uri2fsn: support surrogates. Fixes #6

quodlibet · Jul 26, 2017 · 7878b6c · 7878b6c
1 parent 2a0e095
commit 7878b6c
Show file tree

Hide file tree

Showing 4 changed files with 22 additions and 21 deletions.
diff --git a/senf/_compat.py b/senf/_compat.py
@@ -30,8 +30,8 @@
 if PY2:
     from urlparse import urlparse, urlunparse
     urlparse, urlunparse
-    from urllib import pathname2url, url2pathname, quote, unquote
-    pathname2url, url2pathname, quote, unquote
+    from urllib import quote, unquote
+    quote, unquote
 
     from StringIO import StringIO
     BytesIO = StringIO
@@ -45,8 +45,6 @@
 elif PY3:
     from urllib.parse import urlparse, quote, unquote, urlunparse
     urlparse, quote, unquote, urlunparse
-    from urllib.request import pathname2url, url2pathname
-    pathname2url, url2pathname
 
     from io import StringIO
     StringIO = StringIO

diff --git a/senf/_fsnative.py b/senf/_fsnative.py
@@ -26,8 +26,7 @@
 import codecs
 
 from . import _winapi as winapi
-from ._compat import text_type, PY3, PY2, url2pathname, urlparse, quote, \
-    unquote, urlunparse
+from ._compat import text_type, PY3, PY2, urlparse, quote, unquote, urlunparse
 
 
 is_win = os.name == "nt"
@@ -551,7 +550,18 @@ def uri2fsn(uri):
     uri = urlunparse(parsed)[7:]
 
     if is_win:
-        path = url2pathname(uri)
+        try:
+            drive, rest = uri.split(":", 1)
+        except ValueError:
+            path = ""
+            rest = uri.replace("/", "\\")
+        else:
+            path = drive[-1] + ":"
+            rest = rest.replace("/", "\\")
+        if PY2:
+            path += unquote(rest)
+        else:
+            path += unquote(rest, errors="surrogatepass")
         if netloc:
             path = "\\\\" + path
         if PY2:
@@ -560,11 +570,12 @@ def uri2fsn(uri):
             raise ValueError("embedded null")
         return path
     else:
-        path = unquote(uri)
+        if PY2:
+            path = unquote(uri)
+        else:
+            path = unquote(uri, errors="surrogateescape")
         if "\x00" in path:
             raise ValueError("embedded null")
-        if PY3:
-            path = fsnative(path)
         return path
 
 

diff --git a/tests/test_api.py b/tests/test_api.py
@@ -888,8 +888,7 @@ def test_uri2fsn():
             fsnative(u"/bla:foo@NOPE.com")
         assert uri2fsn("file:///bla?x#b") == fsnative(u"/bla?x#b")
     else:
-        # FIXME
-        # assert uri2fsn("file:///C:/%ED%A0%80") == fsnative(u"C:\\\ud800")
+        assert uri2fsn("file:///C:/%ED%A0%80") == fsnative(u"C:\\\ud800")
         assert uri2fsn("file:///C:/%20") == "C:\\ "
         assert uri2fsn("file:NOPE") == "\\NOPE"
         assert uri2fsn("file:/NOPE") == "\\NOPE"
@@ -967,6 +966,7 @@ def test_uri_roundtrip():
     else:
         path = path2fsn(b"/foo-\xe1\x88\xb4")
 
+        assert uri2fsn(fsn2uri(path2fsn(b"/\x80"))) == path2fsn(b"/\x80")
         assert uri2fsn(fsn2uri(fsnative(u"/foo"))) == "/foo"
         assert uri2fsn(fsn2uri(path)) == path
         assert isinstance(uri2fsn(fsn2uri(path)), fsnative)

diff --git a/tests/test_hypo.py b/tests/test_hypo.py
@@ -44,15 +44,7 @@ def test_anything(pathlike):
     abspath = os.path.abspath(
         fsn.replace(sep, fsnative()).replace(altsep or sep, fsnative()))
     if os.path.isabs(abspath):
-        try:
-            if isinstance(abspath, text_type):
-                abspath.encode("utf-8")
-        except UnicodeEncodeError:
-            # FIXME: url2pathname can't handle surrogates in Python 3,
-            # get rid of it
-            pass
-        else:
-            assert uri2fsn(fsn2uri(abspath)) == abspath
+        assert uri2fsn(fsn2uri(abspath)) == abspath
 
     try:
         # never raises ValueError/TypError