Update imports to import portage.os (with unicode wrappers), and use

_unicode_encode() and _unicode_decode() where appropriate.

svn path=/main/trunk/; revision=14052

7 files changed, 40 insertions, 25 deletions

diff --git a/pym/portage/__init__.py b/pym/portage/__init__.py
index 87863abdb..9d2c03317 100644
--- a/pym/portage/__init__.py
+++ b/pym/portage/__init__.py
@@ -197,6 +197,9 @@ class _unicode_module_wrapper(object):
 			result = _unicode_func_wrapper(result, encoding=encoding)
 		return result
 
+# Assume utf_8 fs encoding everywhere except in merge code.
+_fs_encoding = 'utf_8'
+
 _merge_encoding = sys.getfilesystemencoding()
 
 import os as _os
diff --git a/pym/portage/sets/__init__.py b/pym/portage/sets/__init__.py
index 2d85007ec..f52c7fa1a 100644
--- a/pym/portage/sets/__init__.py
+++ b/pym/portage/sets/__init__.py
@@ -5,7 +5,7 @@
 __all__ = ["SETPREFIX", "get_boolean", "SetConfigError",
 	"SetConfig", "load_default_config"]
 
-import os
+from portage import os
 from ConfigParser import SafeConfigParser, NoOptionError
 from portage import load_mod
 from portage.const import USER_CONFIG_PATH, GLOBAL_CONFIG_PATH
diff --git a/pym/portage/sets/dbapi.py b/pym/portage/sets/dbapi.py
index 0f82bd1c9..5cd6c6ba1 100644
--- a/pym/portage/sets/dbapi.py
+++ b/pym/portage/sets/dbapi.py
@@ -2,6 +2,9 @@
 # Distributed under the terms of the GNU General Public License v2
 # $Id$
 
+import time
+
+from portage import os
 from portage.versions import catpkgsplit, catsplit, pkgcmp, best
 from portage.dep import Atom
 from portage.localization import _
@@ -280,7 +283,6 @@ class AgeSet(EverythingSet):
 		self._age = age
 
 	def _filter(self, atom):
-		import time, os
 	
 		cpv = self._db.match(atom)[0]
 		path = self._db.getpath(cpv, filename="COUNTER")
diff --git a/pym/portage/sets/files.py b/pym/portage/sets/files.py
index b5ad9a05d..acf556b7f 100644
--- a/pym/portage/sets/files.py
+++ b/pym/portage/sets/files.py
@@ -3,10 +3,13 @@
 # $Id$
 
 import errno
-import os
 import re
 from itertools import chain
 
+from portage import os
+from portage import _fs_encoding
+from portage import _unicode_decode
+from portage import _unicode_encode
 from portage.util import grabfile, write_atomic, ensure_dirs, normalize_path
 from portage.const import PRIVATE_PATH, USER_CONFIG_PATH
 from portage.localization import _
@@ -126,35 +129,43 @@ class StaticFileSet(EditablePackageSet):
 			except KeyError:
 				raise SetConfigError(_("Could not find repository '%s'") % match.groupdict()["reponame"])
 
+		try:
+			directory = _unicode_decode(directory,
+				encoding=_fs_encoding, errors='strict')
+			# Now verify that we can also encode it.
+			_unicode_encode(directory,
+				encoding=_fs_encoding, errors='strict')
+		except UnicodeError:
+			directory = _unicode_decode(directory,
+				encoding=_fs_encoding, errors='replace')
+			raise SetConfigError(
+				_("Directory path contains invalid character(s) for encoding '%s': '%s'") \
+				% (_fs_encoding, directory))
+
 		if os.path.isdir(directory):
 			directory = normalize_path(directory)
 
-			if isinstance(directory, unicode):
-				# Avoid UnicodeDecodeError raised from
-				# os.path.join when called by os.walk.
-				directory_unicode = directory
-				directory = directory.encode('utf_8', 'replace')
-			else:
-				directory_unicode = unicode(directory,
-					encoding='utf_8', errors='replace')
-
 			for parent, dirs, files in os.walk(directory):
-				if not isinstance(parent, unicode):
-					parent = unicode(parent,
-						encoding='utf_8', errors='replace')
+				try:
+					parent = _unicode_decode(parent,
+						encoding=_fs_encoding, errors='strict')
+				except UnicodeDecodeError:
+					continue
 				for d in dirs[:]:
 					if d[:1] == '.':
 						dirs.remove(d)
 				for filename in files:
-					if not isinstance(filename, unicode):
-						filename = unicode(filename,
-							encoding='utf_8', errors='replace')
+					try:
+						filename = _unicode_decode(filename,
+							encoding=_fs_encoding, errors='strict')
+					except UnicodeDecodeError:
+						continue
 					if filename[:1] == '.':
 						continue
 					if filename.endswith(".metadata"):
 						continue
 					filename = os.path.join(parent,
-						filename)[1 + len(directory_unicode):]
+						filename)[1 + len(directory):]
 					myname = name_pattern.replace("$name", filename)
 					myname = myname.replace("${name}", filename)
 					rValue[myname] = StaticFileSet(
diff --git a/pym/portage/sets/profiles.py b/pym/portage/sets/profiles.py
index 9c1dcbe86..da3a25e86 100644
--- a/pym/portage/sets/profiles.py
+++ b/pym/portage/sets/profiles.py
@@ -3,7 +3,8 @@
 # $Id$
 
 import logging
-import os
+
+from portage import os
 from portage.util import grabfile_package, stack_lists
 from portage.sets.base import PackageSet
 from portage.sets import get_boolean
diff --git a/pym/portage/sets/security.py b/pym/portage/sets/security.py
index 12503d9a5..115c0f3bd 100644
--- a/pym/portage/sets/security.py
+++ b/pym/portage/sets/security.py
@@ -2,9 +2,7 @@
 # Distributed under the terms of the GNU General Public License v2
 # $Id$
 
-import os
 import portage.glsa as glsa
-from portage.util import grabfile, write_atomic
 from portage.sets.base import PackageSet
 from portage.versions import catpkgsplit, pkgcmp
 from portage.sets import get_boolean
diff --git a/pym/portage/sets/shell.py b/pym/portage/sets/shell.py
index 25ccb03c7..658353ef2 100644
--- a/pym/portage/sets/shell.py
+++ b/pym/portage/sets/shell.py
@@ -3,8 +3,9 @@
 # $Id$
 
 import subprocess
-import os
 
+from portage import os
+from portage import _unicode_decode
 from portage.sets.base import PackageSet
 from portage.sets import SetConfigError
 
@@ -35,8 +36,7 @@ class CommandOutputSet(PackageSet):
 		pipe = subprocess.Popen(self._command, stdout=subprocess.PIPE, shell=True)
 		stdout, stderr = pipe.communicate()
 		if pipe.wait() == os.EX_OK:
-			self._setAtoms(unicode(stdout,
-				encoding='utf_8', errors='replace').splitlines())
+			self._setAtoms(_unicode_decode(stdout).splitlines())
 
 	def singleBuilder(self, options, settings, trees):
 		if not "command" in options: