txchen / eslint-plugin-riot Goto Github PK

Now the logic is to remove all the non-js text and preserve the original line number.

There is one issue: it will violate the multiple blank line rules.

To fix it, only extract the code inside the <script> block.

I use <script></script> to ensure the parser knows where the js starts, but your setting seems to require one of 'es6', 'babel', 'javascript'.

If you replace !attrs.type with attrs.type then assumes 'allows' missing types should be sent to the linter, here: https://github.com/txchen/eslint-plugin-riot/blob/master/src/extract.js#L22

Is this a change you'd make?

(I'm still not sure if this fits in my flow -- I'm using many dirty globals, different than my nodejs code, so I need a totally separate eslint config.)

Lint errors show up fine in the Terminal when we run:

./node_modules/.bin/eslint ./src/*

But does not show with SublimeLinter (for Sublime Text) or linter-eslint (for Atom).

Hi. How do I get the red warning indicators to display in the left margin? I've installed ESLint in Sublime Text 2 and the errors show up in the console, but not in the editor. Thanks.

Thanks for writing this extractor! I was wondering why you return if the script type isn't es6. Is there some linting quirk that ends up happening or something? Or can you simply remove the type and lint the js as usual?

I'm trying to make this plugin work but I keep getting parser error.

% ./node_modules/eslint/bin/eslint.js file.tag

file.tag
  64:20  error  Parsing error: Unexpected token )

.eslintrc

{
    "env": {
        "browser": true,
        "commonjs": true,
        "node": true
    },
    "plugins": ["riot"],
    "extends": "eslint:recommended",
    "rules": {
        "no-console": 1,
        "indent": [
            "error",
            4
        ],
        "linebreak-style": [
            "error",
            "unix"
        ],
        "quotes": [
            "error",
            "single"
        ],
        "semi": [
            "error",
            "always"
        ],
        "no-trailing-spaces": 2
    }
}

file.tag

<script type="es6">
this.likePost = () => {
riot.control.trigger(riot.VE.LIKE_POST, opts.data.postId)
}
</script>

Apologies if I am asking a stupid question...

When I create an empty repository and run npm install eslint-plugin-riot, it seems that the code downloaded doesn't contain the feature enabling to read .riot files.
(It's a simple change to the folder inside the node_modules folder, but I don't think we are supposed to do that?).

Is it possible that the latest npm release doesn't contain the latest commits?

Thanks!

It occurs when compiling with rollup + buble

Sample.
https://github.com/t4y3/example-riot-eslint

It would be cool if this plugin was a tad smarter and linted the HTML attributes and component body. This would probably be a massive undertaking with minimal gain, since most errors will be in the <script> tag.

We're setting up an eslint config for a Riot project at work and need autofix support for how we want to use the linter.

What we'd need to do is translate the source position of every error in the processed code into the source position of every error in the original code. We need two things for this; the source position at the beginning of the script, and the number of characters we're getting rid of at every line when we dedent. The first part is easy --- we can just capture parser.startIndex the first time we hit the onText callback inside the script tag. The dedenting part is the tough bit --- it appears that in general we're not removing the same number of characters from each line in a given file.

After a few hacks at the problem, we ended up just removing the dedenting step, since that doesn't break any of our linter rules; you can see what that looks like at our forked repo. It's working well enough for us, but I didn't want to open a PR that would remove some functionality and probably break people's linters 😄

Thanks!

This plugin does not currently recognize riot 4 files extension.
It has changed from .tag ro .riot.