[pulseaudio-discuss] [Support] TJ's console speech question (fwd)

Jude DaShiell jdashiel at panix.com
Mon Sep 5 03:19:28 UTC 2016


---------- Forwarded message ----------
Date: Sun, 4 Sep 2016 20:00:39
From: kendell clark via Support <support at sonargnulinux.com>
Reply-To: kendell clark <coffeekingms at gmail.com>
To: sonar support list <support at sonargnulinux.com>
Subject: Re: [Support] TJ's console speech question

hi all
TJ, I'm sorry to read about your frustrations with console speech. I've
been trying to solve it for months and kept running up against one
constant issue. We use a sound server called pulse audio. In plain
english, this means that applications connect to pulse audio to get
sound when they need it instead of talking to your hardware. This makes
coding them a lot easier apparently since pulse audio takes care of the
hardware bits. But there is one very serious problem, that was
apparently designed into pulse audio. It runs as the user you're logged
in as, so probably tj for you. Speakup, being built into the linux
kernel, runs as root. That's the master user who has permissions over
every single file on the system and can do anything. It has to or it
couldn't get access to the screen to read it. Pulse audio denies it
access so it can't speak. I don't know how it does this, but it does. A
few weeks ago a friend of mine developed an option you can put into
pulse audio's config file to make it listen on a socket. this means that
apps running as other users can connect and pulse audio won't block
them. I'll attach the file instead of having you write the option out as
it's a bit geekish. Put this file in /home/tj/.config/pulse. You can
shorten this to ~/ if you want, but make sure it goes there. Before you
restart pulse audio, delete the system wide config file. To do this,
enter a terminal and type sudo rm -rf /etc/pulse.  If you forget to do
this you're likely to lose all sound when you restart pulse audio. I
customized the pulse configuration file to make it switch over to new
sound devices when you plug them in, like versions of windows do. This
option is now in the user specific config file I'm attaching and if
they're in both places pulse audio will fail hard. Then reinstall pulse
audio with sudo pacman -S pulseaudio. You will only have to do this
once, and this stuff won't be needed in the next version of sonar. Then
restart pulse audio. To do that, in a terminal, which you can get to by
pressing alt+control+t, type killall pulseaudio. It will start itself on
it's own, and speech should work. I'm putting this in the next version
of sonar, which is due out next month. I'm sorry there's been no alpha
release, I'm still working on it. The distro we're a part of, manjaro,
took about 90 percent of the software I put into sonar out of it until
they could make sure  that the desktop sonar uses was 100 percent
accessible (I already do that) and make sure all the keyboard shortcuts
and such work and don't break (that's over my head and I can't do that).
They will put the software back when they're done and as soon as they do
I'll release the alpha. I've held off because I don't want to give you
guys a version of sonar that's nothing more than a talking desktop with
no software at all not even a browser.
Kendell Clark

Mark Peveto via Support wrote:
> Hi all,
> I read TJ's question about console speech earlier, and i'll do my best to get him going.
> TJ, this may be an involved process, but it does work, so lemme know if you haven't had any luck as of yet.
> Everything happens after coffee!
> Mark Peveto
> Registered Linux user number 600552
> Sent from talking arch using alpine 2.20.15
> Talking arch homepage: https://talkingarch.tk
> Latest version of alpine: git clone http://repo.or.cz/alpine.git
> _______________________________________________
> Support mailing list
> Support at sonargnulinux.com
> http://sonargnulinux.com/mailman/listinfo/support_sonargnulinux.com
-------------- next part --------------
.include /etc/pulse/default.pa
#switch to newly plugged in sound hardware
load-module module-switch-on-connect
#enable console screen readers
load-module module-native-protocol-unix auth-anonymous=1 socket=/tmp/pulse.sock
-------------- next part --------------
Support mailing list
Support at sonargnulinux.com

More information about the pulseaudio-discuss mailing list