Package: speech-dispatcher / 0.8-7

Metadata

Package Version Patches format
speech-dispatcher 0.8-7 3.0 (quilt)

Patch series

view the series file
Patch File delta Description
infinite loop.patch | (download)

src/server/speaking.c | 2 2 + 0 - 0 !
1 file changed, 2 insertions(+)

---
pico generic.patch | (download)

config/modules/Makefile.am | 4 2 + 2 - 0 !
config/modules/pico-generic.conf | 77 77 + 0 - 0 !
config/speechd.conf | 1 1 + 0 - 0 !
3 files changed, 80 insertions(+), 2 deletions(-)

 add support for pico using generic module
configurable module dir.patch | (download)

configure.ac | 5 4 + 1 - 0 !
1 file changed, 4 insertions(+), 1 deletion(-)

 [patch] allow the speech-dispatcher module bin dir to be
 configurable

The speech-dispatcher synthesizer module model is such that any executable
binary or script can perform speech synthesis duties for speech-dispatcher,
as long as it can communicate via the appropriate protocol and file
descriptors. This also extends to using i386 binaries on an amd64 system.

Many distributions now have a model in place that allows the easy use of i386
binaries on an amd64 install. By default, the speech-dispatcher module
directory is placed in ${libdir}, which defaults to a 32-bit or 64-bit library
directory on some distros. Given the above, it should be possible to provide
an architecture independant location for modules to be installed to, and used.

cs docs.patch | (download)

doc/Makefile.am | 6 3 + 3 - 0 !
doc/speech-dispatcher-cs.texi | 889 889 + 0 - 0 !
2 files changed, 892 insertions(+), 3 deletions(-)

 upstream forgot to include czech documentation in the release
fdl.patch | (download)

doc/fdl.texi | 2 1 + 1 - 0 !
1 file changed, 1 insertion(+), 1 deletion(-)

 raise level of subsection in fdl.texi to correct document structure.
pulse default latency.patch | (download)

config/speechd.conf | 2 1 + 1 - 0 !
src/audio/pulse.c | 3 2 + 1 - 0 !
2 files changed, 3 insertions(+), 2 deletions(-)

 speech-dispatcher defaults to extreme low latency
 speech-dispatcher requests a total latency of 1 ms, which in turn gives
 sub-ms latency inside the PulseAudio engine. This causes unnecessary CPU
 consumption, or underruns. I understand that you would want immediate
 feedback, but 20 ms seems more reasonable.
spd_audio Expose dlopened library s symbols to libs .patch | (download)

src/modules/spd_audio.c | 23 22 + 1 - 0 !
1 file changed, 22 insertions(+), 1 deletion(-)

 [patch] spd_audio: expose dlopened library's symbols to libs it
 loads.

This fixes the following issue.
* Libao is loaded as an audio plugin.
* Libao loads its own plugins, using dlopen.
* Those plugins try to call functions defined in libao itself,
  but fail because the symbols aren't globally available.

espeak mbrola generic.conf.patch | (download)

config/modules/espeak-mbrola-generic.conf | 2 1 + 1 - 0 !
1 file changed, 1 insertion(+), 1 deletion(-)

 after enabling the corresponding line in speechd.conf, the
 espeak-mbrola-generic module produces no sound at all.
 .
 The GenericExecuteSynth command in espeak-mbrola-generic.conf seems wrong
 because:
 .
 - the data espeak outputs without the --pho option is not a valid input for
  mbrola;
 .
 - the $VOLUME variable contains a float value suitable for mbrola, while
  espeak -a parameter expects an integer; moreover, the -a parameter has no
  effect at all when espeak produces phoneme data, and can be dropped at
  all since the volume will be determined by mbrola's -v parameter;
 .
 - without the --stdin parameter, espeak will take the text from stdin, but
  will treat each line as a separate sentence;
 .
 - without the -q parameter espeak will try to access audio devices and
  produce audio output by itself;
 .
 - the espeak -p parameter can alter the sound output produced by mbrola and
  can be set to the value contained in the $PITCH env variable.
Bug-Debian: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=665382

fix_spelling_errors_reported_by_lintian.patch | (download)

doc/spd-say.texi | 2 1 + 1 - 0 !
doc/speech-dispatcher.texi | 14 7 + 7 - 0 !
po/cs.po | 2 1 + 1 - 0 !
po/hu.po | 2 1 + 1 - 0 !
src/audio/alsa.c | 2 1 + 1 - 0 !
src/clients/say/options.c | 2 1 + 1 - 0 !
src/modules/cicero.c | 2 1 + 1 - 0 !
src/modules/generic.c | 2 1 + 1 - 0 !
src/modules/module_utils.c | 2 1 + 1 - 0 !
src/server/module.c | 2 1 + 1 - 0 !
src/server/output.c | 8 4 + 4 - 0 !
11 files changed, 20 insertions(+), 20 deletions(-)

 lintian reports several spelling errors.
 This patch fixes those that turn up in the binaries.
update_documentation_with_xdg_path.patch | (download)

doc/speech-dispatcher.texi | 18 10 + 8 - 0 !
doc/ssip.texi | 2 1 + 1 - 0 !
2 files changed, 11 insertions(+), 9 deletions(-)

 in 0.8 the location of the user config/log directory changed
 but the documentation was not updated everywhere.
Fix use after free in src server speaking.c.patch | (download)

src/server/speaking.c | 7 7 + 0 - 0 !
1 file changed, 7 insertions(+)

 [patch 1/2] fix use-after-free in src/server/speaking.c.

When current_message was resumed from the list of paused messages, it
was being freed, even though it had been re-added to the message
queues.

Don t add multiple copies of a message to the paused.patch | (download)

src/server/speaking.c | 5 3 + 2 - 0 !
1 file changed, 3 insertions(+), 2 deletions(-)

 [patch 2/2] don't add multiple copies of a message to the paused
 list.

Fixes another use-after-free in src/server/speaking.c.