Illegal control characters in XML output

Hi, I maintain the R bindings for cmark. One popular use case is converting commonmark to xml for processing the AST.

We are running into a problem when input markdown contains control characters (often captured from a tty), which makes xml output invalid. For example if the markdown text contains \033 and we convert that to xml, we get:

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE document SYSTEM "CommonMark.dtd">
<document xmlns="http://commonmark.org/xml/1.0">
  <paragraph>
    <text xml:space="preserve"></text>
  </paragraph>
</document>

However, trying to parse this with libxml2 fails:

 Error in read_xml.raw(charToRaw(enc2utf8(x)), "UTF-8", ..., as_html = as_html,  : 
  PCDATA invalid Char value 27 [9] 

A real world example is this readme file. This was done with the gfm fork, but I think the problem appears the same.

Is this a bug in cmark, or is markdown text not supposed to contain c0 characters in the first place?

cc @nwellnhof

The report names no source file, test, or entry point. Start by reproducing the shown control-character input through the XML renderer, then locate the renderer and its output tests; determine the expected handling and verify that the resulting XML is accepted by libxml2.