[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: orca for unstable


On Mon, Feb 10, 2014 at 07:40:54PM +0100, Paul Gevers wrote:
> On 10-02-14 10:26, Sebastian Humenda wrote:
> > just to avoid race conditions: I am now looking into the gnome orca package to
> > get it into unstable. Please tell me if anyone of you is working also on this
> > package.
> If I remember correctly, you need a new version of speech-dispatcher for
> this,

The problem may not be the _version_ of speech-dispatcher, but its
default configuration.

In Debian, speech-dispatcher is preconfigured to run on a per-process /
user base, started on demand, which (from my experience) arises multiple
problems with pulseaudio and deadlocks with concurrent audio access for
the sake of "increased security" which I'm unable to fully confirm.

Orca, rather designed to run as a usermode session-based service, will
have to launch its own instance of speech-dispatcher (which is
apparently kind of difficult).

Instead of this default setup (which isn't working well), I'm running
speech-dispatcher as a system-wide service on Knoppix/Adriane, with

SocketPath "/var/run/speech-dispatcher/speechd.sock"

set in /etc/speech-dispatcher/speechd.conf, using libao with alsa
instead of default pulseaudio, and setting a system-wide environment


so all speech-aware programs will now quere their messages to a single
instance of speech-dispatcher.

This eliminates all problems concerning concurrent runs of
speech-dispatcher, and also works fine with orca (currently
unstable/version 3.4.2, used independently with LXDE, KDE and GNOME,
whereas iceweasel, as special case, will only speak in GNOME).


Reply to: