Re-evaluate `ChorusUser::shell`

[bitfl0wer]

https://github.com/polyphony-chat/chorus/blob/ca9c4b8949c021822b7ff974c15d5d3575b53c15/src/instance.rs#L208-L232

Re-evaluate if ChorusUser::shell is still needed, and if so, if it is implemented correctly. I remember that this method was used as a sort of "hack" a couple months ago, but the code has changed a lot since then, so it might not be needed anymore.

[kozabrada123]

It is actually still used in several places:

https://github.com/polyphony-chat/chorus/blob/ca9c4b8949c021822b7ff974c15d5d3575b53c15/src/api/auth/login.rs#L17-L34

https://github.com/polyphony-chat/chorus/blob/ca9c4b8949c021822b7ff974c15d5d3575b53c15/src/api/auth/register.rs#L21-L41

https://github.com/polyphony-chat/chorus/blob/ca9c4b8949c021822b7ff974c15d5d3575b53c15/src/api/users/users.rs#L152-L161

(This one should likely at least support a version with an existing user):

https://github.com/polyphony-chat/chorus/blob/ca9c4b8949c021822b7ff974c15d5d3575b53c15/src/api/users/users.rs#L117-L130

[kozabrada123]

Also, note that despite ChorusUser including a gateway handle

https://github.com/polyphony-chat/chorus/blob/ca9c4b8949c021822b7ff974c15d5d3575b53c15/src/instance.rs#L208-L218

login and register spawn a new gateway, which should probably be fixed :P

https://github.com/polyphony-chat/chorus/blob/ca9c4b8949c021822b7ff974c15d5d3575b53c15/src/api/auth/login.rs#L21-L45

https://github.com/polyphony-chat/chorus/blob/ca9c4b8949c021822b7ff974c15d5d3575b53c15/src/api/auth/register.rs#L25-L54

[bitfl0wer]

I sort of have an issue with the ::shell function even needing to exist in the first place, seems like an architectural blunder from my side. But maybe it's also fine and I am reading too much into this 🥲

[kozabrada123]

If possible, I think it's better not to have sth like this is the first place, though I don't have that much of an issue with using it

[kozabrada123]

either way functions like register and login should probably expand the created shell, instead of using it for one request and then building a new user

[kozabrada123]

Oh dear, I just found this:

login:

• creates shell 1
  • creates gateway 1
• calls get_user()
  • creates shell 2
    • creates gateway 2
  • creates gateway 3

register:

• creates shell 1
  • creates gateway 1
• calls get_user()
  • creates shell 2
    • creates gateway 2
• calls get_settings()
  • creates shell 3
    • creates gateway 3
• creates gateway 4

(In practice this probably isn't as terrible as it sounds though, since only the last gateway is used for authenticating and the others are almost immediately dropped and closed)

[bitfl0wer]

Oh dear lord

what have I done

[bitfl0wer]

To be fair, it's ugly, but it works, and it's not like this has affected the rest of the software architecture or performance in any significant way, so it's "okay"

[kozabrada123]

Sure, but I'd still rather fix it

[bitfl0wer]

Absolutely! Hence the issue

[kozabrada123]

I think the way we use shell is now pretty alright, I don't think we need to change the method itself