Blockquote tag appears after multiline HTML comments not ending with newline

[cyrus-and]

1. What version of goldmark are you using? v1.4.4
2. What version of Go are you using? 1.17
3. What operating system and processor architecture are you using? Darwin 20.6.0 x86_64

4. What did you do?

   package main
   
   import (
       "bytes"
       "fmt"
       "github.com/yuin/goldmark"
   )
   
   func main() {
       input := []byte("<!--\n-->")
       var buffer bytes.Buffer
       if err := goldmark.Convert(input, &buffer); err != nil {
           panic(err)
       }
       fmt.Print(buffer.String())
   }

5. What did you expect to see?

   <!-- raw HTML omitted -->

   or

   <!--
   -->

   or an empty output, certainly not a <blockquote> tag.

6. What did you see instead?

   <!-- raw HTML omitted -->
   <!-- raw HTML omitted -->
   <blockquote>
   </blockquote>

7. Did you confirm your output is different from CommonMark online demo or other official renderer correspond with an extension? Yes:

   <!--
   -->

[teekennedy]

Just ran into this myself while writing unit tests for my markdown renderer for goldmark.

I've narrowed it down to the following conditions:

• As the original commenter mentions, the block must contain multiple lines but not end with a newline.
• In order for an HTML block to not end in a newline, it has to be the final block in the document.
• Applies to HTML block types 1-5. Block types 6 and 7 require a blank line as an end to the block and avoid this bug.

I believe this is caused by an off-by-one error in advancing the source when parsing an HTML block on the last line of the document. It advances up to but not including the final character >, before closing the HTML block node. The parser continues parsing > as another block, generating an empty blockquote node in the process.

Until this is fixed, one can prevent this bug by always ensuring that your markdown document ends with a trailing newline before sending it to goldmark for parsing. It's good unix practice anyway.

[ivanspasov99]

@yuin Ran into the same issue but with pure html

md file with raw html

<pre>
hello
<code>
some code
<code>
<pre>

Receiving at the end "blockquote" tags

Maybe I did not understand how I should workaround it, did not find nothing within the renderer options

[teekennedy]

It's possible this has something to do with the fact that there's no closing tags in your HTML example.

Can you post the output of calling ast.Dump on the parsed source?