Re: Speech Dispatcher via inet_socket

[Sebastian Humenda]

Hi Samuel

Samuel Thibault schrieb am 19.10.2024, 17:54 +0200:
>Sebastian Humenda, le mar. 17 sept. 2024 11:45:30 +0200, a ecrit:
>> I configured in /etc/speech-dispatcher/speechd.conf
>> 
>> ```
>> CommunicationMethod "inet_socket"
>> ```
>> 
>> but clients don't pick up the setting.
>
>Yes, /etc/speech-dispatcher/speechd.conf is read only by the server.

Aw, right. I was assuming it would work similar to Pulseaudio's default.pa
that is read by client and server.

>> I've also tried adding it to
>> $HOME/.config/speech-dispatcher/speechd.conf. However, setting
>> SPEECHD_ADDRESS=inet_socket:127.0.0.1:6560 works fine.
>
>Yes, that's the way to tell clients about it.

While this sounds reasonable that server and client configuration is separate,
it is at the same time a bit unfortunate. For instance, on a Debian system,
BRLTTY runs as admin, Orca as ordinary user. To let both use the same speech
dispatcher, I have to tweak the systemd unit of BRLTTY and the start up script
of Orca to have the environment variable set + kill speech dispatcher after
each boot as something isn't as wished (probably due to Lightdm, etc.).

If Speech Dispatcher clients would pick up the server configuration by some
mechanism, e.g. /etc/speech-dispatcher/client-default.conf, it would be
easier.

I just wanted to spell out the use-case, but admit that I wouldn't have time
to work on it.

Thanks for the clarification.
Sebastian

signature.asc
Description: PGP signature