crybapp / web Goto Github PK

The message shown in the user hover title is inaccurate based on the actual state of the user. See below for examples.

This user does not currently have the controller

I cannot currently steal the controller

The computed property that shows this is located here.

Basically, I joined with two accounts. In one of them, I went back to /home, with its controls.

With my main account, I deleted the room, and it was normal, but in the other account, which was in /home...

this doesn't gets updated, even if you join (which will tell you it doesn't exists) and going back. also gets stuck leaving the room:

Refreshing fixes it, but is not the correct behaviour.

The fancy animation when a VM is running is no longer working. All I see is this even though the VM is running:

Those screenshots should explain themselves.

Seems to be spawning it indefinitely, and failing to connect anywhere, without even cleaning correctly.

Destroying the past object and also adding ratelimits for connections would help on this.

Basic way of getting this: Stop all of Cryb while you're already using a VM in a room.

A indicator to show who the host of the room is would be nice

A whitelist would be really good for some private instances

I'm from germany and have a german keyboard. However, the VM only get's the keycode and has to figure out the input itself and thus special characters that the US Keyboard doesn't have don't work (ÄÖÜ) and all the special characters like @%&/... are mapped to different keys.
Because of this, you also can't paste stuff into the VM. Which might be crucial for some to log in.

My idea would be to handle the input on the client and only send the parsed input (the actual letters or symbols) to the VM.

I think we should work on trying to move all public strings to a directory, maybe l18n/en_US.json that contains all strings that are displayed to the client, so we can work on localisation for users in different languages, etc. Thoughts?

vendors.app.js:606 warn JSMpeg: Possible garbage data. Skipping.
value @ vendors.app.js:606
vendors.app.js:606 warn JSMpeg: Possible garbage data. Skipping.
value @ vendors.app.js:606
wasm-0002793e:22 Uncaught RuntimeError: memory access out of bounds
    at wasm-function[21]:0x121f
    at wasm-function[20]:0xee4
    at wasm-function[19]:0xdbb
    at wasm-function[18]:0xccb
    at MPEG1WASM.decode (https://cryb.app/js/jsmpeg.min.js:3:20525)
    at Player.updateForStreaming (https://cryb.app/js/jsmpeg.min.js:1:8886)
    at Player.update (https://cryb.app/js/jsmpeg.min.js:1:8758)

Pretty much of this - and might be the cause of audio choking.

Pretty self explanatory, see below 👇

In some cases the user icon for a user appears twice. Unknown if this occurs for just the client or everyone in the room, or if the clients icon can only be shown twice or anyones icon can appear twice.

It would be nice if we could just leave and accept the invite in the screen instead of having to go back to the home screen, leaving, and then accepting.

As of right now, if stream is not happening, all you'll get is a white box. A way of detecting if JSMpeg is decoding stuff or not, and then add an animation such as Loading, or information to hang on or report an issue, could be done for this.

I think it's a good idea if there would be a mode where it's invite only for private instances

Related to #5. I joined another (recently created) room then, but then, in the account with the issue:

even though in my main account it shows it is on the room.

By default, the favicon is black which means that it's pretty much invisible if you have dark mode enabled on Google Chrome and other browsers. One solution would be to either change the favicon's color to white or add a white shape around the icon. Google for example is doing that.

Had this happening in Chrome 77. Happens randomly.

Something like Discord and other platforms have. For example, typing :tada: would bring up this emoji: 🎉

When a VM fails for some reason and a user gets disconnected and reconnected to aperture. The browser continues making requests to connect with ALL JWT tokens it has received during the time it's been open.

There is a Google Analytics link which (likely) was used for cryb.app. It would be nice to make this configurable and maybe disabled by default.

If I refresh the tab with an active VM I sometimes get this issue:

And all I'm able to see is this:

Even though the VM is up and running. The only solution is to go back to /home and return to the room from there.

Doesn't seems to happen everytime.

Sometimes this.apertureWs is not getting defined, leading to just a black screen, and no error at all - just it trying indefinitely to connect to https://cryb.app/undefined/?t=token - as an example.

This happens sometimes. All I'm able to see is a white screen. Restarting the VM fixes this issue.

Reported in Cryb Discord server, which was noticed after integrating audio - as after leaving the room, it would still be playing the audio, meaning that the player doesn't stops when it should.

It would be nice if we could control the volume outside of the VM

Possibly daily when users get into web, and possibly if it is noticed that their avatar is 404'ing.

Seems to be an issue with WebSockets. No new chat messages, people entering the room, or the state of the VM spinning up.

Experienced by Pantotone in Cryb Discord server.

We should ask for the user to interact with the page - and definitely not indefinitely try starting up stuff.

This is what it looks like without a video:

And this is what it looks like with the video enabled:

If for example, we've not used Cryb before, and get invited to a room and we're joining, and jsmpeg.min.js hasn't loaded, it will try to use it anyway and then crash as JSMpeg is not defined.

In any case, and to prevent more issues, we should wait until all resources have been loaded in order to start stuff.

Firefox:

Chrome:

It would be nice if the host could kick other people in the room

On logging, this would lead to a blank page without any error:

Also I think we should define better errors as of like, why authentication api-side fails.

This is currently not possible.

Just like in Discord, otherwise it might get really distracting

All stuff gets removed, but you need to click the Cryb logo to go to /home when a room has been destroyed.

Dark mode support for other pages would be nice. We can use the prefers-color-scheme CSS feature. I'm interested to work on this so I'll see what I can do later

Like the title is saying, it would be good to move the brand stuff to .env.example so users can simply tpye git pull to get the newest changes without any errors due to local changes.

Just so we know it's not stuck on anything.

If someone joined the room, the invite at the bottom disappears. I think it would be nice to have a button that will copy the link in the menu in case you need the link again.

I feel this is a network or encoding issue, but not sure. Example:

The host should have the ability to close the room and kick everyone out. Right now, the host can only leave it and the other people will still be inside the room.

If you try to logout, nothing will happen - more like, it'll raise an error.

Leaving this as a note so I don't forget it.

It's currently not deleting the cookie properly nor redirecting.

So you can invite other people without them being able to create a room. They can only stay in the current room and if they leave, they get kicked out completely.

It'd be better if smaller player sizes, such as 640x480 took up the full height and/or width based on the browser size of the player area. See below for an example of what the player currently looks like now at 640x480 👇

This is most likely related to the YouTube Embed API, but I'll see what I can do to see if this issue can be detected and then fixed on the fly.

If I host a room and try to control the mouse it never works. However, if I give it to someone else it works fine.