TODO - external/github.com/brailcom/speechd - Git at Google

 Speech Dispatcher TODO
 ======================

 The release versions are not final, and could change. Targetted release is
 based on demand from users, and difficulty of the work involved.

  * Add pitch as an option for capitalization presentation
    see https://github.com/brailcom/speechd/issues/24
  * Allow setting a synthesis voice in the user config using spd-conf.
  * Add support for Mimic
    see https://github.com/brailcom/speechd/issues/19
  * Drop getting information from other clients
    see https://github.com/brailcom/speechd/issues/335
  (0.11) Migrate to GSettings.
  (0.11) Synthesizer specific settings API.
  (0.11) Use more GLib in the server.
  (0.11) Client audio retrieval API.
  (0.11) Server to module protocol documentation.
  (0.12) Server to module protocol improvements.
  * Move synth modules to plugin architecture with plugin host.
  * Synth plugin API.
  * Allow for building synth plugins out of tree.
  (0.11) Integrate with logind/consolekit.
  (0.11) Properly support system-wide mode.
  * Support spawning the server via Systemd socket activation.


 The above improvements are documented in detail below. If work has started on
 a particular project, a git branch will be noted. These git branches are
 located at https://github.com/TheMuso/speechd-wip.git. To read the most up to
 date copy of this file, please clone the master Speech Dispatcher git
 repository, located at git://git.freebsoft.org/git/speechd.gitand check out
 the master branch.

 Migrate to GSettings
 --------------------

  * Write the GSettings metadata XML file.
  * Migrate the server to GSettings.
  * Listen to GSettings changes.
  * Migrate synthesizer modules to GSettings.
  * Write a program to migrate user settings to GSettings.


 Synthesizer specific settings API
 Depends on: Migration to GSettings
 ---------------------------------

 Background:
  * Currently have API for espeak pitch range in git master, but this is only
    useful for espeak.
  * Espeak module has a config option to show variants along with available
    voices, which can be a very long list and can choak some clients.

  * Implement server to module protocol to support:
    - Request available settings.
    - Request available settings and their current value.
    - Request the value of a setting.
    - Set a setting.
    - Reset a setting to its default.
  * Implement SSIP protocol support.
  * Implement C API, see synthesizer specific settings C API draft.
  * Implement python API.

 Synthesizer specific settings C API draft

 typedef struct {
         char *name;
         char *description; /* This should be localized */
         enum SynthSettingValueType get_type;
         enum SynthSettingValueType set_type;
         int min_value;
         int max_value;
         char **value_list;
         void *cur_value;
 ] SynthSetting;

 In the C API, a NULL terminated array of this structure would be returned for
 all settings a synth offers.

 The SynthSettingValueType enum would look something like this:

 typedef enum {
         SYNTH_SETTING_VALUE_UNKNOWN = 0,
         SYNTH_SETTING_VALUE_NUMBER = 1,
         SYNTH_SETTING_VALUE_STRING = 2,
         SYNTH_SETTING_VALUE_STRING_LIST = 3 /* A list of strings for the user
                                                to choose from, i.e voice variants */
 } SynthSettingValueType;

 C API methods to work with these data types could be as follows:

 SynthSetting **spd_synth_get_settings(SPDConnection *connection);
 int spd_synth_set_setting(SPDConnection *connection, SynthSetting *setting,
 void *value);
 void free_synth_settings(SynthSettings **settings);


 Use more GLib in the server
 ---------------------------

  * Use GLib event loops where possible.
  * use GLib GThreads and GAsyncQueues for thread communication.
  * Use g_spawn calls for executing modules.
  * Support multiple client connection methods, unix socket, inet socket.
  * Use g_debug and other relevant GLib logging facilities for
    messages/logging.
  * Use GThreadedSocketService for handling client connections.
  * Replace custom implementations of parsing buffers with GLib equivalent
    methods where possible.


 Move audio into server
 ----------------------

  * Consider using a separate socket for audio transfer, however this may be
    difficult when attempting to synchronise with index marks. An alternative is
    to send index mark data via the audio socket as well.
  * Rework modules supporting audio output to not use any advanced internal
    playback queueing, and simply send the audio in relatively small buffers to
    the server. Smaller buffers to allow the server to stop/pause the audio more
    responsively.
  * Extend priority system to be either global priority, or priority per audio
    output device.
  * Rework pulseaudio output to use a GLib event loop.
  * Rework other audio output modules to better work within an event loop.


 Client audio retrieval API
 --------------------------

  * Allow client to either request audio directly, or have audio written to a
    designated file on disk.
  * Allow modules to decline the use of direct audio retrieval. I know of one
    speech synth that is not supported by speech dispatcher, who's licensing
    model doesn't allow for direct audio retrieval. If this module is ever
    supported, its code will likely remain closed to prevent people working
    around the implementation, but it would still be nice to support this synth
    in the longer term. (Luke Yelavich)
  * Load a new instance of the requested synth module, and spin up a worker
    thread to handle audio file writing or sending to client, to allow the server
    to dispatch other speech messages, as direct audio retrieval should be
    independant of the priority system.


 Server to module protocol documentation
 ---------------------------------------

 * Similar to the SSIp documentation, write up a texi document that explains
   the server to module protocol, currently over stdin/stdout, but may use other
   IPC in the future.


 Server to module protocol improvements
 --------------------------------------

  * Consider using sockets for IPC, with a dedicated socket per module.
  * Consider implementing shared memory support, particularly for audio data
    transfer, but this may depend on whether GLib has a shared memory API, The
    GMappedFile API may be useful, if the initiator can change the contents of
    the GMappedFile, and the other side can notice changes. Needs investigation.
  * Support the launching of modules via systems other than Speech Dispatcher,
    useful where containers of some sort are being used, and the environment
    requires that any separate processes are run in containers/other kind of
    sandbox, hense the use of sockets as per above.


 Integrate with logind/consolekit
 (Depends on migration to GSettings, GLib main event loops everywhere)
 --------------------------------

  * Query current user, and currently running sessions for that user.
  * Subscribe to tty change events and cork audio playback and synthesis flow
    if none of the user's sessions are active.
  * Allow the enabling/disabling of logind/consolekit via GSettings and at
    runtime, enabled being the default.
  * Allow the disabling of consolekit/logind at build time.
  * Consider abstracting this functionality into plugins, or at the very least
    separate code with an internal API to more easily support any future
    session/seat monitoring systems.


 Properly support system-wide mode
 ---------------------------------

  * Set a default user and group for the system wide instance to run under, at
    build time, and runtime.
  * Add a systemd unit to allow the use of system wide mode, disabled by
    default.


 Support spawning the server via Systemd socket activation
 ---------------------------------------------------------

  * Allow this to be enabled/disabled at build time.


 Copyright (C) 2001 Brailcom, o.p.s
 Copyright (C) 2016 Luke Yelavich <themuso@themuso.com>
 Copyright (C) 2018-2021 Samuel Thibault <samuel.thibault@ens-lyon.org>

 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
 Foundation; either version 2 of the License, or (at your option) any later
 version.

 This program is distributed in the hope that it will be useful, but WITHOUT ANY
 WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A
 PARTICULAR PURPOSE.  See the GNU General Public License for more details (file
 COPYING in the root directory).

 You should have received a copy of the GNU General Public License
 along with this program.  If not, see <https://www.gnu.org/licenses/>.
	Speech Dispatcher TODO
	======================

	The release versions are not final, and could change. Targetted release is
	based on demand from users, and difficulty of the work involved.

	* Add pitch as an option for capitalization presentation
	see https://github.com/brailcom/speechd/issues/24
	* Allow setting a synthesis voice in the user config using spd-conf.
	* Add support for Mimic
	see https://github.com/brailcom/speechd/issues/19
	* Drop getting information from other clients
	see https://github.com/brailcom/speechd/issues/335
	(0.11) Migrate to GSettings.
	(0.11) Synthesizer specific settings API.
	(0.11) Use more GLib in the server.
	(0.11) Client audio retrieval API.
	(0.11) Server to module protocol documentation.
	(0.12) Server to module protocol improvements.
	* Move synth modules to plugin architecture with plugin host.
	* Synth plugin API.
	* Allow for building synth plugins out of tree.
	(0.11) Integrate with logind/consolekit.
	(0.11) Properly support system-wide mode.
	* Support spawning the server via Systemd socket activation.


	The above improvements are documented in detail below. If work has started on
	a particular project, a git branch will be noted. These git branches are
	located at https://github.com/TheMuso/speechd-wip.git. To read the most up to
	date copy of this file, please clone the master Speech Dispatcher git
	repository, located at git://git.freebsoft.org/git/speechd.gitand check out
	the master branch.

	Migrate to GSettings
	--------------------

	* Write the GSettings metadata XML file.
	* Migrate the server to GSettings.
	* Listen to GSettings changes.
	* Migrate synthesizer modules to GSettings.
	* Write a program to migrate user settings to GSettings.


	Synthesizer specific settings API
	Depends on: Migration to GSettings
	---------------------------------

	Background:
	* Currently have API for espeak pitch range in git master, but this is only
	useful for espeak.
	* Espeak module has a config option to show variants along with available
	voices, which can be a very long list and can choak some clients.

	* Implement server to module protocol to support:
	- Request available settings.
	- Request available settings and their current value.
	- Request the value of a setting.
	- Set a setting.
	- Reset a setting to its default.
	* Implement SSIP protocol support.
	* Implement C API, see synthesizer specific settings C API draft.
	* Implement python API.

	Synthesizer specific settings C API draft

	typedef struct {
	char *name;
	char description; / This should be localized */
	enum SynthSettingValueType get_type;
	enum SynthSettingValueType set_type;
	int min_value;
	int max_value;
	char **value_list;
	void *cur_value;
	] SynthSetting;

	In the C API, a NULL terminated array of this structure would be returned for
	all settings a synth offers.

	The SynthSettingValueType enum would look something like this:

	typedef enum {
	SYNTH_SETTING_VALUE_UNKNOWN = 0,
	SYNTH_SETTING_VALUE_NUMBER = 1,
	SYNTH_SETTING_VALUE_STRING = 2,
	SYNTH_SETTING_VALUE_STRING_LIST = 3 /* A list of strings for the user
	to choose from, i.e voice variants */
	} SynthSettingValueType;

	C API methods to work with these data types could be as follows:

	SynthSetting *spd_synth_get_settings(SPDConnection connection);
	int spd_synth_set_setting(SPDConnection connection, SynthSetting setting,
	void *value);
	void free_synth_settings(SynthSettings **settings);


	Use more GLib in the server
	---------------------------

	* Use GLib event loops where possible.
	* use GLib GThreads and GAsyncQueues for thread communication.
	* Use g_spawn calls for executing modules.
	* Support multiple client connection methods, unix socket, inet socket.
	* Use g_debug and other relevant GLib logging facilities for
	messages/logging.
	* Use GThreadedSocketService for handling client connections.
	* Replace custom implementations of parsing buffers with GLib equivalent
	methods where possible.


	Move audio into server
	----------------------

	* Consider using a separate socket for audio transfer, however this may be
	difficult when attempting to synchronise with index marks. An alternative is
	to send index mark data via the audio socket as well.
	* Rework modules supporting audio output to not use any advanced internal
	playback queueing, and simply send the audio in relatively small buffers to
	the server. Smaller buffers to allow the server to stop/pause the audio more
	responsively.
	* Extend priority system to be either global priority, or priority per audio
	output device.
	* Rework pulseaudio output to use a GLib event loop.
	* Rework other audio output modules to better work within an event loop.


	Client audio retrieval API
	--------------------------

	* Allow client to either request audio directly, or have audio written to a
	designated file on disk.
	* Allow modules to decline the use of direct audio retrieval. I know of one
	speech synth that is not supported by speech dispatcher, who's licensing
	model doesn't allow for direct audio retrieval. If this module is ever
	supported, its code will likely remain closed to prevent people working
	around the implementation, but it would still be nice to support this synth
	in the longer term. (Luke Yelavich)
	* Load a new instance of the requested synth module, and spin up a worker
	thread to handle audio file writing or sending to client, to allow the server
	to dispatch other speech messages, as direct audio retrieval should be
	independant of the priority system.


	Server to module protocol documentation
	---------------------------------------

	* Similar to the SSIp documentation, write up a texi document that explains
	the server to module protocol, currently over stdin/stdout, but may use other
	IPC in the future.


	Server to module protocol improvements
	--------------------------------------

	* Consider using sockets for IPC, with a dedicated socket per module.
	* Consider implementing shared memory support, particularly for audio data
	transfer, but this may depend on whether GLib has a shared memory API, The
	GMappedFile API may be useful, if the initiator can change the contents of
	the GMappedFile, and the other side can notice changes. Needs investigation.
	* Support the launching of modules via systems other than Speech Dispatcher,
	useful where containers of some sort are being used, and the environment
	requires that any separate processes are run in containers/other kind of
	sandbox, hense the use of sockets as per above.


	Integrate with logind/consolekit
	(Depends on migration to GSettings, GLib main event loops everywhere)
	--------------------------------

	* Query current user, and currently running sessions for that user.
	* Subscribe to tty change events and cork audio playback and synthesis flow
	if none of the user's sessions are active.
	* Allow the enabling/disabling of logind/consolekit via GSettings and at
	runtime, enabled being the default.
	* Allow the disabling of consolekit/logind at build time.
	* Consider abstracting this functionality into plugins, or at the very least
	separate code with an internal API to more easily support any future
	session/seat monitoring systems.


	Properly support system-wide mode
	---------------------------------

	* Set a default user and group for the system wide instance to run under, at
	build time, and runtime.
	* Add a systemd unit to allow the use of system wide mode, disabled by
	default.


	Support spawning the server via Systemd socket activation
	---------------------------------------------------------

	* Allow this to be enabled/disabled at build time.


	Copyright (C) 2001 Brailcom, o.p.s
	Copyright (C) 2016 Luke Yelavich <themuso@themuso.com>
	Copyright (C) 2018-2021 Samuel Thibault <samuel.thibault@ens-lyon.org>

	This program is free software; you can redistribute it and/or modify it under
	the terms of the GNU General Public License as published by the Free Software
	Foundation; either version 2 of the License, or (at your option) any later
	version.

	This program is distributed in the hope that it will be useful, but WITHOUT ANY
	WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A
	PARTICULAR PURPOSE. See the GNU General Public License for more details (file
	COPYING in the root directory).

	You should have received a copy of the GNU General Public License
	along with this program. If not, see <https://www.gnu.org/licenses/>.