Invalid US-ASCII character "\xE2"

[dalanmiller]

There's quite a few pages on the Internet talking about this problem with Ruby but none of the solutions seem to work with Jekyll. Ultimately, I'm trying to build a Jekyll site in a Docker container and I've tried setting @charset as well as all the locale ENV variables. is there something I can set in _config.yml to solve this?

Configuration file: /www/_config.yml
            Source: /www
       Destination: /www/_site
 Incremental build: enabled
      Generating...
  Conversion error: Jekyll::Converters::Scss encountered an error while converting 'assets/css/styles.scss':
                    Invalid US-ASCII character "\xE2" on line 3
/usr/local/bundle/gems/jekyll-sass-converter-1.3.0/lib/jekyll/converters/scss.rb:97:in `rescue in convert': Invalid US-ASCII character "\xE2" on line 3 (Jekyll::Converters::Scss::SyntaxError)
    from /usr/local/bundle/gems/jekyll-sass-converter-1.3.0/lib/jekyll/converters/scss.rb:95:in `convert'
    from /usr/local/bundle/bundler/gems/jekyll-43a28aed967b/lib/jekyll/convertible.rb:67:in `block in transform'
    from /usr/local/bundle/bundler/gems/jekyll-43a28aed967b/lib/jekyll/convertible.rb:65:in `each'
    from /usr/local/bundle/bundler/gems/jekyll-43a28aed967b/lib/jekyll/convertible.rb:65:in `reduce'
    from /usr/local/bundle/bundler/gems/jekyll-43a28aed967b/lib/jekyll/convertible.rb:65:in `transform'
    from /usr/local/bundle/bundler/gems/jekyll-43a28aed967b/lib/jekyll/convertible.rb:249:in `do_layout'
    from /usr/local/bundle/bundler/gems/jekyll-43a28aed967b/lib/jekyll/page.rb:122:in `render'
    from /usr/local/bundle/bundler/gems/jekyll-43a28aed967b/lib/jekyll/site.rb:302:in `block in render'
    from /usr/local/bundle/bundler/gems/jekyll-43a28aed967b/lib/jekyll/site.rb:301:in `each'
    from /usr/local/bundle/bundler/gems/jekyll-43a28aed967b/lib/jekyll/site.rb:301:in `render'
    from /usr/local/bundle/bundler/gems/jekyll-43a28aed967b/lib/jekyll/site.rb:55:in `process'
    from /usr/local/bundle/bundler/gems/jekyll-43a28aed967b/lib/jekyll/command.rb:28:in `process_site'
    from /usr/local/bundle/bundler/gems/jekyll-43a28aed967b/lib/jekyll/commands/build.rb:58:in `build'
    from /usr/local/bundle/bundler/gems/jekyll-43a28aed967b/lib/jekyll/commands/build.rb:34:in `process'
    from /usr/local/bundle/bundler/gems/jekyll-43a28aed967b/lib/jekyll/commands/build.rb:18:in `block (2 levels) in init_with_program'
    from /usr/local/bundle/gems/mercenary-0.3.5/lib/mercenary/command.rb:220:in `call'
    from /usr/local/bundle/gems/mercenary-0.3.5/lib/mercenary/command.rb:220:in `block in execute'
    from /usr/local/bundle/gems/mercenary-0.3.5/lib/mercenary/command.rb:220:in `each'
    from /usr/local/bundle/gems/mercenary-0.3.5/lib/mercenary/command.rb:220:in `execute'
    from /usr/local/bundle/gems/mercenary-0.3.5/lib/mercenary/program.rb:42:in `go'
    from /usr/local/bundle/gems/mercenary-0.3.5/lib/mercenary.rb:19:in `program'
    from /usr/local/bundle/bundler/gems/jekyll-43a28aed967b/bin/jekyll:20:in `<top (required)>'
    from /usr/local/bundle/bin/jekyll:16:in `load'
    from /usr/local/bundle/bin/jekyll:16:in `<main>'
rake aborted!
Command failed with status (1): [bundle exec jekyll build --trace...]
/www/Rakefile:295:in `jekyll'
/www/Rakefile:97:in `block in <top (required)>'
Tasks: TOP => build
(See full trace by running task with --trace)

[parkr]

Try adding this to your _config.yml

encoding: utf-8

This character is is â. Do you have any non-English characters in assets/css/styles.scss or any of its imports?

[dalanmiller]

Happened again with the _config.yml change. I ack'd through my repo and couldn't find any of those and don't see why I would have any non-English characters in the imports. Also, it says it's occurring here in the styles.scss file:

Conversion error: Jekyll::Converters::Scss encountered an error while converting 'assets/css/styles.scss':
                    Invalid US-ASCII character "\xE2" on line 3
/usr/local/bundle/gems/jekyll-sass-converter-1.3.0/lib/jekyll/converters/scss.rb:97:in `rescue in convert': Invalid US-ASCII character "\xE2" on line 3 (Jekyll::Converters::Scss::SyntaxError)

assets/css/styles.scss

---
---
@charset "utf-8";

// Bourbon, Neat, Bitters, Base styles
@import "normalize";
@import "bourbon";
@import "neat";
@import "base/base";

[dalanmiller]

Also hi @parkr! I ate lunch with you at Github Universe!

[dalanmiller]

For the record, I found this great one-liner for finding non-ASCII characters:

# Linux
grep --color='auto' -P -n "[\x80-\xFF]" file.xml

# Mac OS X after brew install pcre
pcregrep --color='auto' -n "[\x80-\xFF]" file.xml

[pathawks]

If you comment out all the @imports in assets/css/styles.scss do you still get the same error message? If you comment out all but one, can you figure out which one is causing the error message?

[parkr]

Ok, so per the Sass documentation:

Sass follows the CSS spec to determine the encoding of a stylesheet, and falls back to the Ruby string encoding. This means that it first checks the Unicode byte order mark, then the @charset declaration, then the Ruby string encoding. If none of these are set, it will assume the document is in UTF-8.

So it should assume UTF-8, so this character should not be a problem. Interesting issue, this is. Line three would be the line which contains // Bourbon, Neat, Bitters, Base styles, which doesn't appear to contain \xE2. Try @pathawks's solution for debugging this... this is super weird. If you can isolate a given line in your CSS through this method, that would help. We delegate imports entirely to Sass, so they should be able to slurp up everything as UTF-8 by default as specified.

Also hi @parkr! I ate lunch with you at Github Universe!

Hey man, thanks for stopping by Jekyll's issues! :)

[dalanmiller]

Nice to see a friendly face on the Internet now and again @parkr 👍

So just to follow up, I used that unicode character search and found just two lines where I could just replace them with ASCII equivalents. They were both comments anyway so nothing got hurt and I could build my Jekyll docker container.

It'd probably be a lot more helpful for jekyll/jekyll-sass-converter to output the problems lines of the original source files as it traverses the import "tree" versus just the top most line where an @import occurs

[parkr]

It'd probably be a lot more helpful for jekyll/jekyll-sass-converter to output the problems lines of the original source files as it traverses the import "tree" versus just the top most line where an @import occurs

It may be an issue in Sass itself, as we output the line from the Sass error. If Sass passed us back the line in the imported file, then that would help a bit more, of course. Maybe #sass_backtrace would be more helpful? If you have a few minutes, would be great to have you test out the output of that in your failure scenario.

[burntcookie90]

I'm seeing this issue when using octopress on one of my machines but not the other. Haven't been able to figure out which possible variable is causing the issue on the erroring machine.

edit: figured it out, the new linux machine didn't have it's locale set to utf-8 at a system level. After setting that, it works as expected

[pathawks]

Can somebody post the source of a site? I would like to investigate this issue

[parkr]

I believe this issue is fixed in jekyll-sass-converter v1.4.0, which I just released. https://github.com/jekyll/jekyll-sass-converter/releases/tag/v1.4.0

[AngeloAballe]

I'm also having this issue. No problem if hosted via GitHub pages, but when I use docker on my own laptop or server, it has this issue.

The problem is the

---
---

above the scss file. If you remove this, there would be no error but the scss will not be converted. Still don't know how to fix this.

[parkr]

Did you upgrade your version of Jekyll-sass-converter?

[AngeloAballe]

Imgur

I tried. Same error.

[parkr]

@AngeloAballe The YAML front matter (two lines of ---) are not the problem – they merely tell Jekyll to convert the SCSS into CSS. Have you tried outside of Docker? See Dalan's comment about finding these characters:

# Linux
grep --color='auto' -P -n "[\x80-\xFF]" file.xml

# Mac OS X after brew install pcre
pcregrep --color='auto' -n "[\x80-\xFF]" file.xml

[AngeloAballe]

Thanks for the reply @parkr but after searching, I manage to fix it and I'm sharing my solution. I found out that the docker container is using a different charset or locale so you have to set it in you docker file.

I manage to fix mine by adding this on my Dockerfile. No need to upgrade to sass converter 1.4.0.

# Install program to configure locales
RUN apt-get install -y locales
RUN dpkg-reconfigure locales && \
  locale-gen C.UTF-8 && \
  /usr/sbin/update-locale LANG=C.UTF-8

# Install needed default locale for Makefly
RUN echo 'en_US.UTF-8 UTF-8' >> /etc/locale.gen && \
  locale-gen

# Set default locale for the environment
ENV LC_ALL C.UTF-8
ENV LANG en_US.UTF-8
ENV LANGUAGE en_US.UTF-8

Hope this helps everyone!

[envygeeks]

hey @AngeloAballe if you are using the official Docker image can you push a ticket to https://github.com/jekyll/docker and we'll happily add that as a default, I didn't think Alpine lacked the default language vars (but I'm not surprised, it is an embedded OS.)

[mkyung]

Just for the future reference:

I originally turned off Set locale environment variables on startup in my OSX Terminal preference due to other reasons. The error is gone if I check this checkbox.

[michaellee]

I ran into this issue as well while trying to build my project over at GitLab. What was causing the issue was this bit of code:

...
a::before{
  content:'“';
}
a::after{
  content:'”';
}
...

The “ and ” were the culprits. Changing them to be the hex value fixed the issue.

...
a::before{
  content:'\201C';
}
a::after{
  content:'\201D';
}
...

[simonorono]

I'm still getting this error on latest Jekyll. Check the CI log:

https://gitlab.com/simonorono/simonorono.gitlab.io/builds/2904606

[nateberkopec]

I get this problem when building a Jekyll site with wercker and there's Unicode in the SASS. Removing the UTF-8 characters or using @AngeloAballe's solution works.

[MildTomato]

Hey guys!

I'm using Jekyll 3.3.0 and wercker. Getting a Invalid US-ASCII character "\xE2" on Wercker whenever it tries to build.
Seems to be something in foundation-sites which is causing it, but can't find which file is causing the issue.
Only getting the invalid character issue on wercker, runs fine running locally.

[MildTomato]

OK, seemed to be a — in foundation-sites/scss/typography/_base.scss on line 15, one of the comments.

Wercker runs fine now 😆

[Hwesta]

I also ran into this problem, and found this thread very helpful but it didn't cover my case. Hopefully this helps the next person.

For me, setting LANG=en_US.UTF-8 didn't work because the locale hadn't been generated on my system and it was falling back to LANG=C which caused the error. Additionally, the problematic character I had was an en-dash (\xe2\x80\x93) not actually â (\xe2). The fix was to generate a UTF-8 locale and set it. Instructions tested on Arch Linux.

1. Check if locale exists before you try to set it: locale -a
2. If not, generate locale: Uncomment locale name in /etc/locale.gen and run sudo locale-gen
3. Set locale for the system: Edit /etc/locale.conf and log in again for it to take effect
4. Or set locale for the terminal
   • bash: export LANG=en_US.UTF-8
   • fish: set -xg LANG en_US.UTF-8

[dltj]

My jekyll build is being done by AWS CodeBuild, and I found I needed to add these lines to the buildspec.yml file:

  pre_build:
    commands:
      - export LC_ALL="C.UTF-8"
      - export LANG="en_US.UTF-8"
      - export LANGUAGE="en_US.UTF-8"

[IotaSpencer]

If anyone is getting this on GitLab, add the list on the above ^
#4268 (comment)

just put this in a before_script: before everything else

[bsde1245]

I get this error in gitlab, @IotaSpencer Where to add the code. I dont find buildspec.yml file
on sass converter v 1.5.0

Edit: foud it it is in the .gtlab-ci.yml

[reardestani]

For Jekyll on Gitlab, alternative to @IotaSpencer's solution is adding them in variables, like:

variables:
  LC_ALL: "C.UTF-8"
  LANG: "en_US.UTF-8"
  LANGUAGE: "en_US.UTF-8"

[npeirson]

Solution above looks good, I also got around it by putting an env assignment in my rake file.

[pathawks]

It seems this is an environment issue rather than a bug Jekyll could or should do something about. Maybe we could add some documentation regarding this?

Otherwise, this issue should be closed.

/cc: @jekyll/documentation

[stragu]

If anyone on GitLab Pages is wondering why @reardestani 's solution does not work, there is a typo (a missing "). It should be:

variables:
  LC_ALL: "C.UTF-8"
  LANG: "en_US.UTF-8"
  LANGUAGE: "en_US.UTF-8"

Thanks everyone, I added those variables to my .gitlab-ci.yml file and it fixed it (and looked rather clean) :)