I updated to the latest snapshot yesterday and was checking functionality when I noticed none of my Sonos tts commands were playing, though they worked before the upgrade. Checking the log I see THOUSANDS of these messages:
2016-12-27 08:50:36.877 [WARN ] [inding.sonos.internal.SonosXMLParser] - Did not recognise element named narrator
2016-12-27 08:50:36.878 [WARN ] [inding.sonos.internal.SonosXMLParser] - Did not recognise element named albumArtist
Everytime I invoke the “say” command, the log fills up with these warnings but nothing plays.
I have removed the binding and the things and re-added them but that hasn’t changed the errors. I have 7 Sonoses in the system and it doesn’t seem to matter which is the audiosink. The binding version in my paperUI is 0.9.0-SNAPSHOT.
Additionally I noticed that if a user is issuing many say commands without forcing a delayed execution related sounds are not serialized anymore but one is hearing various speeches overlapping.
I really like the old behavior not having to worry about defining Thread::sleep(x) with “x” being speech content specific.
This had also already been working in older releases!
Please try removing all entries in your Sonos queue and try again.
What kind of source (Sonos service) do you have in this queue ?
These warnings should not block the TTS command but the parsing of few metadata fields are apparemtly missing. You mentioned for example narrator. Please give us a complete list of fields that generate such a warning.
The current sound notification system is parsing the content of everything in your queue just to know how many items are in the queue and add a new entry at the end. It could take a lot if time if you have many items in your queue. I think it could be simply avoided. I will try to enhance that.
But please give me the different warnings you see so that I can fix the XML parsing too. Is it only narrator and albumArtist ? I will create another issue for that.
Just try with few items in the queue to check that the TTS is still working.
Note that it is not something that was changed recently.
I created a second issue in Git for the unexpected warnings. I am waiting for your help to get a list of additional fields to consider when parsing the data.
Those were the only warnings in the list and you were right, there were a couple thousand items in the queue. Once I cleared it I get no more warnings and I can hear the say commands again. I think there were so many that it caused the TTS to not play.
For information, I have now fixed the two issues (1 fix is not yet merged). Sound notification is now independent on how many items you have in the queue. You can have several thousands if you like.
Will be available probably in the next official release.
@Lolodomo
Thanks that you are improving the binding. This is really great.
One question: Is there also some progress on “SONOS PLAYBAR support TV optical input #2410” ?
I have been testing Beta 5. And in this build it is not working.
But I am not sure whether last update provided by chaton78 was included in this build.
@Marty56: for the playbar optical input, I have done what I can without the ability to test. As already mentionned, I forgave and let someone with a Playbar fix that issue.
@Mike_Bagdanoff : save/restore is one of the last features I have not yet tested.
I’m not sure if this ever got into the mainline but I updated the binding yesterday and it still happens that if there are items in a queue, the sound will not play
