17830: preserve line endings of original file when updating keywords.

This fixes the test failures on Windows from the new tests, and
includes test fixes as well as the module fix.

diff --git a/Lib/keyword.py b/Lib/keyword.py
index 91528f793a82f900fcd4a053a313ba4db54ad147..6e1e882a91e7006f5699ca9ac6df8292d8b8457f 100755 (executable)
--- a/Lib/keyword.py
+++ b/Lib/keyword.py
@@ -60,6 +60,12 @@ def main():
     if len(args) > 1: optfile = args[1]
     else: optfile = "Lib/keyword.py"
 
+    # load the output skeleton from the target, taking care to preserve its
+    # newline convention.
+    with open(optfile, newline='') as fp:
+        format = fp.readlines()
+    nl = format[0][len(format[0].strip()):] if format else '\n'
+
     # scan the source file for keywords
     with open(iptfile) as fp:
         strprog = re.compile('"([^"]+)"')
@@ -68,25 +74,21 @@ def main():
             if '{1, "' in line:
                 match = strprog.search(line)
                 if match:
-                    lines.append("        '" + match.group(1) + "',\n")
+                    lines.append("        '" + match.group(1) + "'," + nl)
     lines.sort()
 
-    # load the output skeleton from the target
-    with open(optfile) as fp:
-        format = fp.readlines()
-
-    # insert the lines of keywords
+    # insert the lines of keywords into the skeleton
     try:
-        start = format.index("#--start keywords--\n") + 1
-        end = format.index("#--end keywords--\n")
+        start = format.index("#--start keywords--" + nl) + 1
+        end = format.index("#--end keywords--" + nl)
         format[start:end] = lines
     except ValueError:
         sys.stderr.write("target does not contain format markers\n")
         sys.exit(1)
 
     # write the output file
-    with open(optfile, 'w') as fp:
-        fp.write(''.join(format))
+    with open(optfile, 'w', newline='') as fp:
+        fp.writelines(format)
 
 if __name__ == "__main__":
     main()

diff --git a/Lib/test/test_keyword.py b/Lib/test/test_keyword.py
index 45a12b9a98f9030f58a2e717cf463729ab1bd547..af99f52c630d3c5fc002c637557492e5183fd977 100644 (file)
--- a/Lib/test/test_keyword.py
+++ b/Lib/test/test_keyword.py
@@ -9,7 +9,8 @@ import shutil
 import textwrap
 
 KEYWORD_FILE             = support.findfile('keyword.py')
-GRAMMAR_FILE             = os.path.join('..', '..', 'Python', 'graminit.c')
+GRAMMAR_FILE             = os.path.join(os.path.split(__file__)[0],
+                                        '..', '..', 'Python', 'graminit.c')
 TEST_PY_FILE             = 'keyword_test.py'
 GRAMMAR_TEST_FILE        = 'graminit_test.c'
 PY_FILE_WITHOUT_KEYWORDS = 'minimal_keyword.py'
@@ -30,7 +31,7 @@ class Test_iskeyword(unittest.TestCase):
     # preserved for backward compatibility.
     def test_changing_the_kwlist_does_not_affect_iskeyword(self):
         oldlist = keyword.kwlist
-        self.addCleanup(lambda: setattr(keyword, 'kwlist', oldlist))
+        self.addCleanup(setattr, keyword, 'kwlist', oldlist)
         keyword.kwlist = ['its', 'all', 'eggs', 'beans', 'and', 'a', 'slice']
         self.assertFalse(keyword.iskeyword('eggs'))
 
@@ -38,11 +39,12 @@ class Test_iskeyword(unittest.TestCase):
 class TestKeywordGeneration(unittest.TestCase):
 
     def _copy_file_without_generated_keywords(self, source_file, dest_file):
-        with open(source_file) as fp:
+        with open(source_file, 'rb') as fp:
             lines = fp.readlines()
-        with open(dest_file, 'w') as fp:
-            fp.writelines(lines[:lines.index("#--start keywords--\n") + 1])
-            fp.writelines(lines[lines.index("#--end keywords--\n"):])
+        nl = lines[0][len(lines[0].strip()):]
+        with open(dest_file, 'wb') as fp:
+            fp.writelines(lines[:lines.index(b"#--start keywords--" + nl) + 1])
+            fp.writelines(lines[lines.index(b"#--end keywords--" + nl):])
 
     def _generate_keywords(self, grammar_file, target_keyword_py_file):
         proc = subprocess.Popen([sys.executable,
@@ -56,15 +58,15 @@ class TestKeywordGeneration(unittest.TestCase):
                      'test only works from source build directory')
     def test_real_grammar_and_keyword_file(self):
         self._copy_file_without_generated_keywords(KEYWORD_FILE, TEST_PY_FILE)
-        self.addCleanup(lambda: support.unlink(TEST_PY_FILE))
+        self.addCleanup(support.unlink, TEST_PY_FILE)
         self.assertFalse(filecmp.cmp(KEYWORD_FILE, TEST_PY_FILE))
-        self.assertEqual(0, self._generate_keywords(GRAMMAR_FILE,
-                                                    TEST_PY_FILE)[0])
+        self.assertEqual((0, b''), self._generate_keywords(GRAMMAR_FILE,
+                                                           TEST_PY_FILE))
         self.assertTrue(filecmp.cmp(KEYWORD_FILE, TEST_PY_FILE))
 
     def test_grammar(self):
         self._copy_file_without_generated_keywords(KEYWORD_FILE, TEST_PY_FILE)
-        self.addCleanup(lambda: support.unlink(TEST_PY_FILE))
+        self.addCleanup(support.unlink, TEST_PY_FILE)
         with open(GRAMMAR_TEST_FILE, 'w') as fp:
             # Some of these are probably implementation accidents.
             fp.writelines(textwrap.dedent("""\
@@ -86,40 +88,40 @@ class TestKeywordGeneration(unittest.TestCase):
                     {1, 'no good'}
                     {283, 0},
                     {1,  "too many spaces"}"""))
-        self.addCleanup(lambda: support.unlink(GRAMMAR_TEST_FILE))
+        self.addCleanup(support.unlink, GRAMMAR_TEST_FILE)
         self._generate_keywords(GRAMMAR_TEST_FILE, TEST_PY_FILE)
         expected = [
-            "        'This one is tab indented',\n",
-            "        'also legal',\n",
-            "        'continue',\n",
-            "        'crazy but legal',\n",
-            "        'jello',\n",
-            "        'lemon',\n",
-            "        'tomato',\n",
-            "        'turnip',\n",
-            "        'wigii',\n",
+            "        'This one is tab indented',",
+            "        'also legal',",
+            "        'continue',",
+            "        'crazy but legal',",
+            "        'jello',",
+            "        'lemon',",
+            "        'tomato',",
+            "        'turnip',",
+            "        'wigii',",
             ]
         with open(TEST_PY_FILE) as fp:
-            lines = fp.readlines()
-        start = lines.index("#--start keywords--\n") + 1
-        end = lines.index("#--end keywords--\n")
+            lines = fp.read().splitlines()
+        start = lines.index("#--start keywords--") + 1
+        end = lines.index("#--end keywords--")
         actual = lines[start:end]
         self.assertEqual(actual, expected)
 
     def test_empty_grammar_results_in_no_keywords(self):
         self._copy_file_without_generated_keywords(KEYWORD_FILE,
                                                    PY_FILE_WITHOUT_KEYWORDS)
-        self.addCleanup(lambda: support.unlink(PY_FILE_WITHOUT_KEYWORDS))
+        self.addCleanup(support.unlink, PY_FILE_WITHOUT_KEYWORDS)
         shutil.copyfile(KEYWORD_FILE, TEST_PY_FILE)
-        self.addCleanup(lambda: support.unlink(TEST_PY_FILE))
-        self.assertEqual(0, self._generate_keywords(os.devnull,
-                                                    TEST_PY_FILE)[0])
+        self.addCleanup(support.unlink, TEST_PY_FILE)
+        self.assertEqual((0, b''), self._generate_keywords(os.devnull,
+                                                           TEST_PY_FILE))
         self.assertTrue(filecmp.cmp(TEST_PY_FILE, PY_FILE_WITHOUT_KEYWORDS))
 
     def test_keywords_py_without_markers_produces_error(self):
         rc, stderr = self._generate_keywords(os.devnull, os.devnull)
         self.assertNotEqual(rc, 0)
-        self.assertEqual(stderr, b'target does not contain format markers\n')
+        self.assertRegex(stderr, b'does not contain format markers')
 
     def test_missing_grammar_file_produces_error(self):
         rc, stderr = self._generate_keywords(NONEXISTENT_FILE, KEYWORD_FILE)

diff --git a/Misc/NEWS b/Misc/NEWS
index a7d7207c95064467bfe0e255b2844661caed5d68..273f11c92bae14e78c3a0cf8f299a665a22d259f 100644 (file)
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -49,6 +49,9 @@ Core and Builtins
 Library
 -------
 
+- Issue #17830: When keyword.py is used to update a keyword file, it now
+  preserves the line endings of the original file.
+
 - Issue #17272: Making the urllib.request's Request.full_url a descriptor.
   Fixes bugs with assignment to full_url. Patch by Demian Brecht.