PostCSS Conversion Checklist

Starting this issue to track progress on the conversion to using PostCSS.

Misc Tasks:

• merge linting fixes from master
• merge changes from #156, when that is merged
• util module for common operations. see note on then/catch with the parser below.
• write proper value parser
• move to postcss-values-parser

Linters:

• attributeQuotes
• borderZero
• comment
• decimalZero
• depthLevel
• duplicateProperty
• emptyRule
• finalNewline
• hexLength
• hexNotation
• hexValidation
• idSelector
• importantRule
• importPath
• propertyOrdering
• propertyUnits
• qualifyingElement
• selectorNaming
• singleLinePerProperty
• singleLinePerSelector
• spaceAfterPropertyColon
• spaceAfterPropertyName
• spaceAfterPropertyValue
• spaceAroundComma
• spaceAroundOperator
• spaceBeforeBrace
• spaceBetweenParens
• stringQuotes
• trailingSemicolon
• trailingWhitespace
• urlFormat
• urlQuotes
• zeroUnit

Then/Catch

Considering the following code:

parser = getParser(source);
parser.then(function (ast) {
  expect(linter.nodeTypes).to.include(ast.root.first.type);
});

If that expect throws an AssertionError, it’s swallowed by the Promise implementation of PostCSS less. So we’ll essentially get a false-positive for the expect result. Found this while implementing new tests for the nodeTypes property in the comments linter. And so we need to have a .catch(...) for every parse operation. And so we need to return the promise for every it(). We already have a ton of copy/paste’d code in the specs, so a small util lib that wraps this for us is reasonable.

[shellscape]

@jwilsson cool. I’ve got the value parser coding portion done, and I’m working on unit tests now that match postcss-value-parser (minus some really weird rules he was allowing, like unmatched parens and quotes).

let’s get you and @kokarn to knock out a few of those remaining linters that don’t rely on property parsing 😉

[shellscape]

Knocked out 12 of the linters. Will have some more free time towards the weekend.