Allow unicode chars, force_encoding('UTF-8')

[jondkinney]

I was getting errors in some files that had unicode characters in them. This seems to fix/ignore that and let the files be formatted still.

Command failed: erb-format --tailwind-output-path app/assets/builds/tailwind.css --stdin-filename "/Users/jon/Code/Seckret-Project/app/views/home/index.html.erb" --print-width 120
/Users/jon/.local/share/mise/installs/ruby/3.3.4/lib/ruby/gems/3.3.0/gems/erb-formatter-0.7.3/lib/erb/formatter.rb:101:in `gsub!': invalid byte sequence in US-ASCII (ArgumentError)

    @source.gsub!(ERB_PLACEHOLDER) { |tag| build_uid[].tap { |uid| pre_placeholders[uid] = tag } }
                  ^^^^^^^^^^^^^^^
	from /Users/jon/.local/share/mise/installs/ruby/3.3.4/lib/ruby/gems/3.3.0/gems/erb-formatter-0.7.3/lib/erb/formatter.rb:101:in `initialize'
	from /Users/jon/.local/share/mise/installs/ruby/3.3.4/lib/ruby/gems/3.3.0/gems/erb-formatter-0.7.3/lib/erb/formatter/command_line.rb:101:in `new'
	from /Users/jon/.local/share/mise/installs/ruby/3.3.4/lib/ruby/gems/3.3.0/gems/erb-formatter-0.7.3/lib/erb/formatter/command_line.rb:101:in `block in run'
	from /Users/jon/.local/share/mise/installs/ruby/3.3.4/lib/ruby/gems/3.3.0/gems/erb-formatter-0.7.3/lib/erb/formatter/command_line.rb:97:in `each'
	from /Users/jon/.local/share/mise/installs/ruby/3.3.4/lib/ruby/gems/3.3.0/gems/erb-formatter-0.7.3/lib/erb/formatter/command_line.rb:97:in `run'
	from /Users/jon/.local/share/mise/installs/ruby/3.3.4/lib/ruby/gems/3.3.0/gems/erb-formatter-0.7.3/exe/erb-format:5:in `<top (required)>'
	from /Users/jon/.local/share/mise/installs/ruby/3.3.4/bin/erb-format:25:in `load'
	from /Users/jon/.local/share/mise/installs/ruby/3.3.4/bin/erb-format:25:in `<main>'

[elia]

Hey, I tried to add a spec that would fail with utf8 chars, can you please share an example of a template that would fail?
I'll be happy to add it and merge.

[jondkinney]

These characters threw those errors:

The dash: more—effortlessly
The copyright: ©
The quote: Figma and Tailwind ” are used

[elia]

Hey @jondkinney thanks for the report, I was able to sort of reproduce the issue by setting LANG and LC_ALL to C and have ruby used ASCII as the default encoding. I added a few changes to the CI so it will always test in that scenario as well.
I think using UTF-8 for everything is reasonable, but I'm ready to roll it back to a warning if there are incompatible use cases.