Pseudo-elements' content entities are converted

[Ryan Domingue]

Using criticalCSS via the Grunt task, I noticed entities in pseudo-elements' content property are being converted to their associated special characters and, thus, failing to render until the external CSS file has been parsed.

Opening here because it doesn't seem particularly relevant to the Grunt task. Apologies if that's not the case. Let me know if you need additional information.

[Zach Leatherman]

Just repro’d this in a project so 👍 from me.

turned into:

I’d also vote to keep the single colon :after and :before too, since they have a little bit better browser compat but that nit is not a big deal.

[pacdemon]

Any progress on this bug?

[Gary Green]

Wow, bug for 2 years? This is still a problem, just tried with CLI and fontawesome is getting converted. It should keep the value like content: "\f002";

[Paul Bormans]

Is there a workaround to this issue?

[Deleted user]

Unescaping and outputting as utf8 wouldn't be a problem. If only the quotes around it would be preserved.

For now, my workaround is using a regex-replace on the file afterwards. To place quotes around unquoted single-character content: strings in the unicode private use area that fontawesome seemingly uses:

search: /(content:\s*)([\uf000-\uf300])(\s*;)/g
replace: '$1"$2"$3'

[Ryan Domingue]

I had a similar RegEx fix in place, but ultimately decided that was silly and switched to a different critical css package 🤐

[James]

I'm still running into this issue. :/