Feat: Core DCA

[msgmaxim]

Pull Request

Closes: PRO-1531, PRO-1539

Checklist

Please conduct a thorough self-review before opening the PR.

• I am confident that the code works.
• I have updated documentation where appropriate.

Summary

I'm hoping to merge this core functionality early so that @j4m1ef0rd could start working on this in parallel (we can/might also work on this branch ofc, but that's not ideal). This means some important features might come in separate PRs (see below).

• The commits are mostly self contained and can be reviewed one-by-one.
• Added dca_parameters to the request_swap_deposit_address_with_affiliates extrinsic and the the broker rpc method; the parameters are saved on the channel's actions (similarly to refund params). New migration adds None to any existing channels.
• DCA state is stored in RequestState. Now when a swap completes, we check DCA state to see what should happen next: if the swap/chunk is successful, we either schedule the next chunk (recording output) or egress if this was the final chunk; on failure, we refund any unexecuted amount, but still egress output from any previous chunks. No migration is needed since RequestStates don't exist in 1.5.

Out of scope:

• Configurable DCA Limits and parameter validation (PRO-1538)
• Bouncer tests (PRO-1532)
• Broker fee charged per chunk (PRO-1540)

Update: since this is early in review, I pushed another commit that now handles CCM too.

[codecov]

Codecov Report

Attention: Patch coverage is 88.27586% with 85 lines in your changes missing coverage. Please review.

Project coverage is 71%. Comparing base (be0ab82) to head (5e4bd9c).
Report is 1 commits behind head on main.

Files	Patch %	Lines
state-chain/pallets/cf-swapping/src/lib.rs	84%	52 Missing and 9 partials ⚠️
...cf-ingress-egress/src/migrations/add_dca_params.rs	91%	16 Missing ⚠️
state-chain/pallets/cf-ingress-egress/src/lib.rs	78%	4 Missing ⚠️
api/lib/src/lib.rs	0%	3 Missing ⚠️
state-chain/primitives/src/lib.rs	0%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@          Coverage Diff           @@
##            main   #5106    +/-   ##
======================================
- Coverage     72%     71%    -0%     
======================================
  Files        453     454     +1     
  Lines      81473   81768   +295     
  Branches   81473   81768   +295     
======================================
+ Hits       58268   58451   +183     
- Misses     20111   20205    +94     
- Partials    3094    3112    +18     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[kylezs]

https://linear.app/chainflip/issue/PRO-1544/add-params-to-swaprequestparams-in-cli

[dandanlen]

As mentioned in the previous PR, I think it's worth distinguishing more strongly between id types just for peace of mind. I've created a Linear issue PRO-1548.

[dandanlen]

(Will continue reviewing tomorrow)

[dandanlen]

Actually a few more comments here. More to follow tomorrow.

[dandanlen]

Some more comments...

[dandanlen]

Thanks, looking good 🚀 . I'll take a closer look at tests on Monday.

[kylezs]

Migrations / storage versions might be impacted, this PR was required because otherwise CI on main doesn't pass: #5142

[dandanlen]

Test cases look fine to me. Although I feel we can do better to reduce duplication in the tests and make them more readable - they are quite verbose and repetitive (this applies to pre-existing tests too, not just the newly added ones). For example, the use of 'global' constants for test amounts etc makes things a little hard to follow.

I think it's time we added some simple extension traits with common setup/update/check methods that are used across all the swapping tests. I'll raise an issue.

Apart from that I believe the only things that needs changing is what was discussed this morning re. expiry being implicit from the FoK retry delay and number of dca chunks - I think this implies simply updating the refund params for every chunk.