Re: [orca-list] Speech-dispatcher Upgrade Broken on Arch



Hello Peter,

maybe you are missing a few symlinks.

In addition to what Reece wrote here:
https://github.com/espeak-ng/espeak-ng#espeak-compatibility
you need symlinks so that espeakup finds the shared library

You can check this way (s/lib64/lib/ for Arch that is multilib, I presume)

didier[/usr/lib64]$ ls -l libespeak*
-rw-r--r-- 1 root root 881680 janv.  8 19:37 libespeak-ng.a
lrwxrwxrwx 1 root root     22 janv.  8 19:39 libespeak-ng.so -> libespeak-ng.so.1.1.49
lrwxrwxrwx 1 root root     22 janv.  8 19:39 libespeak-ng.so.1 -> libespeak-ng.so.1.1.49
-rwxr-xr-x 1 root root 541416 janv.  8 19:37 libespeak-ng.so.1.1.49
lrwxrwxrwx 1 root root     17 janv.  8 19:39 libespeak.so -> libespeak-ng.so.1
lrwxrwxrwx 1 root root     17 janv.  8 19:00 libespeak.so.1 -> libespeak-ng.so.1
didier[/usr/lib64]$ 

didier[~]$ ls -l /usr/bin/espeak*
lrwxrwxrwx 1 root root     9 janv.  8 19:39 /usr/bin/espeak -> espeak-ng
-rwxr-xr-x 1 root root 23712 janv.  8 19:37 /usr/bin/espeak-ng
-rwxr-xr-x 1 root root 23192 janv. 11 20:05 /usr/bin/espeakup
didier[~]$ 

you also should have /usr/include/espeak/speak_lib.h

I attach the SLKBUILD for Slint, FWIW.

Best,

Didier


On 12/01/2019 09:05, Peter Vágner via orca-list wrote:> Hello,

I have no news.
I just have installed that brand new computer yesterday.
I haven't touched speech-dispatcher config nor pulseaudio config i.e. I'm using the default settings. I've 
installed espeak-ng and it all works.

However I've found out espeakup won't work with espeak-ng, so if you need it there may be additional steps 
needed and I haven't yet figured out on how to make espeakup work.

Greetings

Peter


Dňa January 8, 2019 10:30:59 PM UTC používateľ "Arkadiusz Kozioł" <zywek-mailing nvps pl> napísal:

   Yes, Mate is starting pulseaudio via private autostart service.

   Audacious is really using pulse.

   Yes, i have installed pulseaudio-alsa


   W dniu 08.01.2019 o 21:20, Peter Vágner pisze:
    Hello,

    How does audacious produce sound?
    Are you sure it communicates to pulseaudio?
    Do you also have pulse-audio-alsa installed?
    Is your pulseaudio starting with your user X session via dbus
    activation or is it starting as soon as you are trying to play
    audio either by using speech-dispatcher or audacious?

    Please try to understand what's going on and please try to avoid
    repeating the same thing over and over again. I'm sorry to say
    this but your arguments sound to me the way if you were about to
    say that someone else has broken that for you and your are
    demanding a fix without you doing nothing to help to figure it out.

    Greetings

    Peter


    Dňa January 8, 2019 10:11:08 PM UTC používateľ "Arkadiusz Kozioł"
    <zywek-mailing nvps pl> napísal:

        But sounds are playing, I've tested it with audacious etcetera.

        Only speechd can not use pulse.

        W dniu 08.01.2019 o 21:06, Peter Vágner pisze:
        Hello,

        It should work like this on Arch linux.
        * User starts orca.
        * Orca uses python-speechd to communicate to user specific
        speech-dispatcher instance. If speech-dispatcher is not yet
        running, it will spawn it automagically.
        * Speech-dispatcher will start and try to load either all
        available output modules or those configured in either
        system-wide config or user specific speech-dispatcher config.
        * By default speech-dispatcher is configured to play TTS
        sound via pulseaudio.
        * As soon as speech-dispatcher has to produce audio it tries
        to connect to the user session specific pulseaudio instance
        by using libpulse. On Arch linux auto spawn is disabled in
        the pulseaudio configuration however pulseaudio is inwoked
        via systemd socket activation as soon as it's needed.

        So if you still have issues you need to understand what I'm
        trying to explain to you, poke at your system before starting
        orca, identify your sound devices are working as they should,
        if yes then find out if speech-dispatcher and pulseaudio are
        running prior to starting orca, if they are already running,
        can they produce sound, start orca, poke at your system again
        exploring speech-dispatcher and pulseaudio processes and
        their logs taking into account what we are talking about here.
        Also this is all with the default arch linux configuration
        i.e. no changes to speech-dispatcher and pulseaudio config
        files including system-wide and user specific configs. If you
        have tweaked the configs you need to take into account what
        you have tweaked in there while troubleshooting all this.

        I'm really sorry I can't really diagnose this for you as it
        is all working fine for me on three different Arch linux
        installs I've tried this on so far.
        I will be installing Arch on a brand new machine in a few
        days, so if I discover something usefull I will report it.

        Greetings

        Peter

        Dňa January 8, 2019 9:42:00 AM UTC používateľ Didier Spaier
        <didier slint fr> napísal:

            This just means that either speech-dispatcher doesn't know where is the socket (the file to 
be accessed to 'talk' to pulseaudio), or is not allowed to access pulseaudio

            There can be many reasons for that, but first please answer these already asked questions:
            Which exact command do you type to start speech-dispatcher?
            Why not just let Orca start speech-dispatcher which will then connect itself to pulseaudio 
(or to another sound server, depending of its settings)?
              
            On 08/01/2019 09:11, Arkadiusz Kozioł wrote:

                Just I wrote, that I have an error logs, that
                sd_espeak_ng or another can not connect to the pulse
                audio. WHat more do you need from me to try to help
                me? W dniu 06.01.2019 o 20:09, Mallard pisze:

                    Thank you, Peter. How can I get to the log,
                    please? Ciao, Ollie Il 06/01/2019 20:26, Peter
                    Vágner via orca-list ha scritto:

                        Hello, I'm sorry I really have no further
                        suggestions without looking into the logs as
                        I've tried to explain. greetings Peter Dňa
                        January 6, 2019 5:44:52 PM UTC používateľ
                        "Arkadiusz Kozioł" <zywek-mailing nvps pl>
                        napísal:     I have the newest version of
                        espeak-ng from community repositoryy and I
                            can not use my Arch.     W dniu
                        03.01.2019 o 16:56, Peter Vágner pisze:
                                Hello, The issue with upgrading to
                        speech-dispatcher 0.8.8-3         is that you
                        should run up to date espeak-ng rather than
                        espeak         or possibly out dated version
                        of espeak-ng. Or have you         ensured you
                        have up to date espeak-ng and the issue still
                                persists? Greetings Peter Dňa
                        2019-01-03 13:12 Arkadiusz         Kozioł
                        napísal(a):             I'm using Archlinux,
                        speech dispatcher 0.8.8-2 which it's
                                    working, 0.8.8-3 not works, the
                        newest version of             pulseaudio,
                        sounds are playing, so pulse works, only
                                    speechd can not connect to. W
                        dniu 02.01.2019 o 20:30,             Austin
                        Seraphin pisze:                 Thanks Peter!
                        I didn't realize that espeak-ng made it
                                        into Community. Good!
                        Installing espeak-ng and                
                        updating speechd.conf fixed the problem and
                        things                 sound great. I think
                        it might even act a little                
                        snappier. On 1/1/19 10:37 AM, Peter Vágner
                        wrote:                 Hello,
                        speech-dispatcher 0.8.8-3 and libspeechd
                                        0.8.8-3 are just Arch linux
                        specific rebuilds of the                
                        package. There are no new upstream changes.
                        Arch linux                 specific change is
                        that they officially require                
                        espeak-ng instead of espeak. Is it possible
                        you do                 have some older
                        version of eSpeak-ng installed from
                                        the AUR that may cause
                        incompatibility? If you wish to
                                        live on bleeding edge, please
                        reinstall espeak-ng-git                 from
                        the AUR, otherwise install espeak-ng from the
                                        community repo. sudo pacman
                        -S espeak-ng If this is                 still
                        not working for you, please look at the logs
                                        reading one of these files
                        depending on how you are                
                        starting speech-dispatcher: /run/user/`id
                                       
                        -u`/speech-dispatcher/log/speech-dispatcher.log
                        ~/.cache/speech-dispatcher/log/speech-dispatcher.log
                                        Hopefully by doing this you
                        can find out what is going                 on
                        on your system. For me this is all working
                        great as                 I'm running up to
                        date espeak-ng-git. Greetings Peter
                                        Dňa 1. 1. 2019 o 15:34 José
                        Vilmar Estácio de Souza                
                        napísal(a): I found exactly the same. On
                        12/31/18                 11:30 PM, Austin
                        Seraphin wrote: Hello. I decided to
                                        upgrade my Arch Linux system
                        for the year's end and                 when I
                        rebooted I found that Orca no longer worked.
                                        The spd-say command didn't
                        work either, so I narrowed                 it
                        down to speech-dispatcher. I noticed that
                        this                 upgrade did have a new
                        version, 0.8.8-3. I downgraded
                                        speech-dispatcher and
                        libspeechd to 0.8.8-2 and                
                        rebooted. The system worked. I wonder what
                        has                 changed. I wanted to
                        alert everyone. -- Assinatura


_______________________________________________
orca-list mailing list
orca-list gnome org
https://mail.gnome.org/mailman/listinfo/orca-list
Orca wiki: https://wiki.gnome.org/Projects/Orca
Orca documentation: https://help.gnome.org/users/orca/stable/
GNOME Universal Access guide: https://help.gnome.org/users/gnome-help/stable/a11y.html
Log bugs and feature requests at http://bugzilla.gnome.org

Attachment: SLKBUILD
Description: Binary data



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