Bug #324191 - Add support for FEATURES=compress-build-logs. The causes

all build logs to be compressed while they are being written. Log file
names have an extension that is appropriate for the compression type.
Currently, only gzip(1) compression is supported, so build logs will
have a '.gz' extension when this feature is enabled.

17 files changed, 132 insertions, 171 deletions

diff --git a/bin/misc-functions.sh b/bin/misc-functions.sh
index 3a199bdf7..123edc0ec 100755
--- a/bin/misc-functions.sh
+++ b/bin/misc-functions.sh
@@ -474,10 +474,12 @@ install_qa_check() {
 		)
 		abort="no"
 		i=0
+		local grep_cmd=grep
+		[[ $PORTAGE_LOG_FILE = *.gz ]] && grep_cmd=zgrep
 		while [[ -n ${msgs[${i}]} ]] ; do
 			m=${msgs[$((i++))]}
 			# force C locale to work around slow unicode locales #160234
-			f=$(LC_ALL=C grep "${m}" "${PORTAGE_LOG_FILE}")
+			f=$(LC_ALL=C $grep_cmd "${m}" "${PORTAGE_LOG_FILE}")
 			if [[ -n ${f} ]] ; then
 				vecho -ne '\a\n'
 				eqawarn "QA Notice: Package has poor programming practices which may compile"
@@ -487,8 +489,10 @@ install_qa_check() {
 				abort="yes"
 			fi
 		done
+		local cat_cmd=cat
+		[[ $PORTAGE_LOG_FILE = *.gz ]] && cat_cmd=zcat
 		[[ $reset_debug = 1 ]] && set -x
-		f=$(cat "${PORTAGE_LOG_FILE}" | \
+		f=$($cat_cmd "${PORTAGE_LOG_FILE}" | \
 			"${PORTAGE_PYTHON:-/usr/bin/python}" "$PORTAGE_BIN_PATH"/check-implicit-pointer-usage.py || die "check-implicit-pointer-usage.py failed")
 		if [[ -n ${f} ]] ; then
 
diff --git a/man/make.conf.5 b/man/make.conf.5
index 0827db11a..74b24576b 100644
--- a/man/make.conf.5
+++ b/man/make.conf.5
@@ -231,6 +231,12 @@ A QA\-feature to ensure that a package doesn't overwrite files it doesn't own.
 The \fICOLLISION_IGNORE\fR variable can be used to selectively disable this
 feature. Also see the related \fIprotect\-owned\fR feature.
 .TP
+.B compress\-build\-logs
+The causes all build logs to be compressed while they are being written.
+Log file names have an extension that is appropriate for the compression
+type. Currently, only \fBgzip\fR(1) compression is supported, so build
+logs will have a '.gz' extension when this feature is enabled.
+.TP
 .B digest
 Autogenerate digests for packages when running the \fBemerge\fR(1) command.  If
 the \fIassume\-digests\fR feature is also enabled then existing SRC_URI digests
diff --git a/pym/_emerge/AbstractEbuildProcess.py b/pym/_emerge/AbstractEbuildProcess.py
index aca254191..8c87812f5 100644
--- a/pym/_emerge/AbstractEbuildProcess.py
+++ b/pym/_emerge/AbstractEbuildProcess.py
@@ -183,21 +183,11 @@ class AbstractEbuildProcess(SpawnProcess):
 		phase = self.phase
 		for line in lines:
 			eerror(line, phase=phase, key=self.settings.mycpv, out=out)
-		logfile = self.logfile
-		if logfile is None:
-			logfile = self.settings.get("PORTAGE_LOG_FILE")
 		msg = _unicode_decode(out.getvalue(),
 			encoding=_encodings['content'], errors='replace')
 		if msg:
-			if not self.background:
-				writemsg_stdout(msg, noiselevel=-1)
-			if logfile is not None:
-				log_file = codecs.open(_unicode_encode(logfile,
-					encoding=_encodings['fs'], errors='strict'),
-					mode='a', encoding=_encodings['content'],
-					errors='backslashreplace')
-				log_file.write(msg)
-				log_file.close()
+			self.scheduler.output(msg,
+				log_path=self.settings.get("PORTAGE_LOG_FILE"))
 
 	def _set_returncode(self, wait_retval):
 		SpawnProcess._set_returncode(self, wait_retval)
diff --git a/pym/_emerge/Binpkg.py b/pym/_emerge/Binpkg.py
index 66b7bd424..2201decaf 100644
--- a/pym/_emerge/Binpkg.py
+++ b/pym/_emerge/Binpkg.py
@@ -27,20 +27,8 @@ class Binpkg(CompositeTask):
 		"_image_dir", "_infloc", "_pkg_path", "_tree", "_verify")
 
 	def _writemsg_level(self, msg, level=0, noiselevel=0):
-
-		if not self.background:
-			portage.util.writemsg_level(msg,
-				level=level, noiselevel=noiselevel)
-
-		log_path = self.settings.get("PORTAGE_LOG_FILE")
-		if  log_path is not None:
-			f = codecs.open(_unicode_encode(log_path,
-				encoding=_encodings['fs'], errors='strict'),
-				mode='a', encoding=_encodings['content'], errors='replace')
-			try:
-				f.write(msg)
-			finally:
-				f.close()
+		self.scheduler.output(msg, level=level, noiselevel=noiselevel,
+			log_path=self.settings.get("PORTAGE_LOG_FILE"))
 
 	def _start(self):
 
@@ -146,9 +134,7 @@ class Binpkg(CompositeTask):
 
 		verifier = None
 		if self._verify:
-			logfile = None
-			if self.background:
-				logfile = self.settings.get("PORTAGE_LOG_FILE")
+			logfile = self.settings.get("PORTAGE_LOG_FILE")
 			verifier = BinpkgVerifier(background=self.background,
 				logfile=logfile, pkg=self.pkg)
 			self._start_task(verifier, self._verifier_exit)
diff --git a/pym/_emerge/BinpkgVerifier.py b/pym/_emerge/BinpkgVerifier.py
index 83a02c76c..02a942c94 100644
--- a/pym/_emerge/BinpkgVerifier.py
+++ b/pym/_emerge/BinpkgVerifier.py
@@ -7,7 +7,9 @@ import sys
 import portage
 from portage import os
 from portage import _encodings
+from portage import _unicode_decode
 from portage import _unicode_encode
+from portage import StringIO
 from portage.package.ebuild.fetch import _checksum_failure_temp_file
 import codecs
 
@@ -27,27 +29,10 @@ class BinpkgVerifier(AsynchronousTask):
 		rval = os.EX_OK
 		stdout_orig = sys.stdout
 		stderr_orig = sys.stderr
-		log_file = None
-		if self.background and self.logfile is not None:
-			if sys.hexversion >= 0x3000000:
-				# Since we are replacing the sys.std* streams,
-				# we need to use the normal open() function
-				# so that we get the right class (otherwise our
-				# code that expects the 'buffer' attribute
-				# will break).
-				log_file = open(_unicode_encode(self.logfile,
-					encoding=_encodings['fs'], errors='strict'),
-					mode='a', encoding=_encodings['content'],
-					errors='backslashreplace')
-			else:
-				# For python2, sys.std* are expected to be binary streams.
-				log_file = open(_unicode_encode(self.logfile,
-					encoding=_encodings['fs'], errors='strict'),
-					mode='ab')
+		out = portage.StringIO()
 		try:
-			if log_file is not None:
-				sys.stdout = log_file
-				sys.stderr = log_file
+			sys.stdout = out
+			sys.stderr = out
 			try:
 				bintree.digestCheck(pkg)
 			except portage.exception.FileNotFound:
@@ -69,7 +54,7 @@ class BinpkgVerifier(AsynchronousTask):
 			if rval == os.EX_OK:
 				# If this was successful, discard the log here since otherwise
 				# we'll get multiple logs for the same package.
-				if log_file is not None:
+				if self.logfile is not None:
 					try:
 						os.unlink(self.logfile)
 					except OSError:
@@ -83,8 +68,11 @@ class BinpkgVerifier(AsynchronousTask):
 		finally:
 			sys.stdout = stdout_orig
 			sys.stderr = stderr_orig
-			if log_file is not None:
-				log_file.close()
+
+		msg = _unicode_decode(out.getvalue(),
+			encoding=_encodings['content'], errors='replace')
+		if msg:
+			self.scheduler.output(msg, log_path=self.logfile)
 
 		self.returncode = rval
 		self.wait()
diff --git a/pym/_emerge/EbuildBuild.py b/pym/_emerge/EbuildBuild.py
index b8694df96..5c108fee4 100644
--- a/pym/_emerge/EbuildBuild.py
+++ b/pym/_emerge/EbuildBuild.py
@@ -218,19 +218,8 @@ class EbuildBuild(CompositeTask):
 		if self._issyspkg:
 			msg = ">>> This is a system package, " + \
 				"let's pack a rescue tarball.\n"
-
-			log_path = self.settings.get("PORTAGE_LOG_FILE")
-			if log_path is not None:
-				log_file = codecs.open(_unicode_encode(log_path,
-					encoding=_encodings['fs'], errors='strict'),
-					mode='a', encoding=_encodings['content'], errors='replace')
-				try:
-					log_file.write(msg)
-				finally:
-					log_file.close()
-
-			if not self.background:
-				portage.writemsg_stdout(msg, noiselevel=-1)
+			self.scheduler.output(msg,
+				log_path=self.settings.get("PORTAGE_LOG_FILE"))
 
 		packager = EbuildBinpkg(background=self.background, pkg=self.pkg,
 			scheduler=self.scheduler, settings=self.settings)
diff --git a/pym/_emerge/EbuildFetcher.py b/pym/_emerge/EbuildFetcher.py
index 7f5bc6df4..63e423771 100644
--- a/pym/_emerge/EbuildFetcher.py
+++ b/pym/_emerge/EbuildFetcher.py
@@ -158,19 +158,10 @@ class EbuildFetcher(SpawnProcess):
 		out = portage.StringIO()
 		for line in lines:
 			eerror(line, phase="unpack", key=self.pkg.cpv, out=out)
-		logfile = self.logfile
 		msg = _unicode_decode(out.getvalue(),
 			encoding=_encodings['content'], errors='replace')
 		if msg:
-			if not self.background:
-				writemsg_stdout(msg, noiselevel=-1)
-			if logfile is not None:
-				log_file = codecs.open(_unicode_encode(logfile,
-					encoding=_encodings['fs'], errors='strict'),
-					mode='a', encoding=_encodings['content'],
-					errors='backslashreplace')
-				log_file.write(msg)
-				log_file.close()
+			self.scheduler.output(msg, log_path=self.logfile)
 
 	def _set_returncode(self, wait_retval):
 		SpawnProcess._set_returncode(self, wait_retval)
diff --git a/pym/_emerge/EbuildPhase.py b/pym/_emerge/EbuildPhase.py
index 7fbc66849..d2d4dce3a 100644
--- a/pym/_emerge/EbuildPhase.py
+++ b/pym/_emerge/EbuildPhase.py
@@ -75,24 +75,11 @@ class EbuildPhase(CompositeTask):
 
 		if self.phase == "install":
 			out = portage.StringIO()
-			log_path = self.settings.get("PORTAGE_LOG_FILE")
-			log_file = None
-			if log_path is not None:
-				log_file = codecs.open(_unicode_encode(log_path,
-					encoding=_encodings['fs'], errors='strict'),
-					mode='a', encoding=_encodings['content'], errors='replace')
-			try:
-				_check_build_log(self.settings, out=out)
-				msg = _unicode_decode(out.getvalue(),
-					encoding=_encodings['content'], errors='replace')
-				if msg:
-					if not self.background:
-						writemsg_stdout(msg, noiselevel=-1)
-					if log_file is not None:
-						log_file.write(msg)
-			finally:
-				if log_file is not None:
-					log_file.close()
+			_check_build_log(self.settings, out=out)
+			msg = _unicode_decode(out.getvalue(),
+				encoding=_encodings['content'], errors='replace')
+			self.scheduler.output(msg,
+				log_path=self.settings.get("PORTAGE_LOG_FILE"))
 
 		if fail:
 			self._die_hooks()
@@ -108,15 +95,8 @@ class EbuildPhase(CompositeTask):
 			msg = _unicode_decode(out.getvalue(),
 				encoding=_encodings['content'], errors='replace')
 			if msg:
-				if not self.background:
-					writemsg_stdout(msg, noiselevel=-1)
-				log_path = self.settings.get("PORTAGE_LOG_FILE")
-				if log_path is not None:
-					log_file = codecs.open(_unicode_encode(log_path,
-						encoding=_encodings['fs'], errors='strict'),
-						mode='a', encoding=_encodings['content'], errors='replace')
-					log_file.write(msg)
-					log_file.close()
+				self.scheduler.output(msg,
+					log_path=self.settings.get("PORTAGE_LOG_FILE"))
 
 		post_phase_cmds = _post_phase_cmds.get(self.phase)
 		if post_phase_cmds is not None:
diff --git a/pym/_emerge/MetadataRegen.py b/pym/_emerge/MetadataRegen.py
index f819570d9..6c6dd6b90 100644
--- a/pym/_emerge/MetadataRegen.py
+++ b/pym/_emerge/MetadataRegen.py
@@ -1,4 +1,4 @@
-# Copyright 1999-2009 Gentoo Foundation
+# Copyright 1999-2010 Gentoo Foundation
 # Distributed under the terms of the GNU General Public License v2
 
 import portage
@@ -26,10 +26,6 @@ class MetadataRegen(PollScheduler):
 
 		self._max_jobs = max_jobs
 		self._max_load = max_load
-		self._sched_iface = self._sched_iface_class(
-			register=self._register,
-			schedule=self._schedule_wait,
-			unregister=self._unregister)
 
 		self._valid_pkgs = set()
 		self._cp_set = set()
diff --git a/pym/_emerge/PackageUninstall.py b/pym/_emerge/PackageUninstall.py
index 6f528663a..e806bc154 100644
--- a/pym/_emerge/PackageUninstall.py
+++ b/pym/_emerge/PackageUninstall.py
@@ -42,15 +42,4 @@ class PackageUninstall(AsynchronousTask):
 				portage.util.writemsg_level(msg,
 					level=level, noiselevel=noiselevel)
 		else:
-			if not background:
-				portage.util.writemsg_level(msg,
-					level=level, noiselevel=noiselevel)
-
-			f = codecs.open(_unicode_encode(log_path,
-				encoding=_encodings['fs'], errors='strict'),
-				mode='a', encoding=_encodings['content'], errors='replace')
-			try:
-				f.write(msg)
-			finally:
-				f.close()
-
+			self.scheduler.output(msg, level=level, noiselevel=noiselevel)
diff --git a/pym/_emerge/PollScheduler.py b/pym/_emerge/PollScheduler.py
index 2a77f869e..2c9eb2a8c 100644
--- a/pym/_emerge/PollScheduler.py
+++ b/pym/_emerge/PollScheduler.py
@@ -1,10 +1,13 @@
-# Copyright 1999-2009 Gentoo Foundation
+# Copyright 1999-2010 Gentoo Foundation
 # Distributed under the terms of the GNU General Public License v2
 
+import gzip
 import logging
 import select
 import time
 
+from portage import _encodings
+from portage import _unicode_encode
 from portage import os
 from portage.util import writemsg_level
 
@@ -16,7 +19,7 @@ from _emerge.PollSelectAdapter import PollSelectAdapter
 class PollScheduler(object):
 
 	class _sched_iface_class(SlotObject):
-		__slots__ = ("register", "schedule", "unregister")
+		__slots__ = ("output", "register", "schedule", "unregister")
 
 	def __init__(self):
 		self._max_jobs = 1
@@ -29,6 +32,12 @@ class PollScheduler(object):
 		self._event_handler_id = 0
 		self._poll_obj = create_poll_instance()
 		self._scheduling = False
+		self._background = False
+		self.sched_iface = self._sched_iface_class(
+			output=self._task_output,
+			register=self._register,
+			schedule=self._schedule_wait,
+			unregister=self._unregister)
 
 	def _schedule(self):
 		"""
@@ -228,6 +237,31 @@ class PollScheduler(object):
 
 		return event_handled
 
+	def _task_output(self, msg, log_path=None, level=0, noiselevel=-1):
+		"""
+		Output msg to stdout if not self._background. If log_path
+		is not None then append msg to the log (appends with
+		compression if the filename extension of log_path
+		corresponds to a supported compression type).
+		"""
+
+		if not self._background:
+			writemsg_level(msg, level=level, noiselevel=noiselevel)
+
+		if log_path is not None:
+			f = open(_unicode_encode(log_path,
+				encoding=_encodings['fs'], errors='strict'),
+				mode='ab')
+
+			if log_path.endswith('.gz'):
+				# NOTE: The empty filename argument prevents us from triggering
+				# a bug in python3 which causes GzipFile to raise AttributeError
+				# if fileobj.name is bytes instead of unicode.
+				f =  gzip.GzipFile(filename='', mode='ab', fileobj=f)
+
+			f.write(_unicode_encode(msg))
+			f.close()
+
 _can_poll_device = None
 
 def can_poll_device():
diff --git a/pym/_emerge/QueueScheduler.py b/pym/_emerge/QueueScheduler.py
index 8e1837c03..0e39d6ad3 100644
--- a/pym/_emerge/QueueScheduler.py
+++ b/pym/_emerge/QueueScheduler.py
@@ -1,4 +1,4 @@
-# Copyright 1999-2009 Gentoo Foundation
+# Copyright 1999-2010 Gentoo Foundation
 # Distributed under the terms of the GNU General Public License v2
 
 from _emerge.PollScheduler import PollScheduler
@@ -18,10 +18,6 @@ class QueueScheduler(PollScheduler):
 
 		self._max_jobs = max_jobs
 		self._max_load = max_load
-		self.sched_iface = self._sched_iface_class(
-			register=self._register,
-			schedule=self._schedule_wait,
-			unregister=self._unregister)
 
 		self._queues = []
 		self._schedule_listeners = []
diff --git a/pym/_emerge/Scheduler.py b/pym/_emerge/Scheduler.py
index 1c732c309..aa0a40290 100644
--- a/pym/_emerge/Scheduler.py
+++ b/pym/_emerge/Scheduler.py
@@ -4,6 +4,7 @@
 from __future__ import print_function
 
 import codecs
+import gzip
 import logging
 import shutil
 import sys
@@ -78,7 +79,8 @@ class Scheduler(PollScheduler):
 
 	class _iface_class(SlotObject):
 		__slots__ = ("dblinkEbuildPhase", "dblinkDisplayMerge",
-			"dblinkElog", "dblinkEmergeLog", "fetch", "register", "schedule",
+			"dblinkElog", "dblinkEmergeLog", "fetch",
+			"output", "register", "schedule",
 			"scheduleSetup", "scheduleUnpack", "scheduleYield",
 			"unregister")
 
@@ -209,7 +211,8 @@ class Scheduler(PollScheduler):
 			dblinkDisplayMerge=self._dblink_display_merge,
 			dblinkElog=self._dblink_elog,
 			dblinkEmergeLog=self._dblink_emerge_log,
-			fetch=fetch_iface, register=self._register,
+			fetch=fetch_iface, output=self._task_output,
+			register=self._register,
 			schedule=self._schedule_wait,
 			scheduleSetup=self._schedule_setup,
 			scheduleUnpack=self._schedule_unpack,
@@ -562,22 +565,9 @@ class Scheduler(PollScheduler):
 		installed = type_name == "installed"
 		return self._pkg(cpv, type_name, root_config, installed=installed)
 
-	def _append_to_log_path(self, log_path, msg):
-
-		f = codecs.open(_unicode_encode(log_path,
-			encoding=_encodings['fs'], errors='strict'),
-			mode='a', encoding=_encodings['content'],
-			errors='backslashreplace')
-		try:
-			f.write(_unicode_decode(msg,
-				encoding=_encodings['content'], errors='replace'))
-		finally:
-			f.close()
-
 	def _dblink_elog(self, pkg_dblink, phase, func, msgs):
 
 		log_path = pkg_dblink.settings.get("PORTAGE_LOG_FILE")
-		background = self._background
 		out = StringIO()
 
 		for msg in msgs:
@@ -585,11 +575,7 @@ class Scheduler(PollScheduler):
 
 		out_str = out.getvalue()
 
-		if not background:
-			portage.util.writemsg_stdout(out_str, noiselevel=-1)
-
-		if log_path is not None:
-			self._append_to_log_path(log_path, out_str)
+		self._task_output(out_str, log_path=log_path)
 
 	def _dblink_emerge_log(self, msg):
 		self._logger.log(msg)
@@ -603,10 +589,7 @@ class Scheduler(PollScheduler):
 				portage.util.writemsg_level(msg,
 					level=level, noiselevel=noiselevel)
 		else:
-			if not background:
-				portage.util.writemsg_level(msg,
-					level=level, noiselevel=noiselevel)
-			self._append_to_log_path(log_path, msg)
+			self._task_output(msg, log_path=log_path)
 
 	def _dblink_ebuild_phase(self,
 		pkg_dblink, pkg_dbapi, ebuild_path, phase):
@@ -1100,11 +1083,14 @@ class Scheduler(PollScheduler):
 			log_path = self._locate_failure_log(failed_pkg)
 			if log_path is not None:
 				try:
-					log_file = codecs.open(_unicode_encode(log_path,
-					encoding=_encodings['fs'], errors='strict'),
-					mode='r', encoding=_encodings['content'], errors='replace')
+					log_file = open(_unicode_encode(log_path,
+						encoding=_encodings['fs'], errors='strict'), mode='rb')
 				except IOError:
 					pass
+				else:
+					if log_path.endswith('.gz'):
+						log_file =  gzip.GzipFile(filename='',
+							mode='rb', fileobj=log_file)
 
 			if log_file is not None:
 				try:
diff --git a/pym/_emerge/SpawnProcess.py b/pym/_emerge/SpawnProcess.py
index 6e0586815..aeb206a06 100644
--- a/pym/_emerge/SpawnProcess.py
+++ b/pym/_emerge/SpawnProcess.py
@@ -6,10 +6,13 @@ from _emerge.PollConstants import PollConstants
 import sys
 from portage.cache.mappings import slot_dict_class
 import portage
+from portage import _encodings
+from portage import _unicode_encode
 from portage import os
 import fcntl
 import errno
 import array
+import gzip
 
 class SpawnProcess(SubProcess):
 
@@ -77,7 +80,12 @@ class SpawnProcess(SubProcess):
 			fd_pipes[1] = slave_fd
 			fd_pipes[2] = slave_fd
 
-			files.log = open(logfile, mode='ab')
+			files.log = open(_unicode_encode(logfile,
+				encoding=_encodings['fs'], errors='strict'), mode='ab')
+			if logfile.endswith('.gz'):
+				files.log = gzip.GzipFile(filename='', mode='ab',
+					fileobj=files.log)
+
 			portage.util.apply_secpass_permissions(logfile,
 				uid=portage.portage_uid, gid=portage.portage_gid,
 				mode=0o660)
@@ -188,7 +196,11 @@ class SpawnProcess(SubProcess):
 								fcntl.fcntl(files.stdout.fileno(),
 								fcntl.F_GETFL) ^ os.O_NONBLOCK)
 
-				buf.tofile(files.log)
+				try:
+					buf.tofile(files.log)
+				except TypeError:
+					# array.tofile() doesn't work with GzipFile
+					files.log.write(buf.tostring())
 				files.log.flush()
 			else:
 				self._unregister()
diff --git a/pym/portage/const.py b/pym/portage/const.py
index 62d3ceaf2..bd2c5be8d 100644
--- a/pym/portage/const.py
+++ b/pym/portage/const.py
@@ -85,7 +85,8 @@ EBUILD_PHASES            = ("pretend", "setup", "unpack", "prepare", "configure"
                            "nofetch", "config", "info", "other")
 SUPPORTED_FEATURES       = frozenset([
                            "assume-digests", "buildpkg", "buildsyspkg", "ccache",
-                           "collision-protect", "digest", "distcc", "distlocks",
+                           "collision-protect", "compress-build-logs",
+                           "digest", "distcc", "distlocks",
                            "fakeroot", "fail-clean", "fixpackages", "getbinpkg",
                            "installsources", "keeptemp", "keepwork", "fixlafiles", "lmirror",
                            "metadata-transfer", "mirror", "multilib-strict", "news",
diff --git a/pym/portage/package/ebuild/doebuild.py b/pym/portage/package/ebuild/doebuild.py
index 21f716beb..b340e2759 100644
--- a/pym/portage/package/ebuild/doebuild.py
+++ b/pym/portage/package/ebuild/doebuild.py
@@ -5,6 +5,7 @@ __all__ = ['doebuild', 'doebuild_environment', 'spawn', 'spawnebuild']
 
 import codecs
 import errno
+import gzip
 from itertools import chain
 import logging
 import os as _os
@@ -1179,12 +1180,14 @@ def _check_build_log(mysettings, out=None):
 	if logfile is None:
 		return
 	try:
-		f = codecs.open(_unicode_encode(logfile,
-			encoding=_encodings['fs'], errors='strict'),
-			mode='r', encoding=_encodings['content'], errors='replace')
+		f = open(_unicode_encode(logfile, encoding=_encodings['fs'],
+			errors='strict'), mode='rb')
 	except EnvironmentError:
 		return
 
+	if logfile.endswith('.gz'):
+		f =  gzip.GzipFile(filename='', mode='rb', fileobj=f)
+
 	am_maintainer_mode = []
 	bash_command_not_found = []
 	bash_command_not_found_re = re.compile(
@@ -1212,6 +1215,7 @@ def _check_build_log(mysettings, out=None):
 
 	try:
 		for line in f:
+			line = _unicode_decode(line)
 			if am_maintainer_mode_re.search(line) is not None and \
 				am_maintainer_mode_exclude_re.search(line) is None:
 				am_maintainer_mode.append(line.rstrip("\n"))
diff --git a/pym/portage/package/ebuild/prepare_build_dirs.py b/pym/portage/package/ebuild/prepare_build_dirs.py
index 0ae60342f..25e5beb6e 100644
--- a/pym/portage/package/ebuild/prepare_build_dirs.py
+++ b/pym/portage/package/ebuild/prepare_build_dirs.py
@@ -3,8 +3,8 @@
 
 __all__ = ['prepare_build_dirs']
 
-import codecs
 import errno
+import gzip
 import shutil
 import stat
 import time
@@ -117,15 +117,17 @@ def _adjust_perms_msg(settings, msg):
 
 	if background and log_path is not None:
 		try:
-			log_file = codecs.open(_unicode_encode(log_path,
-				encoding=_encodings['fs'], errors='strict'),
-				mode='a', encoding=_encodings['content'], errors='replace')
+			log_file = open(_unicode_encode(log_path,
+				encoding=_encodings['fs'], errors='strict'), mode='ab')
 		except IOError:
 			def write(msg):
 				pass
 		else:
+			if log_path.endswith('.gz'):
+				log_file =  gzip.GzipFile(filename='',
+					mode='ab', fileobj=log_file)
 			def write(msg):
-				log_file.write(_unicode_decode(msg))
+				log_file.write(_unicode_encode(msg))
 				log_file.flush()
 
 	try:
@@ -281,6 +283,11 @@ def _prepare_workdir(mysettings):
 			writemsg(_("!!! Disabling logging.\n"), noiselevel=-1)
 			while "PORT_LOGDIR" in mysettings:
 				del mysettings["PORT_LOGDIR"]
+
+	compress_log_ext = ''
+	if 'compress-build-logs' in mysettings.features:
+		compress_log_ext = '.gz'
+
 	if "PORT_LOGDIR" in mysettings and \
 		os.access(mysettings["PORT_LOGDIR"], os.W_OK):
 		logid_path = os.path.join(mysettings["PORTAGE_BUILDDIR"], ".logid")
@@ -292,12 +299,14 @@ def _prepare_workdir(mysettings):
 
 		if "split-log" in mysettings.features:
 			mysettings["PORTAGE_LOG_FILE"] = os.path.join(
-				mysettings["PORT_LOGDIR"], "build", "%s/%s:%s.log" % \
-				(mysettings["CATEGORY"], mysettings["PF"], logid_time))
+				mysettings["PORT_LOGDIR"], "build", "%s/%s:%s.log%s" % \
+				(mysettings["CATEGORY"], mysettings["PF"], logid_time,
+				compress_log_ext))
 		else:
 			mysettings["PORTAGE_LOG_FILE"] = os.path.join(
-				mysettings["PORT_LOGDIR"], "%s:%s:%s.log" % \
-				(mysettings["CATEGORY"], mysettings["PF"], logid_time))
+				mysettings["PORT_LOGDIR"], "%s:%s:%s.log%s" % \
+				(mysettings["CATEGORY"], mysettings["PF"], logid_time,
+				compress_log_ext))
 
 		ensure_dirs(os.path.dirname(mysettings["PORTAGE_LOG_FILE"]))
 
@@ -307,4 +316,4 @@ def _prepare_workdir(mysettings):
 		# current policy will allow it to work when a pty is available, but
 		# not through a normal pipe. See bug #162404.
 		mysettings["PORTAGE_LOG_FILE"] = os.path.join(
-			mysettings["T"], "build.log")
+			mysettings["T"], "build.log%s" % compress_log_ext)