Georg Oldenburger

I am not sure about this simple change. Interfaces are different even with current stage. I think we can make Websocket Transport optional for server-side client. And I think we...

Update: problem localized - COOKIE. we have cookies per domain and use cookies to store token - session identifier. There is no such problem with different browsers and incognito mode.

@tshemsedinov please review access syntax. Do we have any blocks in `impress` development, that I should first wait for before start working on this feature?

This is different from low-level implementation, because there will be metadata in a packet as well. So packet will be recognized by metadata. In some cases, we need async messaging,...

`connection.on('message', () =>` is far not the same as `api.interfaceName.processEvent(({ message }) => {})`. There is an `interfaceName` metadata provided withing the packet. So this is an alternative solution, and...

The alternative solution just brought confusion to the topic. I removed it. So speaking about proposed solution. It is just mirroring existing functionality, and of course I see this useful...

Server-side, we have no direct link to a WS connection available. This would break our abstraction layers. So we need to have such functionality implemented. For client, I am sure...

Example: https://github.com/metarhia/Example/blob/6bb335724357510c99cf11c60438bc856ea0e4aa/application/static/console.js#L461 Intorspect method expect interfaceNames. It is not a bug

I created a pull request with 2 tests. First test: single process case with `threadId === 0`, which is `mainThreadId` Second test: case with 3 workers no balancer and 2...