gh-81790: support "UNC" device paths in ntpath.splitdrive()

python · zooba · Jun 10, 2022 · Apr 24, 2022 · Apr 24, 2022 · Apr 24, 2022
commit c97bfbf31fd300d50d60d9e0559d499f11120651
diff --git a/Lib/ntpath.py b/Lib/ntpath.py
@@ -146,17 +146,30 @@ def splitdrive(p):
             sep = b'\\'
             altsep = b'/'
             colon = b':'
+            unc2 = b'\\\\'
+            unc4 = b'\\\\?\\'
+            unc8 = b'\\\\?\\UNC\\'
         else:
             sep = '\\'
             altsep = '/'
             colon = ':'
+            unc2 = '\\\\'
+            unc4 = '\\\\?\\'
+            unc8 = '\\\\?\\UNC\\'
         normp = p.replace(altsep, sep)
-        if (normp[0:2] == sep*2) and (normp[2:3] != sep):
+        if normp[:8] == unc8:
+            normp = sep * 8 + normp[8:]
+            offset = 6
+        elif normp[:4] == unc4:
+            offset = 4
+        else:
+            offset = 0
+        if (normp[offset:offset + 2] == unc2) and (normp[offset + 2:offset + 3] != sep):
             # is a UNC path:
             # vvvvvvvvvvvvvvvvvvvv drive letter or UNC path
             # \\machine\mountpoint\directory\etc\...
             #           directory ^^^^^^^^^^^^^^^
-            index = normp.find(sep, 2)
+            index = normp.find(sep, offset + 2)
             if index == -1:
                 return p[:0], p
             index2 = normp.find(sep, index + 1)
@@ -167,8 +180,8 @@ def splitdrive(p):
             if index2 == -1:
                 index2 = len(p)
             return p[:index2], p[index2:]
-        if normp[1:2] == colon:
-            return p[:2], p[2:]
+        if normp[offset + 1:offset + 2] == colon:
+            return p[:offset + 2], p[offset + 2:]
     return p[:0], p
 
 

@@ -120,9 +120,6 @@ class _WindowsFlavour(_Flavour):
 
     is_supported = (os.name == 'nt')
 
-    drive_letters = set('abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ')
-    ext_namespace_prefix = '\\\\?\\'
-
     reserved_names = (
         {'CON', 'PRN', 'AUX', 'NUL', 'CONIN$', 'CONOUT$'} |
         {'COM%s' % c for c in '123456789\xb9\xb2\xb3'} |
@@ -145,43 +142,11 @@ class _WindowsFlavour(_Flavour):
     #   even with the '\\?\' prefix.
 
     def splitroot(self, part, sep=sep):
-        first = part[0:1]
-        second = part[1:2]
-        if (second == sep and first == sep):
-            # XXX extended paths should also disable the collapsing of "."
-            # components (according to MSDN docs).
-            prefix, part = self._split_extended_path(part)
-            first = part[0:1]
-            second = part[1:2]
+        drv, rest = self.pathmod.splitdrive(part)
+        if drv[:1] == sep or rest[:1] == sep:
+            return drv, sep, rest.lstrip(sep)
         else:
-            prefix = ''
-        third = part[2:3]
-        if (second == sep and first == sep and third != sep):
-            # is a UNC path:
-            # vvvvvvvvvvvvvvvvvvvvv root
-            # \\machine\mountpoint\directory\etc\...
-            #            directory ^^^^^^^^^^^^^^
-            index = part.find(sep, 2)
-            if index != -1:
-                index2 = part.find(sep, index + 1)
-                # a UNC path can't have two slashes in a row
-                # (after the initial two)
-                if index2 != index + 1:
-                    if index2 == -1:
-                        index2 = len(part)
-                    if prefix:
-                        return prefix + part[1:index2], sep, part[index2+1:]
-                    else:
-                        return part[:index2], sep, part[index2+1:]
-        drv = root = ''
-        if second == ':' and first in self.drive_letters:
-            drv = part[:2]
-            part = part[2:]
-            first = third
-        if first == sep:
-            root = first
-            part = part.lstrip(sep)
-        return prefix + drv, root, part
+            return drv, '', rest
 
     def casefold(self, s):
         return s.lower()
@@ -192,16 +157,6 @@ def casefold_parts(self, parts):
     def compile_pattern(self, pattern):
         return re.compile(fnmatch.translate(pattern), re.IGNORECASE).fullmatch
 
-    def _split_extended_path(self, s, ext_prefix=ext_namespace_prefix):
-        prefix = ''
-        if s.startswith(ext_prefix):
-            prefix = s[:4]
-            s = s[4:]
-            if s.startswith('UNC\\'):
-                prefix += s[:3]
-                s = '\\' + s[3:]
-        return prefix, s
-
     def is_reserved(self, parts):
         # NOTE: the rules for reserved names seem somewhat complicated
         # (e.g. r"..\NUL" is reserved but not r"foo\NUL" if "foo" does not