Environment

Using clean-css through Angular CLI 1.7.0 which moved to clean-css. clean-css is imported through webpack plugin as it can seen here: https://github.com/angular/angular-cli/blob/master/packages/@angular/cli/plugins/cleancss-webpack-plugin.ts

Configuration options

 const cleancss = new CleanCSS({
          compatibility: 'ie9',
          level: 2,
          inline: false,
          returnPromise: true,
          sourceMap: this._options.sourceMap,
        });

Input SCSS

Image, since non-breaking space would not show up properly (screenshot taken on VSCode on faulty scss code, with Highlight bad characters VSCode plugin):

This code was transpiled to CSS, which still contains the non-breaking space. Resulting in following error.

Error

/project/node_modules/clean-css/lib/reader/input-source-map-tracker.js:37
if (originalPosition.line === null && line > 1 && selectorFallbacks > 0) {

TypeError: Cannot read property 'line' of undefined
at originalPositionFor (/project/node_modules/clean-css/lib/reader/input-source-map-tracker.js:37:24)
at originalMetadata (/project/node_modules/clean-css/lib/tokenizer/tokenize.js:486:43)
at intoTokens (/project/node_modules/clean-css/lib/tokenizer/tokenize.js:435:68)
at tokenize (/project/node_modules/clean-css/lib/tokenizer/tokenize.js:74:10)
at fromStyles (/project/node_modules/clean-css/lib/reader/read-sources.js:147:12)
at fromString (/project/node_modules/clean-css/lib/reader/read-sources.js:48:10)
at doReadSources (/project/node_modules/clean-css/lib/reader/read-sources.js:33:12)
at readSources (/project/node_modules/clean-css/lib/reader/read-sources.js:24:10)
at /project/node_modules/clean-css/lib/clean.js:99:12
at _combinedTickCallback (internal/process/next_tick.js:131:7)
at process._tickCallback (internal/process/next_tick.js:180:9)

Expected behavior

Presence of non-breaking space should not block the production of css since these kind of spaces should be deleted in the minification process.

Potential fix

We could add a marker for no-break space and replace the character with standard space.

NO_BREAK_SPACE: /\u00A0|\u202F|\uFEFF/g,

then in tokenize:

isNoBreakSpace = Marker.NO_BREAK_SPACE.test(character);
...
else if (isNoBreakSpace) {
      // Replace any kind of no-break-space with standard space
      character = Marker.SPACE;
      buffer.push(character);
    }

Can't open a PR as I don't know how to test that in your test suit.