Hi @brettstimmerman and @bago,

I am using the npm package, juice, on a project for work. The Juice package has a dependency on mensch#v0.3.3, but they reference your github repository instead of the npm package itself. It looks like this could be because you haven't published the release to npm (currently 0.3.1 on npm). If it isn't too much of a hassle, would you mind publishing the latest release to npm? Then I can get them to switch the dependency to npm (my deployment setup doesn't allow me to use git on my prod server).

Thanks for your time. Cheers, Max

Rules parsed with position=true should have the end curly brace be the end position, not the end of the selector/at-rule.

Signed-off-by: Jonathan Horowitz [email protected]

My attempt at fixing https://github.com/brettstimmerman/mensch/issues/15 and more. See new test position.js for fixed expectations.

Moved the col/line computation inside the "skip" method, so to make sure it is always invoked when moving the cursor. NOTE: The method could be optimized and does not support negative movements.

Hi,

version 0.3.3 has an error with stringify indentation, an error that has already been resolved about two years ago !!! but nobody has published a new version (0.3.4) with that solution

Can anyone publish that new version to fix that, please?

The indent method was leaking into the global scope, so this PR cleans that up by adding a file-scoped _level variable. This variable is also initialized inside the stringify function rather than in the level function so that if stringify were to ever throw an exception, subsequent calls would not use the wrong indentation level.

Please let me know if you need any information or need me to do anything to help get this merged in!

Hi Brett,

first of all I want to thank you for this great parser!

In the next weeks I'm going to publish an opensource project (https://github.com/voidlabs/mosaico) depending on my fork for mensch (https://github.com/bago/mensch).

As you can see (from other open issues and PR) I have applied a few patches to mensch, mainly related to "positions" and a few other improvements.

Is there any chance to get write access to your repository or to have you accept my PR anytime soon?

@media tty { 
  i{content:"\";/*" "*/}} a { color: black; } /*";}
}

a {color: white}

The hack itself is handled ok - two strings in the content property

The problem is that the next ruleset is considered part of the media block, while it should be outside of it

Test code:

var mensch = require('mensch');
var css = require('fs').readFileSync('./examples/css.css').toString();
var parsed = mensch.parse(css);
var res = mensch.stringify(parsed, {indentation: '  '})

console.log(res)

Result:

@media tty {
  i {
    content: "\";/*" "*/}} a { color: black; } /*";
  }

  a {
    color: white;
  }
}

body
div
{color: red}

gives

{ type: 'rule',
  selectors: [ 'bodydiv' ],
  declarations: [ { type: 'property', name: 'color', value: 'red' } ] }

I think the selector should be body div

this confuses the parser:

  .sele {
    content: "he\"lo";
  }

this is ok:

  .sele {
    content: "helo";
  }

{ type: 'property', 
  name: 'content', 
  value: '"helo"' }

On line 1 the first column is column 1, on the following lines the first column is column 0. I guess the correct positioning is the one when the first char is col:1, line:1 so the right solution is to have following lines to start with 1 too.

in

body {background:;}
#yo {display: block;}

Expected

body {

}
#yo {
  display: block;
}

Actual

body {
  background: ;#yo display: block;
}

AST

{
  "type": "stylesheet",
  "stylesheet": {
    "rules": [
      {
        "type": "rule",
        "selectors": [
          "body"
        ],
        "declarations": [
          {
            "type": "property",
            "name": "background",
            "value": ";#yo display: block"
          }
        ]
      }
    ]
  }
}

Run this code:

var mensch = require('mensch');
var parsed = mensch.parse('.my-class:not(.class1, .class2) {}', {
  position: true,
  comments: true,
});

The parsed.stylesheet.rules will be an array with two items: ['.my-class:not(.class1', '.class)'], this rules are wrong. It should not be splited by , simply.