Do not pass unicode strings into os.walk calls, since it can cause

internal os.path.join calls to raise UnicodeDecodeError.

svn path=/main/trunk/; revision=13926

8 files changed, 73 insertions, 5 deletions

diff --git a/pym/portage/__init__.py b/pym/portage/__init__.py
index a152293d8..a959207c4 100644
--- a/pym/portage/__init__.py
+++ b/pym/portage/__init__.py
@@ -4901,6 +4901,11 @@ def digestcheck(myfiles, mysettings, strict=0, justmanifest=0):
 	""" epatch will just grab all the patches out of a directory, so we have to
 	make sure there aren't any foreign files that it might grab."""
 	filesdir = os.path.join(pkgdir, "files")
+	if isinstance(filesdir, unicode):
+		# Avoid UnicodeDecodeError raised from
+		# os.path.join when called by os.walk.
+		filesdir = filesdir.encode('utf_8', 'replace')
+
 	for parent, dirs, files in os.walk(filesdir):
 		for d in dirs:
 			if d.startswith(".") or d == "CVS":
@@ -5158,6 +5163,12 @@ def _post_src_install_uid_fix(mysettings):
 		os.system("chflags -R nosunlnk,nouunlnk %s 2>/dev/null" % \
 			(_shell_quote(mysettings["D"]),))
 
+	destdir = mysettings["D"]
+	if isinstance(destdir, unicode):
+		# Avoid UnicodeDecodeError raised from
+		# os.path.join when called by os.walk.
+		destdir = destdir.encode('utf_8', 'replace')
+
 	for parent, dirs, files in os.walk(mysettings["D"]):
 		for fname in chain(dirs, files):
 			fpath = os.path.join(parent, fname)
diff --git a/pym/portage/cache/ebuild_xattr.py b/pym/portage/cache/ebuild_xattr.py
index 83ed8a783..98e848651 100644
--- a/pym/portage/cache/ebuild_xattr.py
+++ b/pym/portage/cache/ebuild_xattr.py
@@ -153,7 +153,14 @@ class database(fs_template.FsBased):
 		return os.path.exists(self.__get_path(cpv))
 
 	def __iter__(self):
-		for root,dirs,files in os.walk(self.portdir):
+
+		portdir = self.portdir
+		if isinstance(portdir, unicode):
+			# Avoid UnicodeDecodeError raised from
+			# os.path.join when called by os.walk.
+			portdir = portdir.encode('utf_8', 'replace')
+
+		for root, dirs, files in os.walk(portdir):
 			for file in files:
 				if file[-7:] == '.ebuild':
 					cat = os.path.basename(os.path.dirname(root))
diff --git a/pym/portage/dbapi/vartree.py b/pym/portage/dbapi/vartree.py
index b9dba438b..c6374cdf9 100644
--- a/pym/portage/dbapi/vartree.py
+++ b/pym/portage/dbapi/vartree.py
@@ -3221,6 +3221,12 @@ class dblink(object):
 		myfilelist = []
 		mylinklist = []
 		paths_with_newlines = []
+
+		if isinstance(srcroot, unicode):
+			# Avoid UnicodeDecodeError raised from
+			# os.path.join when called by os.walk.
+			srcroot = srcroot.encode('utf_8', 'replace')
+
 		srcroot_len = len(srcroot)
 		def onerror(e):
 			raise
diff --git a/pym/portage/env/loaders.py b/pym/portage/env/loaders.py
index 7b4d72721..854304125 100644
--- a/pym/portage/env/loaders.py
+++ b/pym/portage/env/loaders.py
@@ -39,6 +39,12 @@ def RecursiveFileLoader(filename):
 	@rtype: list
 	@returns: List of files to process
 	"""
+
+	if isinstance(filename, unicode):
+		# Avoid UnicodeDecodeError raised from
+		# os.path.join when called by os.walk.
+		filename = filename.encode('utf_8', 'replace')
+
 	try:
 		st = os.stat(filename)
 	except OSError:
diff --git a/pym/portage/manifest.py b/pym/portage/manifest.py
index 80a0c1642..a717d5841 100644
--- a/pym/portage/manifest.py
+++ b/pym/portage/manifest.py
@@ -306,7 +306,14 @@ class Manifest(object):
 		cpvlist = []
 		pn = os.path.basename(self.pkgdir.rstrip(os.path.sep))
 		cat = self._pkgdir_category()
-		for pkgdir, pkgdir_dirs, pkgdir_files in os.walk(self.pkgdir):
+
+		pkgdir = self.pkgdir
+		if isinstance(pkgdir, unicode):
+			# Avoid UnicodeDecodeError raised from
+			# os.path.join when called by os.walk.
+			pkgdir = pkgdir.encode('utf_8', 'replace')
+
+		for pkgdir, pkgdir_dirs, pkgdir_files in os.walk(pkgdir):
 			break
 		for f in pkgdir_files:
 			if f[:1] == ".":
@@ -334,8 +341,15 @@ class Manifest(object):
 				continue
 			self.fhashdict[mytype][f] = perform_multiple_checksums(self.pkgdir+f, self.hashes)
 		recursive_files = []
-		cut_len = len(os.path.join(self.pkgdir, "files") + os.sep)
-		for parentdir, dirs, files in os.walk(os.path.join(self.pkgdir, "files")):
+
+		pkgdir = self.pkgdir
+		if isinstance(pkgdir, unicode):
+			# Avoid UnicodeDecodeError raised from
+			# os.path.join when called by os.walk.
+			pkgdir = pkgdir.encode('utf_8', 'replace')
+
+		cut_len = len(os.path.join(pkgdir, "files") + os.sep)
+		for parentdir, dirs, files in os.walk(os.path.join(pkgdir, "files")):
 			for f in files:
 				full_path = os.path.join(parentdir, f)
 				recursive_files.append(full_path[cut_len:])
diff --git a/pym/portage/sets/files.py b/pym/portage/sets/files.py
index f4ecbae7c..f0b6fed58 100644
--- a/pym/portage/sets/files.py
+++ b/pym/portage/sets/files.py
@@ -125,9 +125,22 @@ class StaticFileSet(EditablePackageSet):
 				directory = self._repopath_sub.sub(trees["porttree"].dbapi.treemap[match.groupdict()["reponame"]], directory)
 			except KeyError:
 				raise SetConfigError(_("Could not find repository '%s'") % match.groupdict()["reponame"])
+
+		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')
+
 		if os.path.isdir(directory):
 			directory = normalize_path(directory)
 			for parent, dirs, files in os.walk(directory):
+				if not isinstance(parent, unicode):
+					parent = unicode(parent,
+						encoding='utf_8', errors='replace')
 				for d in dirs[:]:
 					if d[:1] == '.':
 						dirs.remove(d)
@@ -140,7 +153,7 @@ class StaticFileSet(EditablePackageSet):
 					if filename.endswith(".metadata"):
 						continue
 					filename = os.path.join(parent,
-						filename)[1 + len(directory):]
+						filename)[1 + len(directory_unicode):]
 					myname = name_pattern.replace("$name", filename)
 					myname = myname.replace("${name}", filename)
 					rValue[myname] = StaticFileSet(
diff --git a/pym/portage/update.py b/pym/portage/update.py
index c7001803a..4e309e401 100644
--- a/pym/portage/update.py
+++ b/pym/portage/update.py
@@ -146,6 +146,12 @@ def update_config_files(config_root, protect, protect_mask, update_iter):
 	protect - list of paths from CONFIG_PROTECT
 	protect_mask - list of paths from CONFIG_PROTECT_MASK
 	update_iter - list of update commands as returned from parse_updates()"""
+
+	if isinstance(config_root, unicode):
+		# Avoid UnicodeDecodeError raised from
+		# os.path.join when called by os.walk.
+		config_root = config_root.encode('utf_8', 'replace')
+
 	config_root = normalize_path(config_root)
 	update_files = {}
 	file_contents = {}
diff --git a/pym/portage/util.py b/pym/portage/util.py
index f1bff1d40..5b860218d 100644
--- a/pym/portage/util.py
+++ b/pym/portage/util.py
@@ -780,6 +780,11 @@ def apply_recursive_permissions(top, uid=-1, gid=-1,
 	Returns True if all permissions are applied and False if some are left
 	unapplied."""
 
+	if isinstance(top, unicode):
+		# Avoid UnicodeDecodeError raised from
+		# os.path.join when called by os.walk.
+		top = top.encode('utf_8', 'replace')
+
 	if onerror is None:
 		# Default behavior is to dump errors to stderr so they won't
 		# go unnoticed.  Callers can pass in a quiet instance.