Migrate from codecs.open() to io.open().

The io.open() function is the same as the built-in open() function in
python3, and its implementation is optimized in python-2.7 and later.
In addition to the possible performance improvement, this also allows
us to avoid any future compatibility issues with codecs.open() that
may arise if it is delegated to the built-in open() function as
discussed in PEP 400.

The main caveat involved with io.open() is that TextIOWrapper.write()
raises TypeError if given raw bytes, unlike the streams returned from
codecs.open(). This is mainly an issue for python2 since literal
strings are raw bytes. We handle this by wrapping TextIOWrapper.write()
arguments with our _unicode_decode() function. Also, the
atomic_ofstream class overrides the write() method in python2 so that
it performs automatic coercion to unicode when necessary.

1 files changed, 23 insertions, 18 deletions

diff --git a/bin/egencache b/bin/egencache
index 5307cd5a2..1b4265df1 100755
--- a/bin/egencache
+++ b/bin/egencache
@@ -20,7 +20,7 @@ try:
 except KeyboardInterrupt:
 	sys.exit(128 + signal.SIGINT)
 
-import codecs
+import io
 import logging
 import optparse
 import subprocess
@@ -391,10 +391,10 @@ class GenUseLocalDesc(object):
 					output = open(_unicode_encode(desc_path,
 						encoding=_encodings['fs'], errors='strict'), 'r+b')
 				else:
-					output = codecs.open(_unicode_encode(desc_path,
+					output = io.open(_unicode_encode(desc_path,
 						encoding=_encodings['fs'], errors='strict'),
 						mode='w', encoding=_encodings['repo.content'],
-						errors='replace')
+						errors='backslashreplace')
 			except IOError as e:
 				if not self._preserve_comments or \
 					os.path.isfile(desc_path):
@@ -413,10 +413,10 @@ class GenUseLocalDesc(object):
 					level=logging.WARNING, noiselevel=-1)
 				self._preserve_comments = False
 				try:
-					output = codecs.open(_unicode_encode(desc_path,
+					output = io.open(_unicode_encode(desc_path,
 						encoding=_encodings['fs'], errors='strict'),
 						mode='w', encoding=_encodings['repo.content'],
-						errors='replace')
+						errors='backslashreplace')
 				except IOError as e:
 					writemsg_level(
 						"ERROR: failed to open output file %s: %s\n" \
@@ -437,18 +437,18 @@ class GenUseLocalDesc(object):
 
 			# Finished probing comments in binary mode, now append
 			# in text mode.
-			output = codecs.open(_unicode_encode(desc_path,
+			output = io.open(_unicode_encode(desc_path,
 				encoding=_encodings['fs'], errors='strict'),
 				mode='a', encoding=_encodings['repo.content'],
-				errors='replace')
-			output.write('\n')
+				errors='backslashreplace')
+			output.write(_unicode_decode('\n'))
 		else:
-			output.write('''
+			output.write(_unicode_decode('''
 # This file is deprecated as per GLEP 56 in favor of metadata.xml. Please add
 # your descriptions to your package's metadata.xml ONLY.
 # * generated automatically using egencache *
 
-'''.lstrip())
+'''.lstrip()))
 
 		# The cmp function no longer exists in python3, so we'll
 		# implement our own here under a slightly different name
@@ -522,7 +522,8 @@ class GenUseLocalDesc(object):
 								resatoms = sorted(reskeys, key=cmp_sort_key(atomcmp))
 								resdesc = resdict[reskeys[resatoms[-1]]]
 
-						output.write('%s:%s - %s\n' % (cp, flag, resdesc))
+						output.write(_unicode_decode(
+							'%s:%s - %s\n' % (cp, flag, resdesc)))
 
 		output.close()
 
@@ -609,9 +610,9 @@ class GenChangeLogs(object):
 
 	def generate_changelog(self, cp):
 		try:
-			output = codecs.open('ChangeLog',
+			output = io.open('ChangeLog',
 				mode='w', encoding=_encodings['repo.content'],
-				errors='replace')
+				errors='backslashreplace')
 		except IOError as e:
 			writemsg_level(
 				"ERROR: failed to open ChangeLog for %s: %s\n" % (cp,e,),
@@ -619,7 +620,7 @@ class GenChangeLogs(object):
 			self.returncode |= 2
 			return
 
-		output.write(('''
+		output.write(_unicode_decode('''
 # ChangeLog for %s
 # Copyright 1999-%s Gentoo Foundation; Distributed under the GPL v2
 # $Header: $
@@ -688,10 +689,11 @@ class GenChangeLogs(object):
 			# Reverse the sort order for headers.
 			for c in reversed(changed):
 				if c.startswith('+') and c.endswith('.ebuild'):
-					output.write('*%s (%s)\n' % (c[1:-7], date))
+					output.write(_unicode_decode(
+						'*%s (%s)\n' % (c[1:-7], date)))
 					wroteheader = True
 			if wroteheader:
-				output.write('\n')
+				output.write(_unicode_decode('\n'))
 
 			# strip '<cp>: ', '[<cp>] ', and similar
 			body[0] = re.sub(r'^\W*' + re.escape(cp) + r'\W+', '', body[0])
@@ -711,10 +713,13 @@ class GenChangeLogs(object):
 
 			# don't break filenames on hyphens
 			self._wrapper.break_on_hyphens = False
-			output.write(self._wrapper.fill('%s; %s %s:' % (date, author, ', '.join(changed))))
+			output.write(_unicode_decode(
+				self._wrapper.fill(
+				'%s; %s %s:' % (date, author, ', '.join(changed)))))
 			# but feel free to break commit messages there
 			self._wrapper.break_on_hyphens = True
-			output.write('\n%s\n\n' % '\n'.join([self._wrapper.fill(x) for x in body]))
+			output.write(_unicode_decode(
+				'\n%s\n\n' % '\n'.join(self._wrapper.fill(x) for x in body)))
 
 		output.close()