Decode FORS indices similarly to WOTS

[GillonB]

Update the decoding of the FORS indices as proposed by latest NIST changes. This has already been proposed in the sphincsplus repo (sphincs/sphincsplus#51). Would be good to have the same for liboqs as this change will eventually land. https://groups.google.com/a/list.nist.gov/g/pqc-forum/c/88tuvtb7nN4/m/DA1QCoJWBAAJ

• This changes breaks SPHINCS+ signature compatibility: Signatures generated before cannot be verified correctly with the implementation after this change. Also old implementation cannot verify correctly newly generate signatures.
• No change to available algorithms

relates to #1838

[dstebila]

Because we're pulling out SPHINCS+ implementation from PQClean, we'd prefer to see this change made first in PQClean, and then our automated scripts can pull it in. Would you be able to do that?

[SWilson4]

Closing due to inactivity. Additionally, rather than updating the SPHINCS+ code in place, we will presumably be replacing it with the final version of SLH-DSA whenever it lands upstream.