            } else {
-              this.popCtx();
-            }
-            continue;
-          }
-        }
-
-        // Handle caret-encoded characters only when not in stack context
-        if (!this.inStackContext && letter === '^') {
-          const nextChar = this.scanner.peek(1);
-          if (nextChar) {
-            const code = nextChar.charCodeAt(0);
-            this.scanner.consume(2); // Consume both ^ and the next character
-            if (code === 32) {
-              // Space
-              word += '^';
-            } else if (code === 73) {
-              // Tab
-              if (word) {
-                return [wordToken, word];
-              }
-              return [TokenType.TABULATOR, null];
-            } else if (code === 74) {
-              // Line feed
-              if (word) {
-                return [wordToken, word];
-              }
-              return [TokenType.NEW_PARAGRAPH, null];
-            } else if (code === 77) {
-              // Carriage return
-              // Ignore carriage return
-              continue;
-            } else {
-              word += '▯';
-            }
-            continue;
-          }
-        }
-
-        this.scanner.consume(1);
-        if (letter.charCodeAt(0) >= 32) {
-          word += letter;
-        }
-      }
-
-      if (word) {
-        return [wordToken, word];
-      }
-      return [TokenType.NONE, null];
-    };
-
-    while (true) {
-      const [type, data] = nextToken.call(this);
-      if (type) {
-        yield new MTextToken(type, this.ctxStack.current.copy(), data);
-        if (type === TokenType.NEW_PARAGRAPH && this.resetParagraphParameters) {
-          // Reset paragraph properties and emit PROPERTIES_CHANGED if needed
-          const ctx = this.ctxStack.current;
-          const changed = resetParagraph(ctx);
-          if (this.yieldPropertyCommands && Object.keys(changed).length > 0) {
-            yield new MTextToken(TokenType.PROPERTIES_CHANGED, ctx.copy(), {
-              command: undefined,
-              changes: { paragraph: changed },
-              depth: this.ctxStack.depth,
-            });
-          }
-        }
-        if (followupToken) {
-          yield new MTextToken(
-            followupToken,
-            this.ctxStack.current.copy(),
-            followupData ?? null
-          );
-          followupToken = null;
-          followupData = undefined;
-        }
-      } else {
-        break;
-      }
-    }
-  }
-}
-
-/**
- * Text scanner for parsing MText content
- */
-export class TextScanner {
-  private text: string;
-  private textLen: number;
-  private _index: number;
-
-  /**
-   * Create a new text scanner
-   * @param text - The text to scan
-   */
-  constructor(text: string) {
-    this.text = text;
-    this.textLen = text.length;
-    this._index = 0;
-  }
-
-  /**
-   * Get the current index in the text
-   */
-  get currentIndex(): number {
-    return this._index;
-  }
-
-  /**
-   * Check if the scanner has reached the end of the text
-   */
-  get isEmpty(): boolean {
-    return this._index >= this.textLen;
-  }
-
-  /**
-   * Check if there is more text to scan
-   */
-  get hasData(): boolean {
-    return this._index < this.textLen;
-  }
-
-  /**
-   * Get the next character and advance the index
-   * @returns The next character, or empty string if at end
-   */
-  get(): string {
-    if (this.isEmpty) {
-      return '';
-    }
-    const char = this.text[this._index];
-    this._index++;
-    return char;
-  }
-
-  /**
-   * Advance the index by the specified count
-   * @param count - Number of characters to advance
-   */
-  consume(count: number = 1): void {
-    this._index = Math.max(0, Math.min(this._index + count, this.textLen));
-  }
-
-  /**
-   * Look at a character without advancing the index
-   * @param offset - Offset from current position
-   * @returns The character at the offset position, or empty string if out of bounds
-   */
-  peek(offset: number = 0): string {
-    const index = this._index + offset;
-    if (index >= this.textLen || index < 0) {
-      return '';
-    }
-    return this.text[index];
-  }
-
-  /**
-   * Find the next occurrence of a character
-   * @param char - The character to find
-   * @param escape - Whether to handle escaped characters
-   * @returns Index of the character, or -1 if not found
-   */
-  find(char: string, escape: boolean = false): number {
-    let index = this._index;
-    while (index < this.textLen) {
-      if (escape && this.text[index] === '\\') {
-        if (index + 1 < this.textLen) {
-          if (this.text[index + 1] === char) {
-            return index + 1;
-          }
-          index += 2;
-          continue;
-        }
-        index++;
-        continue;
-      }
-      if (this.text[index] === char) {
-        return index;
-      }
-      index++;
-    }
-    return -1;
-  }
-
-  /**
-   * Get the remaining text from the current position
-   */
-  get tail(): string {
-    return this.text.slice(this._index);
-  }
-
-  /**
-   * Check if the next character is a space
-   */
-  isNextSpace(): boolean {
-    return this.peek() === ' ';
-  }
-
-  /**
-   * Consume spaces until a non-space character is found
-   * @returns Number of spaces consumed
-   */
-  consumeSpaces(): number {
-    let count = 0;
-    while (this.isNextSpace()) {
-      this.consume();
-      count++;
-    }
-    return count;
-  }
-}
-
-/**
- * Class to handle ACI and RGB color logic for MText.
- *
- * This class encapsulates color state for MText, supporting both AutoCAD Color Index (ACI) and RGB color.
- * Only one color mode is active at a time: setting an RGB color disables ACI, and vice versa.
- * RGB is stored as a single 24-bit integer (0xRRGGBB) for efficient comparison and serialization.
- *
- * Example usage:
- * ```ts
- * const color1 = new MTextColor(1); // ACI color
- * const color2 = new MTextColor([255, 0, 0]); // RGB color
- * const color3 = new MTextColor(); // Default (ACI=256, "by layer")
- * ```
- */
-export class MTextColor {
-  /**
-   * The AutoCAD Color Index (ACI) value. Only used if no RGB color is set.
-   * @default 256 ("by layer")
-   */
-  private _aci: number | null = 256;
-  /**
-   * The RGB color value as a single 24-bit integer (0xRRGGBB), or null if not set.
-   * @default null
-   */
-  private _rgbValue: number | null = null; // Store as 0xRRGGBB or null
-
-  /**
-   * Create a new MTextColor instance.
-   * @param color The initial color: number for ACI, [r,g,b] for RGB, or null/undefined for default (ACI=256).
-   */
-  constructor(color?: number | RGB | null) {
-    if (Array.isArray(color)) {
-      this.rgb = color;
-    } else if (typeof color === 'number') {
-      this.aci = color;
-    } else {
-      this.aci = 256;
-    }
-  }
-
-  /**
-   * Get the current ACI color value.
-   * @returns The ACI color (0-256), or null if using RGB.
-   */
-  get aci(): number | null {
-    return this._aci;
-  }
-
-  /**
-   * Set the ACI color value. Setting this disables any RGB color.
-   * @param value The ACI color (0-256), or null to unset.
-   * @throws Error if value is out of range.
-   */
-  set aci(value: number | null) {
-    if (value === null) {
-      this._aci = null;
-    } else if (value >= 0 && value <= 256) {
-      this._aci = value;
-      this._rgbValue = null;
-    } else {
-      throw new Error('ACI not in range [0, 256]');
-    }
-  }
-
-  /**
-   * Get the current RGB color as a tuple [r, g, b], or null if not set.
-   * @returns The RGB color tuple, or null if using ACI.
-   */
-  get rgb(): RGB | null {
-    if (this._rgbValue === null) return null;
-    // Extract R, G, B from 0xRRGGBB
-    const r = (this._rgbValue >> 16) & 0xff;
-    const g = (this._rgbValue >> 8) & 0xff;
-    const b = this._rgbValue & 0xff;
-    return [r, g, b];
-  }
-
-  /**
-   * Set the RGB color. Setting this disables ACI color.
-   * @param value The RGB color tuple [r, g, b], or null to use ACI.
-   */
-  set rgb(value: RGB | null) {
-    if (value) {
-      const [r, g, b] = value;
-      this._rgbValue = ((r & 0xff) << 16) | ((g & 0xff) << 8) | (b & 0xff);
-      this._aci = null;
-    } else {
-      this._rgbValue = null;
-    }
-  }
-
-  /**
-   * Returns true if the color is set by RGB, false if by ACI.
-   */
-  get isRgb(): boolean {
-    return this._rgbValue !== null;
-  }
-
-  /**
-   * Returns true if the color is set by ACI, false if by RGB.
-   */
-  get isAci(): boolean {
-    return this._rgbValue === null && this._aci !== null;
-  }
-
-  /**
-   * Get or set the internal RGB value as a number (0xRRGGBB), or null if not set.
-   * Setting this will switch to RGB mode and set ACI to null.
-   */
-  get rgbValue(): number | null {
-    return this._rgbValue;
-  }
-
-  set rgbValue(val: number | null) {
-    if (val === null) {
-      this._rgbValue = null;
-    } else {
-      this._rgbValue = val & 0xffffff;
-      this._aci = null;
-    }
-  }
-
-  /**
-   * Returns a deep copy of this color.
-   * @returns A new MTextColor instance with the same color state.
-   */
-  copy(): MTextColor {
-    const c = new MTextColor();
-    c._aci = this._aci;
-    c._rgbValue = this._rgbValue;
-    return c;
-  }
-
-  /**
-   * Returns a plain object for serialization.
-   * @returns An object with aci, rgb (tuple), and rgbValue (number or null).
-   */
-  toObject(): { aci: number | null; rgb: RGB | null; rgbValue: number | null } {
-    return { aci: this._aci, rgb: this.rgb, rgbValue: this._rgbValue };
-  }
-
-  /**
-   * Convert the current color to a CSS hex color string (#rrggbb).
-   * Returns null if the color is ACI-based and has no RGB value.
-   */
-  toCssColor(): string | null {
-    if (this._rgbValue !== null) {
-      return colorNumberToHex(this._rgbValue);
-    }
-    return null;
-  }
-
-  /**
-   * Create an MTextColor from a CSS color string.
-   * Supports #rgb, #rrggbb, rgb(...), rgba(...). Returns null if invalid or transparent.
-   */
-  static fromCssColor(value: string | null | undefined): MTextColor | null {
-    const rgbValue = cssColorToRgbValue(value);
-    if (rgbValue === null) return null;
-    const color = new MTextColor();
-    color.rgbValue = rgbValue;
-    return color;
-  }
-
-  /**
-   * Equality check for color.
-   * @param other The other MTextColor to compare.
-   * @returns True if both ACI and RGB values are equal.
-   */
-  equals(other: MTextColor): boolean {
-    return this._aci === other._aci && this._rgbValue === other._rgbValue;
-  }
-}
-
-/**
- * MText context class for managing text formatting state
- */
-export class MTextContext {
-  private _stroke: number = 0;
-  /** Whether to continue stroke formatting */
-  continueStroke: boolean = false;
-  /** Color (ACI or RGB) */
-  color: MTextColor = new MTextColor();
-  /** Line alignment */
-  align: MTextLineAlignment = MTextLineAlignment.BOTTOM;
-  /** Font face properties */
-  fontFace: FontFace = { family: '', style: 'Regular', weight: 400 };
-  /** Capital letter height */
-  private _capHeight: FactorValue = { value: 1.0, isRelative: false };
-  /** Character width factor */
-  private _widthFactor: FactorValue = { value: 1.0, isRelative: false };
-  /**
-   * Character tracking factor a multiplier applied to the default spacing between characters in the MText object.
-   * - Value = 1.0 → Normal spacing.
-   * - Value < 1.0 → Characters are closer together.
-   * - Value > 1.0 → Characters are spaced farther apart.
-   */
-  private _charTrackingFactor: FactorValue = { value: 1.0, isRelative: false };
-  /** Oblique angle */
-  oblique: number = 0.0;
-  /** Paragraph properties */
-  paragraph: ParagraphProperties = {
-    indent: 0,
-    left: 0,
-    right: 0,
-    align: MTextParagraphAlignment.DEFAULT,
-    tabs: [],
-  };
-
-  /**
-   * Get the capital letter height
-   */
-  get capHeight(): FactorValue {
-    return this._capHeight;
-  }
-
-  /**
-   * Set the capital letter height
-   * @param value - Height value
-   */
-  set capHeight(value: FactorValue) {
-    this._capHeight = {
-      value: Math.abs(value.value),
-      isRelative: value.isRelative,
-    };
-  }
-
-  /**
-   * Get the character width factor
-   */
-  get widthFactor(): FactorValue {
-    return this._widthFactor;
-  }
-
-  /**
-   * Set the character width factor
-   * @param value - Width factor value
-   */
-  set widthFactor(value: FactorValue) {
-    this._widthFactor = {
-      value: Math.abs(value.value),
-      isRelative: value.isRelative,
-    };
-  }
-
-  /**
-   * Get the character tracking factor
-   */
-  get charTrackingFactor(): FactorValue {
-    return this._charTrackingFactor;
-  }
-
-  /**
-   * Set the character tracking factor
-   * @param value - Tracking factor value
-   */
-  set charTrackingFactor(value: FactorValue) {
-    this._charTrackingFactor = {
-      value: Math.abs(value.value),
-      isRelative: value.isRelative,
-    };
-  }
-
-  /**
-   * Get the ACI color value
-   */
-  get aci(): number | null {
-    return this.color.aci;
-  }
-
-  /**
-   * Set the ACI color value
-   * @param value - ACI color value (0-256)
-   * @throws Error if value is out of range
-   */
-  set aci(value: number) {
-    this.color.aci = value;
-  }
-
-  /**
-   * Get the RGB color value
-   */
-  get rgb(): RGB | null {
-    return this.color.rgb;
-  }
-
-  /**
-   * Set the RGB color value
-   */
-  set rgb(value: RGB | null) {
-    this.color.rgb = value;
-  }
-
-  /**
-   * Gets whether the current text should be rendered in italic style.
-   * @returns {boolean} True if the font style is 'Italic', otherwise false.
-   */
-  get italic(): boolean {
-    return this.fontFace.style === 'Italic';
-  }
-  /**
-   * Sets whether the current text should be rendered in italic style.
-   * @param value - If true, sets the font style to 'Italic'; if false, sets it to 'Regular'.
-   */
-  set italic(value: boolean) {
-    this.fontFace.style = value ? 'Italic' : 'Regular';
-  }
-
-  /**
-   * Gets whether the current text should be rendered in bold style.
-   * This is primarily used for mesh fonts and affects font selection.
-   * @returns {boolean} True if the font weight is 700 or higher, otherwise false.
-   */
-  get bold(): boolean {
-    return (this.fontFace.weight || 400) >= 700;
-  }
-  /**
-   * Sets whether the current text should be rendered in bold style.
-   * This is primarily used for mesh fonts and affects font selection.
-   * @param value - If true, sets the font weight to 700; if false, sets it to 400.
-   */
-  set bold(value: boolean) {
-    this.fontFace.weight = value ? 700 : 400;
-  }
-
-  /**
-   * Get whether text is underlined
-   */
-  get underline(): boolean {
-    return Boolean(this._stroke & MTextStroke.UNDERLINE);
-  }
-
-  /**
-   * Set whether text is underlined
-   * @param value - Whether to underline
-   */
-  set underline(value: boolean) {
-    this._setStrokeState(MTextStroke.UNDERLINE, value);
-  }
-
-  /**
-   * Get whether text has strike-through
-   */
-  get strikeThrough(): boolean {
-    return Boolean(this._stroke & MTextStroke.STRIKE_THROUGH);
-  }
-
-  /**
-   * Set whether text has strike-through
-   * @param value - Whether to strike through
-   */
-  set strikeThrough(value: boolean) {
-    this._setStrokeState(MTextStroke.STRIKE_THROUGH, value);
-  }
-
-  /**
-   * Get whether text has overline
-   */
-  get overline(): boolean {
-    return Boolean(this._stroke & MTextStroke.OVERLINE);
-  }
-
-  /**
-   * Set whether text has overline
-   * @param value - Whether to overline
-   */
-  set overline(value: boolean) {
-    this._setStrokeState(MTextStroke.OVERLINE, value);
-  }
-
-  /**
-   * Check if any stroke formatting is active
-   */
-  get hasAnyStroke(): boolean {
-    return Boolean(this._stroke);
-  }
-
-  /**
-   * Set the state of a stroke type
-   * @param stroke - The stroke type to set
-   * @param state - Whether to enable or disable the stroke
-   */
-  private _setStrokeState(stroke: MTextStroke, state: boolean = true): void {
-    if (state) {
-      this._stroke |= stroke;
-    } else {
-      this._stroke &= ~stroke;
-    }
-  }
-
-  /**
-   * Create a copy of this context
-   * @returns A new context with the same properties
-   */
-  copy(): MTextContext {
-    const ctx = new MTextContext();
-    ctx._stroke = this._stroke;
-    ctx.continueStroke = this.continueStroke;
-    ctx.color = this.color.copy();
-    ctx.align = this.align;
-    ctx.fontFace = { ...this.fontFace };
-    ctx._capHeight = { ...this._capHeight };
-    ctx._widthFactor = { ...this._widthFactor };
-    ctx._charTrackingFactor = { ...this._charTrackingFactor };
-    ctx.oblique = this.oblique;
-    ctx.paragraph = { ...this.paragraph };
-    return ctx;
-  }
-}
-
-/**
- * Token class for MText parsing
- */
-export class MTextToken {
-  /**
-   * Create a new MText token
-   * @param type - The token type
-   * @param ctx - The text context at this token
-   * @param data - Optional token data
-   */
-  constructor(
-    public type: TokenType,
-    public ctx: MTextContext,
-    public data: TokenData[TokenType]
-  ) {}
-}
+/**
+ * Token types used in MText parsing
+ */
+export enum TokenType {
+  /** No token */
+  NONE = 0,
+  /** Word token with string data */
+  WORD = 1,
+  /** Stack token with [numerator, denominator, type] data */
+  STACK = 2,
+  /** Space token with no data */
+  SPACE = 3,
+  /** Non-breaking space token with no data */
+  NBSP = 4,
+  /** Tab token with no data */
+  TABULATOR = 5,
+  /** New paragraph token with no data */
+  NEW_PARAGRAPH = 6,
+  /** New column token with no data */
+  NEW_COLUMN = 7,
+  /** Wrap at dimension line token with no data */
+  WRAP_AT_DIMLINE = 8,
+  /** Properties changed token with string data (full command) */
+  PROPERTIES_CHANGED = 9,
+  /** AutoCAD percent-sign symbol code (`%%c`, `%%d`, `%%p`, `%%nnn`, `%%%`) */
+  PERCENT_SYMBOL = 10,
+}
+
+/**
+ * AutoCAD percent-sign symbol emitted when {@link MTextParserOptions.yieldPercentSymbols}
+ * is enabled.
+ */
+export type PercentSymbolData =
+  | {
+      kind: 'named';
+      /** Percent code letter: `%%c`, `%%d`, or `%%p`. */
+      code: 'c' | 'd' | 'p';
+      /** Unicode expansion used for display (e.g. `%%c` → `Ø`). */
+      char: string;
+    }
+  | {
+      kind: 'numeric';
+      /**
+       * Decimal code from `%%nnn`, where `nnn` is 1–3 digits (Unicode/ASCII
+       * code point, e.g. `%%34` → `"`, `%%176` → `°`).
+       */
+      charCode: number;
+      /** `String.fromCharCode(charCode)`. */
+      char: string;
+    }
+  | {
+      kind: 'literal';
+      /** Literal percent from `%%%`. */
+      char: '%';
+    };
+
+/**
+ * Represents a factor value that can be either absolute or relative.
+ * Used for properties like height, width, and character tracking in MText formatting.
+ */
+export interface FactorValue {
+  /** The numeric value of the factor */
+  value: number;
+  /** Whether the value is relative (true) or absolute (false) */
+  isRelative: boolean;
+}
+
+/**
+ * Format properties of MText word tokens.
+ * This interface defines all the formatting properties that can be applied to MText content,
+ * including text styling, colors, alignment, font properties, and paragraph formatting.
+ */
+export interface Properties {
+  /** Whether text is underlined */
+  underline?: boolean;
+  /** Whether text has an overline */
+  overline?: boolean;
+  /** Whether text has strike-through */
+  strikeThrough?: boolean;
+  /** AutoCAD Color Index (ACI) color value (0-256), or null if not set */
+  aci?: number | null;
+  /** RGB color tuple [r, g, b], or null if not set */
+  rgb?: RGB | null;
+  /** Line alignment for the text */
+  align?: MTextLineAlignment;
+  /** Font face properties including family, style, and weight */
+  fontFace?: FontFace;
+  /** Capital letter height factor (can be relative or absolute) */
+  capHeight?: FactorValue;
+  /** Character width factor (can be relative or absolute) */
+  widthFactor?: FactorValue;
+  /** Character tracking factor for spacing between characters (can be relative or absolute) */
+  charTrackingFactor?: FactorValue;
+  /** Oblique angle in degrees for text slant */
+  oblique?: number;
+  /** Paragraph formatting properties (partial to allow selective updates) */
+  paragraph?: Partial<ParagraphProperties>;
+}
+
+/**
+ * Represents a change in MText properties, including the command, the changed properties, and the context depth.
+ */
+export interface ChangedProperties {
+  /**
+   * The property command that triggered the change (e.g., 'L', 'C', 'f').
+   * The command will be undefined if it is to restore context.
+   */
+  command: string | undefined;
+  /**
+   * The set of properties that have changed as a result of the command.
+   */
+  changes: Properties;
+  /**
+   * The current context stack depth when the property change occurs.
+   * - 0: The change is global (applies outside of any `{}` block).
+   * - >0: The change is local (applies within one or more nested `{}` blocks).
+   */
+  depth: number; // 0 = global, >0 = local
+}
+
+/**
+ * Type for token data based on token type
+ */
+export type TokenData = {
+  [TokenType.NONE]: null;
+  [TokenType.WORD]: string;
+  [TokenType.STACK]: [string, string, string];
+  [TokenType.SPACE]: null;
+  [TokenType.NBSP]: null;
+  [TokenType.TABULATOR]: null;
+  [TokenType.NEW_PARAGRAPH]: null;
+  [TokenType.NEW_COLUMN]: null;
+  [TokenType.WRAP_AT_DIMLINE]: null;
+  [TokenType.PROPERTIES_CHANGED]: ChangedProperties;
+  [TokenType.PERCENT_SYMBOL]: PercentSymbolData;
+};
+
+/**
+ * Line alignment options for MText
+ */
+export enum MTextLineAlignment {
+  /** Align text to bottom */
+  BOTTOM = 0,
+  /** Align text to middle */
+  MIDDLE = 1,
+  /** Align text to top */
+  TOP = 2,
+}
+
+/**
+ * Paragraph alignment options for MText
+ */
+export enum MTextParagraphAlignment {
+  /** Default alignment */
+  DEFAULT = 0,
+  /** Left alignment */
+  LEFT = 1,
+  /** Right alignment */
+  RIGHT = 2,
+  /** Center alignment */
+  CENTER = 3,
+  /** Justified alignment */
+  JUSTIFIED = 4,
+  /** Distributed alignment */
+  DISTRIBUTED = 5,
+}
+
+/**
+ * Text stroke options for MText
+ */
+export enum MTextStroke {
+  /** No stroke */
+  NONE = 0,
+  /** Underline stroke */
+  UNDERLINE = 1,
+  /** Overline stroke */
+  OVERLINE = 2,
+  /** Strike-through stroke */
+  STRIKE_THROUGH = 4,
+}
+
+/**
+ * RGB color tuple
+ */
+export type RGB = [number, number, number];
+
+/**
+ * Font style type
+ */
+export type FontStyle = 'Regular' | 'Italic';
+
+/**
+ * Font face properties
+ */
+export interface FontFace {
+  /** Font family name */
+  family: string;
+  /** Font style (e.g., 'Regular', 'Italic') */
+  style: FontStyle;
+  /** Font weight (e.g., 400 for normal, 700 for bold) */
+  weight: number;
+}
+
+/**
+ * Paragraph properties
+ */
+export interface ParagraphProperties {
+  /** Indentation value */
+  indent: number;
+  /** Left margin value */
+  left: number;
+  /** Right margin value */
+  right: number;
+  /** Paragraph alignment */
+  align: MTextParagraphAlignment;
+  /** Tab stop positions and types */
+  tabs: (number | string)[];
+}
+
+/**
+ * Special character encoding mapping
+ */
+const SPECIAL_CHAR_ENCODING: Record<string, string> = {
+  c: 'Ø',
+  d: '°',
+  p: '±',
+  '%': '%',
+};
+
+/**
+ * Character to paragraph alignment mapping
+ */
+const CHAR_TO_ALIGN: Record<string, MTextParagraphAlignment> = {
+  l: MTextParagraphAlignment.LEFT,
+  r: MTextParagraphAlignment.RIGHT,
+  c: MTextParagraphAlignment.CENTER,
+  j: MTextParagraphAlignment.JUSTIFIED,
+  d: MTextParagraphAlignment.DISTRIBUTED,
+};
+
+/**
+ * Convert RGB tuple to integer color value
+ * @param rgb - RGB color tuple
+ * @returns Integer color value
+ */
+export function rgb2int(rgb: RGB): number {
+  const [r, g, b] = rgb;
+  return (r << 16) | (g << 8) | b;
+}
+
+/**
+ * Convert integer color value to RGB tuple
+ * @param value - Integer color value
+ * @returns RGB color tuple
+ */
+export function int2rgb(value: number): RGB {
+  const r = (value >> 16) & 0xff;
+  const g = (value >> 8) & 0xff;
+  const b = value & 0xff;
+  return [r, g, b];
+}
+
+function clampColorChannel(value: number): number {
+  return Math.max(0, Math.min(255, Math.round(value)));
+}
+
+function normalizeColorNumber(color: number): number {
+  return Math.max(0, Math.min(0xffffff, Math.round(color)));
+}
+
+function colorNumberToHex(color: number | null): string | null {
+  if (color === null) return null;
+  return `#${normalizeColorNumber(color).toString(16).padStart(6, '0')}`;
+}
+
+function normalizeHexColor(value: string | null | undefined): string | null {
+  if (!value) return null;
+  const normalized = value.trim().toLowerCase();
+  if (/^#[0-9a-f]{6}$/.test(normalized)) return normalized;
+  if (/^[0-9a-f]{6}$/.test(normalized)) return `#${normalized}`;
+  if (/^#[0-9a-f]{3}$/.test(normalized)) {
+    const r = normalized[1];
+    const g = normalized[2];
+    const b = normalized[3];
+    return `#${r}${r}${g}${g}${b}${b}`;
+  }
+  if (/^[0-9a-f]{3}$/.test(normalized)) {
+    const r = normalized[0];
+    const g = normalized[1];
+    const b = normalized[2];
+    return `#${r}${r}${g}${g}${b}${b}`;
+  }
+  return null;
+}
+
+function cssColorToRgbValue(value: string | null | undefined): number | null {
+  if (!value) return null;
+  const raw = value.trim().toLowerCase();
+  if (raw === 'transparent') return null;
+
+  const hex = normalizeHexColor(raw);
+  if (hex) {
+    return normalizeColorNumber(Number.parseInt(hex.slice(1), 16));
+  }
+
+  const fnMatch = raw.match(/^rgba?\((.*)\)$/);
+  if (!fnMatch) return null;
+
+  const parts = fnMatch[1]
+    .replace(/\s*\/\s*/g, ' ')
+    .split(/[,\s]+/)
+    .map(p => p.trim())
+    .filter(Boolean);
+
+  if (parts.length < 3) return null;
+
+  const toChannel = (token: string): number => {
+    if (token.endsWith('%')) {
+      const percent = Number.parseFloat(token.slice(0, -1));
+      return clampColorChannel((percent / 100) * 255);
+    }
+    const num = Number.parseFloat(token);
+    return clampColorChannel(num);
+  };
+
+  const r = toChannel(parts[0]);
+  const g = toChannel(parts[1]);
+  const b = toChannel(parts[2]);
+  return rgb2int([r, g, b]);
+}
+
+/**
+ * Escape DXF line endings
+ * @param text - Text to escape
+ * @returns Escaped text
+ */
+export function escapeDxfLineEndings(text: string): string {
+  return text.replace(/\r\n|\r|\n/g, '\\P');
+}
+
+/**
+ * Check if text contains inline formatting codes
+ * @param text - Text to check
+ * @returns True if text contains formatting codes
+ */
+export function hasInlineFormattingCodes(text: string): boolean {
+  return text.replace(/\\P/g, '').replace(/\\~/g, '').includes('\\');
+}
+
+/**
+ * Extracts all unique font names used in an MText string.
+ * This function searches for font commands in the format \f{fontname}| or \f{fontname}; and returns a set of unique font names.
+ * Font names are converted to lowercase to ensure case-insensitive uniqueness.
+ *
+ * @param mtext - The MText string to analyze for font names
+ * @param removeExtension - Whether to remove font file extensions (e.g., .ttf, .shx) from font names. Defaults to false.
+ * @returns A Set containing all unique font names found in the MText string, converted to lowercase
+ * @example
+ * ```ts
+ * const mtext = "\\fArial.ttf|Hello\\fTimes New Roman.otf|World";
+ * const fonts = getFonts(mtext, true);
+ * // Returns: Set(2) { "arial", "times new roman" }
+ * ```
+ */
+export function getFonts(mtext: string, removeExtension: boolean = false) {
+  const fonts: Set<string> = new Set();
+  const regex = /\\[fF](.*?)[;|]/g;
+
+  [...mtext.matchAll(regex)].forEach(match => {
+    let fontName = match[1].toLowerCase();
+    if (removeExtension) {
+      fontName = fontName.replace(/\.(ttf|otf|woff|shx)$/, '');
+    }
+    fonts.add(fontName);
+  });
+
+  return fonts;
+}
+
+/**
+ * ContextStack manages a stack of MTextContext objects for character-level formatting.
+ *
+ * - Character-level formatting (underline, color, font, etc.) is scoped to `{}` blocks and managed by the stack.
+ * - Paragraph-level formatting (\p) is not scoped, but when a block ends, any paragraph property changes are merged into the parent context.
+ * - On pop, paragraph properties from the popped context are always merged into the new top context.
+ */
+class ContextStack {
+  private stack: MTextContext[] = [];
+
+  /**
+   * Creates a new ContextStack with an initial context.
+   * @param initial The initial MTextContext to use as the base of the stack.
+   */
+  constructor(initial: MTextContext) {
+    this.stack.push(initial);
+  }
+
+  /**
+   * Pushes a copy of the given context onto the stack.
+   * @param ctx The MTextContext to push (copied).
+   */
+  push(ctx: MTextContext) {
+    this.stack.push(ctx);
+  }
+
+  /**
+   * Pops the top context from the stack and merges its paragraph properties into the new top context.
+   * If only one context remains, nothing is popped.
+   * @returns The popped MTextContext, or undefined if the stack has only one context.
+   */
+  pop(): MTextContext | undefined {
+    if (this.stack.length <= 1) return undefined;
+    const popped = this.stack.pop()!;
+    // Merge paragraph properties into the new top context
+    const top = this.stack[this.stack.length - 1];
+    if (JSON.stringify(top.paragraph) !== JSON.stringify(popped.paragraph)) {
+      top.paragraph = { ...popped.paragraph };
+    }
+    return popped;
+  }
+
+  /**
+   * Returns the current (top) context on the stack.
+   */
+  get current(): MTextContext {
+    return this.stack[this.stack.length - 1];
+  }
+
+  /**
+   * Returns the current stack depth (number of nested blocks), not counting the root context.
+   */
+  get depth(): number {
+    return this.stack.length - 1;
+  }
+
+  /**
+   * Returns the root (bottom) context, which represents the global formatting state.
+   * Used for paragraph property application.
+   */
+  get root(): MTextContext {
+    return this.stack[0];
+  }
+
+  /**
+   * Replaces the current (top) context with the given context.
+   * @param ctx The new context to set as the current context.
+   */
+  setCurrent(ctx: MTextContext) {
+    this.stack[this.stack.length - 1] = ctx;
+  }
+}
+
+/**
+ * Configuration options for the MText parser.
+ * These options control how the parser behaves during tokenization and property handling.
+ */
+export interface MTextParserOptions {
+  /**
+   * Whether to yield PROPERTIES_CHANGED tokens when formatting properties change.
+   * When true, the parser will emit tokens whenever properties like color, font, or alignment change.
+   * When false, property changes are applied silently to the context without generating tokens.
+   * @default false
+   */
+  yieldPropertyCommands?: boolean;
+  /**
+   * Whether to reset paragraph parameters when encountering a new paragraph token.
+   * When true, paragraph properties (indent, margins, alignment, tab stops) are reset to defaults
+   * at the start of each new paragraph.
+   * @default false
+   */
+  resetParagraphParameters?: boolean;
+  /**
+   * Whether to emit {@link TokenType.PERCENT_SYMBOL} tokens for AutoCAD `%%` symbol
+   * codes instead of expanding them into {@link TokenType.WORD} characters.
+   * @default false
+   */
+  yieldPercentSymbols?: boolean;
+  /**
+   * Custom decoder function for MIF (Multibyte Interchange Format) codes.
+   * If provided, this function will be used instead of the default decodeMultiByteChar.
+   * The function receives the hex code string and should return the decoded character.
+   * @param hex - Hex code string (e.g., "C4E3" or "1A2B3")
+   * @returns Decoded character or empty square (▯) if invalid
+   * @default undefined (uses default decoder)
+   */
+  mifDecoder?: (hex: string) => string;
+  /**
+   * The length of MIF hex codes to parse. MIF codes in AutoCAD can vary in length
+   * depending on the specific SHX big font used (typically 4 or 5 digits).
+   * If not specified, the parser will try to auto-detect the length by attempting
+   * to match 4 digits first, then 5 digits if needed.
+   * @default undefined (auto-detect)
+   */
+  mifCodeLength?: 4 | 5 | 'auto';
+}
+
+/**
+ * Main parser class for MText content
+ */
+export class MTextParser {
+  private scanner: TextScanner;
+  private ctxStack: ContextStack;
+  private continueStroke: boolean = false;
+  private yieldPropertyCommands: boolean;
+  private resetParagraphParameters: boolean;
+  private yieldPercentSymbols: boolean;
+  private inStackContext: boolean = false;
+  private mifDecoder: (hex: string) => string;
+  private mifCodeLength: 4 | 5 | 'auto';
+
+  /**
+   * Creates a new MTextParser instance
+   * @param content - The MText content to parse
+   * @param ctx - Optional initial MText context
+   * @param options - Parser options
+   */
+  constructor(content: string, ctx?: MTextContext, options: MTextParserOptions = {}) {
+    this.scanner = new TextScanner(content);
+    const initialCtx = ctx ?? new MTextContext();
+    this.ctxStack = new ContextStack(initialCtx);
+    this.yieldPropertyCommands = options.yieldPropertyCommands ?? false;
+    this.resetParagraphParameters = options.resetParagraphParameters ?? false;
+    this.yieldPercentSymbols = options.yieldPercentSymbols ?? false;
+    this.mifDecoder = options.mifDecoder ?? this.decodeMultiByteChar.bind(this);
+    this.mifCodeLength = options.mifCodeLength ?? 'auto';
+  }
+
+  /**
+   * Decode multi-byte character from hex code
+   * @param hex - Hex code string (e.g. "C4E3" or "1A2B3")
+   * @returns Decoded character or empty square if invalid
+   */
+  private decodeMultiByteChar(hex: string): string {
+    try {
+      // For 5-digit codes, return placeholder directly
+      if (hex.length === 5) {
+        const prefix = hex[0];
+
+        // Notes:
+        // I know AutoCAD uses prefix 1 for Shift-JIS, 2 for big5, and 5 for gbk.
+        // But I don't know whether there are other prefixes and their meanings.
+        let encoding = 'gbk';
+        if (prefix === '1') {
+          encoding = 'shift-jis';
+        } else if (prefix === '2') {
+          encoding = 'big5';
+        }
+        const bytes = new Uint8Array([
+          parseInt(hex.substr(1, 2), 16),
+          parseInt(hex.substr(3, 2), 16),
+        ]);
+        const decoder = new TextDecoder(encoding);
+        const result = decoder.decode(bytes);
+        return result;
+      } else if (hex.length === 4) {
+        // For 4-digit hex codes, decode as 2-byte character
+        const bytes = new Uint8Array([
+          parseInt(hex.substr(0, 2), 16),
+          parseInt(hex.substr(2, 2), 16),
+        ]);
+
+        // Try GBK first
+        const gbkDecoder = new TextDecoder('gbk');
+        const gbkResult = gbkDecoder.decode(bytes);
+        if (gbkResult !== '▯') {
+          return gbkResult;
+        }
+
+        // Try BIG5 if GBK fails
+        const big5Decoder = new TextDecoder('big5');
+        const big5Result = big5Decoder.decode(bytes);
+        if (big5Result !== '▯') {
+          return big5Result;
+        }
+      }
+
+      return '▯';
+    } catch {
+      return '▯';
+    }
+  }
+
+  /**
+   * Extract MIF hex code from scanner
+   * @param length - The length of the hex code to extract (4 or 5), or 'auto' to detect
+   * @returns The extracted hex code, or null if not found
+   */
+  private extractMifCode(length: 4 | 5 | 'auto'): string | null {
+    if (length === 'auto') {
+      // Try 5 digits first if available, then fall back to 4 digits
+      const code5 = this.scanner.tail.match(/^[0-9A-Fa-f]{5}/)?.[0];
+      if (code5) {
+        return code5;
+      }
+      const code4 = this.scanner.tail.match(/^[0-9A-Fa-f]{4}/)?.[0];
+      if (code4) {
+        return code4;
+      }
+      return null;
+    } else {
+      const code = this.scanner.tail.match(new RegExp(`^[0-9A-Fa-f]{${length}}`))?.[0];
+      return code ?? null;
+    }
+  }
+
+  /**
+   * Parse an AutoCAD `\U+nnnn` escape after the leading `U` has been consumed.
+   *
+   * @remarks
+   * Exactly four hexadecimal digits are read. A high surrogate followed by
+   * another `\U+nnnn` low surrogate is combined into one supplementary-plane
+   * character (AutoCAD's encoding for code points above U+FFFF).
+   *
+   * @returns The decoded character, or `null` if the escape is incomplete
+   */
+  private parseUnicodeEscape(): string | null {
+    if (this.scanner.peek() !== '+') {
+      return null;
+    }
+
+    this.scanner.consume(1); // Consume the '+'
+    const hexMatch = this.scanner.tail.match(/^[0-9A-Fa-f]{4}/);
+    if (!hexMatch) {
+      this.scanner.consume(-1); // Rewind '+'
+      return null;
+    }
+
+    const codeUnit = parseInt(hexMatch[0], 16);
+    this.scanner.consume(hexMatch[0].length);
+
+    // High surrogate: combine with a following \U+ low-surrogate escape
+    if (codeUnit >= 0xd800 && codeUnit <= 0xdbff) {
+      const pairMatch = this.scanner.tail.match(/^\\U\+([0-9A-Fa-f]{4})/);
+      if (pairMatch) {
+        const lowUnit = parseInt(pairMatch[1], 16);
+        if (lowUnit >= 0xdc00 && lowUnit <= 0xdfff) {
+          this.scanner.consume(pairMatch[0].length);
+          const codePoint = ((codeUnit - 0xd800) << 10) + (lowUnit - 0xdc00) + 0x10000;
+          try {
+            return String.fromCodePoint(codePoint);
+          } catch {
+            return '▯';
+          }
+        }
+      }
+    }
+
+    // BMP code unit, or an unpaired surrogate kept as a UTF-16 unit
+    return String.fromCharCode(codeUnit);
+  }
+
+  /**
+   * Push current context onto the stack
+   */
+  private pushCtx(): void {
+    this.ctxStack.push(this.ctxStack.current);
+  }
+
+  /**
+   * Pop context from the stack
+   */
+  private popCtx(): void {
+    this.ctxStack.pop();
+  }
+
+  /**
+   * Parse stacking expression (numerator/denominator)
+   * @returns Tuple of [TokenType.STACK, [numerator, denominator, type]]
+   */
+  private parseStacking(): [TokenType, [string, string, string]] {
+    const scanner = new TextScanner(this.extractExpression(true));
+    let numerator = '';
+    let denominator = '';
+    let stackingType = '';
+
+    const getNextChar = (): [string, boolean] => {
+      let c = scanner.peek();
+      let escape = false;
+      if (c.charCodeAt(0) < 32) {
+        c = ' ';
+      }
+      if (c === '\\') {
+        escape = true;
+        scanner.consume(1);
+        c = scanner.peek();
+      }
+      scanner.consume(1);
+      return [c, escape];
+    };
+
+    const parseNumerator = (): [string, string] => {
+      let word = '';
+      while (scanner.hasData) {
+        const [c, escape] = getNextChar();
+        // Check for stacking operators first
+        if (!escape && (c === '/' || c === '#' || c === '^')) {
+          return [word, c];
+        }
+        word += c;
+      }
+      return [word, ''];
+    };
+
+    const parseDenominator = (skipLeadingSpace: boolean): string => {
+      let word = '';
+      let skipping = skipLeadingSpace;
+      while (scanner.hasData) {
+        const [c, escape] = getNextChar();
+        if (skipping && c === ' ') {
+          continue;
+        }
+        skipping = false;
+        // Stop at terminator unless escaped
+        if (!escape && c === ';') {
+          break;
+        }
+        word += c;
+      }
+      return word;
+    };
+
+    [numerator, stackingType] = parseNumerator();
+    if (stackingType) {
+      // Only skip leading space for caret divider
+      denominator = parseDenominator(stackingType === '^');
+    }
+
+    // Special case for \S^!/^?;
+    if (numerator === '' && denominator.includes('I/')) {
+      return [TokenType.STACK, [' ', ' ', '/']];
+    }
+
+    // Handle caret as a stacking operator
+    if (stackingType === '^') {
+      return [TokenType.STACK, [numerator, denominator, '^']];
+    }
+
+    return [TokenType.STACK, [numerator, denominator, stackingType]];
+  }
+
+  /**
+   * Parse MText properties
+   * @param cmd - The property command to parse
+   * @returns Property changes if yieldPropertyCommands is true and changes occurred
+   */
+  private parseProperties(cmd: string): TokenData[TokenType.PROPERTIES_CHANGED] | void {
+    const prevCtx = this.ctxStack.current.copy();
+    const newCtx = this.ctxStack.current.copy();
+    switch (cmd) {
+      case 'L':
+        newCtx.underline = true;
+        this.continueStroke = true;
+        break;
+      case 'l':
+        newCtx.underline = false;
+        if (!newCtx.hasAnyStroke) {
+          this.continueStroke = false;
+        }
+        break;
+      case 'O':
+        newCtx.overline = true;
+        this.continueStroke = true;
+        break;
+      case 'o':
+        newCtx.overline = false;
+        if (!newCtx.hasAnyStroke) {
+          this.continueStroke = false;
+        }
+        break;
+      case 'K':
+        newCtx.strikeThrough = true;
+        this.continueStroke = true;
+        break;
+      case 'k':
+        newCtx.strikeThrough = false;
+        if (!newCtx.hasAnyStroke) {
+          this.continueStroke = false;
+        }
+        break;
+      case 'A':
+        this.parseAlign(newCtx);
+        break;
+      case 'C':
+        this.parseAciColor(newCtx);
+        break;
+      case 'c':
+        this.parseRgbColor(newCtx);
+        break;
+      case 'H':
+        this.parseHeight(newCtx);
+        break;
+      case 'W':
+        this.parseWidth(newCtx);
+        break;
+      case 'Q':
+        this.parseOblique(newCtx);
+        break;
+      case 'T':
+        this.parseCharTracking(newCtx);
+        break;
+      case 'p':
+        this.parseParagraphProperties(newCtx);
+        break;
+      case 'f':
+      case 'F':
+        this.parseFontProperties(newCtx);
+        break;
+      default:
+        throw new Error(`Unknown command: ${cmd}`);
+    }
+
+    // Update continueStroke based on current stroke state
+    this.continueStroke = newCtx.hasAnyStroke;
+    newCtx.continueStroke = this.continueStroke;
+    // Use setCurrent to replace the current context
+    this.ctxStack.setCurrent(newCtx);
+
+    if (this.yieldPropertyCommands) {
+      const changes = this.getPropertyChanges(prevCtx, newCtx);
+      if (Object.keys(changes).length > 0) {
+        return {
+          command: cmd,
+          changes,
+          depth: this.ctxStack.depth,
+        };
+      }
+    }
+  }
+
+  /**
+   * Get property changes between two contexts
+   * @param oldCtx - The old context
+   * @param newCtx - The new context
+   * @returns Object containing changed properties
+   */
+  private getPropertyChanges(
+    oldCtx: MTextContext,
+    newCtx: MTextContext
+  ): TokenData[TokenType.PROPERTIES_CHANGED]['changes'] {
+    const changes: TokenData[TokenType.PROPERTIES_CHANGED]['changes'] = {};
+
+    if (oldCtx.underline !== newCtx.underline) {
+      changes.underline = newCtx.underline;
+    }
+    if (oldCtx.overline !== newCtx.overline) {
+      changes.overline = newCtx.overline;
+    }
+    if (oldCtx.strikeThrough !== newCtx.strikeThrough) {
+      changes.strikeThrough = newCtx.strikeThrough;
+    }
+    if (oldCtx.color.aci !== newCtx.color.aci) {
+      changes.aci = newCtx.color.aci;
+    }
+    if (oldCtx.color.rgbValue !== newCtx.color.rgbValue) {
+      changes.rgb = newCtx.color.rgb;
+    }
+    if (oldCtx.align !== newCtx.align) {
+      changes.align = newCtx.align;
+    }
+    if (JSON.stringify(oldCtx.fontFace) !== JSON.stringify(newCtx.fontFace)) {
+      changes.fontFace = newCtx.fontFace;
+    }
+    if (
+      oldCtx.capHeight.value !== newCtx.capHeight.value ||
+      oldCtx.capHeight.isRelative !== newCtx.capHeight.isRelative
+    ) {
+      changes.capHeight = newCtx.capHeight;
+    }
+    if (
+      oldCtx.widthFactor.value !== newCtx.widthFactor.value ||
+      oldCtx.widthFactor.isRelative !== newCtx.widthFactor.isRelative
+    ) {
+      changes.widthFactor = newCtx.widthFactor;
+    }
+    if (
+      oldCtx.charTrackingFactor.value !== newCtx.charTrackingFactor.value ||
+      oldCtx.charTrackingFactor.isRelative !== newCtx.charTrackingFactor.isRelative
+    ) {
+      changes.charTrackingFactor = newCtx.charTrackingFactor;
+    }
+    if (oldCtx.oblique !== newCtx.oblique) {
+      changes.oblique = newCtx.oblique;
+    }
+    if (JSON.stringify(oldCtx.paragraph) !== JSON.stringify(newCtx.paragraph)) {
+      // Only include changed paragraph properties
+      const changedProps: Partial<ParagraphProperties> = {};
+      if (oldCtx.paragraph.indent !== newCtx.paragraph.indent) {
+        changedProps.indent = newCtx.paragraph.indent;
+      }
+      if (oldCtx.paragraph.align !== newCtx.paragraph.align) {
+        changedProps.align = newCtx.paragraph.align;
+      }
+      if (oldCtx.paragraph.left !== newCtx.paragraph.left) {
+        changedProps.left = newCtx.paragraph.left;
+      }
+      if (oldCtx.paragraph.right !== newCtx.paragraph.right) {
+        changedProps.right = newCtx.paragraph.right;
+      }
+      if (JSON.stringify(oldCtx.paragraph.tabs) !== JSON.stringify(newCtx.paragraph.tabs)) {
+        changedProps.tabs = newCtx.paragraph.tabs;
+      }
+      if (Object.keys(changedProps).length > 0) {
+        changes.paragraph = changedProps;
+      }
+    }
+
+    return changes;
+  }
+
+  /**
+   * Parse alignment property
+   * @param ctx - The context to update
+   */
+  private parseAlign(ctx: MTextContext): void {
+    const char = this.scanner.get();
+    if ('012'.includes(char)) {
+      ctx.align = parseInt(char) as MTextLineAlignment;
+    } else {
+      ctx.align = MTextLineAlignment.BOTTOM;
+    }
+    this.consumeOptionalTerminator();
+  }
+
+  /**
+   * Parse height property
+   * @param ctx - The context to update
+   */
+  private parseHeight(ctx: MTextContext): void {
+    const expr = this.extractFloatExpression(true);
+    if (expr) {
+      try {
+        if (expr.endsWith('x')) {
+          // For height command, treat x suffix as relative value
+          ctx.capHeight = {
+            value: parseFloat(expr.slice(0, -1)),
+            isRelative: true,
+          };
+        } else {
+          ctx.capHeight = {
+            value: parseFloat(expr),
+            isRelative: false,
+          };
+        }
+      } catch {
+        // If parsing fails, treat the entire command as literal text
+        this.scanner.consume(-expr.length); // Rewind to before the expression
+        return;
+      }
+    }
+    this.consumeOptionalTerminator();
+  }
+
+  /**
+   * Parse width property
+   * @param ctx - The context to update
+   */
+  private parseWidth(ctx: MTextContext): void {
+    const expr = this.extractFloatExpression(true);
+    if (expr) {
+      try {
+        if (expr.endsWith('x')) {
+          // For width command, treat x suffix as relative value
+          ctx.widthFactor = {
+            value: parseFloat(expr.slice(0, -1)),
+            isRelative: true,
+          };
+        } else {
+          ctx.widthFactor = {
+            value: parseFloat(expr),
+            isRelative: false,
+          };
+        }
+      } catch {
+        // If parsing fails, treat the entire command as literal text
+        this.scanner.consume(-expr.length); // Rewind to before the expression
+        return;
+      }
+    }
+    this.consumeOptionalTerminator();
+  }
+
+  /**
+   * Parse character tracking property
+   * @param ctx - The context to update
+   */
+  private parseCharTracking(ctx: MTextContext): void {
+    const expr = this.extractFloatExpression(true);
+    if (expr) {
+      try {
+        if (expr.endsWith('x')) {
+          // For tracking command, treat x suffix as relative value
+          ctx.charTrackingFactor = {
+            value: Math.abs(parseFloat(expr.slice(0, -1))),
+            isRelative: true,
+          };
+        } else {
+          ctx.charTrackingFactor = {
+            value: Math.abs(parseFloat(expr)),
+            isRelative: false,
+          };
+        }
+      } catch {
+        // If parsing fails, treat the entire command as literal text
+        this.scanner.consume(-expr.length); // Rewind to before the expression
+        return;
+      }
+    }
+    this.consumeOptionalTerminator();
+  }
+
+  /**
+   * Parse float value or factor
+   * @param value - Current value to apply factor to
+   * @returns New value
+   */
+  private parseFloatValueOrFactor(value: number): number {
+    const expr = this.extractFloatExpression(true);
+    if (expr) {
+      if (expr.endsWith('x')) {
+        const factor = parseFloat(expr.slice(0, -1));
+        value *= factor; // Allow negative factors
+      } else {
+        value = parseFloat(expr); // Allow negative values
+      }
+    }
+    return value;
+  }
+
+  /**
+   * Parse oblique angle property
+   * @param ctx - The context to update
+   */
+  private parseOblique(ctx: MTextContext): void {
+    const obliqueExpr = this.extractFloatExpression(false);
+    if (obliqueExpr) {
+      ctx.oblique = parseFloat(obliqueExpr);
+    }
+    this.consumeOptionalTerminator();
+  }
+
+  /**
+   * Parse ACI color property
+   * @param ctx - The context to update
+   */
+  private parseAciColor(ctx: MTextContext): void {
+    const aciExpr = this.extractIntExpression();
+    if (aciExpr) {
+      const aci = parseInt(aciExpr);
+      if (aci < 257) {
+        ctx.color.aci = aci;
+      }
+    }
+    this.consumeOptionalTerminator();
+  }
+
+  /**
+   * Parse RGB color property
+   * @param ctx - The context to update
+   */
+  private parseRgbColor(ctx: MTextContext): void {
+    const rgbExpr = this.extractIntExpression();
+    if (rgbExpr) {
+      const value = parseInt(rgbExpr) & 0xffffff;
+      ctx.color.rgbValue = value;
+    }
+    this.consumeOptionalTerminator();
+  }
+
+  /**
+   * Extract float expression from scanner
+   * @param relative - Whether to allow relative values (ending in 'x')
+   * @returns Extracted expression
+   */
+  private extractFloatExpression(relative: boolean = false): string {
+    const pattern = relative
+      ? /^[+-]?(?:\d+(?:\.\d*)?|\.\d+)(?:[eE][+-]?\d+)?x?/
+      : /^[+-]?(?:\d+(?:\.\d*)?|\.\d+)(?:[eE][+-]?\d+)?/;
+    const match = this.scanner.tail.match(pattern);
+    if (match) {
+      const result = match[0];
+      this.scanner.consume(result.length);
+      return result;
+    }
+    return '';
+  }
+
+  /**
+   * Extract integer expression from scanner
+   * @returns Extracted expression
+   */
+  private extractIntExpression(): string {
+    const match = this.scanner.tail.match(/^\d+/);
+    if (match) {
+      const result = match[0];
+      this.scanner.consume(result.length);
+      return result;
+    }
+    return '';
+  }
+
+  /**
+   * Extract expression until semicolon or end
+   * @param escape - Whether to handle escaped semicolons
+   * @returns Extracted expression
+   */
+  private extractExpression(escape: boolean = false): string {
+    const stop = this.scanner.find(';', escape);
+    if (stop < 0) {
+      const expr = this.scanner.tail;
+      this.scanner.consume(expr.length);
+      return expr;
+    }
+    // Check if the semicolon is escaped by looking at the previous character
+    const prevChar = this.scanner.peek(stop - this.scanner.currentIndex - 1);
+    const isEscaped = prevChar === '\\';
+    const expr = this.scanner.tail.slice(0, stop - this.scanner.currentIndex + (isEscaped ? 1 : 0));
+    this.scanner.consume(expr.length + 1);
+    return expr;
+  }
+
+  /**
+   * Parse font properties
+   * @param ctx - The context to update
+   */
+  private parseFontProperties(ctx: MTextContext): void {
+    const parts = this.extractExpression().split('|');
+    if (parts.length > 0 && parts[0]) {
+      const name = parts[0];
+      let style: FontStyle = 'Regular';
+      let weight = 400;
+
+      for (const part of parts.slice(1)) {
+        if (part.startsWith('b1')) {
+          weight = 700;
+        } else if (part === 'i' || part.startsWith('i1')) {
+          style = 'Italic';
+        } else if (part === 'i0' || part.startsWith('i0')) {
+          style = 'Regular';
+        }
+      }
+
+      ctx.fontFace = {
+        family: name,
+        style,
+        weight,
+      };
+    }
+  }
+
+  /**
+   * Parse paragraph properties from the MText content
+   * Handles properties like indentation, alignment, and tab stops
+   * @param ctx - The context to update
+   */
+  private parseParagraphProperties(ctx: MTextContext): void {
+    const scanner = new TextScanner(this.extractExpression());
+    /** Current indentation value */
+    let indent = ctx.paragraph.indent;
+    /** Left margin value */
+    let left = ctx.paragraph.left;
+    /** Right margin value */
+    let right = ctx.paragraph.right;
+    /** Current paragraph alignment */
+    let align = ctx.paragraph.align;
+    /** Array of tab stop positions and types */
+    let tabStops: (number | string)[] = [];
+
+    /**
+     * Parse a floating point number from the scanner's current position.
+     * Accepts an optional sign, a leading decimal point, and scientific notation,
+     * matching {@link extractFloatExpression}.
+     * On a match, consumes the number and any immediately following commas.
+     * @returns The parsed value, or 0 when the scanner is not on a number.
+     *          A miss consumes nothing, so looping callers must advance themselves.
+     */
+    const parseFloatValue = (): number => {
+      const match = scanner.tail.match(/^[+-]?(?:\d+(?:\.\d*)?|\.\d+)(?:[eE][+-]?\d+)?/);
+      if (match) {
+        const value = parseFloat(match[0]);
+        scanner.consume(match[0].length);
+        while (scanner.peek() === ',') {
+          scanner.consume(1);
+        }
+        return value;
+      }
+      return 0;
+    };
+
+    while (scanner.hasData) {
+      const cmd = scanner.get();
+      switch (cmd) {
+        case 'i': // Indentation
+          indent = parseFloatValue();
+          break;
+        case 'l': // Left margin
+          left = parseFloatValue();
+          break;
+        case 'r': // Right margin
+          right = parseFloatValue();
+          break;
+        case 'x': // Skip
+          break;
+        case 'q': {
+          // Alignment
+          const adjustment = scanner.get();
+          align = CHAR_TO_ALIGN[adjustment] || MTextParagraphAlignment.DEFAULT;
+          while (scanner.peek() === ',') {
+            scanner.consume(1);
+          }
+          break;
+        }
+        case 't': // Tab stops
+          tabStops = [];
+          while (scanner.hasData) {
+            const indexBefore = scanner.currentIndex;
+            const type = scanner.peek();
+            if (type === 'r' || type === 'c') {
+              scanner.consume(1);
+              const valueIndex = scanner.currentIndex;
+              const value = parseFloatValue();
+              // 0 is also the miss result, so record a right/center stop only
+              // when a number was consumed. Otherwise `\ptrz` becomes a fake
+              // right tab at 0.
+              if (scanner.currentIndex !== valueIndex) {
+                tabStops.push(type + value.toString());
+              }
+            } else {
+              const value = parseFloatValue();
+              if (scanner.currentIndex !== indexBefore) {
+                tabStops.push(value);
+              }
+            }
+            // AutoCAD clears tab stops with `\pi0,l0,tz;`. `z` matches no
+            // number and leaves the scanner in place; consume it or this
+            // loop pushes zeros until it runs out of memory.
+            if (scanner.currentIndex === indexBefore) {
+              scanner.consume(1);
+            }
+          }
+          break;
+      }
+    }
+
+    ctx.paragraph = {
+      indent,
+      left,
+      right,
+      align,
+      tabs: tabStops,
+    };
+  }
+
+  /**
+   * Consume optional terminator (semicolon)
+   */
+  private consumeOptionalTerminator(): void {
+    if (this.scanner.peek() === ';') {
+      this.scanner.consume(1);
+    }
+  }
+
+  /**
+   * Builds {@link PercentSymbolData} for a recognized `%%` code letter.
+   */
+  private buildPercentSymbolData(code: string, specialChar: string): PercentSymbolData | null {
+    if (code === 'c' || code === 'd' || code === 'p') {
+      return { kind: 'named', code, char: specialChar };
+    }
+    if (code === '%') {
+      return { kind: 'literal', char: '%' };
+    }
+    return null;
+  }
+
+  /**
+   * Parse MText content into tokens
+   * @yields MTextToken objects
+   */
+  *parse(): Generator<MTextToken> {
+    const wordToken = TokenType.WORD;
+    const spaceToken = TokenType.SPACE;
+    let followupToken: TokenType | null = null;
+    let followupData: TokenData[TokenType] | undefined;
+
+    function resetParagraph(ctx: MTextContext): Partial<ParagraphProperties> {
+      const prev = { ...ctx.paragraph };
+      ctx.paragraph = {
+        indent: 0,
+        left: 0,
+        right: 0,
+        align: MTextParagraphAlignment.DEFAULT,
+        tabs: [],
+      };
+      const changed: Partial<ParagraphProperties> = {};
+      if (prev.indent !== 0) changed.indent = 0;
+      if (prev.left !== 0) changed.left = 0;
+      if (prev.right !== 0) changed.right = 0;
+      if (prev.align !== MTextParagraphAlignment.DEFAULT)
+        changed.align = MTextParagraphAlignment.DEFAULT;
+      if (JSON.stringify(prev.tabs) !== JSON.stringify([])) changed.tabs = [];
+      return changed;
+    }
+
+    const nextToken = (): [TokenType, TokenData[TokenType]] => {
+      let word = '';
+      while (this.scanner.hasData) {
+        let escape = false;
+        let letter = this.scanner.peek();
+        const cmdStartIndex = this.scanner.currentIndex;
+
+        // Handle control characters first. Flush any pending word before emitting
+        // paragraph/tab tokens — the same pattern used for spaces and `\P` — so
+        // LibreDWG/DXF converters that expand `\P` to raw `\n` do not drop text.
+        if (letter.charCodeAt(0) < 32) {
+          if (letter === '\t') {
+            this.scanner.consume(1);
+            if (word) {
+              followupToken = TokenType.TABULATOR;
+              followupData = null;
+              return [wordToken, word];
+            }
+            return [TokenType.TABULATOR, null];
+          }
+          if (letter === '\n') {
+            this.scanner.consume(1);
+            if (word) {
+              followupToken = TokenType.NEW_PARAGRAPH;
+              followupData = null;
+              return [wordToken, word];
+            }
+            return [TokenType.NEW_PARAGRAPH, null];
+          }
+          // Other C0 controls behave like a space delimiter.
+          this.scanner.consume(1);
+          if (word) {
+            followupToken = spaceToken;
+            return [wordToken, word];
+          }
+          return [spaceToken, null];
+        }
+
+        if (letter === '\\') {
+          if ('\\{}'.includes(this.scanner.peek(1))) {
+            escape = true;
+            this.scanner.consume(1);
+            letter = this.scanner.peek();
+          } else {
+            if (word) {
+              return [wordToken, word];
+            }
+            this.scanner.consume(1);
+            const cmd = this.scanner.get();
+            switch (cmd) {
+              case '~':
+                return [TokenType.NBSP, null];
+              case 'P':
+                return [TokenType.NEW_PARAGRAPH, null];
+              case 'N':
+                return [TokenType.NEW_COLUMN, null];
+              case 'X':
+                return [TokenType.WRAP_AT_DIMLINE, null];
+              case 'S': {
+                this.inStackContext = true;
+                const result = this.parseStacking();
+                this.inStackContext = false;
+                return result;
+              }
+              case 'm':
+              case 'M':
+                // Handle multi-byte character encoding (MIF)
+                if (this.scanner.peek() === '+') {
+                  this.scanner.consume(1); // Consume the '+'
+                  const hexCode = this.extractMifCode(this.mifCodeLength);
+                  if (hexCode) {
+                    this.scanner.consume(hexCode.length);
+                    const decodedChar = this.mifDecoder(hexCode);
+                    if (word) {
+                      return [wordToken, word];
+                    }
+                    return [wordToken, decodedChar];
+                  }
+                  // If no valid hex code found, rewind the '+' character
+                  this.scanner.consume(-1);
+                }
+                // If not a valid multi-byte code, treat as literal text
+                word += '\\M';
+                continue;
+              case 'U': {
+                // AutoCAD \U+nnnn uses exactly four hex digits. Characters
+                // outside the BMP are encoded as two consecutive \U+XXXX
+                // UTF-16 surrogate escapes.
+                const decodedChar = this.parseUnicodeEscape();
+                if (decodedChar !== null) {
+                  if (word) {
+                    return [wordToken, word];
+                  }
+                  return [wordToken, decodedChar];
+                }
+                // If not a valid Unicode code, treat as literal text
+                word += '\\U';
+                continue;
+              }
+              default:
+                if (cmd) {
+                  try {
+                    const propertyChanges = this.parseProperties(cmd);
+                    if (this.yieldPropertyCommands && propertyChanges) {
+                      return [TokenType.PROPERTIES_CHANGED, propertyChanges];
+                    }
+                    // After processing a property command, continue with normal parsing
+                    continue;
+                  } catch {
+                    const commandText = this.scanner.tail.slice(
+                      cmdStartIndex,
+                      this.scanner.currentIndex
+                    );
+                    word += commandText;
+                  }
+                }
+            }
+            continue;
+          }
+        }
+
+        if (letter === '%' && this.scanner.peek(1) === '%') {
+          const code = this.scanner.peek(2).toLowerCase();
+          const specialChar = SPECIAL_CHAR_ENCODING[code];
+          if (specialChar) {
+            this.scanner.consume(3);
+            if (this.yieldPercentSymbols) {
+              const symbolData = this.buildPercentSymbolData(code, specialChar);
+              if (symbolData) {
+                if (word) {
+                  followupToken = TokenType.PERCENT_SYMBOL;
+                  followupData = symbolData;
+                  return [wordToken, word];
+                }
+                return [TokenType.PERCENT_SYMBOL, symbolData];
+              }
+            }
+            word += specialChar;
+            continue;
+          } else {
+            /**
+             * Supports Control Codes: `%%nnn`, where nnn is 1–3 decimal digits
+             * for the character's Unicode/ASCII code point (e.g. `%%34` → `"`,
+             * `%%176` → `°`). Digit count is not fixed at three.
+             *
+             * Reference: https://help.autodesk.com/view/ACD/2026/ENU/?guid=GUID-968CBC1D-BA99-4519-ABDD-88419EB2BF92
+             */
+            const digitChars: string[] = [];
+            for (let i = 0; i < 3; i++) {
+              const d = this.scanner.peek(2 + i);
+              if (d >= '0' && d <= '9') {
+                digitChars.push(d);
+              } else {
+                break;
+              }
+            }
+
+            if (digitChars.length > 0) {
+              const charCode = Number.parseInt(digitChars.join(''), 10);
+              this.scanner.consume(2 + digitChars.length);
+              if (this.yieldPercentSymbols) {
+                const symbolData: PercentSymbolData = {
+                  kind: 'numeric',
+                  charCode,
+                  char: String.fromCharCode(charCode),
+                };
+                if (word) {
+                  followupToken = TokenType.PERCENT_SYMBOL;
+                  followupData = symbolData;
+                  return [wordToken, word];
+                }
+                return [TokenType.PERCENT_SYMBOL, symbolData];
+              }
+              word += String.fromCharCode(charCode);
+            } else {
+              // Skip invalid special character codes (`%%` + non-digit)
+              this.scanner.consume(3);
+            }
+
+            continue;
+          }
+        }
+
+        if (letter === ' ') {
+          if (word) {
+            this.scanner.consume(1);
+            followupToken = spaceToken;
+            return [wordToken, word];
+          }
+          this.scanner.consume(1);
+          return [spaceToken, null];
+        }
+
+        if (!escape) {
+          if (letter === '{') {
+            if (word) {
+              return [wordToken, word];
+            }
+            this.scanner.consume(1);
+            this.pushCtx();
+            continue;
+          } else if (letter === '}') {
+            if (word) {
+              return [wordToken, word];
+            }
+            this.scanner.consume(1);
+            // Context restoration with yieldPropertyCommands
+            if (this.yieldPropertyCommands) {
+              const prevCtx = this.ctxStack.current;
+              this.popCtx();
+              const changes = this.getPropertyChanges(prevCtx, this.ctxStack.current);
+              if (Object.keys(changes).length > 0) {
+                return [
+                  TokenType.PROPERTIES_CHANGED,
+                  { command: undefined, changes, depth: this.ctxStack.depth },
+                ];
+              }
+            } else {
+              this.popCtx();
+            }
+            continue;
+          }
+        }
+
+        // Handle caret-encoded characters only when not in stack context
+        if (!this.inStackContext && letter === '^') {
+          const nextChar = this.scanner.peek(1);
+          if (nextChar) {
+            const code = nextChar.charCodeAt(0);
+            this.scanner.consume(2); // Consume both ^ and the next character
+            if (code === 32) {
+              // Space
+              word += '^';
+            } else if (code === 73) {
+              // Tab
+              if (word) {
+                return [wordToken, word];
+              }
+              return [TokenType.TABULATOR, null];
+            } else if (code === 74) {
+              // Line feed
+              if (word) {
+                return [wordToken, word];
+              }
+              return [TokenType.NEW_PARAGRAPH, null];
+            } else if (code === 77) {
+              // Carriage return
+              // Ignore carriage return
+              continue;
+            } else {
+              word += '▯';
+            }
+            continue;
+          }
+        }
+
+        this.scanner.consume(1);
+        if (letter.charCodeAt(0) >= 32) {
+          word += letter;
+        }
+      }
+
+      if (word) {
+        return [wordToken, word];
+      }
+      return [TokenType.NONE, null];
+    };
+
+    const maybeResetAfterParagraph = (tokenType: TokenType) => {
+      if (tokenType !== TokenType.NEW_PARAGRAPH || !this.resetParagraphParameters) {
+        return;
+      }
+      // Reset paragraph properties and emit PROPERTIES_CHANGED if needed
+      const ctx = this.ctxStack.current;
+      const changed = resetParagraph(ctx);
+      if (this.yieldPropertyCommands && Object.keys(changed).length > 0) {
+        return new MTextToken(TokenType.PROPERTIES_CHANGED, ctx.copy(), {
+          command: undefined,
+          changes: { paragraph: changed },
+          depth: this.ctxStack.depth,
+        });
+      }
+      return;
+    };
+
+    while (true) {
+      const [type, data] = nextToken.call(this);
+      if (type) {
+        yield new MTextToken(type, this.ctxStack.current.copy(), data);
+        const resetToken = maybeResetAfterParagraph(type);
+        if (resetToken) {
+          yield resetToken;
+        }
+        if (followupToken) {
+          yield new MTextToken(followupToken, this.ctxStack.current.copy(), followupData ?? null);
+          const followupResetToken = maybeResetAfterParagraph(followupToken);
+          if (followupResetToken) {
+            yield followupResetToken;
+          }
+          followupToken = null;
+          followupData = undefined;
+        }
+      } else {
+        break;
+      }
+    }
+  }
+}
+
+/**
+ * Text scanner for parsing MText content
+ */
+export class TextScanner {
+  private text: string;
+  private textLen: number;
+  private _index: number;
+
+  /**
+   * Create a new text scanner
+   * @param text - The text to scan
+   */
+  constructor(text: string) {
+    this.text = text;
+    this.textLen = text.length;
+    this._index = 0;
+  }
+
+  /**
+   * Get the current index in the text
+   */
+  get currentIndex(): number {
+    return this._index;
+  }
+
+  /**
+   * Check if the scanner has reached the end of the text
+   */
+  get isEmpty(): boolean {
+    return this._index >= this.textLen;
+  }
+
+  /**
+   * Check if there is more text to scan
+   */
+  get hasData(): boolean {
+    return this._index < this.textLen;
+  }
+
+  /**
+   * Get the next character and advance the index
+   * @returns The next character, or empty string if at end
+   */
+  get(): string {
+    if (this.isEmpty) {
+      return '';
+    }
+    const char = this.text[this._index];
+    this._index++;
+    return char;
+  }
+
+  /**
+   * Advance the index by the specified count
+   * @param count - Number of characters to advance
+   */
+  consume(count: number = 1): void {
+    this._index = Math.max(0, Math.min(this._index + count, this.textLen));
+  }
+
+  /**
+   * Look at a character without advancing the index
+   * @param offset - Offset from current position
+   * @returns The character at the offset position, or empty string if out of bounds
+   */
+  peek(offset: number = 0): string {
+    const index = this._index + offset;
+    if (index >= this.textLen || index < 0) {
+      return '';
+    }
+    return this.text[index];
+  }
+
+  /**
+   * Find the next occurrence of a character
+   * @param char - The character to find
+   * @param escape - Whether to handle escaped characters
+   * @returns Index of the character, or -1 if not found
+   */
+  find(char: string, escape: boolean = false): number {
+    let index = this._index;
+    while (index < this.textLen) {
+      if (escape && this.text[index] === '\\') {
+        if (index + 1 < this.textLen) {
+          if (this.text[index + 1] === char) {
+            return index + 1;
+          }
+          index += 2;
+          continue;
+        }
+        index++;
+        continue;
+      }
+      if (this.text[index] === char) {
+        return index;
+      }
+      index++;
+    }
+    return -1;
+  }
+
+  /**
+   * Get the remaining text from the current position
+   */
+  get tail(): string {
+    return this.text.slice(this._index);
+  }
+
+  /**
+   * Check if the next character is a space
+   */
+  isNextSpace(): boolean {
+    return this.peek() === ' ';
+  }
+
+  /**
+   * Consume spaces until a non-space character is found
+   * @returns Number of spaces consumed
+   */
+  consumeSpaces(): number {
+    let count = 0;
+    while (this.isNextSpace()) {
+      this.consume();
+      count++;
+    }
+    return count;
+  }
+}
+
+/**
+ * Class to handle ACI and RGB color logic for MText.
+ *
+ * This class encapsulates color state for MText, supporting both AutoCAD Color Index (ACI) and RGB color.
+ * Only one color mode is active at a time: setting an RGB color disables ACI, and vice versa.
+ * RGB is stored as a single 24-bit integer (0xRRGGBB) for efficient comparison and serialization.
+ *
+ * Example usage:
+ * ```ts
+ * const color1 = new MTextColor(1); // ACI color
+ * const color2 = new MTextColor([255, 0, 0]); // RGB color
+ * const color3 = new MTextColor(); // Default (ACI=256, "by layer")
+ * ```
+ */
+export class MTextColor {
+  /**
+   * The AutoCAD Color Index (ACI) value. Only used if no RGB color is set.
+   * @default 256 ("by layer")
+   */
+  private _aci: number | null = 256;
+  /**
+   * The RGB color value as a single 24-bit integer (0xRRGGBB), or null if not set.
+   * @default null
+   */
+  private _rgbValue: number | null = null; // Store as 0xRRGGBB or null
+
+  /**
+   * Create a new MTextColor instance.
+   * @param color The initial color: number for ACI, [r,g,b] for RGB, or null/undefined for default (ACI=256).
+   */
+  constructor(color?: number | RGB | null) {
+    if (Array.isArray(color)) {
+      this.rgb = color;
+    } else if (typeof color === 'number') {
+      this.aci = color;
+    } else {
+      this.aci = 256;
+    }
+  }
+
+  /**
+   * Get the current ACI color value.
+   * @returns The ACI color (0-256), or null if using RGB.
+   */
+  get aci(): number | null {
+    return this._aci;
+  }
+
+  /**
+   * Set the ACI color value. Setting this disables any RGB color.
+   * @param value The ACI color (0-256), or null to unset.
+   * @throws Error if value is out of range.
+   */
+  set aci(value: number | null) {
+    if (value === null) {
+      this._aci = null;
+    } else if (value >= 0 && value <= 256) {
+      this._aci = value;
+      this._rgbValue = null;
+    } else {
+      throw new Error('ACI not in range [0, 256]');
+    }
+  }
+
+  /**
+   * Get the current RGB color as a tuple [r, g, b], or null if not set.
+   * @returns The RGB color tuple, or null if using ACI.
+   */
+  get rgb(): RGB | null {
+    if (this._rgbValue === null) return null;
+    // Extract R, G, B from 0xRRGGBB
+    const r = (this._rgbValue >> 16) & 0xff;
+    const g = (this._rgbValue >> 8) & 0xff;
+    const b = this._rgbValue & 0xff;
+    return [r, g, b];
+  }
+
+  /**
+   * Set the RGB color. Setting this disables ACI color.
+   * @param value The RGB color tuple [r, g, b], or null to use ACI.
+   */
+  set rgb(value: RGB | null) {
+    if (value) {
+      const [r, g, b] = value;
+      this._rgbValue = ((r & 0xff) << 16) | ((g & 0xff) << 8) | (b & 0xff);
+      this._aci = null;
+    } else {
+      this._rgbValue = null;
+    }
+  }
+
+  /**
+   * Returns true if the color is set by RGB, false if by ACI.
+   */
+  get isRgb(): boolean {
+    return this._rgbValue !== null;
+  }
+
+  /**
+   * Returns true if the color is set by ACI, false if by RGB.
+   */
+  get isAci(): boolean {
+    return this._rgbValue === null && this._aci !== null;
+  }
+
+  /**
+   * Get or set the internal RGB value as a number (0xRRGGBB), or null if not set.
+   * Setting this will switch to RGB mode and set ACI to null.
+   */
+  get rgbValue(): number | null {
+    return this._rgbValue;
+  }
+
+  set rgbValue(val: number | null) {
+    if (val === null) {
+      this._rgbValue = null;
+    } else {
+      this._rgbValue = val & 0xffffff;
+      this._aci = null;
+    }
+  }
+
+  /**
+   * Returns a deep copy of this color.
+   * @returns A new MTextColor instance with the same color state.
+   */
+  copy(): MTextColor {
+    const c = new MTextColor();
+    c._aci = this._aci;
+    c._rgbValue = this._rgbValue;
+    return c;
+  }
+
+  /**
+   * Returns a plain object for serialization.
+   * @returns An object with aci, rgb (tuple), and rgbValue (number or null).
+   */
+  toObject(): { aci: number | null; rgb: RGB | null; rgbValue: number | null } {
+    return { aci: this._aci, rgb: this.rgb, rgbValue: this._rgbValue };
+  }
+
+  /**
+   * Convert the current color to a CSS hex color string (#rrggbb).
+   * Returns null if the color is ACI-based and has no RGB value.
+   */
+  toCssColor(): string | null {
+    if (this._rgbValue !== null) {
+      return colorNumberToHex(this._rgbValue);
+    }
+    return null;
+  }
+
+  /**
+   * Create an MTextColor from a CSS color string.
+   * Supports #rgb, #rrggbb, rgb(...), rgba(...). Returns null if invalid or transparent.
+   */
+  static fromCssColor(value: string | null | undefined): MTextColor | null {
+    const rgbValue = cssColorToRgbValue(value);
+    if (rgbValue === null) return null;
+    const color = new MTextColor();
+    color.rgbValue = rgbValue;
+    return color;
+  }
+
+  /**
+   * Equality check for color.
+   * @param other The other MTextColor to compare.
+   * @returns True if both ACI and RGB values are equal.
+   */
+  equals(other: MTextColor): boolean {
+    return this._aci === other._aci && this._rgbValue === other._rgbValue;
+  }
+}
+
+/**
+ * MText context class for managing text formatting state
+ */
+export class MTextContext {
+  private _stroke: number = 0;
+  /** Whether to continue stroke formatting */
+  continueStroke: boolean = false;
+  /** Color (ACI or RGB) */
+  color: MTextColor = new MTextColor();
+  /** Line alignment */
+  align: MTextLineAlignment = MTextLineAlignment.BOTTOM;
+  /** Font face properties */
+  fontFace: FontFace = { family: '', style: 'Regular', weight: 400 };
+  /** Capital letter height */
+  private _capHeight: FactorValue = { value: 1.0, isRelative: false };
+  /** Character width factor */
+  private _widthFactor: FactorValue = { value: 1.0, isRelative: false };
+  /**
+   * Character tracking factor a multiplier applied to the default spacing between characters in the MText object.
+   * - Value = 1.0 → Normal spacing.
+   * - Value < 1.0 → Characters are closer together.
+   * - Value > 1.0 → Characters are spaced farther apart.
+   */
+  private _charTrackingFactor: FactorValue = { value: 1.0, isRelative: false };
+  /** Oblique angle */
+  oblique: number = 0.0;
+  /** Paragraph properties */
+  paragraph: ParagraphProperties = {
+    indent: 0,
+    left: 0,
+    right: 0,
+    align: MTextParagraphAlignment.DEFAULT,
+    tabs: [],
+  };
+
+  /**
+   * Get the capital letter height
+   */
+  get capHeight(): FactorValue {
+    return this._capHeight;
+  }
+
+  /**
+   * Set the capital letter height
+   * @param value - Height value
+   */
+  set capHeight(value: FactorValue) {
+    this._capHeight = {
+      value: Math.abs(value.value),
+      isRelative: value.isRelative,
+    };
+  }
+
+  /**
+   * Get the character width factor
+   */
+  get widthFactor(): FactorValue {
+    return this._widthFactor;
+  }
+
+  /**
+   * Set the character width factor
+   * @param value - Width factor value
+   */
+  set widthFactor(value: FactorValue) {
+    this._widthFactor = {
+      value: Math.abs(value.value),
+      isRelative: value.isRelative,
+    };
+  }
+
+  /**
+   * Get the character tracking factor
+   */
+  get charTrackingFactor(): FactorValue {
+    return this._charTrackingFactor;
+  }
+
+  /**
+   * Set the character tracking factor
+   * @param value - Tracking factor value
+   */
+  set charTrackingFactor(value: FactorValue) {
+    this._charTrackingFactor = {
+      value: Math.abs(value.value),
+      isRelative: value.isRelative,
+    };
+  }
+
+  /**
+   * Get the ACI color value
+   */
+  get aci(): number | null {
+    return this.color.aci;
+  }
+
+  /**
+   * Set the ACI color value
+   * @param value - ACI color value (0-256)
+   * @throws Error if value is out of range
+   */
+  set aci(value: number) {
+    this.color.aci = value;
+  }
+
+  /**
+   * Get the RGB color value
+   */
+  get rgb(): RGB | null {
+    return this.color.rgb;
+  }
+
+  /**
+   * Set the RGB color value
+   */
+  set rgb(value: RGB | null) {
+    this.color.rgb = value;
+  }
+
+  /**
+   * Gets whether the current text should be rendered in italic style.
+   * @returns {boolean} True if the font style is 'Italic', otherwise false.
+   */
+  get italic(): boolean {
+    return this.fontFace.style === 'Italic';
+  }
+  /**
+   * Sets whether the current text should be rendered in italic style.
+   * @param value - If true, sets the font style to 'Italic'; if false, sets it to 'Regular'.
+   */
+  set italic(value: boolean) {
+    this.fontFace.style = value ? 'Italic' : 'Regular';
+  }
+
+  /**
+   * Gets whether the current text should be rendered in bold style.
+   * This is primarily used for mesh fonts and affects font selection.
+   * @returns {boolean} True if the font weight is 700 or higher, otherwise false.
+   */
+  get bold(): boolean {
+    return (this.fontFace.weight || 400) >= 700;
+  }
+  /**
+   * Sets whether the current text should be rendered in bold style.
+   * This is primarily used for mesh fonts and affects font selection.
+   * @param value - If true, sets the font weight to 700; if false, sets it to 400.
+   */
+  set bold(value: boolean) {
+    this.fontFace.weight = value ? 700 : 400;
+  }
+
+  /**
+   * Get whether text is underlined
+   */
+  get underline(): boolean {
+    return Boolean(this._stroke & MTextStroke.UNDERLINE);
+  }
+
+  /**
+   * Set whether text is underlined
+   * @param value - Whether to underline
+   */
+  set underline(value: boolean) {
+    this._setStrokeState(MTextStroke.UNDERLINE, value);
+  }
+
+  /**
+   * Get whether text has strike-through
+   */
+  get strikeThrough(): boolean {
+    return Boolean(this._stroke & MTextStroke.STRIKE_THROUGH);
+  }
+
+  /**
+   * Set whether text has strike-through
+   * @param value - Whether to strike through
+   */
+  set strikeThrough(value: boolean) {
+    this._setStrokeState(MTextStroke.STRIKE_THROUGH, value);
+  }
+
+  /**
+   * Get whether text has overline
+   */
+  get overline(): boolean {
+    return Boolean(this._stroke & MTextStroke.OVERLINE);
+  }
+
+  /**
+   * Set whether text has overline
+   * @param value - Whether to overline
+   */
+  set overline(value: boolean) {
+    this._setStrokeState(MTextStroke.OVERLINE, value);
+  }
+
+  /**
+   * Check if any stroke formatting is active
+   */
+  get hasAnyStroke(): boolean {
+    return Boolean(this._stroke);
+  }
+
+  /**
+   * Set the state of a stroke type
+   * @param stroke - The stroke type to set
+   * @param state - Whether to enable or disable the stroke
+   */
+  private _setStrokeState(stroke: MTextStroke, state: boolean = true): void {
+    if (state) {
+      this._stroke |= stroke;
+    } else {
+      this._stroke &= ~stroke;
+    }
+  }
+
+  /**
+   * Create a copy of this context
+   * @returns A new context with the same properties
+   */
+  copy(): MTextContext {
+    const ctx = new MTextContext();
+    ctx._stroke = this._stroke;
+    ctx.continueStroke = this.continueStroke;
+    ctx.color = this.color.copy();
+    ctx.align = this.align;
+    ctx.fontFace = { ...this.fontFace };
+    ctx._capHeight = { ...this._capHeight };
+    ctx._widthFactor = { ...this._widthFactor };
+    ctx._charTrackingFactor = { ...this._charTrackingFactor };
+    ctx.oblique = this.oblique;
+    ctx.paragraph = { ...this.paragraph };
+    return ctx;
+  }
+}
+
+/**
+ * Token class for MText parsing
+ */
+export class MTextToken {
+  /**
+   * Create a new MText token
+   * @param type - The token type
+   * @param ctx - The text context at this token
+   * @param data - Optional token data
+   */
+  constructor(
+    public type: TokenType,
+    public ctx: MTextContext,
+    public data: TokenData[TokenType]
+  ) {}
+}