[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
WIP audio in server
|
From: |
Trevor Saunders |
|
Subject: |
WIP audio in server |
|
Date: |
Thu, 13 Aug 2015 18:43:32 -0400 |
On Mon, Aug 10, 2015 at 08:21:00PM -0600, Jeremy Whiting wrote:
> Ok, last patch of the day. Undone items:
>
> 1. Audio socket cleanup. Not sure what needs to be done here. Should
> the socket files get deleted during shutdown, etc.
is there some reason to use a socket instead of just pipe(2) then you
wouldn't need to deal with this at all?
> 2. Stopping audio (probably can be done from parse_stop in parse.c
> 3. Play command use which is only used in generic.c
iirc a couple other modules ivona and festival I think do there own
thing where the synth is a separate process and does its own audio I
think. I don't really have any great ideas here.
Have you considered using a process separate from the main server for
audio output? I guess its not all that complicated, but it might allow
us to sandbox a bit more of speech dispatcher.
Trev
>
> BR,
> Jeremy
>
> On Mon, Aug 10, 2015 at 7:05 PM, Jeremy Whiting <jpwhiting at kde.org> wrote:
> > Ok, another update. This time audio parameters are coming from the
> > user's config (I think, GlobalFDSet getting initialized is a mystery
> > to me so far because of macros and dotconf callbacks. Seems to work
> > here though.
> >
> > BR,
> > Jeremy
> >
> >
> > On Mon, Aug 10, 2015 at 5:43 PM, Jeremy Whiting <jpwhiting at kde.org>
> > wrote:
> >> Ok, here's a working patch. A few things I'll fix before this is ready
> >> for master though.
> >>
> >> 1. Audio initialization needs to come from the config files again.
> >> 2. Audio socket cleanup.
> >> 3. Documentation changes for this big change in how spd works.
> >> 4. How to request the server stop playing audio (or maybe it knows
> >> because it's telling the modules the same thing...
> >> 5. Audio file playback in generic.c needs to open the file and send
> >> the audio on the socket.
> >>
> >> BR,
> >> Jeremy
> >>
> >>
> >> On Wed, Aug 5, 2015 at 5:42 PM, Jeremy Whiting <jpwhiting at kde.org>
> >> wrote:
> >>> Ok, the audio in the server is in it's own thread now, and mostly all
> >>> code is in server/audio.c to keep it separate from the other file
> >>> descriptor handling for clients. I'm still getting pauses for some
> >>> reason, but it is threaded now at least and works when run with -d
> >>> also. I'll try to figure out where the pauses are coming from.
> >>>
> >>> BR,
> >>> Jeremy
> >>>
> >>> On Fri, Jul 31, 2015 at 6:47 PM, Jeremy Whiting <jpwhiting at kde.org>
> >>> wrote:
> >>>> I've spent a bit of time wrestling with this code today and have found
> >>>> the following.
> >>>>
> >>>> If I don't initialize pulse by calling _pulse_open but only initialize
> >>>> once I have data to set the format it works mostly, though there are
> >>>> pauses of silence in long phrases from espeak still.
> >>>> If I do initialize pulse by calling _pulse_open in pulse_open and also
> >>>> reinitializing in pulse_play as required for audio rate changes etc.
> >>>> it doesn't play anything somehow (it's getting stuck in pa_simple_free
> >>>> on a mutex somehow).
> >>>>
> >>>> Do I need to run a thread for every audio socket we attach in the
> >>>> server? If so what should the thread's start_routine look like, just
> >>>> while(1) and the callback audio_process_incoming is called when we get
> >>>> traffic on the fd ? Once this is working (and I clean up how we
> >>>> initialize the audio to not hard coded values, but get the config from
> >>>> the config file) I can look at making the audio use the pulse async
> >>>> api, but I want to get the proof of concept working as a first step.
> >>>> My knowledge of pthreads seems to be blocking me at the moment though.
> >>>>
> >>>> thanks,
> >>>> Jeremy
> >>>>
> >>>>
> >>>>
> >>>> On Thu, Jul 30, 2015 at 7:42 PM, Jeremy Whiting <jpwhiting at kde.org>
> >>>> wrote:
> >>>>> Oops, I didn't see this until just now. At any rate I got that issue
> >>>>> solved, now only playback failing (I verified I get samples in the
> >>>>> server and I think it's initializing the AudioId properly since it's
> >>>>> not NULL here. It seems to think it's playing from all the return
> >>>>> values, but I hear no audio yet. I also wasn't sure what to do about
> >>>>> sending the audio parameters just yet, so hard coded some based on my
> >>>>> config here for now.
> >>>>>
> >>>>> BR,
> >>>>> Jeremy
> >>>>>
> >>>>> On Thu, Jul 30, 2015 at 6:40 PM, Luke Yelavich
> >>>>> <luke.yelavich at canonical.com> wrote:
> >>>>>> On Fri, Jul 31, 2015 at 10:40:04AM AEST, Luke Yelavich wrote:
> >>>>>>> On Fri, Jul 31, 2015 at 07:27:27AM AEST, Jeremy Whiting wrote:
> >>>>>>> > Hey all,
> >>>>>>> >
> >>>>>>> > I'm implementing moving audio from the modules to the server (and
> >>>>>>> > modules will send audio data to the server on a unix socket). I've
> >>>>>>> > got
> >>>>>>> > the socket creation, and seem to have the ability to connect to the
> >>>>>>> > socket in the modules but it's hanging here when I try to run
> >>>>>>> > spd-say
> >>>>>>> > hello. Also I'm getting this in my speech-dispatcher.log as if it's
> >>>>>>> > trying to open a second audio connection from sd_espeak for some
> >>>>>>> > reason when it hangs (and no log output after this):
> >>>>>>> >
> >>>>>>> > [Thu Jul 30 15:03:15 2015 : 829380] speechd: Adding audio
> >>>>>>> > connection on socket 4
> >>>>>>> > [Thu Jul 30 15:03:29 2015 : 105629] speechd: Adding module on fd
> >>>>>>> > 28
> >>>>>>> > [Thu Jul 30 15:03:29 2015 : 105654] speechd: Adding audio
> >>>>>>> > connection on socket 4
> >>>>>>> >
> >>>>>>> > I'm probably doing something obviously wrong, but can't seem to see
> >>>>>>> > what at the moment though I've been beating my head against it for a
> >>>>>>> > while and debugging. Can you see anything obvious in my changes?
> >>>>>>>
> >>>>>>> Well, I wonder if what you have in speechd_audio_connection_new is
> >>>>>>> correct. You make reference to module sockets and the server socket
> >>>>>>> where clients connect, and not the audio socket.
> >>>>>>
> >>>>>> Helps if I attach the diff.
> >>>>>>
> >>>>>> Luke
> From 51316a1237c09b50a1c84a44348a278c5982a056 Mon Sep 17 00:00:00 2001
> From: Jeremy Whiting <jpwhiting at kde.org>
> Date: Wed, 29 Jul 2015 18:28:03 -0600
> Subject: [PATCH] Moving audio from modules to the server.
>
> Audio socket is created in the server.
> Audio system is initialized in the server.
> Modules connect to audio socket and send audio data (metadata first separated
> by :, then samples).
> Server receives AudioTrack deserializes the metadata and plays AudioTrack.
> Updated speech-dispatcher.texi to remove section about AUDIO command handling
> as it's gone.
> ---
> doc/speech-dispatcher.texi | 5 -
> include/speechd_defines.h | 19 +-
> src/Makefile.am | 2 +-
> src/modules/Makefile.am | 30 +--
> src/modules/espeak.c | 34 +--
> src/modules/festival.c | 7 +-
> src/modules/generic.c | 4 +-
> src/modules/module_main.c | 16 +-
> src/modules/module_utils.c | 184 +++++--------
> src/modules/module_utils.h | 4 +-
> src/modules/spd_audio.c | 321 ----------------------
> src/modules/spd_audio.h | 46 ----
> src/server/Makefile.am | 8 +-
> src/server/audio.c | 647
> +++++++++++++++++++++++++++++++++++++++++++++
> src/server/audio.h | 60 +++++
> src/server/module.c | 12 -
> src/server/msg.h | 3 +-
> src/server/output.c | 23 --
> src/server/output.h | 1 -
> src/server/server.c | 1 +
> src/server/set.c | 14 +
> src/server/set.h | 1 +
> src/server/speechd.c | 21 +-
> src/server/speechd.h | 9 +
> 24 files changed, 877 insertions(+), 595 deletions(-)
> delete mode 100644 src/modules/spd_audio.c
> delete mode 100644 src/modules/spd_audio.h
> create mode 100644 src/server/audio.c
> create mode 100644 src/server/audio.h
>
> diff --git a/doc/speech-dispatcher.texi b/doc/speech-dispatcher.texi
> index 88ca972..2d1a0ad 100755
> --- a/doc/speech-dispatcher.texi
> +++ b/doc/speech-dispatcher.texi
> @@ -2752,11 +2752,6 @@ punctuation_some=NULL
> 203 OK SETTINGS RECEIVED
> @end example
>
> - at item AUDIO
> -Audio has exactly the same structure as @code{SET}, but is transmitted
> -only once immediatelly after @code{INIT} to transmit the requested audio
> -parameters and tell the output module to open the audio device.
> -
> @item QUIT
> Terminates the output module. It should send the response, deallocate
> all the resources, close all descriptors, terminate all child
> diff --git a/include/speechd_defines.h b/include/speechd_defines.h
> index 3a544f0..d540c29 100644
> --- a/include/speechd_defines.h
> +++ b/include/speechd_defines.h
> @@ -22,14 +22,15 @@
> #ifndef SPEECHD_DEFINES_H
> #define SPEECHD_DEFINES_H
>
> -#define SPD_ALLCLIENTS "ALL"
> -#define SPD_SELF "SELF"
> -#define SPD_VOLUME "VOLUME"
> -#define SPD_PITCH "PITCH"
> -#define SPD_PITCH_RANGE "PITCH_RANGE"
> -#define SPD_RATE "RATE"
> -#define SPD_LANGUAGE "LANGUAGE"
> -#define SPD_OUTPUT_MODULE "OUTPUT_MODULE"
> -#define SPD_SYNTHESIS_VOICE "SYNTHESIS_VOICE"
> +#define NEWLINE "\r\n"
> +#define SPD_ALLCLIENTS "ALL"
> +#define SPD_SELF "SELF"
> +#define SPD_VOLUME "VOLUME"
> +#define SPD_PITCH "PITCH"
> +#define SPD_PITCH_RANGE "PITCH_RANGE"
> +#define SPD_RATE "RATE"
> +#define SPD_LANGUAGE "LANGUAGE"
> +#define SPD_OUTPUT_MODULE "OUTPUT_MODULE"
> +#define SPD_SYNTHESIS_VOICE "SYNTHESIS_VOICE"
>
> #endif /* not ifndef SPEECHD_DEFINES_H */
> diff --git a/src/Makefile.am b/src/Makefile.am
> index 81d0690..8690889 100644
> --- a/src/Makefile.am
> +++ b/src/Makefile.am
> @@ -1,4 +1,4 @@
> ## Process this file with automake to produce Makefile.in
>
> -SUBDIRS=common server audio modules api clients tests
> +SUBDIRS=common audio server modules api clients tests
>
> diff --git a/src/modules/Makefile.am b/src/modules/Makefile.am
> index 9012a4b..7010e39 100644
> --- a/src/modules/Makefile.am
> +++ b/src/modules/Makefile.am
> @@ -1,84 +1,74 @@
> ## Process this file with automake to produce Makefile.in
>
> inc_local = -I$(top_srcdir)/include
> -audio_SOURCES = spd_audio.c spd_audio.h
> common_SOURCES = module_main.c module_utils.c module_utils.h
> common_LDADD = $(SNDFILE_LIBS) $(DOTCONF_LIBS) $(GLIB_LIBS) $(GTHREAD_LIBS)
>
> AM_CFLAGS = $(ERROR_CFLAGS)
> AM_CPPFLAGS = $(inc_local) -DDATADIR=\"$(snddatadir)\" -D_GNU_SOURCE \
> - -DPLUGIN_DIR="\"$(audiodir)\"" \
> $(DOTCONF_CFLAGS) $(GLIB_CFLAGS) $(GTHREAD_CFLAGS) \
> $(ibmtts_include) $(SNDFILE_CFLAGS)
>
> modulebin_PROGRAMS = sd_dummy sd_generic sd_festival sd_cicero
>
> -sd_dummy_SOURCES = dummy.c $(audio_SOURCES) $(common_SOURCES) \
> +sd_dummy_SOURCES = dummy.c $(common_SOURCES) \
> module_utils_addvoice.c
> sd_dummy_LDADD = $(top_builddir)/src/common/libcommon.la \
> - $(audio_dlopen_modules) \
> $(common_LDADD)
> dist_snddata_DATA = dummy-message.wav
>
> sd_festival_SOURCES = festival.c festival_client.c festival_client.h \
> - $(audio_SOURCES) $(common_SOURCES)
> + $(common_SOURCES)
> sd_festival_LDADD = $(top_builddir)/src/common/libcommon.la \
> - $(audio_dlopen_modules) \
> $(common_LDADD) $(EXTRA_SOCKET_LIBS)
>
> -sd_generic_SOURCES = generic.c $(audio_SOURCES) $(common_SOURCES) \
> +sd_generic_SOURCES = generic.c $(common_SOURCES) \
> module_utils_addvoice.c
> sd_generic_LDADD = $(top_builddir)/src/common/libcommon.la \
> - $(audio_dlopen_modules) \
> $(common_LDADD)
>
> -sd_cicero_SOURCES = cicero.c $(audio_SOURCES) $(common_SOURCES)
> +sd_cicero_SOURCES = cicero.c $(common_SOURCES)
> sd_cicero_LDADD = $(top_builddir)/src/common/libcommon.la \
> - $(audio_dlopen_modules) \
> $(common_LDADD)
>
> if flite_support
> modulebin_PROGRAMS += sd_flite
> -sd_flite_SOURCES = flite.c $(audio_SOURCES) $(common_SOURCES)
> +sd_flite_SOURCES = flite.c $(common_SOURCES)
> sd_flite_LDADD = $(top_builddir)/src/common/libcommon.la \
> - $(audio_dlopen_modules) \
> $(flite_kal) $(flite_basic) \
> $(common_LDADD)
> endif
>
> if ibmtts_support
> modulebin_PROGRAMS += sd_ibmtts
> -sd_ibmtts_SOURCES = ibmtts.c $(audio_SOURCES) $(common_SOURCES) \
> +sd_ibmtts_SOURCES = ibmtts.c $(common_SOURCES) \
> module_utils_addvoice.c
> sd_ibmtts_LDADD = $(top_builddir)/src/common/libcommon.la \
> - $(audio_dlopen_modules) \
> -libmeci \
> $(common_LDADD)
> endif
>
> if espeak_support
> modulebin_PROGRAMS += sd_espeak
> -sd_espeak_SOURCES = espeak.c $(audio_SOURCES) $(common_SOURCES)
> +sd_espeak_SOURCES = espeak.c $(common_SOURCES)
> sd_espeak_LDADD = $(top_builddir)/src/common/libcommon.la \
> - $(audio_dlopen_modules) \
> -lespeak $(EXTRA_ESPEAK_LIBS) \
> $(common_LDADD)
> endif
>
> if ivona_support
> modulebin_PROGRAMS += sd_ivona
> -sd_ivona_SOURCES = ivona.c ivona_client.c ivona_client.h $(audio_SOURCES) \
> +sd_ivona_SOURCES = ivona.c ivona_client.c ivona_client.h \
> $(common_SOURCES)
> sd_ivona_LDADD = $(top_builddir)/src/common/libcommon.la \
> - $(audio_dlopen_modules) \
> -ldumbtts \
> $(common_LDADD)
> endif
>
> if pico_support
> modulebin_PROGRAMS += sd_pico
> -sd_pico_SOURCES = pico.c $(audio_SOURCES) $(common_SOURCES)
> +sd_pico_SOURCES = pico.c $(common_SOURCES)
> sd_pico_LDADD = $(top_builddir)/src/common/libcommon.la \
> - $(audio_dlopen_modules) -lttspico \
> + -lttspico \
> $(common_LDADD)
> endif
> diff --git a/src/modules/espeak.c b/src/modules/espeak.c
> index 48fc743..6fb9ffe 100644
> --- a/src/modules/espeak.c
> +++ b/src/modules/espeak.c
> @@ -43,7 +43,6 @@
> #endif
>
> /* Speech Dispatcher includes. */
> -#include "spd_audio.h"
> #include <speechd_types.h>
> #include "module_utils.h"
>
> @@ -558,22 +557,23 @@ static void *_espeak_stop_or_pause(void *nothing)
> pthread_cond_broadcast(&playback_queue_condition);
> pthread_mutex_unlock(&playback_queue_mutex);
>
> - if (module_audio_id) {
> - DBG("Espeak: Stopping audio.");
> - ret = spd_audio_stop(module_audio_id);
> - DBG_WARN(ret == 0,
> - "spd_audio_stop returned non-zero value.");
> - while (is_thread_busy(&espeak_play_suspended_mutex)) {
> - ret = spd_audio_stop(module_audio_id);
> - DBG_WARN(ret == 0,
> - "spd_audio_stop returned non-zero
> value.");
> - g_usleep(5000);
> - }
> - } else {
> - while (is_thread_busy(&espeak_play_suspended_mutex)) {
> - g_usleep(5000);
> - }
> - }
> + /* TODO: Add a way to request the server stop audio playback */
> +// if (module_audio_id) {
> +// DBG("Espeak: Stopping audio.");
> +// ret = spd_audio_stop(module_audio_id);
> +// DBG_WARN(ret == 0,
> +// "spd_audio_stop returned non-zero value.");
> +// while (is_thread_busy(&espeak_play_suspended_mutex)) {
> +// ret = spd_audio_stop(module_audio_id);
> +// DBG_WARN(ret == 0,
> +// "spd_audio_stop returned non-zero
> value.");
> +// g_usleep(5000);
> +// }
> +// } else {
> +// while (is_thread_busy(&espeak_play_suspended_mutex)) {
> +// g_usleep(5000);
> +// }
> +// }
>
> DBG("Espeak: Waiting for synthesis to stop.");
> ret = espeak_Cancel();
> diff --git a/src/modules/festival.c b/src/modules/festival.c
> index df6d58a..39ab49e 100644
> --- a/src/modules/festival.c
> +++ b/src/modules/festival.c
> @@ -431,9 +431,10 @@ int module_stop(void)
> if (!festival_stop) {
> pthread_mutex_lock(&sound_output_mutex);
> festival_stop = 1;
> - if (festival_speaking && module_audio_id) {
> - spd_audio_stop(module_audio_id);
> - }
> + // TODO: Add a way for modules to request speech stop
> maybe ?
> +// if (festival_speaking && module_audio_id) {
> +// spd_audio_stop(module_audio_id);
> +// }
> pthread_mutex_unlock(&sound_output_mutex);
> }
> }
> diff --git a/src/modules/generic.c b/src/modules/generic.c
> index 172b6ec..0360f0b 100644
> --- a/src/modules/generic.c
> +++ b/src/modules/generic.c
> @@ -388,8 +388,8 @@ void *_generic_speak(void *nothing)
> homedir =
> g_strdup("UNKNOWN_HOME_DIRECTORY");
>
> - play_command =
> - spd_audio_get_playcmd(module_audio_id);
> +// play_command =
> +// spd_audio_get_playcmd(module_audio_id);
> if (play_command == NULL) {
> DBG("This audio backend has no default
> play command; using \"play\"\n");
> play_command = "play";
> diff --git a/src/modules/module_main.c b/src/modules/module_main.c
> index f53b053..49b73bc 100644
> --- a/src/modules/module_main.c
> +++ b/src/modules/module_main.c
> @@ -31,7 +31,6 @@
> #include <pthread.h>
> #include <glib.h>
> #include <dotconf.h>
> -#include <ltdl.h>
>
> #include <spd_utils.h>
> #include "module_utils.h"
> @@ -77,10 +76,7 @@ int main(int argc, char *argv[])
> char *configfilename = NULL;
> char *status_info = NULL;
>
> - /* Initialize ltdl's list of preloaded audio backends. */
> - LTDL_SET_PRELOADED_SYMBOLS();
> module_num_dc_options = 0;
> - module_audio_id = 0;
>
> if (argc >= 2) {
> configfilename = g_strdup(argv[1]);
> @@ -149,6 +145,16 @@ int main(int argc, char *argv[])
> exit(1);
> }
>
> + ret = module_audio_init(&status_info);
> +
> + if (ret != 0) {
> + printf("399-%s\n", status_info);
> + printf("%s\n", "399 ERR CANT INIT AUDIO");
> + g_free(status_info);
> + module_close();
> + exit(1);
> + }
> +
> printf("299-%s\n", status_info);
> ret = printf("%s\n", "299 OK LOADED SUCCESSFULLY");
>
> @@ -190,8 +196,6 @@ int main(int argc, char *argv[])
> else
> PROCESS_CMD(SET, do_set)
> else
> - PROCESS_CMD(AUDIO, do_audio)
> - else
> PROCESS_CMD(LOGLEVEL, do_loglevel)
> else
> PROCESS_CMD_W_ARGS(DEBUG, do_debug)
> diff --git a/src/modules/module_utils.c b/src/modules/module_utils.c
> index 298274a..bde2257 100644
> --- a/src/modules/module_utils.c
> +++ b/src/modules/module_utils.c
> @@ -26,17 +26,22 @@
> #endif
>
> #include <sndfile.h>
> +#include <sys/types.h>
> +#include <sys/socket.h>
> +#include <sys/un.h>
>
> #include <fdsetconv.h>
> #include <spd_utils.h>
> #include "module_utils.h"
> -
> -static char *module_audio_pars[10];
> +#include <speechd_defines.h>
>
> extern char *module_index_mark;
>
> pthread_mutex_t module_stdout_mutex = PTHREAD_MUTEX_INITIALIZER;
>
> +/* Socket for sending audio data to */
> +int audio_socket;
> +
> char *do_message(SPDMessageType msgtype)
> {
> int ret;
> @@ -95,11 +100,6 @@ char *do_message(SPDMessageType msgtype)
> msg_settings_old.voice_type = -1;
> }
>
> - /* Volume is controlled by the synthesizer. Always play at normal on
> audio device. */
> - if (spd_audio_set_volume(module_audio_id, 85) < 0) {
> - DBG("Can't set volume. audio not initialized?");
> - }
> -
> ret = module_speak(msg->str, strlen(msg->str), msgtype);
>
> g_string_free(msg, 1);
> @@ -267,91 +267,12 @@ char *do_set(void)
> return g_strdup("401 ERROR INTERNAL"); /* Can't be reached */
> }
>
> -#define SET_AUDIO_STR(name,idx) \
> - if(!strcmp(cur_item, #name)){ \
> - g_free(module_audio_pars[idx]); \
> - if(!strcmp(cur_value, "NULL")) module_audio_pars[idx] = NULL; \
> - else module_audio_pars[idx] = g_strdup(cur_value); \
> - }
> -
> -char *do_audio(void)
> -{
> - char *cur_item = NULL;
> - char *cur_value = NULL;
> - char *line = NULL;
> - int ret;
> - size_t n;
> - int err = 0; /* Error status */
> - char *status = NULL;
> - char *msg;
> -
> - printf("207 OK RECEIVING AUDIO SETTINGS\n");
> - fflush(stdout);
> -
> - while (1) {
> - line = NULL;
> - n = 0;
> - ret = spd_getline(&line, &n, stdin);
> - if (ret == -1) {
> - err = 1;
> - break;
> - }
> - if (!strcmp(line, ".\n")) {
> - g_free(line);
> - break;
> - }
> - if (!err) {
> - cur_item = strtok(line, "=");
> - if (cur_item == NULL) {
> - err = 1;
> - continue;
> - }
> - cur_value = strtok(NULL, "\n");
> - if (cur_value == NULL) {
> - err = 1;
> - continue;
> - }
> -
> - SET_AUDIO_STR(audio_output_method, 0)
> - else
> - SET_AUDIO_STR(audio_oss_device, 1)
> - else
> - SET_AUDIO_STR(audio_alsa_device, 2)
> - else
> - SET_AUDIO_STR(audio_nas_server, 3)
> - else
> - SET_AUDIO_STR(audio_pulse_server, 4)
> - else
> - SET_AUDIO_STR(audio_pulse_min_length, 5)
> - else
> - err = 2; /* Unknown parameter */
> - }
> - g_free(line);
> - }
> -
> - if (err == 1)
> - return g_strdup("302 ERROR BAD SYNTAX");
> - if (err == 2)
> - return g_strdup("303 ERROR INVALID PARAMETER OR VALUE");
> -
> - err = module_audio_init(&status);
> -
> - if (err == 0)
> - msg = g_strdup_printf("203 OK AUDIO INITIALIZED");
> - else
> - msg = g_strdup_printf("300-%s\n300 UNKNOWN ERROR", status);
> -
> - g_free(status);
> - return msg;
> -}
> -
> #define SET_LOGLEVEL_NUM(name, cond) \
> if(!strcmp(cur_item, #name)){ \
> number = strtol(cur_value, &tptr, 10); \
> if(!(cond)){ err = 2; continue; } \
> if (tptr == cur_value){ err = 2; continue; } \
> log_level = number; \
> - spd_audio_set_loglevel(module_audio_id, number); \
> }
>
> char *do_loglevel(void)
> @@ -516,9 +437,6 @@ void do_quit(void)
> printf("210 OK QUIT\n");
> fflush(stdout);
>
> - spd_audio_close(module_audio_id);
> - module_audio_id = NULL;
> -
> module_close();
> return;
> }
> @@ -988,50 +906,74 @@ void *module_get_ht_option(GHashTable * hash_table,
> const char *key)
> return option;
> }
>
> -int module_audio_init(char **status_info)
> +/* Determine address for the unix socket */
> +static char *_get_default_audio_unix_socket_name(void)
> {
> - char *error = 0;
> - gchar **outputs;
> - int i = 0;
> + GString *socket_filename;
> + char *h;
> + const char *rundir = g_get_user_runtime_dir();
> + socket_filename = g_string_new("");
> + g_string_printf(socket_filename, "%s/speech-dispatcher/audio.sock",
> + rundir);
> + // Do not return glib string, but glibc string...
> + h = strdup(socket_filename->str);
> + g_string_free(socket_filename, 1);
> + return h;
> +}
>
> - DBG("Openning audio output system");
> - if (NULL == module_audio_pars[0]) {
> - *status_info =
> +int module_audio_init(char **status_info)
> +{
> + /* Open connection to audio socket */
> + char *str;
> + char *socket_filename = _get_default_audio_unix_socket_name();
> + int len;
> + struct sockaddr_un server;
> +
> + if ((audio_socket = socket(AF_UNIX, SOCK_STREAM, 0)) == -1) {
> + *status_info =
> g_strdup
> - ("Sound output method specified in configuration not
> supported. "
> - "Please choose 'oss', 'alsa', 'nas', 'libao' or 'pulse'.");
> + ("Unable to create socket to send audio data");
> return -1;
> }
> -
> - outputs = g_strsplit(module_audio_pars[0], ",", 0);
> - while (NULL != outputs[i]) {
> - module_audio_id =
> - spd_audio_open(outputs[i], (void **)&module_audio_pars[1],
> - &error);
> - if (module_audio_id) {
> - DBG("Using %s audio output method", outputs[i]);
> - g_strfreev(outputs);
> - *status_info =
> - g_strdup("audio initialized successfully.");
> - return 0;
> - }
> - i++;
> +
> + server.sun_family = AF_UNIX;
> + strcpy(server.sun_path, socket_filename);
> + len = strlen(server.sun_path) + sizeof(server.sun_family);
> + if (connect(audio_socket, (struct sockaddr *)&server, len) == -1) {
> + *status_info =
> + g_strdup_printf
> + ("Unable to connect to server socket at %s",
> socket_filename);
> + return -1;
> }
>
> - *status_info =
> - g_strdup_printf("Opening sound device failed. Reason: %s. ", error);
> - g_free(error); /* g_malloc'ed, in spd_audio_open. */
> -
> - g_strfreev(outputs);
> - return -1;
> + str = g_strdup("ACK"NEWLINE);
> + if (send(audio_socket, str, strlen(str), 0) == -1) {
> + g_free (str);
> + *status_info =
> + g_strdup_printf
> + ("Unable to send ACK on audio socket %s", socket_filename);
> + return -1;
> + }
> + g_free(str);
>
> + return 0;
> }
>
> int module_tts_output(AudioTrack track, AudioFormat format)
> {
> -
> - if (spd_audio_play(module_audio_id, track, format) < 0) {
> - DBG("Can't play track for unknown reason.");
> + /* Send audiotrack data to the socket */
> + char *metadata = g_strdup_printf("%d:%d:%d:%d:%d"NEWLINE,
> + format,
> + track.bits,
> + track.num_channels,
> + track.sample_rate,
> + track.num_samples);
> + if (send(audio_socket, metadata, strlen(metadata), 0) == -1) {
> + DBG("Can't send audiotrack metadata for some reason.");
> + return -1;
> + }
> + if (send(audio_socket, track.samples, track.num_samples * sizeof(signed
> short), 0) == -1) {
> + DBG("Can't send audio samples for some reason.");
> return -1;
> }
> return 0;
> diff --git a/src/modules/module_utils.h b/src/modules/module_utils.h
> index 7483930..895db80 100644
> --- a/src/modules/module_utils.h
> +++ b/src/modules/module_utils.h
> @@ -41,12 +41,10 @@
> #include <sys/ipc.h>
>
> #include <speechd_types.h>
> -#include "spd_audio.h"
> +#include <spd_audio_plugin.h>
>
> int log_level;
>
> -AudioID *module_audio_id;
> -
> SPDMsgSettings msg_settings;
> SPDMsgSettings msg_settings_old;
>
> diff --git a/src/modules/spd_audio.c b/src/modules/spd_audio.c
> deleted file mode 100644
> index 9bf8e37..0000000
> --- a/src/modules/spd_audio.c
> +++ /dev/null
> @@ -1,321 +0,0 @@
> -
> -/*
> - * spd_audio.c -- Spd Audio Output Library
> - *
> - * Copyright (C) 2004, 2006 Brailcom, o.p.s.
> - *
> - * This is free software; you can redistribute it and/or modify it under the
> - * terms of the GNU Lesser General Public License as published by the Free
> - * Software Foundation; either version 2.1, or (at your option) any later
> - * version.
> - *
> - * This software 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.
> - *
> - * You should have received a copy of the GNU Lesser General Public License
> - * along with this package; see the file COPYING. If not, write to the Free
> - * Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
> - * 02110-1301, USA.
> - *
> - * $Id: spd_audio.c,v 1.21 2008-06-09 10:29:12 hanke Exp $
> - */
> -
> -/*
> - * spd_audio is a simple realtime audio output library with the capability of
> - * playing 8 or 16 bit data, immediate stop and synchronization. This library
> - * currently provides OSS, NAS, ALSA and PulseAudio backend. The available
> backends are
> - * specified at compile-time using the directives WITH_OSS, WITH_NAS,
> WITH_ALSA,
> - * WITH_PULSE, WITH_LIBAO but the user program is allowed to switch between
> them at run-time.
> - */
> -
> -#ifdef HAVE_CONFIG_H
> -#include <config.h>
> -#endif
> -
> -#include "spd_audio.h"
> -
> -#include <stdio.h>
> -#include <string.h>
> -#include <fcntl.h>
> -#include <sys/ioctl.h>
> -#include <sys/time.h>
> -#include <time.h>
> -#include <unistd.h>
> -#include <errno.h>
> -
> -#include <pthread.h>
> -
> -#include <glib.h>
> -#include <ltdl.h>
> -
> -static int spd_audio_log_level;
> -static lt_dlhandle lt_h;
> -
> -/* Dynamically load a library with RTLD_GLOBAL set.
> -
> - This is needed when a dynamically-loaded library has its own plugins
> - that call into the parent library.
> - Most of the credit for this function goes to Gary Vaughan.
> -*/
> -static lt_dlhandle my_dlopenextglobal(const char *filename)
> -{
> - lt_dlhandle handle = NULL;
> - lt_dladvise advise;
> -
> - if (lt_dladvise_init(&advise))
> - return handle;
> -
> - if (!lt_dladvise_ext(&advise) && !lt_dladvise_global(&advise))
> - handle = lt_dlopenadvise(filename, advise);
> -
> - lt_dladvise_destroy(&advise);
> - return handle;
> -}
> -
> -/* Open the audio device.
> -
> - Arguments:
> - type -- The requested device. Currently AudioOSS or AudioNAS.
> - pars -- and array of pointers to parameters to pass to
> - the device backend, terminated by a NULL pointer.
> - See the source/documentation of each specific backend.
> - error -- a pointer to the string where error description is
> - stored in case of failure (returned AudioID == NULL).
> - Otherwise will contain NULL.
> -
> - Return value:
> - Newly allocated AudioID structure that can be passed to
> - all other spd_audio functions, or NULL in case of failure.
> -
> -*/
> -AudioID *spd_audio_open(char *name, void **pars, char **error)
> -{
> - AudioID *id;
> - spd_audio_plugin_t const *p;
> - spd_audio_plugin_t *(*fn) (void);
> - gchar *libname;
> - int ret;
> -
> - /* now check whether dynamic plugin is available */
> - ret = lt_dlinit();
> - if (ret != 0) {
> - *error = (char *)g_strdup_printf("lt_dlinit() failed");
> - return (AudioID *) NULL;
> - }
> -
> - ret = lt_dlsetsearchpath(PLUGIN_DIR);
> - if (ret != 0) {
> - *error = (char *)g_strdup_printf("lt_dlsetsearchpath() failed");
> - return (AudioID *) NULL;
> - }
> -
> - libname = g_strdup_printf(SPD_AUDIO_LIB_PREFIX "%s", name);
> - lt_h = my_dlopenextglobal(libname);
> - g_free(libname);
> - if (NULL == lt_h) {
> - *error =
> - (char *)g_strdup_printf("Cannot open plugin %s. error: %s",
> - name, lt_dlerror());
> - return (AudioID *) NULL;
> - }
> -
> - fn = lt_dlsym(lt_h, SPD_AUDIO_PLUGIN_ENTRY_STR);
> - if (NULL == fn) {
> - *error = (char *)g_strdup_printf("Cannot find symbol %s",
> - SPD_AUDIO_PLUGIN_ENTRY_STR);
> - return (AudioID *) NULL;
> - }
> -
> - p = fn();
> - if (p == NULL || p->name == NULL) {
> - *error = (char *)g_strdup_printf("plugin %s not found", name);
> - return (AudioID *) NULL;
> - }
> -
> - id = p->open(pars);
> - if (id == NULL) {
> - *error =
> - (char *)g_strdup_printf("Couldn't open %s plugin", name);
> - return (AudioID *) NULL;
> - }
> -
> - id->function = p;
> -#if defined(BYTE_ORDER) && (BYTE_ORDER == BIG_ENDIAN)
> - id->format = SPD_AUDIO_BE;
> -#else
> - id->format = SPD_AUDIO_LE;
> -#endif
> -
> - *error = NULL;
> -
> - return id;
> -}
> -
> -/* Play a track on the audio device (blocking).
> -
> - Arguments:
> - id -- the AudioID* of the device returned by spd_audio_open
> - track -- a track to play (see spd_audio.h)
> -
> - Return value:
> - 0 if everything is ok, a non-zero value in case of failure.
> - See the particular backend documentation or source for the
> - meaning of these non-zero values.
> -
> - Comment:
> - spd_audio_play() is a blocking function. It returns exactly
> - when the given track stopped playing. However, it's possible
> - to safely interrupt it using spd_audio_stop() described below.
> - (spd_audio_stop() needs to be called from another thread, obviously.)
> -
> -*/
> -int spd_audio_play(AudioID * id, AudioTrack track, AudioFormat format)
> -{
> - int ret;
> -
> - if (id && id->function->play) {
> - /* Only perform byte swapping if the driver in use has given us
> audio in
> - an endian format other than what the running CPU supports. */
> - if (format != id->format) {
> - unsigned char *out_ptr, *out_end, c;
> - out_ptr = (unsigned char *)track.samples;
> - out_end =
> - out_ptr +
> - track.num_samples * 2 * track.num_channels;
> - while (out_ptr < out_end) {
> - c = out_ptr[0];
> - out_ptr[0] = out_ptr[1];
> - out_ptr[1] = c;
> - out_ptr += 2;
> - }
> - }
> - ret = id->function->play(id, track);
> - } else {
> - fprintf(stderr, "Play not supported on this device\n");
> - return -1;
> - }
> -
> - return ret;
> -}
> -
> -/* Stop playing the current track on device id
> -
> -Arguments:
> - id -- the AudioID* of the device returned by spd_audio_open
> -
> -Return value:
> - 0 if everything is ok, a non-zero value in case of failure.
> - See the particular backend documentation or source for the
> - meaning of these non-zero values.
> -
> -Comment:
> - spd_audio_stop() safely interrupts spd_audio_play() when called
> - from another thread. It shouldn't cause any clicks or unwanted
> - effects in the sound output.
> -
> - It's safe to call spd_audio_stop() even if the device isn't playing
> - any track. In that case, it does nothing. However, there is a danger
> - when using spd_audio_stop(). Since you must obviously do it from
> - another thread than where spd_audio_play is running, you must make
> - yourself sure that the device is still open and the id you pass it
> - is valid and will be valid until spd_audio_stop returns. In other words,
> - you should use some mutex or other synchronization device to be sure
> - spd_audio_close isn't called before or during spd_audio_stop execution.
> -*/
> -
> -int spd_audio_stop(AudioID * id)
> -{
> - int ret;
> - if (id && id->function->stop) {
> - ret = id->function->stop(id);
> - } else {
> - fprintf(stderr, "Stop not supported on this device\n");
> - return -1;
> - }
> - return ret;
> -}
> -
> -/* Close the audio device id
> -
> -Arguments:
> - id -- the AudioID* of the device returned by spd_audio_open
> -
> -Return value:
> - 0 if everything is ok, a non-zero value in case of failure.
> -
> -Comments:
> -
> - Please make sure no other spd_audio function with this device id
> - is running in another threads. See spd_audio_stop() for detailed
> - description of possible problems.
> -*/
> -int spd_audio_close(AudioID * id)
> -{
> - int ret = 0;
> - if (id && id->function->close) {
> - ret = (id->function->close(id));
> - }
> -
> - if (NULL != lt_h) {
> - lt_dlclose(lt_h);
> - lt_h = NULL;
> - lt_dlexit();
> - }
> -
> - return ret;
> -}
> -
> -/* Set volume for playing tracks on the device id
> -
> -Arguments:
> - id -- the AudioID* of the device returned by spd_audio_open
> - volume -- a value in the range <-100:100> where -100 means the
> - least volume (probably silence), 0 the default volume
> - and +100 the highest volume possible to make on that
> - device for a single flow (i.e. not using mixer).
> -
> -Return value:
> - 0 if everything is ok, a non-zero value in case of failure.
> - See the particular backend documentation or source for the
> - meaning of these non-zero values.
> -
> -Comments:
> -
> - In case of /dev/dsp, it's not possible to set volume for
> - the particular flow. For that reason, the value 0 means
> - the volume the track was recorded on and each smaller value
> - means less volume (since this works by deviding the samples
> - in the track by a constant).
> -*/
> -int spd_audio_set_volume(AudioID * id, int volume)
> -{
> - if ((volume > 100) || (volume < -100)) {
> - fprintf(stderr, "Requested volume out of range");
> - return -1;
> - }
> - if (id == NULL) {
> - fprintf(stderr, "audio id is NULL in spd_audio_set_volume\n");
> - return -1;
> - }
> - id->volume = volume;
> - return 0;
> -}
> -
> -void spd_audio_set_loglevel(AudioID * id, int level)
> -{
> - if (level) {
> - spd_audio_log_level = level;
> - if (id != 0 && id->function != 0)
> - id->function->set_loglevel(level);
> - }
> -}
> -
> -char const *spd_audio_get_playcmd(AudioID * id)
> -{
> - if (id != 0 && id->function != 0) {
> - return id->function->get_playcmd();
> - }
> - return NULL;
> -}
> diff --git a/src/modules/spd_audio.h b/src/modules/spd_audio.h
> deleted file mode 100644
> index f9452e8..0000000
> --- a/src/modules/spd_audio.h
> +++ /dev/null
> @@ -1,46 +0,0 @@
> -
> -/*
> - * spd_audio.h -- The SPD Audio Library Header
> - *
> - * Copyright (C) 2004 Brailcom, o.p.s.
> - *
> - * This is free software; you can redistribute it and/or modify it under the
> - * terms of the GNU Lesser General Public License as published by the Free
> - * Software Foundation; either version 2.1, or (at your option) any later
> - * version.
> - *
> - * This software 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.
> - *
> - * You should have received a copy of the GNU Lesser General Public License
> - * along with this package; see the file COPYING. If not, write to the Free
> - * Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
> - * 02110-1301, USA.
> - *
> - * $Id: spd_audio.h,v 1.21 2008-10-15 17:28:17 hanke Exp $
> - */
> -
> -#ifndef __SPD_AUDIO_H
> -#define __SPD_AUDIO_H
> -
> -#include <spd_audio_plugin.h>
> -
> -#define SPD_AUDIO_LIB_PREFIX "spd_"
> -
> -AudioID *spd_audio_open(char *name, void **pars, char **error);
> -
> -int spd_audio_play(AudioID * id, AudioTrack track, AudioFormat format);
> -
> -int spd_audio_stop(AudioID * id);
> -
> -int spd_audio_close(AudioID * id);
> -
> -int spd_audio_set_volume(AudioID * id, int volume);
> -
> -void spd_audio_set_loglevel(AudioID * id, int level);
> -
> -char const *spd_audio_get_playcmd(AudioID * id);
> -
> -#endif /* ifndef #__SPD_AUDIO_H */
> diff --git a/src/server/Makefile.am b/src/server/Makefile.am
> index 607cf8e..bad5a61 100644
> --- a/src/server/Makefile.am
> +++ b/src/server/Makefile.am
> @@ -9,12 +9,14 @@ speech_dispatcher_SOURCES = speechd.c speechd.h server.c
> server.h \
> parse.c parse.h set.c set.h msg.h alloc.c alloc.h \
> compare.c compare.h speaking.c speaking.h options.c options.h \
> output.c output.h sem_functions.c sem_functions.h \
> - index_marking.c index_marking.h
> + index_marking.c index_marking.h audio.c audio.h
> speech_dispatcher_CFLAGS = $(ERROR_CFLAGS)
> speech_dispatcher_CPPFLAGS = $(inc_local) $(DOTCONF_CFLAGS) $(GLIB_CFLAGS) \
> $(GMODULE_CFLAGS) $(GTHREAD_CFLAGS) -DSYS_CONF=\"$(spdconfdir)\" \
> -DSND_DATA=\"$(snddatadir)\" -DMODULEBINDIR=\"$(modulebindir)\" \
> - -D_GNU_SOURCE -DDEFAULT_AUDIO_METHOD=\"$(default_audio_method)\"
> + -D_GNU_SOURCE -DDEFAULT_AUDIO_METHOD=\"$(default_audio_method)\" \
> + -DPLUGIN_DIR=\"$(audiodir)\"
> speech_dispatcher_LDFLAGS = $(RDYNAMIC)
> speech_dispatcher_LDADD = $(lib_common) $(DOTCONF_LIBS) $(GLIB_LIBS) \
> - $(GMODULE_LIBS) $(GTHREAD_LIBS) $(EXTRA_SOCKET_LIBS)
> + $(GMODULE_LIBS) $(GTHREAD_LIBS) $(EXTRA_SOCKET_LIBS) \
> + $(audio_dlopen_modules)
> diff --git a/src/server/audio.c b/src/server/audio.c
> new file mode 100644
> index 0000000..08a6567
> --- /dev/null
> +++ b/src/server/audio.c
> @@ -0,0 +1,647 @@
> +
> +/*
> + * spd_audio.c -- Spd Audio Output Library
> + *
> + * Copyright (C) 2004, 2006 Brailcom, o.p.s.
> + *
> + * This is free software; you can redistribute it and/or modify it under the
> + * terms of the GNU Lesser General Public License as published by the Free
> + * Software Foundation; either version 2.1, or (at your option) any later
> + * version.
> + *
> + * This software 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.
> + *
> + * You should have received a copy of the GNU Lesser General Public License
> + * along with this package; see the file COPYING. If not, write to the Free
> + * Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
> + * 02110-1301, USA.
> + *
> + * $Id: spd_audio.c,v 1.21 2008-06-09 10:29:12 hanke Exp $
> + */
> +
> +/*
> + * spd_audio is a simple realtime audio output library with the capability of
> + * playing 8 or 16 bit data, immediate stop and synchronization. This library
> + * currently provides OSS, NAS, ALSA and PulseAudio backend. The available
> backends are
> + * specified at compile-time using the directives WITH_OSS, WITH_NAS,
> WITH_ALSA,
> + * WITH_PULSE, WITH_LIBAO but the user program is allowed to switch between
> them at run-time.
> + */
> +
> +#ifdef HAVE_CONFIG_H
> +#include <config.h>
> +#endif
> +
> +#include "audio.h"
> +
> +#include <stdio.h>
> +#include <string.h>
> +#include <fcntl.h>
> +#include <sys/ioctl.h>
> +#include <sys/time.h>
> +#include <time.h>
> +#include <unistd.h>
> +#include <errno.h>
> +
> +#include <pthread.h>
> +
> +#include <glib.h>
> +#include <glib/gstdio.h>
> +#include <ltdl.h>
> +
> +#include "speechd.h"
> +#include "speechd_defines.h"
> +#include "set.h"
> +
> +static int spd_audio_log_level;
> +static lt_dlhandle lt_h;
> +
> +/* Server audio socket file descriptor */
> +int audio_server_socket;
> +
> +AudioID *audio_id;
> +static char *audio_pars[10]; /* Audio module parameters */
> +
> +static pthread_t audio_thread;
> +static sem_t audio_play_semaphore;
> +
> +static gboolean audio_close_requested = FALSE;
> +
> +/* Dynamically load a library with RTLD_GLOBAL set.
> +
> + This is needed when a dynamically-loaded library has its own plugins
> + that call into the parent library.
> + Most of the credit for this function goes to Gary Vaughan.
> +*/
> +static lt_dlhandle my_dlopenextglobal(const char *filename)
> +{
> + lt_dlhandle handle = NULL;
> + lt_dladvise advise;
> +
> + if (lt_dladvise_init(&advise))
> + return handle;
> +
> + if (!lt_dladvise_ext(&advise) && !lt_dladvise_global(&advise))
> + handle = lt_dlopenadvise(filename, advise);
> +
> + lt_dladvise_destroy(&advise);
> + return handle;
> +}
> +
> +/* Open the audio device.
> +
> + Arguments:
> + type -- The requested device. Currently AudioOSS or AudioNAS.
> + pars -- and array of pointers to parameters to pass to
> + the device backend, terminated by a NULL pointer.
> + See the source/documentation of each specific backend.
> + error -- a pointer to the string where error description is
> + stored in case of failure (returned AudioID == NULL).
> + Otherwise will contain NULL.
> +
> + Return value:
> + Newly allocated AudioID structure that can be passed to
> + all other spd_audio functions, or NULL in case of failure.
> +
> +*/
> +AudioID *spd_audio_open(char *name, void **pars, char **error)
> +{
> + MSG(5, "spd_audio_open called with name %s", name);
> + AudioID *id;
> + spd_audio_plugin_t const *p;
> + spd_audio_plugin_t *(*fn) (void);
> + gchar *libname;
> + int ret;
> +
> + /* now check whether dynamic plugin is available */
> + ret = lt_dlinit();
> + if (ret != 0) {
> + *error = (char *)g_strdup_printf("lt_dlinit() failed");
> + return (AudioID *) NULL;
> + }
> +
> + ret = lt_dlsetsearchpath(PLUGIN_DIR);
> + if (ret != 0) {
> + *error = (char *)g_strdup_printf("lt_dlsetsearchpath() failed");
> + return (AudioID *) NULL;
> + }
> +
> + libname = g_strdup_printf(SPD_AUDIO_LIB_PREFIX "%s", name);
> + lt_h = my_dlopenextglobal(libname);
> + g_free(libname);
> + if (NULL == lt_h) {
> + *error =
> + (char *)g_strdup_printf("Cannot open plugin %s. error: %s",
> + name, lt_dlerror());
> + return (AudioID *) NULL;
> + }
> +
> + fn = lt_dlsym(lt_h, SPD_AUDIO_PLUGIN_ENTRY_STR);
> + if (NULL == fn) {
> + *error = (char *)g_strdup_printf("Cannot find symbol %s",
> + SPD_AUDIO_PLUGIN_ENTRY_STR);
> + return (AudioID *) NULL;
> + }
> +
> + MSG(5, "calling init function");
> + p = fn();
> + if (p == NULL || p->name == NULL) {
> + *error = (char *)g_strdup_printf("plugin %s not found", name);
> + return (AudioID *) NULL;
> + }
> +
> + MSG(5, "calling open function");
> + id = p->open(pars);
> + if (id == NULL) {
> + *error =
> + (char *)g_strdup_printf("Couldn't open %s plugin", name);
> + return (AudioID *) NULL;
> + }
> +
> + id->function = p;
> +#if defined(BYTE_ORDER) && (BYTE_ORDER == BIG_ENDIAN)
> + id->format = SPD_AUDIO_BE;
> +#else
> + id->format = SPD_AUDIO_LE;
> +#endif
> +
> + *error = NULL;
> +
> + return id;
> +}
> +
> +/* Play a track on the audio device (blocking).
> +
> + Arguments:
> + id -- the AudioID* of the device returned by spd_audio_open
> + track -- a track to play (see spd_audio.h)
> +
> + Return value:
> + 0 if everything is ok, a non-zero value in case of failure.
> + See the particular backend documentation or source for the
> + meaning of these non-zero values.
> +
> + Comment:
> + spd_audio_play() is a blocking function. It returns exactly
> + when the given track stopped playing. However, it's possible
> + to safely interrupt it using spd_audio_stop() described below.
> + (spd_audio_stop() needs to be called from another thread, obviously.)
> +
> +*/
> +int spd_audio_play(AudioID * id, AudioTrack track, AudioFormat format)
> +{
> + int ret;
> +
> + if (id && id->function->play) {
> + /* Only perform byte swapping if the driver in use has given us
> audio in
> + an endian format other than what the running CPU supports. */
> + if (format != id->format) {
> + unsigned char *out_ptr, *out_end, c;
> + out_ptr = (unsigned char *)track.samples;
> + out_end =
> + out_ptr +
> + track.num_samples * 2 * track.num_channels;
> + while (out_ptr < out_end) {
> + c = out_ptr[0];
> + out_ptr[0] = out_ptr[1];
> + out_ptr[1] = c;
> + out_ptr += 2;
> + }
> + }
> + MSG(5, "playing audio on audio_id %d", id);
> + ret = id->function->play(id, track);
> + } else {
> + fprintf(stderr, "Play not supported on this device\n");
> + return -1;
> + }
> +
> + return ret;
> +}
> +
> +/* Stop playing the current track on device id
> +
> +Arguments:
> + id -- the AudioID* of the device returned by spd_audio_open
> +
> +Return value:
> + 0 if everything is ok, a non-zero value in case of failure.
> + See the particular backend documentation or source for the
> + meaning of these non-zero values.
> +
> +Comment:
> + spd_audio_stop() safely interrupts spd_audio_play() when called
> + from another thread. It shouldn't cause any clicks or unwanted
> + effects in the sound output.
> +
> + It's safe to call spd_audio_stop() even if the device isn't playing
> + any track. In that case, it does nothing. However, there is a danger
> + when using spd_audio_stop(). Since you must obviously do it from
> + another thread than where spd_audio_play is running, you must make
> + yourself sure that the device is still open and the id you pass it
> + is valid and will be valid until spd_audio_stop returns. In other words,
> + you should use some mutex or other synchronization device to be sure
> + spd_audio_close isn't called before or during spd_audio_stop execution.
> +*/
> +
> +int spd_audio_stop(AudioID * id)
> +{
> + int ret;
> + if (id && id->function->stop) {
> + ret = id->function->stop(id);
> + } else {
> + fprintf(stderr, "Stop not supported on this device\n");
> + return -1;
> + }
> + return ret;
> +}
> +
> +/* Close the audio device id
> +
> +Arguments:
> + id -- the AudioID* of the device returned by spd_audio_open
> +
> +Return value:
> + 0 if everything is ok, a non-zero value in case of failure.
> +
> +Comments:
> +
> + Please make sure no other spd_audio function with this device id
> + is running in another threads. See spd_audio_stop() for detailed
> + description of possible problems.
> +*/
> +int spd_audio_close(AudioID * id)
> +{
> + int ret = 0;
> + if (id && id->function->close) {
> + ret = (id->function->close(id));
> + }
> +
> + if (NULL != lt_h) {
> + lt_dlclose(lt_h);
> + lt_h = NULL;
> + lt_dlexit();
> + }
> +
> + return ret;
> +}
> +
> +/* Set volume for playing tracks on the device id
> +
> +Arguments:
> + id -- the AudioID* of the device returned by spd_audio_open
> + volume -- a value in the range <-100:100> where -100 means the
> + least volume (probably silence), 0 the default volume
> + and +100 the highest volume possible to make on that
> + device for a single flow (i.e. not using mixer).
> +
> +Return value:
> + 0 if everything is ok, a non-zero value in case of failure.
> + See the particular backend documentation or source for the
> + meaning of these non-zero values.
> +
> +Comments:
> +
> + In case of /dev/dsp, it's not possible to set volume for
> + the particular flow. For that reason, the value 0 means
> + the volume the track was recorded on and each smaller value
> + means less volume (since this works by deviding the samples
> + in the track by a constant).
> +*/
> +int spd_audio_set_volume(AudioID * id, int volume)
> +{
> + if ((volume > 100) || (volume < -100)) {
> + fprintf(stderr, "Requested volume out of range");
> + return -1;
> + }
> + if (id == NULL) {
> + fprintf(stderr, "audio id is NULL in spd_audio_set_volume\n");
> + return -1;
> + }
> + id->volume = volume;
> + return 0;
> +}
> +
> +void spd_audio_set_loglevel(AudioID * id, int level)
> +{
> + if (level) {
> + spd_audio_log_level = level;
> + if (id != 0 && id->function != 0)
> + id->function->set_loglevel(level);
> + }
> +}
> +
> +char const *spd_audio_get_playcmd(AudioID * id)
> +{
> + if (id != 0 && id->function != 0) {
> + return id->function->get_playcmd();
> + }
> + return NULL;
> +}
> +
> +void speechd_audio_socket_init(void)
> +{
> + /* For now use unix socket for audio. Maybe later we can add inet
> socket support */
> + GString *audio_socket_filename;
> + audio_socket_filename = g_string_new("");
> + if (SpeechdOptions.runtime_speechd_dir) {
> + g_string_printf(audio_socket_filename, "%s/audio.sock",
> + SpeechdOptions.runtime_speechd_dir);
> + } else {
> + FATAL
> + ("Socket name file not set and user has no runtime
> directory");
> + }
> + g_free(SpeechdOptions.audio_socket_path);
> + SpeechdOptions.audio_socket_path =
> g_strdup(audio_socket_filename->str);
> + g_string_free(audio_socket_filename, 1);
> +
> + MSG(1, "Creating audio socket at %s", SpeechdOptions.audio_socket_path);
> +
> + /* Audio data is only using unix sockets for now, possibly adapt to use
> + * inet sockets also later? */
> + if (g_file_test(SpeechdOptions.audio_socket_path, G_FILE_TEST_EXISTS))
> + if (g_unlink(SpeechdOptions.audio_socket_path) == -1)
> + FATAL
> + ("Local socket file for audio exists but impossible
> to delete. Wrong permissions?");
> + /* Connect and start listening on local unix socket */
> + audio_server_socket =
> make_local_socket(SpeechdOptions.audio_socket_path);
> +}
> +
> +/* play the audio data on _fd_ if we got some activity. */
> +int play_audio(int fd)
> +{
> + size_t bytes = 0; /* Number of bytes we got */
> + int buflen = BUF_SIZE;
> + char *buf = (char *)g_malloc(buflen + 1);
> + AudioTrack track;
> + AudioFormat format;
> + gchar ** metadata;
> + int bytes_read;
> +
> + /* Read data from socket */
> + /* Read exactly one complete line, the `parse' routine relies on it */
> + {
> + while (1) {
> + int n = read(fd, buf + bytes, 1);
> + if (n <= 0) {
> + MSG(5, "ERROR: Read 0 bytes from fd");
> + g_free(buf);
> + return -1;
> + }
> + /* Note, bytes is a 0-based index into buf. */
> + if ((buf[bytes] == '\n')
> + && (bytes >= 1) && (buf[bytes - 1] == '\r')) {
> + buf[++bytes] = '\0';
> + break;
> + }
> + if (buf[bytes] == '\0')
> + buf[bytes] = '?';
> + if ((++bytes) == buflen) {
> + buflen *= 2;
> + buf = g_realloc(buf, buflen + 1);
> + }
> + }
> + }
> +
> + /* Parse the data and read the reply */
> + MSG2(5, "protocol", "%d:DATA:|%s| (%d)", fd, buf, bytes);
> + if (strcmp(buf, "ACK"NEWLINE) == 0) {
> + g_free(buf);
> + return 0;
> + }
> + /* parse the AudioTrack information from buf */
> + metadata = g_strsplit(buf, ":", 5);
> + if (metadata == NULL || metadata[0] == NULL
> + || metadata[1] == NULL || metadata[2] == NULL
> + || metadata[3] == NULL || metadata[4] == NULL) {
> + MSG(5, "Error: Unable to read Audiotrack metadata!");
> + return -1;
> + }
> + format = strtol(metadata[0], NULL, 10);
> + track.bits = strtol(metadata[1], NULL, 10);
> + track.num_channels = strtol(metadata[2], NULL, 10);
> + track.sample_rate = strtol(metadata[3], NULL, 10);
> + track.num_samples = strtol(metadata[4], NULL, 10);
> +
> + MSG(5, "Track num samples is %d", track.num_samples);
> +
> + if (track.num_samples <= 0) {
> + MSG(5, "Error: num_samples is invalid");
> + return -1;
> + }
> + /* then free buf */
> + g_free(buf);
> + /* Get the rest of the data */
> + track.samples = g_malloc0_n(track.num_samples, sizeof(signed short));
> + bytes_read = read(fd, track.samples, track.num_samples * sizeof(signed
> short));
> +
> + if (bytes_read != track.num_samples * sizeof(signed short)) {
> + MSG(5, "Error: num_samples %d doesn't match bytes read %d",
> track.num_samples, bytes_read);
> + return -1;
> + }
> +
> + MSG(5, "Going to play audio on audio with id %d", audio_id);
> +
> + /* And play the AudioTrack */
> + if (spd_audio_play(audio_id, track, format) < 0) {
> + MSG(5, "Error: unable to play audio");
> + return -1;
> + }
> +
> + return 0;
> +}
> +
> +static gboolean audio_socket_process_incoming (gint fd,
> + GIOCondition condition,
> + gpointer data)
> +{
> + int ret;
> + ret = speechd_audio_connection_new(fd);
> + if (ret != 0) {
> + MSG(2, "Error: Failed to add new module audio!");
> + if (SPEECHD_DEBUG) {
> + FATAL("Failed to add new module audio!");
> + }
> + }
> +
> + return TRUE;
> +}
> +
> +static gboolean audio_process_incoming (gint fd,
> + GIOCondition condition,
> + gpointer data)
> +{
> + MSG(5, "audio_process_incoming called for fd %d", fd);
> + int nread;
> +
> + ioctl(fd, FIONREAD, &nread);
> +
> + if (nread == 0) {
> + /* module has gone */
> + MSG(2, "Info: Module has gone.");
> + return FALSE;
> + }
> +
> + MSG(5, "read %d bytes from fd %d", nread, fd);
> +
> + /* client sends some commands or data */
> + if (play_audio(fd) == -1) {
> + MSG(2, "Error: Failed to serve client on fd %d!", fd);
> + }
> +
> + return TRUE;
> +}
> +
> +/* Playback thread. */
> +static void *_speechd_play(void *nothing)
> +{
> + char *error = 0;
> + gchar **outputs;
> + int i = 0;
> + gboolean found_audio_module = FALSE;
> +
> + MSG(1, "Playback thread starting.......");
> +
> + /* TODO: Use real values from config rather than these hard coded test
> values */
> + if (GlobalFDSet.audio_oss_device != NULL)
> + audio_pars[1] = g_strdup(GlobalFDSet.audio_oss_device);
> + else
> + audio_pars[1] = NULL;
> +
> + if (GlobalFDSet.audio_alsa_device != NULL)
> + audio_pars[2] = g_strdup(GlobalFDSet.audio_alsa_device);
> + else
> + audio_pars[2] = NULL;
> +
> +
> + if (GlobalFDSet.audio_nas_server != NULL)
> + audio_pars[3] = g_strdup(GlobalFDSet.audio_nas_server);
> + else
> + audio_pars[3] = NULL;
> +
> + if (GlobalFDSet.audio_pulse_server != NULL)
> + audio_pars[4] = g_strdup(GlobalFDSet.audio_pulse_server);
> + else
> + audio_pars[4] = NULL;
> +
> + if (GlobalFDSet.audio_pulse_min_length != NULL)
> + audio_pars[5] = g_strdup_printf("%d",
> GlobalFDSet.audio_pulse_min_length);
> + else
> + audio_pars[5] = NULL;
> +
> + MSG(1, "Openning audio output system");
> + if (GlobalFDSet.audio_output_method == NULL) {
> + MSG(1, "Sound output method specified in configuration not
> supported. "
> + "Please choose 'oss', 'alsa', 'nas', 'libao' or 'pulse'.");
> + return 0;
> + }
> +
> + outputs = g_strsplit(GlobalFDSet.audio_output_method, ",", 0);
> + while (NULL != outputs[i]) {
> + audio_id =
> + spd_audio_open(outputs[i], (void **)&audio_pars[1],
> + &error);
> + if (audio_id) {
> + spd_audio_set_loglevel(audio_id,
> SpeechdOptions.log_level);
> + MSG(5, "Using %s audio output method with log level
> %d", outputs[i], SpeechdOptions.log_level);
> +
> + /* Volume is controlled by the synthesizer. Always play
> at normal on audio device. */
> + if (spd_audio_set_volume(audio_id, 85) < 0) {
> + MSG(2, "Can't set volume. audio not
> initialized?");
> + }
> +
> + g_strfreev(outputs);
> + MSG(5, "audio initialized successfully.");
> + found_audio_module = TRUE;
> + break;
> + }
> + i++;
> + }
> +
> + if (!found_audio_module) {
> + MSG(1, "Opening sound device failed. Reason: %s. ", error);
> + g_free(error); /* g_malloc'ed, in spd_audio_open. */
> + }
> +
> + /* Connect to the server socket */
> + g_unix_fd_add(audio_server_socket, G_IO_IN,
> + audio_socket_process_incoming, NULL);
> +
> + /* Block all signals to this thread. */
> +// set_speaking_thread_parameters();
> +
> + while (!audio_close_requested) {
> + /* If semaphore not set, set suspended lock and suspend until
> it is signaled. */
> + if (0 != sem_trywait(&audio_play_semaphore)) {
> + sem_wait(&audio_play_semaphore);
> + }
> + MSG(5, "Playback semaphore on.");
> + if (audio_close_requested)
> + break;
> + }
> +
> + MSG(1, "Playback thread ended.......");
> + return 0;
> +}
> +
> +void speechd_audio_init()
> +{
> + int ret = 0;
> +
> + audio_id = 0;
> + sem_init(&audio_play_semaphore, 0, 0);
> +
> + ret = pthread_create(&audio_thread, NULL, _speechd_play, NULL);
> + if (ret != 0)
> + FATAL("Audio thread failed!\n");
> +}
> +
> +/* activity is on audio_server_socket (request for a new connection) */
> +int speechd_audio_connection_new(int audio_server_socket)
> +{
> + MSG(5, "Adding audio connection on socket %d", audio_server_socket);
> + TAudioFDSetElement *new_fd_set;
> + struct sockaddr_in module_address;
> + unsigned int module_len = sizeof(module_address);
> + int module_socket;
> +
> + module_socket =
> + accept(audio_server_socket, (struct sockaddr *)&module_address,
> + &module_len);
> +
> + if (module_socket == -1) {
> + MSG(2,
> + "Error: Can't handle connection request of a module for
> audio");
> + return -1;
> + }
> +
> + /* We add the associated client_socket to the descriptor set. */
> + if (module_socket > SpeechdStatus.max_fd)
> + SpeechdStatus.max_fd = module_socket;
> + MSG(4, "Adding module on fd %d", module_socket);
> +
> + /* Create a record in fd_settings */
> + new_fd_set = (TAudioFDSetElement *) default_audio_fd_set();
> + if (new_fd_set == NULL) {
> + MSG(2,
> + "Error: Failed to create a record in fd_settings for the
> module for audio");
> + if (SpeechdStatus.max_fd == module_socket)
> + SpeechdStatus.max_fd--;
> + return -1;
> + }
> + new_fd_set->fd = module_socket;
> + new_fd_set->fd_source = g_unix_fd_add(module_socket, G_IO_IN,
> audio_process_incoming, NULL);
> +
> + return 0;
> +}
> +
> +void speechd_audio_cleanup(void)
> +{
> + if (close(audio_server_socket) == -1)
> + MSG(2, "close() audio server socket failed: %s",
> strerror(errno));
> +
> + MSG(2, "Closing audio output...");
> + spd_audio_close(audio_id);
> + audio_id = NULL;
> +}
> diff --git a/src/server/audio.h b/src/server/audio.h
> new file mode 100644
> index 0000000..d371c50
> --- /dev/null
> +++ b/src/server/audio.h
> @@ -0,0 +1,60 @@
> +
> +/*
> + * spd_audio.h -- The SPD Audio Library Header
> + *
> + * Copyright (C) 2004 Brailcom, o.p.s.
> + *
> + * This is free software; you can redistribute it and/or modify it under the
> + * terms of the GNU Lesser General Public License as published by the Free
> + * Software Foundation; either version 2.1, or (at your option) any later
> + * version.
> + *
> + * This software 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.
> + *
> + * You should have received a copy of the GNU Lesser General Public License
> + * along with this package; see the file COPYING. If not, write to the Free
> + * Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
> + * 02110-1301, USA.
> + *
> + * $Id: spd_audio.h,v 1.21 2008-10-15 17:28:17 hanke Exp $
> + */
> +
> +#ifndef __SPD_AUDIO_H
> +#define __SPD_AUDIO_H
> +
> +#include <spd_audio_plugin.h>
> +
> +#define SPD_AUDIO_LIB_PREFIX "spd_"
> +
> +AudioID *spd_audio_open(char *name, void **pars, char **error);
> +
> +int spd_audio_play(AudioID * id, AudioTrack track, AudioFormat format);
> +
> +int spd_audio_stop(AudioID * id);
> +
> +int spd_audio_close(AudioID * id);
> +
> +int spd_audio_set_volume(AudioID * id, int volume);
> +
> +void spd_audio_set_loglevel(AudioID * id, int level);
> +
> +char const *spd_audio_get_playcmd(AudioID * id);
> +
> +/* Speech dispatcher server functions */
> +
> +/* speechd_play_audio() reads audio data and plays it */
> +int speechd_play_audio(int fd);
> +
> +int speechd_audio_connection_new(int audio_socket);
> +/* Initialize the audio socket */
> +void speechd_audio_socket_init(void);
> +/* Initialize the audio backend based on user's settings in a new thread */
> +void speechd_audio_init(void);
> +
> +/* Clean up audio socket and module */
> +void speechd_audio_cleanup(void);
> +
> +#endif /* ifndef #__SPD_AUDIO_H */
> diff --git a/src/server/module.c b/src/server/module.c
> index 3932e6d..b27727d 100644
> --- a/src/server/module.c
> +++ b/src/server/module.c
> @@ -313,18 +313,6 @@ OutputModule *load_output_module(char *mod_name, char
> *mod_prog,
> output_module_debug(module);
> }
>
> - /* Initialize audio settings */
> - ret = output_send_audio_settings(module);
> - if (ret != 0) {
> - MSG(1,
> - "ERROR: Can't initialize audio in output module, see reason
> above.");
> - module->working = 0;
> - kill(module->pid, 9);
> - waitpid(module->pid, NULL, WNOHANG);
> - destroy_module(module);
> - return NULL;
> - }
> -
> /* Send log level configuration setting */
> ret = output_send_loglevel_setting(module);
> if (ret != 0) {
> diff --git a/src/server/msg.h b/src/server/msg.h
> index 67c07a9..a8f5048 100644
> --- a/src/server/msg.h
> +++ b/src/server/msg.h
> @@ -24,7 +24,8 @@
> #ifndef MSG_H
> #define MSG_H
>
> -#define NEWLINE "\r\n"
> +#include <speechd_defines.h>
> +
> #define OK_LANGUAGE_SET "201 OK
> LANGUAGE SET" NEWLINE
> #define OK_PRIORITY_SET "202 OK
> PRIORITY SET" NEWLINE
> #define OK_RATE_SET "203 OK RATE
> SET" NEWLINE
> diff --git a/src/server/output.c b/src/server/output.c
> index 0d3e40b..cfa02bc 100644
> --- a/src/server/output.c
> +++ b/src/server/output.c
> @@ -482,29 +482,6 @@ int output_send_settings(TSpeechDMessage * msg,
> OutputModule * output)
> g_string_append_printf(set_str, #name"=NULL\n"); \
> }
>
> -int output_send_audio_settings(OutputModule * output)
> -{
> - GString *set_str;
> - int err;
> -
> - MSG(4, "Module set parameters.");
> - set_str = g_string_new("");
> - ADD_SET_STR(audio_output_method);
> - ADD_SET_STR(audio_oss_device);
> - ADD_SET_STR(audio_alsa_device);
> - ADD_SET_STR(audio_nas_server);
> - ADD_SET_STR(audio_pulse_server);
> - ADD_SET_INT(audio_pulse_min_length);
> -
> - SEND_CMD_N("AUDIO");
> - SEND_DATA_N(set_str->str);
> - SEND_CMD_N(".");
> -
> - g_string_free(set_str, 1);
> -
> - return 0;
> -}
> -
> int output_send_loglevel_setting(OutputModule * output)
> {
> GString *set_str;
> diff --git a/src/server/output.h b/src/server/output.h
> index 10bbe80..d48602b 100644
> --- a/src/server/output.h
> +++ b/src/server/output.h
> @@ -41,7 +41,6 @@ GString *output_read_reply(OutputModule * output);
> char *output_read_reply2(OutputModule * output);
> int output_send_data(char *cmd, OutputModule * output, int wfr);
> int output_send_settings(TSpeechDMessage * msg, OutputModule * output);
> -int output_send_audio_settings(OutputModule * output);
> int output_send_loglevel_setting(OutputModule * output);
> int output_module_is_speaking(OutputModule * output, char **index_mark);
> int waitpid_with_timeout(pid_t pid, int *status_ptr, int options,
> diff --git a/src/server/server.c b/src/server/server.c
> index 6bdc78e..cf72880 100644
> --- a/src/server/server.c
> +++ b/src/server/server.c
> @@ -31,6 +31,7 @@
> #include "speaking.h"
> #include "sem_functions.h"
> #include "history.h"
> +#include "speechd_defines.h"
>
> int last_message_id = 0;
>
> diff --git a/src/server/set.c b/src/server/set.c
> index d0c2305..23e2a9c 100644
> --- a/src/server/set.c
> +++ b/src/server/set.c
> @@ -537,6 +537,20 @@ TFDSetElement *default_fd_set(void)
> return (new);
> }
>
> +TAudioFDSetElement *default_audio_fd_set(void)
> +{
> + TAudioFDSetElement *new;
> +
> + new = (TAudioFDSetElement *) g_malloc(sizeof(TAudioFDSetElement));
> +
> + /* Fill with the global settings values */
> + /* We can't use global_fdset copy as this
> + returns static structure and we need dynamic */
> + new->output_module = g_strdup(GlobalFDSet.output_module);
> +
> + return (new);
> +}
> +
> int get_client_uid_by_fd(int fd)
> {
> int *uid;
> diff --git a/src/server/set.h b/src/server/set.h
> index 08866ab..2b7f488 100644
> --- a/src/server/set.h
> +++ b/src/server/set.h
> @@ -94,6 +94,7 @@ int set_debug_all(int debug);
> int set_debug_destination_all(char *debug_destination);
>
> TFDSetElement *default_fd_set(void);
> +TAudioFDSetElement *default_audio_fd_set(void);
>
> char *set_param_str(char *parameter, char *value);
>
> diff --git a/src/server/speechd.c b/src/server/speechd.c
> index d790c3f..21ca5c1 100644
> --- a/src/server/speechd.c
> +++ b/src/server/speechd.c
> @@ -31,8 +31,10 @@
> #include <sys/stat.h>
> #include <sys/socket.h>
> #include <sys/un.h>
> +#include <ltdl.h>
>
> #include "speechd.h"
> +#include "audio.h"
>
> /* Declare dotconf functions and data structures*/
> #include "configuration.h"
> @@ -76,6 +78,10 @@ static gboolean client_process_incoming (gint fd,
> GIOCondition condition,
> gpointer data);
>
> +static gboolean audio_process_incoming (gint fd,
> + GIOCondition condition,
> + gpointer data);
> +
> void check_client_count(void);
>
> #ifdef __SUNPRO_C
> @@ -867,6 +873,7 @@ int make_local_socket(const char *filename)
> FATAL("listen() failed for local socket");
> }
>
> + MSG(5, "Successfully opened local socket at %s", filename);
> return sock;
> }
>
> @@ -911,6 +918,7 @@ int make_inet_socket(const int port)
> ("listen() failed for inet socket, another Speech
> Dispatcher running?");
> }
>
> + MSG(5, "Successfully opened inet socket on port %d", port);
> return server_socket;
> }
>
> @@ -981,6 +989,7 @@ int main(int argc, char *argv[])
> char *spawn_communication_method = NULL;
> int spawn_port = 0;
> char *spawn_socket_path = NULL;
> + char *status_info;
>
> /* Strip all permisions for 'others' from the files created */
> umask(007);
> @@ -991,6 +1000,9 @@ int main(int argc, char *argv[])
> custom_logfile = NULL;
> custom_log_kind = NULL;
>
> + /* Initialize ltdl's list of preloaded audio backends. */
> + LTDL_SET_PRELOADED_SYMBOLS();
> +
> /* initialize i18n support */
> i18n_init();
>
> @@ -1116,8 +1128,11 @@ int main(int argc, char *argv[])
> exit(1);
> }
>
> + /* We need this first since modules will connect to it */
> + speechd_audio_socket_init();
> +
> speechd_init();
> -
> +
> /* Handle socket_path 'default' */
> // TODO: This is a hack, we should do that at appropriate places...
> if (!strcmp(SpeechdOptions.socket_path, "default")) {
> @@ -1237,6 +1252,8 @@ int main(int argc, char *argv[])
> g_unix_signal_add(SIGUSR1, speechd_reload_dead_modules, NULL);
> (void)signal(SIGPIPE, SIG_IGN);
>
> + speechd_audio_init();
> +
> MSG(4, "Creating new thread for speak()");
> ret = pthread_create(&speak_thread, NULL, speak, NULL);
> if (ret != 0)
> @@ -1280,6 +1297,8 @@ int main(int argc, char *argv[])
> if (close(server_socket) == -1)
> MSG(2, "close() failed: %s", strerror(errno));
>
> + speechd_audio_cleanup();
> +
> MSG(4, "Removing pid file");
> destroy_pid_file();
>
> diff --git a/src/server/speechd.h b/src/server/speechd.h
> index e5e620b..0e80d6f 100644
> --- a/src/server/speechd.h
> +++ b/src/server/speechd.h
> @@ -111,6 +111,12 @@ typedef struct {
> } TFDSetElement;
>
> typedef struct {
> + int fd; /* File descriptor the module is on. */
> + guint fd_source; /* Used to store the GSource ID for watching fd
> activity in the main loop */
> + char *output_module; /* Output module name. (e.g. "festival",
> "flite", "apollo", ...) */
> +} TAudioFDSetElement;
> +
> +typedef struct {
> char *pattern;
> TFDSetElement val;
> } TFDSetClientSpecific;
> @@ -153,6 +159,7 @@ struct {
> char *communication_method;
> int communication_method_set;
> char *socket_path;
> + char *audio_socket_path;
> int socket_path_set;
> int port, port_set;
> int localhost_access_only, localhost_access_only_set;
> @@ -256,6 +263,8 @@ int isanum(const char *str);
> absolute (starting with slash) or relative. */
> char *spd_get_path(char *filename, char *startdir);
>
> +int make_local_socket(const char *filename);
> +
> /* Functions used in speechd.c only */
> int speechd_connection_new(int server_socket);
> int speechd_connection_destroy(int fd);
> --
> 2.5.0
>
> _______________________________________________
> Speechd mailing list
> Speechd at lists.freebsoft.org
> http://lists.freebsoft.org/mailman/listinfo/speechd