Re: [orca-list] SELinux notification and speech-dispatcher on fedora?

[Vsevolod Popov <sevapopov13 gmail com>]

Hello,
Restarting speech dispatcher's service is what was written in one of 
guides that I read about building RHVoice,
Thank you for your explanation. I saw that speech-dispatcher's unit 
isn't enabled at boot so I guessed that Orca enables speech-dispatcher 
when it's started and you actually proofed my guesses.
On 9/1/22 21:02, Rastislav Kish via orca-list wrote:
> Hi,
>
> sorry, if my question is noob, I'm not a Fedora user and I don't use RH
> voices either.
>
> But, is speech dispatcher actually supposed to run as a systemd unit?
>
>
> I don't see any configuration files for it on my system, I ripgrepped
> through systemctl list-units and didn't find anything either.
>
> I checked also my Fedora Mate 36 VM, with the same result.
>
>
> I did some speechd related work on multiple occasions, like building my
> C# speechd library.
>
> As far as I understand, the server is by design not supposed to act as a
> full time daemon. Clients (like Orca) spawn a temporal speechd service
> (not a systemd one, just a detached program), that lives until at least
> one connection is active.
>
>
> When there are no more connections, it automatically shuts down after 5
> seconds unless a new connection is made.
>
>
> You can actually see this behavior in the terminal by typing pgrep
> speech-dispatch.
>
> With running orca, you'll get the server's PID. If you turn Orca off and
> wait for 5 seconds, then re-activate the command by pressing up arrow
> and enter, you'll get nothing listed.
>
> After turning orca on, issuing the command will give you a PID again,
> but different from the previous-one.
>
>
> With this model, as far as the libraries support autospawning, a systemd
> service is not necessary, as the server should be started when needed.

--
regards,
Vsevolod
https://github.com/sevapopov2/