i don't mind merging this since i think doing this is pretty silly anyways, but are people going to be annoying with this? there are some people who want the CSS to be exact, including spacing.

some hacks require madness like this but whether or not they are relevant anymore who knows, I stay away from that stuff haha

As much as I understand this bug and pull request, isn't it better removing all comments throughout (including comments outside properties and values) before unleashing the parser? Are comments gracefully handled when appearing outside properties and values? sorry, I haven't checked.

That would be just two lines of code and one pass over the entire input string for removing all comments from the whole input string (as much as javascript's replace function is efficient). Something like:

  # Regex replace: remove all CSS comments (of the form /* anything */) 
  # First back-slash escapes the string, not the regex
  regex = new RegExp('/\\*.*?\\*/', 'g') 
  string = string.replace(regex, '')

...excuse my coffeescript ;-)

As a user of this repo, I'd probably not favor the extra cycles for parsing out comments on each pass, as in this pull request. Maybe it's better not to parse each and every element, notwithstanding that it also matters what is the treatment of comments elsewhere in the CSS input.

i'm +1 to that but people are inevitably going to complain. plus, i don't see performance an issue with this library.

Removing comments before parsing would break position information and so error reporting and source maps.

The correct regex for CSS comments is /\/\*[^*]*\*+([^/*][^*]*\*+)*\//g btw. This is from http://www.w3.org/TR/CSS21/grammar.html

hah. thank god for that regex! this PR's test was actually incorrect. thank god for not pulling it :)

@jonathanong is there a release planned that includes this change?

yeah v2