Auto removal of trailing whitespace to skip binary file correctly

main/githooks.py

@@ -5,8 +5,10 @@
 '''
 
 from collections import defaultdict
+from io import StringIO
 from pathlib import Path
 from tempfile import NamedTemporaryFile
+from unittest.mock import patch
 import os
 import platform
 import re
@@ -276,6 +278,7 @@ def check_eol(files):
             with open(filename, 'rb') as fileobj:
                 data = fileobj.read().decode()
         except UnicodeDecodeError:
+            _skip(filename, 'File is not UTF-8 encoded')
             continue
 
         # Skip binary file
@@ -290,8 +293,12 @@ def check_eol(files):
 
 def check_do_not_merge_in_file(filename, new_file=False):
     '''Check for "do not merge" in a filename'''
-    with open(filename, 'rb') as fileobj:
-        lines = fileobj.read().decode().splitlines(True)
+    try:
+        with open(filename, 'rb') as fileobj:
+            lines = fileobj.read().decode().splitlines(True)
+    except UnicodeDecodeError:
+        _skip(filename, 'File is not UTF-8 encoded')
+        return 0
 
     if new_file:
         line_nums = [f'1-{len(lines)}']
@@ -377,7 +384,8 @@ def trim_trailing_whitespace_in_file(filename, new_file, dry_run,
         with open(filename, 'rb') as fileobj:
             lines = fileobj.read().decode().splitlines(True)
     except UnicodeDecodeError:
-        return
+        _skip(filename, 'File is not UTF-8 encoded')
+        return 0
 
     if new_file:
         line_nums = [f'1-{len(lines)}']
@@ -398,7 +406,7 @@ def trim_trailing_whitespace_in_file(filename, new_file, dry_run,
             if dry_run:
                 modified_lines.append(str(line_num))
             else:
-                print(f'   Fixed line {line_num}')
+                print(f'   Fixed line {filename}:{line_num}')
                 modified_file = True
                 lines[line_num-1] = after
 
@@ -436,6 +444,14 @@ def test_trim_trailing_whitespace(self):
             retval = trim_trailing_whitespace_in_file(tmp.name, True, True)
             self.assertEqual(retval, 0)
 
+    def test_decodeerror(self):
+        # A text file that is not utf-8 encoded - report and skip
+        test_file = Path(__file__).parent / '../test/decode_error.txt'
+        with patch('sys.stdout', new=StringIO()) as tmp_stdout:
+            retval = trim_trailing_whitespace_in_file(test_file, True, True)
+            self.assertEqual(retval, 0)
+            self.assertEqual(tmp_stdout.getvalue().strip(), f'SKIP {test_file}: File is not UTF-8 encoded')
+
 
 def remove_trailing_white_space(files, new_files=False, dry_run=False):
     '''Remove trailing white spaces in all new and modified lines