trailing-whitespace hook: restoring original file in case of failure - fixes #134

pre_commit_hooks/trailing_whitespace_fixer.py

diff --git a/pre_commit_hooks/trailing_whitespace_fixer.py b/pre_commit_hooks/trailing_whitespace_fixer.py
index c159071..22c41ba 100644
--- a/pre_commit_hooks/trailing_whitespace_fixer.py
+++ b/pre_commit_hooks/trailing_whitespace_fixer.py
@@ -9,7 +9,7 @@
 
 
 def _fix_file(filename, markdown=False):
-    for line in fileinput.input([filename], inplace=True):
+    for line in fileinput.input([filename], inplace=True, backup='.bak'):
         # preserve trailing two-space for non-blank lines in markdown files
         if markdown and (not line.isspace()) and (line.endswith("  \n")):
             line = line.rstrip(' \n')
@@ -64,14 +64,20 @@
                 .format(ext)
             )
 
-    if bad_whitespace_files:
-        for bad_whitespace_file in bad_whitespace_files:
-            print('Fixing {0}'.format(bad_whitespace_file))
-            _, extension = os.path.splitext(bad_whitespace_file.lower())
+    return_code = 0
+    for bad_whitespace_file in bad_whitespace_files:
+        print('Fixing {0}'.format(bad_whitespace_file))
+        _, extension = os.path.splitext(bad_whitespace_file.lower())
+        try:
             _fix_file(bad_whitespace_file, all_markdown or extension in md_exts)
-        return 1
-    else:
-        return 0
+            return_code = 1
+        # pylint: disable=broad-except
+        except Exception as error:  # pragma: no cover
+            # e.g. error can be a UnicodeDecodeError in Python 3
+            print('Ignoring {} that caused a {}'.format(bad_whitespace_file, error.__class__))
+            os.remove(bad_whitespace_file)
+            os.rename(bad_whitespace_file + '.bak', bad_whitespace_file)
+    return return_code
 
 
 if __name__ == '__main__':

tests/trailing_whitespace_fixer_test.py

diff --git a/tests/trailing_whitespace_fixer_test.py b/tests/trailing_whitespace_fixer_test.py
index 6f4fdfd..78e6e73 100644
--- a/tests/trailing_whitespace_fixer_test.py
+++ b/tests/trailing_whitespace_fixer_test.py
@@ -1,6 +1,8 @@
 from __future__ import absolute_import
 from __future__ import unicode_literals
 
+import sys
+
 import pytest
 
 from pre_commit_hooks.trailing_whitespace_fixer import fix_trailing_whitespace
@@ -103,3 +105,11 @@
 
 def test_returns_zero_for_no_changes():
     assert fix_trailing_whitespace([__file__]) == 0
+
+
+def test_preserve_non_utf8_file(tmpdir):
+    path = tmpdir.join('file.txt')
+    path.write_binary(b'<a>\xe9 \n</a>')
+    ret = fix_trailing_whitespace([path.strpath])
+    assert ret == (1 if sys.version_info[0] < 3 else 0)  # a UnicodeDecodeError is only triggered in Python 3
+    assert path.size() > 0