When writing the manifest under Python 3, skip filenames that cannot be encoded to UTF-8.

--HG--
branch : distribute
extra : rebase_source : f1b439267fce37754aac49af15a9e26346950a26

2 files changed, 14 insertions, 2 deletions

diff --git a/setuptools/command/egg_info.py b/setuptools/command/egg_info.py
index 9955c8ef..6e3d67af 100755
--- a/setuptools/command/egg_info.py
+++ b/setuptools/command/egg_info.py
@@ -323,6 +323,18 @@ class manifest_maker(sdist):
         by 'add_defaults()' and 'read_template()') to the manifest file
         named by 'self.manifest'.
         """
+        # The manifest must be UTF-8 encodable. See #303.
+        if sys.version_info >= (3,):
+            files = []
+            for file in self.filelist.files:
+                try:
+                    file.encode("utf-8")
+                except UnicodeEncodeError:
+                    log.warn("'%s' not UTF-8 encodable -- skipping" % file)
+                else:
+                    files.append(file)
+            self.filelist.files = files
+
         files = self.filelist.files
         if os.sep!='/':
             files = [f.replace(os.sep,'/') for f in files]
@@ -360,7 +372,7 @@ def write_file (filename, contents):
     """
     contents = "\n".join(contents)
     if sys.version_info >= (3,):
-        contents = contents.encode("utf-8", "surrogateescape")
+        contents = contents.encode("utf-8")
     f = open(filename, "wb")        # always write POSIX-style manifest
     f.write(contents)
     f.close()
diff --git a/setuptools/command/sdist.py b/setuptools/command/sdist.py
index 42558143..d5259c2b 100755
--- a/setuptools/command/sdist.py
+++ b/setuptools/command/sdist.py
@@ -283,7 +283,7 @@ class sdist(_sdist):
         manifest = open(self.manifest, 'rbU')
         for line in manifest:
             if sys.version_info >= (3,):
-                line = line.decode('UTF-8', 'surrogateescape')
+                line = line.decode('UTF-8')
             # ignore comments and blank lines
             line = line.strip()
             if line.startswith('#') or not line: