Expose shortid method for SessionContext

[benalleng]

By exposing the shortid method for SessionContext it allows us to be able to track the id across all psbt intermediate steps and allows us to not need this method duplicated in every step of the process.

Closes #697

[coveralls]

Pull Request Test Coverage Report for Build 15047681959

Details

• 12 of 15 (80.0%) changed or added relevant lines in 3 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage decreased (-0.005%) to 82.242%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
payjoin/src/receive/v2/persist.rs	1	2	50.0%
payjoin/src/receive/multiparty/mod.rs	4	6	66.67%

Totals
Change from base Build 15031183812:	-0.005%
Covered Lines:	5474
Relevant Lines:	6656

---

💛 - Coveralls

[benalleng]

Curious if you think you are satified with the changes to the method slightly to address #612 here too

[benalleng]

@nothingmuch I remember hearing you said you were planning on removing the id() method would that change ultimately make this change moot?

[nothingmuch]

@nothingmuch I remember hearing you said you were planning on removing the id() method would that change ultimately make this change moot?

hmm, i don't think so

If i understand the code correctly, .id() is only used in #642 for error reporting, the full SessionContexts are compared to each other directly, so this is just a way of reporting the sessions that this error condition is related to?

the idea of computing the ShortId derived from the receiver key associated with the session still make sense, we just wanted to move away from relying on the same short ID as the database key, as in here:

rust-payjoin/payjoin/src/receive/v2/persist.rs

Line 9 in e2b66a2

pub struct ReceiverToken(ShortId);

i was referring to this method:

rust-payjoin/payjoin/src/receive/v2/mod.rs

Line 343 in e2b66a2

pub fn id(&self) -> ShortId { id(&self.context.s) }

but honestly i kinda misspoke, i didn't consider if we need it for other reasons i just assumed we don't, and what i really meant was just to no longer use this for keying the database

[arminsabouri]

Simpler, cleaner and removes duplicate code. Seems like a win to me!

[arminsabouri]

One thing did just come to mind. Does this need to be pub? IIRC only internal types need access to this?
Seems like it was pub previously as well.

[DanGould]

should be pub(crate) because it's on SessionContext and not receiver.id()

[DanGould]

If i understand the code correctly, .id() is only used in #642 for error reporting, the full SessionContexts are compared to each other directly, so this is just a way of reporting the sessions that this error condition is related to?

-- @nothingmuch

See BullBitcoin source id() usage. receiver.id() is used to identify the saved thing in the DB downstream until we get SessionEventLog. That's why this is pub.