bytes32(string) fix

[rjnrohit]

closes #892

[swapnilraj]

the tests pass so it looks good but could explain the bug and how your change fixes it?

[rjnrohit]

the tests pass so it looks good but could explain the bug and how your change fixes it?

This fixes the bug because previously arg.kind remains as a string all the way upto cairo writer, And in case of strings to fixedBytes conversion, the cairoWriter writes the expression as a single felt value using node.value despite node having typestring uint256 . Since, all FixedBytes conversions are later converted into number/hexNumbers, the, arg.kind for string conversion shouldn't be stay as string

warp/src/cairoWriter/writers/literalWriter.ts

Line 8 in a158b06

switch (node.kind) {

switch (node.kind) {
      case LiteralKind.Number:
        switch (primitiveTypeToCairo(node.typeString)) {
          case 'Uint256': {
            const [high, low] = divmod(BigInt(node.value), BigInt(Math.pow(2, 128)));
            return [`Uint256(low=${low}, high=${high})`];
          }
          case 'felt':
            return [node.value];
          default:
            throw new TranspileFailedError('Attempted to write unexpected cairo type');
        }
      case LiteralKind.Bool:
        return [node.value === 'true' ? '1' : '0'];
      case LiteralKind.String:
      case LiteralKind.UnicodeString: {
        if (
          node.value.length === node.hexValue.length / 2 &&
          node.value.length < 32 &&
          node.value.split('').every((v) => v.charCodeAt(0) < 127)
        ) {
          return [`'${node.value}'`];
        }
        return [`0x${node.hexValue}`];
      }

[swapnilraj]

Nice catch and good explanation! thanks