Gnought

Do you have a external persistence session storage layer in mosquito server? @robertsLando

Hmm, this is same as the current Aedes behaviour. The obsolete LWT is left in the persistent storage of mosquito.

I mean the ClientId should be within the mentioned characeter set `0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ` No spaces, no hypens, no dash.

Lots of aedes codes are in callback-based, is it possible to switch to promise-based?

@robertsLando Yes, it's inspired by #362

Please hold a bit, I’m going to make a big change. This PR may be closed after that if my work is successful. Btw, I prefer to use rebase rather...

@robertsLando will keep you posted. my upcoming changes will be - replace the packet queue before connected event - redesign the _nextBatch

The upcoming changes have been ready in my local working copy. I will submit a PR after the following PR has been merged into the main branch. https://github.com/moscajs/aedes/pull/409 https://github.com/moscajs/aedes/pull/408 https://github.com/moscajs/aedes/pull/416...

@robertsLando Here is an early access (https://github.com/gnought/aedes/commits/next) what my next PR will be after migrating #408, #416 and #418 into the main branch. The main changes are introducing a `mqtt-stream-parser`...

@robertsLando when we pause the stream, NodeJS will do buffering itself (https://nodejs.org/api/stream.html#stream_buffering), and we could still get those buffered data after resume. Constructing our own queue is a rewheel their...