Python 3 uses unicode, and is not compatible with subprocess pipes

[Brianetta]

I'm using pyandoc inside a Python 3 CGI script. I'm having to convert strings in the following manner to avoid getting an exception from Python:

doc.markdown = "This is a *really* simple example".encode('latin-1')

print(doc.html.decode('UTF-8'))

Failure to do this causes some messy (and difficult to understand) errors and an aborted script. Perhaps this should be handled explicitly, or at least caught and explained?

[slestak]

This appears to affect python 2.7.2 as well.

update: I see my problem, you have PANDOC_PATH hardcoded in core.py.

['/Users/kreitz/.cabal/bin//pandoc', '--from=markdown', '--to=rst']

[kennethreitz]

You have to pass in byte strings in Python 3.