chore: bump filecoin-proofs-api to v19.0

[rjan90]

chore: bump filecoin-proofs-api to v19.0

Breaking Change Note:

Users will need to update their calls from:

// Before
ClearCache(sectorSize, cacheDirPath)
ClearSyntheticProofs(sectorSize, cacheDirPath)

// After  
ClearCache(cacheDirPath)
ClearSyntheticProofs(cacheDirPath)

The changes are consistent across all layers and properly remove the unused sector_size parameter completely from the API.

[Copilot]

Pull Request Overview

This PR updates the filecoin-proofs-api dependency from version 18.1 to 19.0, representing a major version bump that may include breaking changes.

• Upgraded filecoin-proofs-api from v18.1 to v19.0

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.
You can also share your feedback on Copilot code review for a chance to win a $100 gift card. Take the survey.}

[rvagg]

looks like we removed sector_size from a couple of calls that'll need to be updated here

[rvagg]

oh .. those sectorSize arguments go all the way up to the Go API, we should remove it there as well and this would become a breaking change.

[rjan90]

oh .. those sectorSize arguments go all the way up to the Go API, we should remove it there as well and this would become a breaking change.

Ah, just pushed underscoring the parameters, as I believed we still would need them for the Go-code. But I will look into removing them there

[vmx]

See #513 for how I've planned it. Though a breaking change is even better if that's possible.

[rjan90]

@vmx I was also going down that route in the first couple of commits:

• 1a66f3c
• a823243

But the build checks were still failing after those changes. At which point I switched to remove the sectorSize from ClearCache and ClearSyntheticProof.

[vmx]

• a823243

The error on that one was a timeout. I restarted it and it now passes.

[rvagg]

I think we can just break these.

In Lotus they're isolated to here: https://github.com/filecoin-project/lotus/blob/master/storage/sealer/ffiwrapper/sealer_cgo.go

In Curio they're in https://github.com/filecoin-project/curio/blob/main/lib/ffi/sdr_funcs.go and one in https://github.com/filecoin-project/curio/blob/main/lib/ffi/snap_funcs.go (cc @LexLuthr).

Not having to have sector size at some of these callsites might afford some additional cleanup opportunities, but I haven't looked at it in any detail.

So as long as we're bumping a version here to signal breakage, and documenting it, adapting to it upstream shouldn't be hard.