Fixed #158. Now properly encoding output to stdout.

This is another try at this problem. The trick is geting code that works
with both Python 2 and Python 3. I think this does it. The only improvment
I can see now is to catch any errors and customize the error message to sugg
that the user set the environment variable PYTHONIOENCODING to the desired
encoding before calling the commandline script.

1 files changed, 7 insertions, 8 deletions

diff --git a/markdown/__init__.py b/markdown/__init__.py
index cda6b76..1b71295 100644
--- a/markdown/__init__.py
+++ b/markdown/__init__.py
@@ -379,15 +379,14 @@ class Markdown:
                 output_file.write(html)
                 # Don't close here. User may want to write more.
         else:
-            if sys.stdout.encoding:
-                # If we are in Python 3 or if we are not piping output:
+            # Encode manually and write bytes to stdout. 
+            html = html.encode(encoding, errors="xmlcharrefreplace")
+            try:
+                # Write bytes directly to buffer (Python 3).
+                sys.stdout.buffer.write(html)
+            except AttributeError:
+                # Probably Python 2, which works with bytes by default.
                 sys.stdout.write(html)
-            else:
-                # In python 2.x if you pipe output on command line,
-                # sys.stdout.encoding is None. So lets set it:
-                writer = codecs.getwriter(encoding)
-                stdout = writer(sys.stdout, errors="xmlcharrefreplace")
-                stdout.write(html)
 
         return self