issue w/ for statement

[mattiaerre]

Description

missing a semicolon if for statement body is not included in braces.

Source

  function count(string[] data) public pure returns (uint) {
    uint result;
    for(uint i = 0; i < data.length; i++)
      result += 1;
    return result;
  }

Prettified

  function count(string[] data) public pure returns(uint) {
    uint result;
    for (uint i = 0; i < data.length; i++) result += 1
    return result;
  }

N.B. there is a missing ; at the end of the for body

[fvictorio]

Just repeating what we've already said elsewhere. The problem seems to be here:

prettier-plugin-solidity/src/printer.js

Lines 164 to 170 in 6428414

	case 'ExpressionStatement': {
	const addSemicolon = path.getParentNode().type !== 'ForStatement';
	return concat([
	node.expression ? path.call(print, 'expression') : '',
	addSemicolon ? ';' : ''
	]);
	}

where we add a semicolon only if the parent node is not a for statement. But if we don't do this, then an extra semicolon is added to the loop condition, because that is an ExpressionStatement too.

An extremely ugly fix would be to mutate the node so that the print function knows if it should add the semicolon. I don't like this, but it could work.

Another idea is to never add a semicolon in the ExpressionStatement printer, and make that a responsibility of the parent node.

[mattiaerre]

I like the second option much more but if we can have something that still fixes the problem I really don't mind if it's really terrible tbh; happy if you want to create a PR that fixes this issue btw

[fvictorio]

The more I think about this, the more convinced I am that this is a problem upstream, so I opened federicobond/solidity-parser-antlr#39 to see if it can be changed. But this may be a breaking change, so I don't know if it will be modified/fixed.

I say we do this: if the parser is modified and that allows us to do a proper fix, we do that. If not, let's do the easy/mutating approach, and revisit it later if needed.

[mattiaerre]

shall we decorate our parser so that it returns the structure we think it's more beneficial for us? that be also another option what do you think?

[fvictorio]

That's a great idea. Let's do that for now.