Description of problem:
In current Fedora 28, asciidoc ships with broken a2x binary (calling non-existent decode method on string object). As far as I can tell, this issue is fixed in F29+ with the asciidoc-python3-a2x-decode-fix.patch.

Version-Release number of selected component (if applicable):
asciidoc-8.6.10-0.1.20180605git986f99d.fc28

How reproducible:
Try to convert document to xhtml:

$ a2x -f xhtml -D doc/xhtml doc/source.txt

Actual results:
Traceback (most recent call last):
  File "/usr/bin/a2x", line 952, in <module>
    a2x.execute()
  File "/usr/bin/a2x", line 373, in execute
    self.__getattribute__('to_'+self.format)()
  File "/usr/bin/a2x", line 650, in to_xhtml
    self.copy_resources(xhtml_file, src_dir, self.destination_dir)
  File "/usr/bin/a2x", line 582, in copy_resources
    lambda attrs: attrs.get('type') == 'text/css')
  File "/usr/bin/a2x", line 263, in find_resources
    parser.feed(contents.decode(encoding))
AttributeError: 'str' object has no attribute 'decode'

Expected results:
No error, document is converted.