Merge commit 'release-0.16.2'

Conflicts:
	Makefile.am
	NEWS
	configure.ac

Makefile.am

@@ -886,6 +886,7 @@ test_run_input_LDADD = $(MPD_LIBS) \
 	$(INPUT_LIBS) \
 	$(GLIB_LIBS)
 test_run_input_SOURCES = test/run_input.c \
+	test/stdbin.h \
 	src/conf.c src/tokenizer.c src/utils.c src/string_util.c\
 	src/tag.c src/tag_pool.c src/tag_save.c \
 	src/fd_util.c \
@@ -933,6 +934,7 @@ test_run_decoder_LDADD = $(MPD_LIBS) \
 	$(INPUT_LIBS) $(DECODER_LIBS) \
 	$(GLIB_LIBS)
 test_run_decoder_SOURCES = test/run_decoder.c \
+	test/stdbin.h \
 	src/conf.c src/tokenizer.c src/utils.c src/string_util.c src/log.c \
 	src/tag.c src/tag_pool.c \
 	src/replay_gain_info.c \
@@ -973,6 +975,7 @@ test_run_filter_LDADD = $(MPD_LIBS) \
 	$(SAMPLERATE_LIBS) \
 	$(GLIB_LIBS)
 test_run_filter_SOURCES = test/run_filter.c \
+	test/stdbin.h \
 	src/filter_plugin.c \
 	src/filter_registry.c \
 	src/conf.c src/tokenizer.c src/utils.c src/string_util.c \
@@ -995,6 +998,7 @@ endif
 if ENABLE_ENCODER
 noinst_PROGRAMS += test/run_encoder
 test_run_encoder_SOURCES = test/run_encoder.c \
+	test/stdbin.h \
 	src/conf.c src/tokenizer.c \
 	src/utils.c src/string_util.c \
 	src/tag.c src/tag_pool.c \
@@ -1002,12 +1006,15 @@ test_run_encoder_SOURCES = test/run_encoder.c \
 	src/audio_format.c \
 	src/audio_parser.c \
 	$(ENCODER_SRC)
+test_run_encoder_CPPFLAGS = $(AM_CPPFLAGS) \
+	$(ENCODER_CFLAGS)
 test_run_encoder_LDADD = $(MPD_LIBS) \
 	$(ENCODER_LIBS) \
 	$(GLIB_LIBS)
 endif
 
 test_software_volume_SOURCES = test/software_volume.c \
+	test/stdbin.h \
 	src/audio_check.c \
 	src/audio_parser.c \
 	src/pcm_volume.c
@@ -1015,6 +1022,7 @@ test_software_volume_LDADD = \
 	$(GLIB_LIBS)
 
 test_run_normalize_SOURCES = test/run_normalize.c \
+	test/stdbin.h \
 	src/audio_check.c \
 	src/audio_parser.c \
 	src/AudioCompress/compress.c
@@ -1052,6 +1060,7 @@ test_run_output_LDADD = $(MPD_LIBS) \
 	$(OUTPUT_LIBS) \
 	$(GLIB_LIBS)
 test_run_output_SOURCES = test/run_output.c \
+	test/stdbin.h \
 	src/conf.c src/tokenizer.c src/utils.c src/string_util.c src/log.c \
 	src/audio_check.c \
 	src/audio_format.c \

NEWS

@@ -15,6 +15,21 @@ ver 0.17 (2011/??/??)
 * state_file: add option "restore_paused"
 
 
+ver 0.16.2 (2011/03/18)
+* configure.ac:
+  - fix bashism in tremor test
+* decoder:
+  - tremor: fix configure test
+  - gme: detect end of song
+* encoder:
+  - vorbis: reset the Ogg stream after flush
+* output:
+  - httpd: fix uninitialized variable
+  - httpd: include sys/socket.h
+  - oss: AFMT_S24_PACKED is little-endian
+  - oss: disable 24 bit playback on FreeBSD
+
+
 ver 0.16.1 (2011/01/09)
 * audio_check: fix parameter in prototype
 * add void casts to suppress "result unused" warnings (clang)
@@ -145,9 +160,13 @@ ver 0.16 (2010/12/11)
 * make single mode 'sticky'
 
 
-ver 0.15.16 (2010/??/??)
+ver 0.15.16 (2011/03/13)
+* output:
+  - ao: initialize the ao_sample_format struct
+  - jack: fix crash with mono playback
 * encoders:
   - lame: explicitly configure the output sample rate
+* update: log all file permission problems
 
 
 ver 0.15.15 (2010/11/08)

configure.ac

@@ -660,7 +660,7 @@ fi
 AM_CONDITIONAL(ENABLE_CDIO_PARANOIA, test x$enable_cdio_paranoia = xyes)
 
 dnl ---------------------------------- libogg ---------------------------------
-if test x$with_tremor == xno || test -z $with_tremor; then
+if test x$with_tremor = xno || test -z $with_tremor; then
 	PKG_CHECK_MODULES(OGG, [ogg], enable_ogg=yes, enable_ogg=no)
 fi
 
@@ -959,13 +959,19 @@ if test x$enable_tremor = xyes; then
 	ac_save_LIBS="$LIBS"
 	CFLAGS="$CFLAGS $TREMOR_CFLAGS"
 	LIBS="$LIBS $TREMOR_LIBS"
-	AC_CHECK_LIB(vorbisidec,ov_read,enable_vorbis=yes,enable_vorbis=no;
+	AC_CHECK_LIB(vorbisidec,ov_read,,enable_tremor=no;
 		AC_MSG_WARN([vorbisidec lib needed for ogg support with tremor -- disabling ogg support]))
 	CFLAGS="$ac_save_CFLAGS"
 	LIBS="$ac_save_LIBS"
+fi
 
+if test x$enable_tremor = xyes; then
 	AC_DEFINE(HAVE_TREMOR,1,
 		[Define to use tremor (libvorbisidec) for ogg support])
+	AC_DEFINE(ENABLE_VORBIS_DECODER, 1, [Define for Ogg Vorbis support]),
+else
+	TREMOR_CFLAGS=
+	TREMOR_LIBS=
 fi
 
 AC_SUBST(TREMOR_CFLAGS)
@@ -1005,7 +1011,7 @@ if test x$enable_vorbis = xyes; then
 	fi
 fi
 
-AM_CONDITIONAL(ENABLE_VORBIS_DECODER, test x$enable_vorbis = xyes)
+AM_CONDITIONAL(ENABLE_VORBIS_DECODER, test x$enable_vorbis = xyes || test x$enable_tremor = xyes)
 
 dnl --------------------------------- sidplay ---------------------------------
 found_sidplay=$HAVE_CXX

src/audio_format.h

@@ -22,6 +22,7 @@
 
 #include <stdint.h>
 #include <stdbool.h>
+#include <assert.h>
 
 enum sample_format {
 	SAMPLE_FORMAT_UNDEFINED = 0,
@@ -219,6 +220,9 @@ static inline void
 audio_format_mask_apply(struct audio_format *af,
 			const struct audio_format *mask)
 {
+	assert(audio_format_valid(af));
+	assert(audio_format_mask_valid(mask));
+
 	if (mask->sample_rate != 0)
 		af->sample_rate = mask->sample_rate;
 
@@ -227,6 +231,8 @@ audio_format_mask_apply(struct audio_format *af,
 
 	if (mask->channels != 0)
 		af->channels = mask->channels;
+
+	assert(audio_format_valid(af));
 }
 
 /**

src/audio_parser.c

@@ -192,6 +192,7 @@ audio_format_parse(struct audio_format *dest, const char *src,
 	}
 
 	audio_format_init(dest, rate, sample_format, channels);
+	assert(audio_format_valid(dest));
 
 	return true;
 }

src/command.c

@@ -763,7 +763,7 @@ handle_load(struct client *client, G_GNUC_UNUSED int argc, char *argv[])
 	result = playlist_open_into_queue(argv[1], &g_playlist,
 					  client->player_control, true);
 	if (result != PLAYLIST_RESULT_NO_SUCH_LIST)
-		return result;
+		return print_playlist_result(client, result);
 
 	result = playlist_load_spl(&g_playlist, client->player_control,
 				   argv[1]);

src/decoder/gme_decoder_plugin.c

@@ -153,6 +153,9 @@ gme_file_decode(struct decoder *decoder, const char *path_fs)
 	if((gme_err = gme_start_track(emu, song_num)) != NULL)
 		g_warning("%s", gme_err);
 
+	if(ti->length > 0)
+		gme_set_fade(emu, ti->length);
+
 	/* play */
 	do {
 		gme_err = gme_play(emu, GME_BUFFER_SAMPLES, buf);

src/encoder/vorbis_encoder.c

@@ -276,6 +276,8 @@ vorbis_encoder_flush(struct encoder *_encoder, G_GNUC_UNUSED GError **error)
 	vorbis_analysis_init(&encoder->vd, &encoder->vi);
 	vorbis_block_init(&encoder->vd, &encoder->vb);
 
+	ogg_stream_reset(&encoder->os);
+
 	encoder->flush = true;
 	return true;
 }

src/output/ao_plugin.c

@@ -26,6 +26,9 @@
 #undef G_LOG_DOMAIN
 #define G_LOG_DOMAIN "ao"
 
+/* An ao_sample_format, with all fields set to zero: */
+static const ao_sample_format OUR_AO_FORMAT_INITIALIZER;
+
 static unsigned ao_output_ref;
 
 struct ao_data {
@@ -167,7 +170,7 @@ static bool
 ao_output_open(void *data, struct audio_format *audio_format,
 	       GError **error)
 {
-	ao_sample_format format;
+	ao_sample_format format = OUR_AO_FORMAT_INITIALIZER;
 	struct ao_data *ad = (struct ao_data *)data;
 
 	switch (audio_format->format) {

src/output/httpd_output_plugin.c

@@ -36,6 +36,7 @@
 #include <errno.h>
 
 #ifdef HAVE_LIBWRAP
+#include <sys/socket.h> /* needed for AF_UNIX */
 #include <tcpd.h>
 #endif
 
@@ -123,6 +124,7 @@ httpd_output_init(G_GNUC_UNUSED const struct audio_format *audio_format,
 
 	/* initialize metadata */
 	httpd->metadata = NULL;
+	httpd->unflushed_input = 0;
 
 	/* initialize encoder */
 

src/output/jack_output_plugin.c

@@ -40,7 +40,7 @@ enum {
 	MAX_PORTS = 16,
 };
 
-static const size_t sample_size = sizeof(jack_default_audio_sample_t);
+static const size_t jack_sample_size = sizeof(jack_default_audio_sample_t);
 
 struct jack_data {
 	/**
@@ -103,9 +103,9 @@ mpd_jack_available(const struct jack_data *jd)
 			min = current;
 	}
 
-	assert(min % sample_size == 0);
+	assert(min % jack_sample_size == 0);
 
-	return min / sample_size;
+	return min / jack_sample_size;
 }
 
 static int
@@ -123,7 +123,7 @@ mpd_jack_process(jack_nframes_t nframes, void *arg)
 		const jack_nframes_t available = mpd_jack_available(jd);
 		for (unsigned i = 0; i < jd->audio_format.channels; ++i)
 			jack_ringbuffer_read_advance(jd->ringbuffer[i],
-						     available * sample_size);
+						     available * jack_sample_size);
 
 		/* generate silence while MPD is paused */
 
@@ -144,7 +144,7 @@ mpd_jack_process(jack_nframes_t nframes, void *arg)
 	for (unsigned i = 0; i < jd->audio_format.channels; ++i) {
 		out = jack_port_get_buffer(jd->ports[i], nframes);
 		jack_ringbuffer_read(jd->ringbuffer[i],
-				     (char *)out, available * sample_size);
+				     (char *)out, available * jack_sample_size);
 
 		for (jack_nframes_t f = available; f < nframes; ++f)
 			/* ringbuffer underrun, fill with silence */
@@ -675,7 +675,7 @@ mpd_jack_play(void *data, const void *chunk, size_t size, GError **error_r)
 				space = space1;
 		}
 
-		if (space >= frame_size)
+		if (space >= jack_sample_size)
 			break;
 
 		/* XXX do something more intelligent to
@@ -683,7 +683,7 @@ mpd_jack_play(void *data, const void *chunk, size_t size, GError **error_r)
 		g_usleep(1000);
 	}
 
-	space /= sample_size;
+	space /= jack_sample_size;
 	if (space < size)
 		size = space;
 

src/output/oss_plugin.c

@@ -41,6 +41,15 @@
 # include <sys/soundcard.h>
 #endif /* !(defined(__OpenBSD__) || defined(__NetBSD__) */
 
+/* We got bug reports from FreeBSD users who said that the two 24 bit
+   formats generate white noise on FreeBSD, but 32 bit works.  This is
+   a workaround until we know what exactly is expected by the kernel
+   audio drivers. */
+#ifndef __linux__
+#undef AFMT_S24_PACKED
+#undef AFMT_S24_NE
+#endif
+
 struct oss_data {
 	int fd;
 	const char *device;
@@ -461,6 +470,12 @@ oss_setup_sample_format(int fd, struct audio_format *audio_format,
 			break;
 
 		audio_format->format = mpd_format;
+
+#ifdef AFMT_S24_PACKED
+		if (oss_format == AFMT_S24_PACKED)
+			audio_format->reverse_endian =
+				G_BYTE_ORDER != G_LITTLE_ENDIAN;
+#endif
 		return true;
 
 	case ERROR:
@@ -502,6 +517,12 @@ oss_setup_sample_format(int fd, struct audio_format *audio_format,
 				break;
 
 			audio_format->format = mpd_format;
+
+#ifdef AFMT_S24_PACKED
+			if (oss_format == AFMT_S24_PACKED)
+				audio_format->reverse_endian =
+					G_BYTE_ORDER != G_LITTLE_ENDIAN;
+#endif
 			return true;
 
 		case ERROR:

src/output_control.c

@@ -139,6 +139,7 @@ audio_output_open(struct audio_output *ao,
 {
 	bool open;
 
+	assert(audio_format_valid(audio_format));
 	assert(mp != NULL);
 
 	if (ao->fail_timer != NULL) {

src/output_thread.c

@@ -96,6 +96,8 @@ ao_filter_open(struct audio_output *ao,
 	       struct audio_format *audio_format,
 	       GError **error_r)
 {
+	assert(audio_format_valid(audio_format));
+
 	/* the replay_gain filter cannot fail here */
 	if (ao->replay_gain_filter != NULL)
 		filter_open(ao->replay_gain_filter, audio_format, error_r);
@@ -137,6 +139,7 @@ ao_open(struct audio_output *ao)
 	assert(!ao->open);
 	assert(ao->pipe != NULL);
 	assert(ao->chunk == NULL);
+	assert(audio_format_valid(&ao->in_audio_format));
 
 	if (ao->fail_timer != NULL) {
 		/* this can only happen when this
@@ -165,6 +168,8 @@ ao_open(struct audio_output *ao)
 		return;
 	}
 
+	assert(audio_format_valid(filter_audio_format));
+
 	ao->out_audio_format = *filter_audio_format;
 	audio_format_mask_apply(&ao->out_audio_format,
 				&ao->config_audio_format);

src/pipe.h

@@ -20,9 +20,9 @@
 #ifndef MPD_PIPE_H
 #define MPD_PIPE_H
 
-#ifndef NDEBUG
 #include <stdbool.h>
 
+#ifndef NDEBUG
 struct audio_format;
 #endif
 

src/update_walk.c

@@ -300,6 +300,9 @@ stat_directory(const struct directory *directory, struct stat *st)
 	if (path_fs == NULL)
 		return -1;
 	ret = stat(path_fs, st);
+	if (ret < 0)
+		g_warning("Failed to stat %s: %s", path_fs, g_strerror(errno));
+
 	g_free(path_fs);
 	return ret;
 }
@@ -316,6 +319,9 @@ stat_directory_child(const struct directory *parent, const char *name,
 		return -1;
 
 	ret = stat(path_fs, st);
+	if (ret < 0)
+		g_warning("Failed to stat %s: %s", path_fs, g_strerror(errno));
+
 	g_free(path_fs);
 	return ret;
 }
@@ -557,6 +563,7 @@ directory_child_access(const struct directory *directory,
 	/* access() is useless on WIN32 */
 	(void)directory;
 	(void)name;
+	(void)mode;
 	return true;
 #else
 	char *path = map_directory_child_fs(directory, name);