[v2] mpv stops randomly on start

[vankasteelj]

When spawning a new player, sometimes (way to often, maybe 1/10) the load doesn't work.

mpv.start().then(() => {
   console.log('mpv started');
   return mpv.load('<my file>');
}).then(() => {
   console.log('file loaded');
});

The player does spawn, even the file is being read by MPV, yet the module fired a "stopped" event in between and the MPV binary is "free", your module doesnt have access to it.

The "stopped" event does occur after the console log "Connected to socket '\.\pipe\mpvserver'" though.

[vankasteelj]

When the above happens, the mpv.mpvPlayer.connected is false, but there is no exit code (it's set to default null). The exit code does update to 0 after a manual closing of the MPV spawned window

[vankasteelj]

So, it does seem that the issue is that mpv sends back (through ipcInterface) an "idle" event before playing the actual file. Causing the player to disconnect.

[vankasteelj]

Disabling the this.emit('stopped') in case of "idle" message from ipc (_events.js) corrects the issue. Now the file loads correctly everytime. Of course, it also disables the auto-close feature at the end of the file: the mpv player does close, but this module isnt made aware of it, logically because I presume mpv sends a "idle" message at that time. MMmmh

[vankasteelj]

Maybe this is an issue with how I use your module? I interpret the "stop" event as a "the file playback has stopped", when it is, as I see in your code, "the player idles". Anyway, I worked around it, but it still is wierd that it happens randomly.

[AxelTerizaki]

Which version of MPV are you using ? I know I've had some weird issues depending on which version I used.

[vankasteelj]

0.29 (31/07/2018)

[j-holub]

Thanks for letting me know. At the moment university keeps me really really busy that's why I haven't done much for this project in the past month. I'll try to have a look at it soon. Thanks for the very detailed feedback I think that will help a lot.

[j-holub]

I cannot reproduce this. I have done some changes to start() lately, which should make it more robust. Can you see if it still occurs on your case?

[j-holub]

Closing this for now as I cannot reproduce the error. Feel free to reopen it if the problem still prevails.