If hugo have some chinese attributes for asciidoctor, asciidoctor throw error

[zkaip]

if hugo have markup attribute like the follow

The error is happened as follow

then I edit lib/asciidoctor/cli/options.rb
from
attr = attr.encode UTF_8 unless attr.encoding == UTF_8

to
attr = attr.force_encoding UTF_8 unless attr.encoding == UTF_8

the bug is fixed!

Please reference my fixed to upgrade asciidoctor!

[mojavelinux]

An AsciiDoc processor assumes that all input is either UTF-8 encoded or properly labeled. See https://docs.asciidoctor.org/asciidoc/latest/normalization/. It's likely that when Hugo runs the asciidoctor command using the system call, it ends up causing mislabeled data to be passed to Asciidoctor. The proposed change would be reverting the change made for #2764. You can find information about how to relabel the strings in the following comment: #2764 (comment)

Here's what I would be willing to do. We could try .encode and if that fails, fall back to force_encoding. In other words:

            unless attr.encoding == UTF_8
              begin
                attr = attr.encode UTF_8 
              rescue ::Encoding::EncodingError
                attr = attr.force_encoding UTF_8
              end
            end

Can you try that logic and see if it works for you?

[zkaip]

I solved this problem by modifying the source code, but I hope that this issue can also be updated in the main asciidoctor version, because, if the user passes a non-utf_8 string, then compatibility should also be done

[mojavelinux]

Can you please test with the code I provided?

[zkaip]

@mojavelinux you code not worked...

if I changed the code as the follow

then the result is follow

[mojavelinux]

Thank you for testing. Let me update the code and I'll get back to you once I have something new to test.

The reason I cannot accept the code as you wrote it is that it would break compatibility with strings that are properly labeled. But I think we can find a way to support both cases.

[mojavelinux]

Can you try this code instead? It looks like I had the name of the error class wrong

 unless attr.encoding == UTF_8
  begin
    attr = attr.encode UTF_8 
  rescue ::EncodingError
    attr = attr.force_encoding UTF_8
  end
end

[zkaip]

@mojavelinux Yes, this code resolved the problem.

unless attr.encoding == UTF_8
  begin
    attr = attr.encode UTF_8 
  rescue ::EncodingError
    attr = attr.force_encoding UTF_8
  end
end

[mojavelinux]

Thanks for verifying! I'll go ahead and proceed with that change.