Add __all__ (suggested by Raymond Hettinger).
Rename 'whitespace' global to '_whitespace' -- it's not part of the
public interface.
diff --git a/Lib/textwrap.py b/Lib/textwrap.py
index de07c8d..c8a1dac 100644
--- a/Lib/textwrap.py
+++ b/Lib/textwrap.py
@@ -12,6 +12,8 @@
import string, re
+__all__ = ['TextWrapper', 'wrap', 'fill']
+
# Hardcode the recognized whitespace characters to the US-ASCII
# whitespace characters. The main reason for doing this is that in
# ISO-8859-1, 0xa0 is non-breaking whitespace, so in certain locales
@@ -20,7 +22,7 @@
# same as any other whitespace char, which is clearly wrong (it's a
# *non-breaking* space), 2) possibly cause problems with Unicode,
# since 0xa0 is not in range(128).
-whitespace = '\t\n\x0b\x0c\r '
+_whitespace = '\t\n\x0b\x0c\r '
class TextWrapper:
"""
@@ -58,11 +60,11 @@
be broken, and some lines might be longer than 'width'.
"""
- whitespace_trans = string.maketrans(whitespace, ' ' * len(whitespace))
+ whitespace_trans = string.maketrans(_whitespace, ' ' * len(_whitespace))
unicode_whitespace_trans = {}
uspace = ord(u' ')
- for x in map(ord, whitespace):
+ for x in map(ord, _whitespace):
unicode_whitespace_trans[x] = uspace
# This funky little regex is just the trick for splitting