rojects/source/components/code-panel/CodeMirrorLexerMode.ts

import { BrowserLexer } from './BrowserLexer.js';

export interface MultiLineBlock
{
  type:    string;
  start:   RegExp;
  end:     RegExp;
  cmClass: string;
}

interface KeywordSet
{
  id:         string;
  words:      Set<string>;
  cmClass:    string;
  tokenTypes: string[];
}

export class CodeMirrorLexerMode
{
  private _lexer:           BrowserLexer;
  private _typeToClass:     Record<string, string | null>;
  private _multiLineBlocks: MultiLineBlock[];
  private _multiLineTypes:  Set<string>;
  private _keywordSets:     KeywordSet[] = [];
  private _cmModeCache:     any = null;

  name: string = '';

  constructor(
    lexer:           BrowserLexer,
    typeToClass:     Record<string, string | null>,
    multiLineBlocks: MultiLineBlock[] = []
  )
  {
    this._lexer           = lexer;
    this._typeToClass     = typeToClass;
    this._multiLineBlocks = multiLineBlocks;
    this._multiLineTypes  = new Set( multiLineBlocks.map( b => b.type ) );
  }

  setKeywordSet( id: string, words: string[], cmClass: string, tokenTypes: string[] = [ 'CWORD' ] ): this
  {
    const existing = this._keywordSets.find( ks => ks.id === id );

    if ( existing )
    {
      existing.words      = new Set( words );
      existing.cmClass    = cmClass;
      existing.tokenTypes = tokenTypes;
    }
    else
    {
      this._keywordSets.push( { id, words: new Set( words ), cmClass, tokenTypes } );
    }

    return this;
  }

  removeKeywordSet( id: string ): this
  {
    this._keywordSets = this._keywordSets.filter( ks => ks.id !== id );
    return this;
  }

  refresh( cm: any ): void
  {
    cm.setOption( 'mode', this.name );
  }

  get cmMode(): any
  {
    if ( this._cmModeCache ) return this._cmModeCache;

    const self = this;

    this._cmModeCache =
    {
      startState: () => ( { multiLine: null as string | null } ),

      token( stream: any, state: any ): string | null
      {
        // Ongoing multi-line block
        if ( state.multiLine !== null )
        {
          const block     = self._multiLineBlocks.find( b => b.type === state.multiLine )!;
          const remaining = stream.string.slice( stream.pos );
          const endMatch  = block.end.exec( remaining );

          if ( endMatch !== null )
          {
            const len = endMatch.index + endMatch[ 0 ].length;
            for ( let i = 0; i < len; i++ ) stream.next();
            state.multiLine = null;
          }
          else
          {
            stream.skipToEnd();
          }

          return block.cmClass;
        }

        // Multi-line block starts
        for ( const block of self._multiLineBlocks )
        {
          const remaining  = stream.string.slice( stream.pos );
          const startMatch = block.start.exec( remaining );

          if ( startMatch && startMatch.index === 0 )
          {
            const startLen  = startMatch[ 0 ].length;
            const afterStart = remaining.slice( startLen );
            const endMatch  = block.end.exec( afterStart );

            if ( endMatch !== null )
            {
              const totalLen = startLen + endMatch.index + endMatch[ 0 ].length;
              for ( let i = 0; i < totalLen; i++ ) stream.next();
            }
            else
            {
              for ( let i = 0; i < startLen; i++ ) stream.next();
              stream.skipToEnd();
              state.multiLine = block.type;
            }

            return block.cmClass;
          }
        }

        // Regular matchers
        const matchers = self._lexer.modes.get( 'default' ) ?? [];

        for ( const matcher of matchers )
        {
          if ( self._multiLineTypes.has( matcher.type ) ) continue;

          const startPos = stream.pos;
          const len      = matcher.matchLength( stream.string, startPos );

          if ( len > 0 )
          {
            for ( let i = 0; i < len; i++ ) stream.next();

            const word = stream.string.slice( startPos, stream.pos );

            for ( const ks of self._keywordSets )
            {
              if ( ks.tokenTypes.includes( matcher.type ) && ks.words.has( word ) )
              {
                return ks.cmClass;
              }
            }

            return self._typeToClass[ matcher.type ] ?? null;
          }
        }

        stream.next();
        return null;
      }
    };

    return this._cmModeCache;
  }
}