Merge pull request #44 from nexB/posix-safe-filename

Make safe filename safe to use on POSIX

CHANGELOG.rst

@@ -1,10 +1,14 @@
 Release notes
 =============
 
-Version (next) 
+Version (next)
 ------------------------------
 
-TBD.
+- Add ``posix_only`` option to ``commoncode.paths.portable_filename`` and
+  ``commoncode.paths.safe_path``. This option prevents
+  ``commoncode.paths.portable_filename`` and ``commoncode.paths.safe_path`` from
+  replacing filenames and punctuation in filenames that are valid on POSIX
+  operating systems, but not Windows.
 
 Version 31.0.0 - (2022-05-16)
 ------------------------------
@@ -50,7 +54,7 @@ This is a major version with API-breaking changes in the resource module.
   otherwise missing from files path list.
   In particular this behaviour changed when you create a VirtualCodebase from
   a previous Codebase created with a "full_root" argument. Previously, the
-  missing paths of a "full_root" Codebase were kept unchanged. 
+  missing paths of a "full_root" Codebase were kept unchanged.
   Note that the VirtualCodebase has always ignored the "full_root" argument.
 
 - The Codebase and VirtualCodebase are now iterable. Iterating on a codebase
@@ -80,7 +84,7 @@ Other changes:
 
 - Remove Python upper version limit.
 - Merge latest skeleton
-- fileutils.parent_directory() now accepts a "with_trail" argument. 
+- fileutils.parent_directory() now accepts a "with_trail" argument.
   The returned directory has a trailing path separator unless with_trail is False.
   The default is True and the default behaviour is unchanged.
 

src/commoncode/paths.py

@@ -26,7 +26,7 @@
 # Build OS-portable and safer paths
 
 
-def safe_path(path, posix=False, preserve_spaces=False):
+def safe_path(path, posix=False, preserve_spaces=False, posix_only=False):
     """
     Convert `path` to a safe and portable POSIX path usable on multiple OSes.
     The returned path is an ASCII-only byte string, resolved for relative
@@ -52,7 +52,13 @@ def safe_path(path, posix=False, preserve_spaces=False):
     _pathmod, path_sep = path_handlers(path, posix)
 
     segments = [s.strip() for s in path.split(path_sep) if s.strip()]
-    segments = [portable_filename(s, preserve_spaces=preserve_spaces) for s in segments]
+    segments = [
+        portable_filename(
+            s,
+            preserve_spaces=preserve_spaces,
+            posix_only=posix_only
+        ) for s in segments
+    ]
 
     if not segments:
         return '_'
@@ -134,17 +140,34 @@ def resolve(path, posix=True):
     return path
 
 
-legal_punctuation = r"!\#$%&\(\)\+,\-\.;\=@\[\]_\{\}\~"
-legal_spaces = r" "
-legal_chars = r'A-Za-z0-9' + legal_punctuation
+legal_punctuation = r'!\#$%&\(\)\+,\-\.;\=@\[\]_\{\}\~'
+legal_spaces = r' '
+legal_alphanumeric = r'A-Za-z0-9'
+legal_chars = legal_alphanumeric + legal_punctuation
 legal_chars_inc_spaces = legal_chars + legal_spaces
 illegal_chars_re = r'[^' + legal_chars + r']'
 illegal_chars_exc_spaces_re = r'[^' + legal_chars_inc_spaces + r']'
 replace_illegal_chars = re.compile(illegal_chars_re).sub
 replace_illegal_chars_exc_spaces = re.compile(illegal_chars_exc_spaces_re).sub
 
 
-def portable_filename(filename, preserve_spaces=False):
+posix_legal_punctuation = r'<:"/>\|\*\^\\\'`\?' + legal_punctuation
+posix_legal_chars = legal_alphanumeric + posix_legal_punctuation
+posix_legal_chars_inc_spaces = posix_legal_chars + legal_spaces
+posix_illegal_chars_re = r'[^' + posix_legal_chars + r']'
+posix_illegal_chars_exc_spaces_re = r'[^' + posix_legal_chars_inc_spaces + r']'
+replace_illegal_posix_chars = re.compile(posix_illegal_chars_re).sub
+replace_illegal_posix_chars_exc_spaces = re.compile(posix_illegal_chars_exc_spaces_re).sub
+
+
+ILLEGAL_WINDOWS_NAMES = set([
+    'com1', 'com2', 'com3', 'com4', 'com5', 'com6', 'com7', 'com8', 'com9',
+    'lpt1', 'lpt2', 'lpt3', 'lpt4', 'lpt5', 'lpt6', 'lpt7', 'lpt8', 'lpt9',
+    'aux', 'con', 'nul', 'prn'
+])
+
+
+def portable_filename(filename, preserve_spaces=False, posix_only=False):
     """
     Return a new name for `filename` that is portable across operating systems.
 
@@ -170,22 +193,21 @@ def portable_filename(filename, preserve_spaces=False):
     if not filename:
         return '_'
 
-    if preserve_spaces:
-        filename = replace_illegal_chars_exc_spaces('_', filename)
+    if posix_only:
+        if preserve_spaces:
+            filename = replace_illegal_posix_chars_exc_spaces('_', filename)
+        else:
+            filename = replace_illegal_posix_chars('_', filename)
     else:
-        filename = replace_illegal_chars('_', filename)
-
-    # these are illegal both upper and lowercase and with or without an extension
-    # we insert an underscore after the base name.
-    windows_illegal_names = set([
-        'com1', 'com2', 'com3', 'com4', 'com5', 'com6', 'com7', 'com8', 'com9',
-        'lpt1', 'lpt2', 'lpt3', 'lpt4', 'lpt5', 'lpt6', 'lpt7', 'lpt8', 'lpt9',
-        'aux', 'con', 'nul', 'prn'
-    ])
+        if preserve_spaces:
+            filename = replace_illegal_chars_exc_spaces('_', filename)
+        else:
+            filename = replace_illegal_chars('_', filename)
 
-    basename, dot, extension = filename.partition('.')
-    if basename.lower() in windows_illegal_names:
-        filename = ''.join([basename, '_', dot, extension])
+    if not posix_only:
+        basename, dot, extension = filename.partition('.')
+        if basename.lower() in ILLEGAL_WINDOWS_NAMES:
+            filename = ''.join([basename, '_', dot, extension])
 
     # no name made only of dots.
     if set(filename) == set(['.']):
@@ -198,6 +220,7 @@ def portable_filename(filename, preserve_spaces=False):
 
     return filename
 
+
 #
 # paths comparisons, common prefix and suffix extraction
 #

tests/test_paths.py

@@ -93,6 +93,14 @@ def test_safe_path_posix_style_many_dots(self):
         expected = 'dotdot/dotdot/dotdot/webform.components.inc'
         assert test == expected
 
+    def test_safe_path_posix_only(self):
+        test_path = 'var/lib/dpkg/info/libgsm1:amd64.list'
+        test = paths.safe_path(test_path)
+        expected = 'var/lib/dpkg/info/libgsm1_amd64.list'
+        assert test == expected
+        test = paths.safe_path(test_path, posix_only=True)
+        assert test == test_path
+
     def test_resolve_mixed_slash(self):
         test = paths.resolve('C:\\..\\./drupal.js')
         expected = 'C/drupal.js'
@@ -140,6 +148,24 @@ def test_portable_filename(self):
         expected = 'This_contain_UMLAUT_umlauts.txt'
         assert paths.portable_filename(u'This contain UMLAUT \xfcml\xe4uts.txt') == expected
 
+        # Check to see if illegal Windows filenames are properly handled
+        for illegal_window_name in paths.ILLEGAL_WINDOWS_NAMES:
+            # Rename files with names that are illegal on Windows
+            expected = f'{illegal_window_name}_'
+            assert paths.portable_filename(illegal_window_name) == expected
+
+            # Allow files with names that are illegal on Windows
+            assert paths.portable_filename(illegal_window_name, posix_only=True) == illegal_window_name
+
+        # Check to see if the posix_only option does and does not replace
+        # punctuation characters that are illegal in Windows filenames
+        for valid_posix_path_char in paths.posix_legal_punctuation:
+            test_name = f'test{valid_posix_path_char}'
+            assert paths.portable_filename(test_name, posix_only=True) == test_name
+            if valid_posix_path_char not in paths.legal_punctuation:
+                expected = f'test_'
+                assert paths.portable_filename(test_name) == expected
+
 
 class TestCommonPath(TestCase):