Handle v2 error generic

[DanGould]

Make the Receiver state machine generic so that it the error handling function can always provide context to an error

Extract JSON OHTTP request to be returned to the directory

This is a Draft because I'm not sure which design is going to make the most sense yet. Sharing my exploration of way to close #468

[coveralls]

Pull Request Test Coverage Report for Build 12739793767

Details

• 84 of 174 (48.28%) changed or added relevant lines in 5 files are covered.
• 5 unchanged lines in 3 files lost coverage.
• Overall coverage decreased (-0.6%) to 60.261%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
payjoin/src/receive/v2/error.rs	0	3	0.0%
payjoin-cli/src/db/v2.rs	0	4	0.0%
payjoin/src/receive/error.rs	0	6	0.0%
payjoin-cli/src/app/v2.rs	0	29	0.0%
payjoin/src/receive/v2/mod.rs	84	132	63.64%

Files with Coverage Reduction	New Missed Lines	%
payjoin-cli/src/app/v2.rs	1	0.0%
payjoin-cli/src/db/v2.rs	1	0.0%
payjoin/src/receive/v2/mod.rs	3	80.48%

Totals
Change from base Build 12677131350:	-0.6%
Covered Lines:	2910
Relevant Lines:	4829

---

💛 - Coveralls

[DanGould]

This works.

However,

after updating, I realize that in the implementation only the Receiver<InitialState> needed to use the pub fn extract_err_req, process_err_res if the error is propagated up to the caller. So I think I'm going to rework this not to be generic and to simply handle the error at that higher level.

[arminsabouri]

This works.

However,

after updating, I realize that in the implementation only the Receiver<InitialState> needed to use the pub fn extract_err_req, process_err_res if the error is propagated up to the caller. So I think I'm going to rework this not to be generic and to simply handle the error at that higher level.

I'm not sure if you going to revert your generic work in 3534835 based on this comment. This work would potentially reduce the code footprint in multiparty things. Especially if we make sender context an associated type.

[DanGould]

Yes @0xBEEFCAF3, I made a more minimal implementation of this idea in #474. Which side of the fence would "potentially reduce the code footprint in multiplarty things," and how?

Could you explain what you mean about making sender context an associated type? Or do you mean SessionContext?

[DanGould]

The generic part of this doesn't even come into play with the implementation, so I'm going to close this for now. If it turns out that making a Receiver accept an associated SessionContext from either v2 or v3 I'm cool to reopen this

[arminsabouri]

Yes @0xBEEFCAF3, I made a more minimal implementation of this idea in #474. Which side of the fence would "potentially reduce the code footprint in multiplarty things," and how?

Could you explain what you mean about making sender context an associated type? Or do you mean SessionContext?

As I was working on the multiparty state machine I noticed that its essentially the same as the v2 reciever state machine (so far... I havent worked on any real validation so it may be the case that there needs to be additional / fewer states). The only real difference is the SessionContext type on the v2 reciever state structs (which needs to be a list of sender session contexts)
If we used traits and associated types that allows me to re-use the v2 recievers and just redefine SessionContext. I would say the demand for this level of change pretty low. I can still get what I want done and doing this refactor would just be a pre-optimization