pass Cycle to the cycle recovery function

[mtshiba]

Currently, the cycle recovery function receives the following arguments:

• db: &dyn Db
• id: Id,
• last_provisional_value: &Output,
• value: Output,
• iteration: u32,
• (input: Input, ...)

After some consideration, we realized that we needed to expose the IDs of all cycle heads, which meant an additional argument was added, and the current approach of passing all information about the cycle to the cycle recovery function one by one would no longer be desirable. Cycle information that can be passed independently of the query function signature should be passed as a single context structure.

In this PR, we will expose this context as Cycle and pass it to the cycle recovery function.

[netlify]

✅ Deploy Preview for salsa-rs canceled.

Name	Link
🔨 Latest commit	ada3ed8
🔍 Latest deploy log	https://app.netlify.com/projects/salsa-rs/deploys/691fdae0699934000814e7c2

[codspeed-hq]

CodSpeed Performance Report

Merging #1028 will degrade performances by 5.01%

_{Comparing mtshiba:expose-cycle-heads (ada3ed8) with master (a885bb4)}

Summary

❌ 1 (👁 1) regression
✅ 12 untouched

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
👁	amortized[SupertypeInput]	2.8 µs	2.9 µs	-5.01%

[mtshiba]

Hmm, it seems like I'm experiencing a consistent regression, but I'm not sure what the direct cause is.

[Veykril]

I think you can ignore those, the regressions are in benchmarks that should be irrelevant to your changes so its likely just noise from code location changes

[MichaReiser]

This looks good to me, I only have a few nit comments