2016-06-28  Alec Leamas  <leamas.alec@gmail.com>

	Merge branch 'release-0_9_4' into release

	Update for 0.9.4a

	systemd: Fix syntax errors in systemd unit files (rhbz: 1350750)

2016-06-06  Alec Leamas  <leamas.alec@gmail.com>

	irrecord: Enhance "Don't use devinput" message (#180).

2016-06-01  Alec Leamas  <leamas.alec@gmail.com>

	plugins/uirt2_raw: Reset device to UIR mode on deinit.
	If the UIRT is left in RAW mode, it does not trigger wake from S3.
	Setting the UIRT back to UIR mode on deinit allows wake from S3 to
	work properly.

2016-06-01  Bengt Martensson  <barf@bengt-martensson.de>

	plugins/girs.c: Enable reception after sending.

2016-06-01  Christoph Bartelmus  <cb@nowhere.net>

	lirc.org: Update

2016-05-28  Alec Leamas  <leamas.alec@gmail.com>

	lircd, liblirc_client: Fix freeaddrinfo handling (#195).
	freeaddrinfo() was used the wrong wa which caused a segfault using
	--connect  and during a boot race condition. Fixed the bad usage
	of freeaddrinf() in two places + some cleanup.

2016-05-27  Alec Leamas  <leamas.alec@gmail.com>

	lirc_options.conf: debug => loglevel (#199).

2016-05-25  Thomas Petazzoni  <thomas.petazzoni@free-electrons.com>

	lib: use proper linking method to avoid parallel build issue
	Using <foo>_LDFLAGS = -l<library> is correct when <library> is an
	external library. However, when it is built by the same package, and
	especially in the same directory, this is wrong and can cause parallel
	build issues. In lib/Makefile.am, there was:

	libirrecord_la_LDFLAGS      = -llirc

	But the liblirc library is built in the same directory. Or, due to the
	using of <foo>_LDFLAGS, make is not aware of the build dependency
	between libirrecord and liblirc.

	To solve this, <foo>_LIBADD should be used instead, as follows:

	libirrecord_la_LIBADD       = liblirc.la

	This fixes parallel build issues seen by automated build tests
	conducted by the Buildroot project, such as:

	  http://autobuild.buildroot.org/results/eb4/eb47d57de8182d25b1dacbf0ac3726ed20063d04/build-end.log

2016-05-25  Baruch Siach  <baruch@tkos.co.il>

	tools: make_rel_symlink.py can also use python2
	The make_rel_symlink.py script is compatible with both python2 and python3.
	Don't hard code a requirement for python3.

2016-05-24  Alec Leamas  <leamas.alec@gmail.com>

	plugins/devinput: Make the list-devices support available.

	configure.ac: Bugfix (BSD setting reported wrong).

	Don't build commandir unless we have usb.h (#191).

	lirc.org: Bugfix.

2016-05-23  Alec Leamas  <leamas.alec@gmail.com>

	Revert "Build: Handle alternative usb.h locations."
	This reverts commit 46fd216d0862cf3ed745fca5499f992779be0433.

	Revert "Build: Handle alternative usb.h locations."
	This reverts commit 46fd216d0862cf3ed745fca5499f992779be0433, which
	is a plain error.

	Merge branch 'release-0_9_4' into release

	Build: Handle alternative usb.h locations.

2016-05-23  Alec Leamas  <leamas.alec@gmail.com>

	Build: Fix some BSD portability issues (#155).
	  - Make implicit header sys/wait.h explicit.
	  - cp -ar => cp -pr and sed => $(SED)
	  - #!/bin/bash =>  #!/usr/bin/env bash
	  - Handle missing modinfo by skipping kernel drivers.
	  - Don't build plugins requiring linux-specific stuff.
	  - and some more...

	The bug is really about OSX, but this is what we do for now.
	Package builds on FreeBSD 10.3, but cannot run 'make dist'
	due to some files generated when doing 'make dist' requiring
	linux tools.

2016-05-22  Alec Leamas  <leamas.alec@gmail.com>

	Update for 0.9.4

	lib: Update version-info age part (added interfaces).

	NEWS: update

	Documentation fixes.

	plugins/default: Support the "auto" device.
	Making the default and devinput driver support the "auto" device should
	cope with a large number of trivial configurations.

	lirc-setup: Add support for drvctl device enumeration.

	lirc-setup: Fix bad imports (python gi lib update).

	plugins/audio_alsa: Use specific waitfordata (#169).

2016-05-18  Alec Leamas  <leamas.alec@gmail.com>

	Debian build: Kill some trivial lintian warnings.

2016-05-13  Alec Leamas  <leamas.alec@gmail.com>

	gitignore: update.

	plugins: Don't add trailing line on pluginlist.am.

	README: update.

	doc: Manpages index fixes, try to fix broken html for lircd-uinput.8.

2016-05-10  Alec Leamas  <leamas.alec@gmail.com>

	contrib: release process update.

	doc: Add example udev rule for fixed /dev/lirc* links.

2016-05-09  Alec Leamas  <leamas.alec@gmail.com>

	build: Handle brain-dead Ubuntu systemd configuration.

2016-05-06  Alec Leamas  <leamas.alec@gmail.com>

	NEWS: update-

	NEWS: update-

	Merge branch 'release-0_9_4' into release

	Update for lirc-0.9.4pre2

	NEWS: update.

	irrecord (lirc-make-devinput): Handle new include files (#187).

2016-05-05  Alec Leamas  <leamas.alec@gmail.com>

	lirc-lsplugins: Handle plugins with errors in --yaml mode.

	tests: bugfix.

	Makefile.am: bugfixes.

	plugins/file: Fix compiler warning.

	plugins/default: Add device enumeration.

	plugins/devinput: Add device enumeration.

	lib/driver: Add some glob_t convenience stuff.

	Build: Add udev library detection.

2016-05-03  Alec Leamas  <leamas.alec@gmail.com>

	Update for lirc-0.9.4pre2

	tests: bugfix.

	Makefile.am: bugfixes.

	NEWS: update.

	plugins/default: Add device enumeration.

	plugins/devinput: Add device enumeration.

	lib/driver: Add some glob_t convenience stuff.

2016-04-27  Alec Leamas  <leamas.alec@gmail.com>

	git-tools/checkpatch.pl: Kill unescaped brace warning.

	NEWS: update.

2016-04-27  William Manley  <will@williammanley.net>

	plugins: Add new ftdix driver + docs
	Mode2 transmitter using the bitbang mode of an FTDI USB-to-serial chip
	such as the FT230X.  The standard FTDI driver with the FT232R is
	unreliable having a IR transmit failure rate of ~2%.  The FT232R
	also becomes unusable if you set the tx_baud_rate to values other than
	65536.

	This driver works differently modifying its tx_baud_rate according to
	the carrier frequency set.  This is ok because the FT230X does not have
	the timing bugs present in the FT232R.  This allows much lower data
	rates over USB and so improves reliablity, particularly if other
	devices are connect to the same USB port.

	Receive is not included to significantly simplify design of this driver.

2016-04-27  Alec Leamas  <leamas.alec@gmail.com>

	plugins/file: Fix bad test for LIRCCODE cases (#171).

	lirs-lsremotes: Don't use non-standard GLOB_BRACE (#174).

	doc: mode2.1 manpage update.

2016-04-26  Alec Leamas  <leamas.alec@gmail.com>

	NEWS: update.

2016-04-26  Helen Foster  <helen.foster@somewhere.net>

	Fix  misleading logprintf comments (#179)

2016-04-26  William Manley  <will@williammanley.net>

	plugins/ftdi: Tear down subprocess if parent dies
	If the lircd parent process is killed with SIGKILL it will be unable
	to call `hwftdi_deinit` and so the subprocess it spawned will continue
	living, keeping the ftdi device busy.  This change uses receiving EOF
	on the pipe from the parent process as a signal that the parent process
	has died and will shut itself down accordingly too.

2016-04-26  William Manley  <will@williammanley.net>

	plugins/ftdi: Conform to lirc driver API v2 lifecycle expectations
	As documented in a1c63fe.  Specifically:

	> * At some later point, lirc calls the deinit() func. This puts the
	>   driver in a stand-by state, from which it is a reasonably fast
	>   operation to call `init_func()` again.
	> * As a last operation, lirc might call the `close()` func. This
	>   relinquishes all resources allocated by the driver.
	> * functions are not guaranteed to be executed in the order described.
	>   In general, the driver should not break if e. g., the `init()` or
	>   `deinit()` function is called twice.

2016-04-26  Alec Leamas  <leamas.alec@gmail.com>

	tests: Fix logging tests.

	build: Distcheck bugfix.

	doc: Remove upstreamed lirc.4 manpage.

	lib: Clean up loglevel handling (#178).

2016-04-25  Bengt Martensson  <barf@bengt-martensson.de>

	plugins:  girs.c bugfixes.
	See https://sourceforge.net/p/lirc/mailman/message/34982405/

2016-04-25  Alec Leamas  <leamas.alec@gmail.com>

	Fix typo causing lost deinit_func() calls (#179).

2016-01-26  Alec Leamas  <leamas.alec@gmail.com>

	doc: lircd.conf: Document suppress_repeat parameter.

2016-01-25  Alec Leamas  <leamas.alec@gmail.com>

	Makefile.am: Fix typo.

	plugins/default: Work around kernel bug (#172).

2016-01-24  Alec Leamas  <leamas.alec@gmail.com>

	irpipe: Bugfix.

	plugins/file: Bugfix dumping non-LIRCCODE data (#171).

2016-01-23  Alec Leamas  <leamas.alec@gmail.com>

	Add explicit default -L${libdir} linker flag.
	Adding an explicit -L${libdir} is redundant on single-arch hosts
	but is essential e. g., when building a 32-bit library on a 64-bit host.

2016-01-17  Alec Leamas  <leamas.alec@gmail.com>

	build: Fix cleaning.

	systemd: Remove space in key definitions
	Spaces generates valid packaging warnings  on Debian (lintian).

2016-01-12  Alec Leamas  <leamas.alec@gmail.com>

	drivers: Update iguanair plugin docs.

	docs: update.

	doc: Add a "Default Configuration" chapter.

	Fix generation of plugin docs.

2016-01-08  Alec Leamas  <leamas.alec@gmail.com>

	lircd-uinput.8: Fix bad section.

	Fix Debian FTBS compilation error.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	doc: Updates.

	Update for lirc 0.9.4pre1 release

	docs: update.

	README: update.

	Fix --options-file option parsing (#165).

	doc: Makefile.am distcheck bugfix.

	Use kernel lirc.h if available (#163).

	Make --uinput a separate service lircd-uinput (#162-part1).

	Upstream the lirc.4 manpage (#164).

	doc: lirc.4 upstream review remarks fixes.
	This version has been merged to upstream
	http://git.kernel.org/cgit/docs/man-pages/man-pages.git

	Build: Fix clean errors.

	tools: Teach irpipe to handle LIRCCODE drivers.

	plugins: file: Add sending of LIRCCODE files.

	lib: log: Remove unused cruft.

	irpipe: Fix compiler option (breaks w recent kernel).

2016-01-06  Nicolas  <https://sourceforge.net/u/nadenislamarre/profile/send_message>

	Temporay fix for uinput device setup (#161).
	We normally init the driver when the first client connects to
	the socket, and calle deinit() whenthe last client disconnects.
	Using uinput we connect at startup, and never disconnects.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	Revert use of CLOCK_MONOTONIC_RAW (#158).

	Fix wrong device in 03e378f (#150, again).

	build: Fix getgroups(2) configuration (#155).

	lirc-config: pylint/pep8 cleanup + bad bugfix.

	test: Update plugin count.

	lircd: Add tentative Darwin timer handling.

	build, lircd: Use CLOCK_MONOTONIC_RAW if available.

2016-01-06  Jarkko Sonninen  <kasper@iki.fi>

	lircd: Fix sending gap between repeats (#158).

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	plugins: Move iguanair driver out of tree.

	Remove in-source debian packaging files.

	debian: Update.
	- Follow-ups to kernel-drivers.dist handling.
	- Be strict when checking for libftdi > 1.0.

	All: MacOS compilation fixes (#155).
	- Affects all parts

	Create kernel-driver.yaml data in runtime (#157).

	log: Fix bad useconds typing in 9f8c440 (#159).

	doc: Restore manpages destroyed by bdc7525.

	debian: Fix some remarks from debian-mentors.

	debian: Bugfixes (#160).

	NEWS: drivers OOT move.

	drivers: Update yard2 docs.

2016-01-06  Jarkko Sonninen  <kasper@iki.fi>

	lib: Add microseconds to log time (#159).

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	plugins: Move irman driver out of tree.
	This removes all libirman dependencies from the main build tree.
	The driver can be built and installed out of tree, after that
	lirc has been installed.

	plugins: Move yard2 driver out of tree.

	doc: Add yard2 plugin documentation.

	debian: bugfixes.

	doc: Fix broken links.

	doc: Add some missing files.

	doc: Update (includes fixes to #154).
	  - New substitution variable @website@ for manpage use.
	  - irexec manpage overhaul.
	  - New lircrc examples added to contrib.
	  - New FTDI driver documentation.
	  - New lircrc manpage.

	Add a system-wide irexec service (#156).

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	Add plugin device enumeration.
	The only reliable way to enumerate devices, in particular the USB ones
	is done by the driver. This adds a new option in the configs file +
	a mode2 CLI option to display the values listed by the driver. The
	actual code is a DRVCTL so it does not really change the driver
	interface.

	Implement in atilibusb and ftdi, the latter untested. The ftdi code
	fixes an existing error in the drvctl_func.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	build: Fix OOT plugin docs generation, update example code.

	doc: Update.

	contrib: Add a new "blacklist all" catch all rule, clean-up.

	lirc-setup: Bugfixes
	Fix some glitches which surfaced when starting on Debian.

	Create a devinput-based default setup (#106).
	Using the devinput driver with the 'auto' device as default setup works
	if there is a single remote with rc-based driver. While certainly not
	covering all cases, it's a reasonable attempt - there is actually a
	chance this "works" out of the box for some users.

	irrecord:  Fix error messages (#152).

	plugins: Add "No such file" logging to default (#151).

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	Clean up generating of devinput.lircd.conf (#153).
	The existing logic was completely broken. The idea to distribute
	input_map.inc is wrong since it reflects the kernel on the machine
	where it was created. Better then to just distribute a best-effort
	devinput.lircd.conf + simple tool to re-generate it from the target
	system kernel.

	After patch there is just one single script used both when building and
	when users re-generate devinput.lircd.conf.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	build:  Debian packaging build wrap-up.
	Locating modinfo is causing some pain on Debian system which havn't
	/sbin:/usr/sbin in $PATH. Locate using autoconf and use that value.
	While we are on it, rename paths.py to config.py, which reflects the
	purpose.

	A VERSION file is added, installed as a document. Git version data
	is updated at "make dist".

	The debian packaging is updated, and outside the package the Fedora
	debian toolchain has been updated so that source packages can be built.

	Various bugfixes and clean-up.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	lib: irrecord: Re-factoring bugfix (#147).

	doc: Add missing lircd-setup manpage.

	doc: Add missing pronto2lirc manpage.

2016-01-06  Alec Leamas  <leamas@nowhere.net>

	contrib: Remove outdated start script lircs.

	contrib: Remove init.d cruft (trust the packagers).

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	Revert walk-around kernel bug fixes (#119, #148)
	Reverting  "default: Conditionalize kernel bug walk-around"
	(68e49ce4a) and "plugins: default: Work around kernel bug (#119)."
	(19b01d0fc). The fix is broken, and the upstream bug is fixed
	from kernel 4.1.6.

2016-01-06  Bengt Martensson  <barf@bengt-martensson.de>

	plugins:girs: Minor fixes (#150).

2016-01-06  Jon Bergli Heier  <snakebite@jvnv.net>

	lib: receive: Return 0 on poll() timeout.

	lib: receive: Fixed timeout value for poll().

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	NEWS: update

	Sanitize the for_each_plugin() signature.

	doc, tools: Database distcheck fixes.

	doc: Mega update.

	doc: Provide a configuration-guide symlink to plugindocs.

	lib, tools: Streamline application logging.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	lirc-setup: Update: tests, documentation links, bugfixes
	- Added options for running mode2 and lircd+irw for selected config.
	- Added documentation links for driver
	- Support for /bin/sh driver hints scripts
	- Added paths.py to substitute installation paths.
	- Tons of bugfixes.

	The UI has more or less collapsed and need an overhaul. However,
	this seems at a glance funtionally complete.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	irrecord: Fix some messages.

	plugins: Generate devinput.lircd.conf, fix obsolete file warning.
	Add code to generate devinput.lircd.conf on current system. Fix warning
	in devinput plugin for obsolete file.

	plugins: Add some complicated device_hints.

	configs: Fix yaml generation, bugfixes.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	lirc-lsplugins: Quote some YAML output.
	The YAML parser is unhappy with more complicated shell commands in
	the device_hint string. Write string as a double-quoted string,
	quoting all \ to \\. The more YAML-friendly single quoting doesn't
	work since "'" is part of command shell commands.

	Also adds the info string to the generated output.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	Revert "build: Add selinux policy."
	Lets make an upstream policy patch instead:
	https://github.com/fedora-selinux/selinux-policy/pull/41

	irrecord: Fix un-decodable buttons handling (#146).

	atilibusb: Fix crash for non-existing device (#145).

	lirc-lsplugins: Fix --plugindir not being respected (#144).

	tests: fix  expected driver count (girs => +1).

	Retire the caraca driver.
	The driver depends on a library which isn't packaged in any major
	distribution. The code is since 2002, and does not build. Keeping
	the driver in attic for the time being.

	NEWS:update

	doc, tools: distcheck database fixes.

	girs: Mute some checkfiles warnings.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	build: Add selinux policy.
	The selinux policy is built if the libselinux-devel files (really
	libselinux.pc) is present. The set of policies (mls, targeted) is
	hardcoded in Makefile.am.  Although polices are built they are not
	installed during 'make install'.

	The long time solution is to add this policy to the upstream generic
	one. However, using a package policy yields a faster turn-around in
	the initial incarnations. There is also some problems with the
	module naming:

	https://github.com/fedora-selinux/selinux-policy/issues/38

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	build: Generate devinput.lircd.conf.
	The devinput.lircd.conf reflects the host (i. e., the kernel) rather
	than any external hardware. Thus it makes sense to rebuild it
	as part of the build so an actual up-to-date version always is
	available.

	doc: Fix install-time paths in manpages (#140).
	Add build mechanics to patch installation paths in manpages and
	apply them for /var/run (@varrundir@) and /etc (@etcdir@).

	irrecord: From mailing list: improve some text.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	tools: irrecord, mode2, xmode2: Options and default values.
	As pointed out in the mailing list, using the lircd values for
	driver and device are reasonable fallbacks when no value
	exists in the [irrecord] section. While we are on it, fix
	debuglevel and plugindir as well. This fix:

	 - Makes all three tools handle --device, --driver, --debug
	   --plugindir, --options-file and --keep-root in a uniform way.
	 - Makes all three tools fall back to the [lircd] section for
	   --driver, --device, --plugindir and --debug if not defined
	   elsewhere.
	 - Adds feedback for effective driver and device.
	 - Makes a manpage overhaul.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	plugins: Add document reference in .info.
	For those drivers which have documentation, linking to that seems to
	be a Good Thing. Moving the plugindocs dir declared in lirc-driver.pc
	fomr /usr/share to /usr/share/doc. This is a border case, the driver
	works without the docs but contains a link to it.

	doc: Add out-of-tree plugin docs generation.

	doc: Adding irman driver doc sketch.

	doc: Driver API logging update.

	logging: Macroize logperror.

	logging: Macroize standard logprintf usecases.
	Using macros is a better interface, and opens up for addding e. g.,
	file and line number to messages.

	logging: replace LOGPERROR ->logperror.

	logging: Replace LOGPRINTF -> logprintf (#129).

	configs: bugfixes, README update.

	Lower priority of duplicate codes messages.
	The messages about multiple definitions or multiple values are
	far from severe. In many cases the codes are defined this way
	deliberately. NOTICE is more appropriate.

	mode2: Clean up stdout/stderr output(#142).
	Old code puts error messages on stdout which is a plain bug.
	Fix this. While we're on it: fix some message definitions.
	Raise default loglevel to DEBUG.

	girs: Add configs/ file.
	From a user perspective, the girs driver is as of now Arduino
	support and this config file presents it as such. It is perfectly
	possible to add another one named "girs" presenting the driver as
	a generic tool should the need arise.

	girs: Add device_hint, update to API version 3.

2016-01-06  Bengt Martensson  <barf@bengt-martensson.de>

	Girs driver: completed code, documented.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	contrib: Add generic USB devices permissions rule

	Documentation: updates for configuration tools.

	datatable: Adapt to new database format.

	data2hwdb: Adapt to new data model.

	lirc-setup: Handle local files, fix some deprecations.

	lirc-setup: Adapting to new data model.

	build: Add mechanics to build *yaml files.

	Adding new lirc-kerneldrivers tool.

	configs/README: Provisionary config file documentation.

	configs: Update to new standards, resurrect most in attic.
	The open question is how many configs using the 'default' driver that
	actually are supported by the kernel. The first one who actually
	tries will know, there isn't much of an alternative (?).

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	configs: Nove obsolete configs to attic
	The lirc_gpio driver is retired since long. Moving configs
	depneding on this driver to new "attic" place so they can
	still be found if someone wants to resurrect them.

	The wpc8769l driver didn't make it when the lirc drivers were
	upstreamed to the kernel. Moving the config to the attic, and
	re-adding the kernel driver (which doesn't build as-is) if
	someone want's to resurrect them.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	configs: First stab on kernel-drivers.dist.

	irdb-get: Add support for creating confs_by_driver yaml.
	   - New yaml-config subcommand creates confs_by_driver.yaml
	   - Add support for XDG_CACHE_HOME for cached directory
	   - Add testing environment variables.
	   - Add -v/--version support

	lirc-lsplugins: Add support for yaml listing.

	plugins: Show real device in error feedback.
	For drivers which autodetects devices the probed the device should
	be made available as early as possible in driver.device so that
	error messages can reflect the probed device.

	plugins: Add configuration data, API version bump to 3.

	plugins: devinput: Handle larger events (kernel interface update).
	kernel input events are now 24 bytes, breaking some hard-coded
	limits in both this driver and in mode2. Bump driver interface
	version.

	NEWS: update.

	mode2: Make read buffer size dynamic.
	New input size from the /dev/input events breaks hard-coded limits.
	Instead of just raising the limits makes the size dynamic.

	plugins: devinput: Add DRVCTL_GET_RAW_CODELENGTH support.
	If mode2 should work, it needs to read the correct number of bytes
	when reading from /dev/input devices. This is required, since
	this driver does not implement readdata().

	lib: Add new DRVCTL_GET_RAW_CODELENGTH drvctl command.

	lib: Use getaddrinfo instead of gethostbyname (#122).

	lircd: Use getaddrinfo instead of gethostbyname (#122).
	Also: refactor some code.

	NEWS:update.

	doc: Major update.
	  Configuration guide:
	     - New appendix "Handling USB HID devices".
	     - New appendix "Using irrecord".
	     - New appendix "Running as a regular user".
	  Fix rpmlint manpage warnings
	  Add note on bundled header in lirc(4).
	  irsend: Fix help message + manpage glitch.
	  Various bugfixes.

	Add lircd setup tool (#121).

	plugins/default: Just check lirc protocol if read-only access (#133).
	When assuring the the used protocol is 'lirc' in the current rc device,
	first check if it's already set. If OK, the code works without write
	access.

	plugins/devinput: Add an "auto" device (#134).

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	Clean up documentation (#136).
	Remove the manpages under html/ and mode everything to lirc.org.
	Clean up some links and symlinks. Remaining symlinks reflects
	the different view from api-docs/ and lirc.org/html.

	The symlinks can now be easily expanded using tar --dereference.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	checkfiles: Mute needless kernel.space check (#135).

	Revert "tools: Add missing file lirc-setup/database.py."
	This reverts commit 034bc3b198a057427d79d4bd93a40234936be1cc, which is
	part of a yet-not-merged feature branch.

	default: Conditionalize kernel bug walk-around
	The fix introduced in "19b01d plugins: default: Work around kernel bug
	(#119)" does not play with the regression tests.  Add a simple fix
	which disables the fix if LIRC_REGRESSION_TEST is defined in the
	environment for now, better long-term strategies are needed later,
	and update the test script.

	contrib: Update release process.

	tools: Add missing file lirc-setup/database.py.

	plugins: Update to checkpatch standards.

	lirc_options.conf: Clean-up.

	lirc-config-tool: Update some heuristics.

	Updating to 0.9.4-devel

	Link a single large doc file to keep dist archives sane.
	WHile it's good thing to have everything in the archive (and versioned)
	this single seldom read PDF doubles the distribution size. Store
	it on sourceforge outside the git repo and link to it.

	Documentation (README, Install.txt, lirc.org/index.html) fixed.

	lircd: Fix crash in debug output (#126).

	lib: Add explicit default value for lircd.conf:frequency (#132).

	Build: fix dependencies.

	doc: Update

	doc: driver-api: Add permissions text.

	irrecord-ui: Handle remotes without gap (#98).

	lib: irrecord: Rewrite insane waitfordata loop.

	tools: Pylint/pep8 fixes.

	tools: Fix bad typo creating lirc_lsremotes, not lirc-lsremotes.

	tests: Adjust # plugins.

	lirc-lsplugins: Use puts when applicable.

	lircd: --effective-user: Set also gid (RH bz 1253907).

	doc: Add missing files.
	This should teach me to run distcheck from a fresh clone.

	doc: Updates, bugfixes.
	 - NEWS updated
	 - Driver api: Refer to manpage instead of duplicate.
	 - Bundle and link NEWS in docs,
	 - All sorts of typos etc.
	 - Clean up some links.
	 - All sorts of updates & fixes.

	tools: lirc-lsplugins: Patch legend info, fix irw formatting.

	lib: Typo fix.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	lircd: Set supplementary groups (RH bz 1253907).
	Setting effective user id isn't enough, we also need the supplementary
	groups (e. g., to access /var/lock files).

	Downstream: https://bugzilla.redhat.com/show_bug.cgi?id=1253907

	This is perhaps no final solution, but definitely makes the semantics
	better.

2016-01-06  Alec Leamas  <leamas.alec@gmail.com>

	plugins: default: Work around kernel bug (#119).

	irrecord-ui: More user feedback, formatting.

	lircrcd: poll() bugfix.

	lircd: poll() bugfixes.

	tools: Clean up perror output using perrorf()

	daemons: Clean up perror output using perrorf()

	lib: Add perrorf() function, formatted perror() support.

	Remove links on uninstall (#113).
	Surprise: 'make distcheck' does not check for left-over links
	after uninstall...

	No designated initializers in C++ (#111).

	doc: bugfix.

	doc: Move som stuff to configuration guide, more bugfix.

	build: distcheck/VPATH fixes.

	doc: Update, fix links, reorganize.
	- Update install info.
	- Configure -> Configuration reference, bugfixes.
	- Make a separate Program Overview.
	- Manpages macro usage cleanup.
	- Add som info on drivers.
	- New iguanair driver doc.
	- Fix broken links and missing images (lot's of them).

	doc: distcheck fix.

	NEWS:update

	doc: Move things from install and configure, clean up.

	doc: Add driver support files info to driver API.
	Clean up some things in technical and index.

	doc: Move protocol info to manpages (lircd, lircrcd, lircmd).
	The protocol info in technical.html is not updated, many (most?)
	commands are missing. Moving to manpage to create better chance
	for folks updating the code to also update the docs. Update manpages
	to current status.

	doc: Remove duplicated config tools text.

	doc: Add sending to client manual.

	doc: Expand api-docs on install and link to them.

	doc: Add irpipe manpage.

	doc: index: Clean up manpage TOC, add programs overview.

	doc: install, technical: refresh, remove obsolete text.

	doc: Makefile.am: Mute some garbage, bugfixes.

	lirc.org: Update for 0.9.3.

	daemons: Remove irsimsend and irsimreceive.
	These adds more complexity to lircd, and we don't need that. The
	new irsimsend + irpipe tools provides better substitutes.

	doc: Configuration Guide: html bugfix.

	tools/irpipe: Nuke some compiler warnings.

	tools: Makefile.am distcheck fix.

	Merge lirc.4 and lirc_ioctl.4 manpages into lirc.4.
	Also fix some review remarks fromm Michael Kerrick, Linux manpages
	project.

	doc: Add some missing manpage links.

	doc: lircd manpage update.

	doc: Remove bad links to lirc.org.

	doc: Configuration Guide: Add tools and files appendixes.

	doc: Add table.html to lirc.org.

	doc: Configuration Guide update.

	Update lirc-codec-regression-test to use irpipe
	New test does not use lircd.simsend or lircd.simreceive but are based
	on the irpipe kernel driver.

	tools: irpipe test tool added.
	Userspace tool to handle the irpipe kernel driver.

	Adding irpipe kernel test driver.

	plugins: zotac: Use poll() instead of select().

	plugins: uirt2_common: Use poll() instead of select().

	plugins: uirt_common: Use poll() instead of select().

	plugins: tira: Use poll() instead of select().

	plugins:irtoy:Use poll() instead of select().

	plugins: irlink: Use poll() instead of select().

	plugins: IguanaIR: Use poll() instead of select().

	plugins: atwf83: Use poll() instead of select().

	lib: receive: Use poll() instead of select().

	irpty: Use poll() instead of select().

	irrecord: Use poll() instead of select().

	lircrcd: Use poll() instead of select().

	lircd: Use poll() instead of select().

	irw: Make a clean exit on SIGUSR1.

	NEWS: update

	lirc-lsplugins: Display log after parsing file(s).

	irsimsend: Fix argument parsing -> -U/--pluginpath bugfix.

	irsimsend: bugfix.

	Updating version to 0.9.4

2016-01-06  Bengt Martensson  <barf@bengt-martensson.de>

	plugins: irtoy: Added support for debug LEDs (RA2,...,RA5), new docs.
	Original blinking when IR received caused problems, removed.
	Use receivePin when listening (compatiblity with IrScrutinizer).

	Docs added (irtoy_driver.html).

2015-10-14  Alec Leamas  <leamas.alec@gmail.com>

	irrecord: Fix crash when invoked without arguments.

	Update for 0.9.3a

	NEWS: update

2015-10-10  Alec Leamas  <leamas.alec@gmail.com>

	Revert walk-around kernel bug fixes (#119, #148)
	Reverting  "default: Conditionalize kernel bug walk-around"
	(68e49ce4a) and "plugins: default: Work around kernel bug (#119)."
	(19b01d0fc). The fix is broken, and the upstream bug is fixed
	from kernel 4.1.6.

2015-10-04  Jon Bergli Heier  <snakebite@jvnv.net>

	lib: receive: Return 0 on poll() timeout.

	lib: receive: Fixed timeout value for poll().

2015-10-02  Alec Leamas  <leamas.alec@gmail.com>

	irrecord: Fix un-decodable buttons handling (#146).

2015-10-01  Alec Leamas  <leamas.alec@gmail.com>

	atilibusb: Fix crash for non-existing device (#145).

	lirc-lsplugins: Fix --plugindir not being respected (#144).

2015-09-29  Alec Leamas  <leamas.alec@gmail.com>

	lib: Fix parsing of multiple files (#141).
	In certain combinations of including multiple files which contains
	concatened remotes the concatenated remotes were silently dropped.

2015-09-25  Alec Leamas  <leamas.alec@gmail.com>

	irrecord: Fix get_toggle_mask infinite loop (#139).

2015-09-24  Alec Leamas  <leamas.alec@gmail.com>

	irrecord: Fix insane driver default being devinput.

2015-09-20  Alec Leamas  <leamas.alec@gmail.com>

	lib: Use correct groups when dropping CLI privileges.

	plugins: file.c: Fix off-by-one sscanf error (#138).

2015-09-09  Alec Leamas  <leamas.alec@gmail.com>

	lib: Fix missing --lirc flag.
	This is not strictly required, the linker resolves the symbols anyway.
	However, there is packaging noise without it.

2015-09-07  Alec Leamas  <leamas.alec@gmail.com>

	default: Conditionalize kernel bug walk-around
	The fix introduced in "19b01d plugins: default: Work around kernel bug
	(#119)" does not play with the regression tests.  Add a simple fix
	which disables the fix if LIRC_REGRESSION_TEST is defined in the
	environment for now, better long-term strategies are needed later,
	and update the test script.
