Crude transformer for bison grammars to use named references

transform.js

/**
 * Transform a single bison item to named references
 * 
 * For example to transform only "c_decl"
 * 
 * node transform.js Source/CParse/parser.y c_decl Source/CParse/parser.y
 */

const fs = require('fs');

let input = fs.readFileSync(process.argv[2], 'utf-8');

console.log(`Transforming ${process.argv[3]}`);

const identifier = new RegExp(`\\b(${process.argv[3]})\\b\\s*:`, 'gm');
console.log(identifier)
const elements_first = /\s*(([\<\>a-zA-Z0-9_]+\s*)+)/g;
const elements_next = /\s*\|\s*(([\<\>a-zA-Z0-9_]+\s*)+)/g;
const single_element = /\b([a-zA-Z0-9_]+)\b/g;

let m;
while ((m = identifier.exec(input)) !== null) {
  console.log(`Found ( ${m[1]} ) section`);

  let idx = m.index + m[0].length;
  let elements = elements_first;
  while (true) {
    //console.log(`\tnext block ( ${input.slice(idx, idx + 200)} )`);
    if (input.slice(idx).match(/^\s*;/))
      break;
    elements.lastIndex = 0;
    const args = elements.exec(input.slice(idx));
    if (!args) break;
    idx += args.index + args[0].length;

    let arg;
    console.log(`\tvariant ( ${args[1]} )`);
    let narg = 1;
    let replacements = {};
    while ((arg = single_element.exec(args[1])) !== null) {
      replacements[`$${narg}`] = `$${arg[1]}`;
      narg++;
    }
    console.log(replacements);

    let end_of_block = idx, start_of_block = idx;
    let braces = 0;
    while(true) {
      switch (input[end_of_block++]) {
        case '{': braces++; break;
        case '}': braces--; break;
      }
      if (braces === 0) break;
    }
    let block = input.slice(idx, end_of_block);
    //console.log('block', block);
    for (const r of Object.keys(replacements)) {
      block = block.replaceAll(new RegExp(`\\${r}\\b`, 'g'), replacements[r]);
    }
    input = input.slice(0, start_of_block) + block + input.slice(end_of_block);

    elements = elements_next;
    idx = start_of_block + block.length;
  }
}

if (process.argv[4])
  fs.writeFileSync(process.argv[4], input, 'utf-8');