Split resume session relays

[benalleng]

Closes #1391

This commit ensures that each resumed payjoin-cli session is using a
separate instance of the RelayManager to then check the ohttp connection
independently. This fixes a bug where resuming would converge all
existing sessions to one ohttp relay.

Also, the incompatible_msrv allow seems not needed anymore?

Big pickle wrote most of this code. The free open weight models aren't too shabby now either. Though it did want to recreate the entire config instead of just making the app mutable. I thought this seemed a bit easier to parse as its clear everything else is the same.

Pull Request Checklist

Please confirm the following before requesting review:

• I have disclosed my use of
  AI
  in the body of this PR.
• I have read CONTRIBUTING.md and rebased my branch to produce hygienic commits.

[coveralls]

Pull Request Test Coverage Report for Build 23145192620

Details

• 3 of 6 (50.0%) changed or added relevant lines in 1 file are covered.
• 1 unchanged line in 1 file lost coverage.
• Overall coverage decreased (-0.01%) to 84.114%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
payjoin-cli/src/app/v2/mod.rs	3	6	50.0%

Files with Coverage Reduction	New Missed Lines	%
payjoin-cli/src/app/v2/mod.rs	1	53.89%

Totals
Change from base Build 23145123909:	-0.01%
Covered Lines:	10648
Relevant Lines:	12659

---

💛 - Coveralls

[arminsabouri]

CAck cf8b0b7

big pickle? Cool.
Should we / do we already ensure that there is > 2 ohttp relays ?

[spacebear21]

This PR also includes the commit from #1390, is it meant to supersede that PR? I left a comment there that I think we should address holistically (i.e. is ohttp_keys "cache" config field relevant at all since we refetch every time anyways?)

[benalleng]

I should add blocked as this simply builds on top of #1390

This PR would have simply needed to also add the commit from #1390 as the ohttp_keys have been "cached" by that point so relay selection is skipped and resume would just choose ohttp_relay[0] no matter what. #1390 (comment)

[bc1cindy]

tACK 5062488

minimal fix, each resumed session gets a fresh RelayManager, isolating relay state between sessions

but I think removing state from RelayManager and moving randomization into the struct as a method could simplify the design further, as commented in #1385 (comment)