Related to #2, wouldn't it make sense to make most of these options true by default? I could understand making certain options default to false (specifically those which might be considered a bit extreme). But to have all options default to false makes the grunt task completely ineffectual out of the box.

When I look at the issues, there hasn't been a single response by the author. So: Is this plugin supported? If not: Would you consider handing it over to someone else who likes to maintain it?

When running grunt htmlhint and it reports an error Grunt believes that the task has failed however, it's actually that an error with the HTML was reported.

There are no formal releases or tags for this repository (github tab is empty). It would be very handy to add them.

Thanks!

Describe the bug
Travis-CI build fails with grunt --force
See Travis build status https://travis-ci.org/htmlhint/grunt-htmlhint/jobs/445465000

To Reproduce
Steps to reproduce the behavior:

1. Review Travis-CI build errors, https://travis-ci.org/htmlhint/grunt-htmlhint/jobs/445465000

NOTE: See use of grunt versus the package.json settings:

  "scripts": {
    "test": "grunt --force"
  },

Per my testing, it seems like it could be fixed by using:

  "scripts": {
    "test": "grunt -f"
  },

Expected behavior
Successful Travis-CI build.

我用到grunt中，发现都不报错的。

自己配置了

'tag-pair': true,
'attr-value-double-quotes':true

之后才报这两个的检查错误，请问默认设置是怎么样的？难道全部是false的？

Hey folks,

Your script does have serious problems while fetching "https://github.com/ariya/esprima/tarball/master" locally and on CI like Travis. It used to cause errored builds and took me days to find out what is happening.

Please upgrade your nice task to latest dependencies and let us see if this helps.

Thank you

The grunt-htmlhint, I can guess that it has a function of reporter.
Plz tell me whether to output in xml How do.

Would you consider changing your options list implementation to acting on an
option_name: false,
setting ?

Right now, it appears to accept settings of either true or false as an indication to apply the option. Mere presence in the list - not the true/false setting - causes the rule to be applied.

...or, update the documentation to more strongly indicate that an option should be left off the list if its application is not desired. Expanding on the current statement,

If options is empty, task will scan nothing.

would be very helpful.

I keep receiving the following warning whenever I do anything with NPM after installing grunt-htmlhint:

npm WARN [email protected] requires a peer of grunt@~0.4.1 but none was installed.

OS: Windows 10.0.10586
Grunt CLI: v1.2.0
Grunt: v1.0.1

When running the task, I would expect the task to fail when any error is found within the html. (according to the validation rules I made up).

In my case, there are errors in the markup, which are also shown in the task results, but the task itself does not fail.

Here is my task setup

{
    options : {
        htmlhintrc : 'config/.htmlhintrc'
    },
    raw : [
        '<%= config.dev.data.raw_html %>/**/*.html'
    ]
}

which runs fine and results in the following output:

Running "htmlhint:raw" (htmlhint) task
   dev/data/raw/brand/de/stada.html
      L3 |</div>
          ^ Tag must be paired, missing: [ </h1> ], start tag match failed [ <h1> ] on line 2. (tag-pair)

   dev/data/raw/brand/en/stada.html
      L3 |</div>
          ^ Tag must be paired, missing: [ </h1> ], start tag match failed [ <h1> ] on line 2. (tag-pair)

>> 14 files lint free.

Done, without errors.

But there are errors! And 2 of 14 files are not lint free.

Shouldn't the task fail? Is it a bug?

I am using [email protected] and [email protected]

Do I understand correct that this Grunt task can't load custom lint rules?

Currently if I try to run grunt task for templates (e.g. micro templates or ejs templates) it throws errors for <% tags.

The readme says this utility is to "Validate html files with htmlhint". It took some research before I discovered that's htmlhint is not a validator at all, but a linter. (Or more precisely a code style checker.)

I would expect something that bills itself as an HTML validator to ensure spec-compliant HTML. This is a great tool in addition to an HTML validator. But it is not, itself, one.

I think the readme ought to be updated to read something along the lines of "Lint html files with htmlhint" or "Check for HTML code-style violations with htmlhint" or similar.

I have some .PHP files that have <script>tags that, doesn't have a <head> or <body> tag, because they are just included in another .PHP file. So, when i run the task 'htmlhint' with the option 'Head script disabled' activate, it shows some errors, because it dont find these mentioned tags. But actually, this is not an error. The error must be showed only if there is a <script>tag inside the tag <head>. Please, include an option to ignore it if no tag is founded.

P.S.: (Sorry about the english=))

like https://developer.mozilla.org/en/docs/Web/SVG/Element/clipPath

The html element name of [ clipPath ] must be in lowercase. (tagname-lowercase)

clippath is clipPath in html

Hello,

any chance that htmlhint could detect non-printable characters? Got a link with invisible non-printable characters, which was thus 404 whereas it seemed ok for a human eye.

Thanks :-)

It would be great to have an option to log into file not only console.
Similar option exists in "jshint" and is called reporter. It also has checkstyle option.

Thanks in advance.

Describe the bug
npm audit generates security warnings

To Reproduce
Steps to reproduce the behavior:

1. npm install
2. npm audit
3. See warnings:

found 8 vulnerabilities (5 low, 3 high) in 264 scanned packages
  run `npm audit fix` to fix 6 of them.
  2 vulnerabilities require semver-major dependency updates.

Expected behavior
No high risk vulnerabilities and few, if any, low risk vulnerabilities.

Screenshots
N/A

Desktop (please complete the following information):

• OS: macOS 10.14
• node 11.0.0
• npm 6.4.1

Since a lot of people are using grunt to validate source code —and not built files, such as minified javascript— it would be nice if grunt-htmlhint would also support linting url's, instead of only files.

Say you're using a local node.js server with a template engine like Swig, and you'd want to check your html before the build, so you can fail the build if necessary. Like refusing to optimize require.js with r.js if there are JSHint errors. frntndr is an example of such a project. Who has the full html for a single page inside one file nowadays?

For this to work, I'd be nice if grunt-htmlhint could validate url's as well. Since running htmlhint over swig files will naturally fail, there needs to be a way of validating combined html from several template files. Like the grunt-htmllint-http plugin does. But ideally you wouldn't want another grunt plugin for that if you ask me.

I could see an example config like this:

htmlhint: {
  options: {
    'tag-pair': true
  },
  dist: {
    src: [
      'http://localhost/',
      'http://localhost/articles.html'
    ]
  }
}

But this is of course only a proposal, maybe a different way of specifying url's makes more sense?

Edit: I could help with the development of this feature, if you agree this would be a nice addition.