Lonami

I recommend squashing, but don't merge it yet because it seems it only works for the moon. Breaks as soon as I introduce more text.

I guess the algorithm was right and it was only missing adding the surrogates because JavaScript represents strings in a different way to Python.

Ah, my bad, I missed that. Would it be worthwhile to put it in `Feed` as well, or does having it there make not much sense?

Might be from the time where `Client` and `ClientHandle` were separate things (there's no such separation in the current git version). Please be welcome to send a pull request removing...

Perhaps it was semantics (e.g. if you mark a message as read, it makes sense that the message was mutated so the method would require mutable access). But it's kind...

I have added the following code to the [`dialogs.rs` example](https://github.com/Lonami/grammers/blob/e5252e02f57cc90820c75f27fc23bac6c7d9971b/lib/grammers-client/examples/dialogs.rs) (inside the loop which prints the chats): ```rust let repacked = PackedChat::from_bytes(&dialog.chat().pack().to_bytes()).unwrap(); assert_eq!(repacked, dialog.chat().pack()); if let Some(input_user) = repacked.try_to_input_user() {...

Also, if I debug the values of `repacked`, they are the same, even if I restart the program.

Are you fetching the sender from a private conversation, a small group chat, or a megagroup? If it's a megagroup, how many members does it have? If possible, what do...

If I try to print a sender from a private conversation, the hash looks normal. I can print it, embed it, restart the application, and `get_full_user` works. I then tried...

The packed values you posted seem fine (access hash is non-zero). I really don't know why Telegram refuses to use those values. Does it work if you use `get_full_user` immediately...