Merge pull request #46 from jpyams/python3

Add Python 3 support

11 files changed, 72 insertions, 15 deletions

diff --git a/.travis.yml b/.travis.yml
index b90b600..2ece8ef 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -1,5 +1,6 @@
 language: python
 python:
   - "2.7"
+  - "3.3"
 install: pip install -r requirements.txt
 script: bin/test
diff --git a/README.markdown b/README.markdown
index ce453af..445ad1c 100644
--- a/README.markdown
+++ b/README.markdown
@@ -76,3 +76,7 @@ After this, the `screenplain` command will use the working copy of your code.
 To run unit tests and style checks, run:
 
     bin/test
+
+For developing for Python 3, instead use:
+
+    mkvirtualenv --no-site-packages --python=$(which python3) screenplain-py3
diff --git a/bin/test b/bin/test
index 2d35baf..025248e 100755
--- a/bin/test
+++ b/bin/test
@@ -1,3 +1,8 @@
 #!/bin/bash
-nosetests --nocapture --with-doctest --doctest-tests $* && \
-  pep8 --ignore=E402 screenplain tests
+if [[ $(python -V 2>&1) =~ "Python 2" ]]
+then
+    nosetests --nocapture --with-doctest --doctest-tests -I ^test.py $* && \
+      pep8 --ignore=E402 screenplain tests
+else
+    python test.py && pep8 --ignore=E402 screenplain tests
+fi
diff --git a/requirements.txt b/requirements.txt
index 321e747..50f908b 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -2,3 +2,4 @@ reportlab
 unittest2
 nose
 pep8
+six
diff --git a/screenplain/export/html.py b/screenplain/export/html.py
index 1f88e64..81c0f8b 100644
--- a/screenplain/export/html.py
+++ b/screenplain/export/html.py
@@ -18,19 +18,22 @@ class tag(object):
     >>> import sys
     >>> from __future__ import with_statement
     >>> with tag(sys.stdout, 'div'):
-    ...     sys.stdout.write('hello')
+    ...     print('hello')
     ...
-    <div>hello</div>
+    <div>hello
+    </div>
 
     Adding classes to the element is possible:
 
     >>> with tag(sys.stdout, 'div', classes=['action']):
-    ...     sys.stdout.write('hello')
-    <div class="action">hello</div>
+    ...     print('hello')
+    <div class="action">hello
+    </div>
 
     >>> with tag(sys.stdout, 'div', classes=['action', 'centered']):
-    ...     sys.stdout.write('hello')
-    <div class="action centered">hello</div>
+    ...     print('hello')
+    <div class="action centered">hello
+    </div>
 
     """
     def __init__(self, out, tag, classes=None):
diff --git a/screenplain/parsers/fountain.py b/screenplain/parsers/fountain.py
index 3290dd2..5434c1a 100644
--- a/screenplain/parsers/fountain.py
+++ b/screenplain/parsers/fountain.py
@@ -5,6 +5,7 @@
 import itertools
 from itertools import takewhile
 import re
+from six import next
 
 from screenplain.types import (
     Slug, Action, Dialog, DualDialog, Transition, Section, PageBreak,
@@ -269,7 +270,7 @@ def parse_title_page(lines):
 
     it = iter(lines)
     try:
-        line = it.next()
+        line = next(it)
         while True:
             key_match = title_page_key_re.match(line)
             if not key_match:
@@ -278,7 +279,7 @@ def parse_title_page(lines):
             if value:
                 # Single line key/value
                 result.setdefault(key, []).append(value)
-                line = it.next()
+                line = next(it)
             else:
                 for line in it:
                     value_match = title_page_value_re.match(line)
diff --git a/screenplain/richstring.py b/screenplain/richstring.py
index e24db1a..ad667d1 100644
--- a/screenplain/richstring.py
+++ b/screenplain/richstring.py
@@ -4,10 +4,21 @@
 
 import re
 import cgi
+import six
 
 _magic_re = re.compile(u'[\ue700-\ue705]')
 
 
+def _escape(s):
+    """Replaces special HTML characters like <
+    and non-ascii characters with ampersand escapes.
+
+    """
+    encoded = cgi.escape(s).encode('ascii', 'xmlcharrefreplace')
+    # In Py3, encoded is bytes type, so convert it to a string
+    return encoded.decode('ascii')
+
+
 class RichString(object):
     """A sequence of segments where each segment can have its own style."""
 
@@ -20,7 +31,10 @@ class RichString(object):
         return ' + '.join(repr(s) for s in self.segments)
 
     def __unicode__(self):
-        return ''.join(unicode(s) for s in self.segments)
+        return ''.join(six.text_type(s) for s in self.segments)
+
+    def __str__(self):
+        return self.__unicode__()
 
     def startswith(self, string):
         """Checks if the first segment in this string starts with a
@@ -93,6 +107,9 @@ class Segment(object):
     def __unicode__(self):
         return self.text
 
+    def __str__(self):
+        return self.text
+
     def __eq__(self, other):
         return (
             isinstance(other, Segment) and
@@ -116,7 +133,7 @@ class Segment(object):
             re.sub(
                 '  +',  # at least two spaces
                 lambda m: '&nbsp;' * (len(m.group(0)) - 1) + ' ',
-                cgi.escape(self.text).encode('ascii', 'xmlcharrefreplace'),
+                _escape(self.text),
             ) +
             ''.join(style.end_html for style in reversed(ordered_styles))
         )
diff --git a/test.py b/test.py
new file mode 100644
index 0000000..c4e5887
--- /dev/null
+++ b/test.py
@@ -0,0 +1,24 @@
+"""Runs nosetest after preparing the test cases.
+
+Removes the leading u from unicode literals so
+Python 3 doctests won't fail.
+
+"""
+
+from screenplain import richstring
+import re
+
+unicode_literal = re.compile(r'u([\'"])')
+
+for n in (
+    'parse_emphasis',
+    '_unescape',
+    '_demagic_literals',
+):
+    attr = getattr(richstring, n)
+    old_doc = getattr(attr, '__doc__')
+    new_doc = unicode_literal.sub(r'\1', old_doc)
+    setattr(attr, '__doc__', new_doc)
+
+import nose
+nose.main(argv='nosetests --nocapture --with-doctest --doctest-tests'.split())
diff --git a/tests/fdx_test.py b/tests/fdx_test.py
index b6e2bc9..497f908 100644
--- a/tests/fdx_test.py
+++ b/tests/fdx_test.py
@@ -3,7 +3,7 @@
 # http://www.opensource.org/licenses/mit-license.php
 
 from testcompat import TestCase
-from StringIO import StringIO
+from six import StringIO
 
 from screenplain.export.fdx import write_text
 from screenplain.richstring import plain, bold, italic
diff --git a/tests/fountain_test.py b/tests/fountain_test.py
index fff1e48..c31cc66 100644
--- a/tests/fountain_test.py
+++ b/tests/fountain_test.py
@@ -9,7 +9,7 @@ from screenplain.types import (
     Slug, Action, Dialog, DualDialog, Transition, Section, PageBreak
 )
 from screenplain.richstring import plain, italic, empty_string
-from StringIO import StringIO
+from six import StringIO
 
 
 def parse(lines):
diff --git a/tests/richstring_test.py b/tests/richstring_test.py
index f2bed5e..ee42087 100644
--- a/tests/richstring_test.py
+++ b/tests/richstring_test.py
@@ -3,6 +3,7 @@
 # http://www.opensource.org/licenses/mit-license.php
 
 from testcompat import TestCase
+import six
 from screenplain.richstring import (
     RichString, Segment,
     Bold, Italic,
@@ -38,7 +39,7 @@ class RichStringOperatorTests(TestCase):
         s = bold('Hello') + plain(' there ') + bold('folks')
         self.assertEquals(
             u'Hello there folks',
-            unicode(s)
+            six.text_type(s)
         )
 
     def test_eq(self):