Bugs in 3.0.5971.0 on Debian 8

[mooks 0]

I updated Emby Server from the official Repo to 3.0.5971.0.

Now 2 bugs are occuring for users. 

The first one: every user is now {0}:

 

 

The second: I have two users: Mooks (me) and Magg.

When Magg is watching a Video Emby shows that me (Mooks) is watching that video on dashboard. But I'm not watching it. So Emby sees Magg as Mooks. Looks like also a account management problem.

 

I restarted Emby Server, ran database clean up and deleted Magg with creating the account new. The errors are remaining.

 

I am using Emby behind an Apache Reverse Proxy. Before the update it worked with this without any errors.

 

In server log I see that error:

2016-05-25 11:24:20.6533 Error NotFoundHttpHandler: 127.0.0.1:8920 Request not found: /embywebsocket?api_key=bcfac5a7d35e47028391511147297063&deviceId=163d93e2476586cebfeea0e807bce747 

Also tested with truncating the cache and disabling "caching web answers" (in developer settings, I have it in German, dunno the correct English phrase). 

 

PS:

For me it appears that Emby Server uses the browser for user identification now. 

Another user uses Firefox, he is separated shown when streaming a video. Mooks and Magg are using both Chrome of same version, but different IP and ISP.

 

PPS:

Okay, in the Android app it shows the client of 'Magg' (Chrome browser) as remote device which is sticked to 'Mooks'. Maybe that causes the error

Edited May 25, 2016 by mooks

[mooks 0]

Okay, crawled it down. The problem is, with the new Emby version there is an error on how Emby detects a device. So every Chrome browser is seen as the same device. Only one profile will be created in 

/var/lib/emby-server/data/devices/

The Chrome Browser device is:

# cat device.json
{"ReportedName":"Chrome 50.0.2661.102","Name":"Chrome 50.0.2661.102","Id":"163d93e2476586cebfeea0e807bce747","LastUserName":"Mooks","AppName":"Emby Mobile","AppVersion":"3.0.5971.0","LastUserId":"05ec6f28b04d4012b5a6101334bd2d68","DateLastModified":"2016-05-25T14:09:02.2510000Z",

Can this bug be resolved in the near time and pushed to the repo mirror? Sorry, but I don't do .NET programming as I would do it on my own if I can.

[Luke 37178]

what action are you doing that has the {0} in the prompt?

[mooks 0]

what action are you doing that has the {0} in the prompt?

 

I was deleting a user.

[Luke 37178]

ok, i just reviewed this and it's because the translation for your language is just a little behind.

[mooks 0]

ok, i just reviewed this and it's because the translation for your language is just a little behind.

 

Where can I contribute? I could bring it to the latest version tomorrow.

 

Btw, did you look at the second error? I gave some details in an issue on GitHub.

Edited May 25, 2016 by mooks

[Luke 37178]

what language is it? i can invite you to our onesky translation site.

[Luke 37178]

the second one is caused by the same device id getting detected for two different browsers. we are using fingerprint2js to build unique identifier. we switched to that a couple months ago and this is actually the very first report of that ever happening. so i'll have to review that.

[mooks 0]

what language is it? i can invite you to our onesky translation site.

 

German. Would be nice.

 

the second one is caused by the same device id getting detected for two different browsers. we are using fingerprint2js to build unique identifier. we switched to that a couple months ago and this is actually the very first report of that ever happening. so i'll have to review that.

 

Okay, we tested our browser on http://valve.github.io/fingerprintjs2/. The problem is, it has to be ran two times with the button, so the function gets called two times. After that the fingerprint id won't change again and we have seperate ones.