Maximum call stack size exceeded

[MichalBryxi]

Somewhere between v5.0.1 and v5.0.3 we got an issue that manifests as follows (log taken from ember-template-lint-plugin-tailwindcss):

    +     "message": "Maximum call stack size exceeded",
    +     "severity": 2,
    +     "source": "RangeError: Maximum call stack size exceeded
    +     at Object.get [as print] (/Users/michal/ember/ember-template-lint-plugin-tailwindcss/node_modules/ember-template-recast/node_modules/@glimmer/syntax/dist/commonjs/es2017/index.js:78:17)
    +     at print (/Users/michal/ember/ember-template-lint-plugin-tailwindcss/node_modules/ember-template-recast/src/index.ts:21:12)
    +     at Object.override (/Users/michal/ember/ember-template-lint-plugin-tailwindcss/node_modules/ember-template-recast/src/index.ts:25:18)
    +     at Printer.print (/Users/michal/ember/ember-template-lint-plugin-tailwindcss/node_modules/ember-template-recast/node_modules/@glimmer/syntax/packages/@glimmer/syntax/lib/generation/printer.ts:539:20)
    +     at Object.build (/Users/michal/ember/ember-template-lint-plugin-tailwindcss/node_modules/ember-template-recast/node_modules/@glimmer/syntax/packages/@glimmer/syntax/lib/generation/print.ts:13:10)
    +     at print (/Users/michal/ember/ember-template-lint-plugin-tailwindcss/node_modules/ember-template-recast/src/index.ts:21:12)
    +     at Object.override (/Users/michal/ember/ember-template-lint-plugin-tailwindcss/node_modules/ember-template-recast/src/index.ts:25:18)
    +     at Printer.print (/Users/michal/ember/ember-template-lint-plugin-tailwindcss/node_modules/ember-template-recast/node_modules/@glimmer/syntax/packages/@glimmer/syntax/lib/generation/printer.ts:539:20)

[rwjblue]

Do you happen to know the steps to repro? Our internal tests don't seem to hit this, it's going to be hard to figure out what is wrong without some high level understanding of what manipulation causes it... 🤔

Maybe you could copy a test like this one, to replicate?

ember-template-recast/src/parse-result.test.ts

Lines 16 to 26 in ec7c070

	test('updating attributes on a non-self-closing void element', function () {
	let template = `<img src="{{something}}">`;
	let ast = parse(template);
	let element = ast.body[0] as AST.ElementNode;
	let attribute = element.attributes[0] as AST.AttrNode;
	let concat = attribute.value as AST.ConcatStatement;
	(concat.parts[0] as AST.MustacheStatement).path = builders.path('this.something');
	expect(print(ast)).toEqual(`<img src="{{this.something}}">`);
	});

[MichalBryxi]

No idea yet. It might even be a complete red herring as only when I shake the yarn.lock I get the problematic behaviour. I will try to get a failing test for you.

[MichalBryxi]

I could trace it to this recursive print() function. When I try to console.log(ast) for my case that trips over, I get something like:

{
        type: 'ConcatStatement',
        parts: [
          {
            type: 'TextNode',
            chars: '\n' +
              '      bazDoo\n' +
              '      gap-4 grid grid-cols-2\n' +
              '      lg:grid-cols-5 md:grid-cols-3 sm:grid-cols-3 xl:grid-cols-6\n' +
              '      ',
            loc: [Object]
          },
          {
            type: 'MustacheStatement',
            path: [Object],
            params: [Array],
            hash: [Object],
            escaped: true,
            loc: [Object],
            strip: [Object]
          },
          { type: 'TextNode', chars: '\n      ', loc: [Object] },
          {
            type: 'MustacheStatement',
            path: [Object],
            params: [Array],
            hash: [Object],
            escaped: true,
            loc: [Object],
            strip: [Object]
          },
          { type: 'TextNode', chars: '\n    ', loc: [Object] }
        ],
        loc: {
          source: null,
          start: { line: 3, column: 10 },
          end: { line: 13, column: 5 }
        }
}

Which seems reasonable to me. But I don't understand why it's recursing infinitely nor what's the logic of that function in the first place.

[MichalBryxi]

I think I found a repro: #589

Seems like when I create new concat with parts from previous concat bad things will happen. cc: @rwjblue

[MichalBryxi]

Sounds like this one was resolved with #671