Incorrect service_account storage_key

[emielsebastiaan]

Incorrect ServiceAccount StorageKey

We found an inconsistency between our [{PyJAMaz}] implementation and the JAM-Duna implementation in the test data for the Assurances mode.
Reference:

jamtestnet/data/assurances/state_transitions/1_008.json

Line 126 in 06a2842

"0xffa64892e8000000000000000000000000000000000000000000000000000000",

Analysis

Our implementation cannot retrieve a service_account_id for service 3901900966 where we expect to find it in the state K/V-database.
It seems that the state key of JAM-Duna data is not conform the Graypaper specification in equation D.1. (https://graypaper.fluffylabs.dev/#/5f542d7/38f00038fb00)

Difference

JAM Duna Data

0xffa64892e8000000000000000000000000000000000000000000000000000000

[{PyJAMaz}] Data

0xffa60048009200e8000000000000000000000000000000000000000000000000

[danicuki]

The service with id 3901900966 seems to be introduced in block 0004 (migration to state 0005), but I don't see how it gets into the state. I am trying to pass block 0004 in my tests and this is one of the inconsistencies.

[sourabhniyogi]

We will publish a new dataset with a hopefully 0.5/0.6 GP compliance bootstrap/fib/... service to address #74 and provide a derivation of the service id and new service key. Stay tuned!

[sourabhniyogi]

The derivation we published on this commit:

s = 0
eta_0 = 592170100ab4055c92a269faf579e29e7e453d13549c0dc4f0120fc670ff357d
timeslot = 16
encode(s, eta_0, timeslot)_ = 0000000020592170100ab4055c92a269faf579e29e7e453d13549c0dc4f0120fc670ff357d10000000
hash(encoded) = a64792e833b594bab9bd59fabb3415502ea07256a3744896bf44847ee899efc4
hashed[:4] = a64792e8
decode(hashed[:4]) = 3901900710
decoded mod 4294966784 + 256= 3901900966

We agree on the 3901900966 (hex a64792e8), great!

Then with GP 0.6.2 C(255,s)

C(255, s=3901900966 (hex=e89248a6))=0xffa60048009200e8000000000000000000000000000000000000000000000000

It changed in GP 0.4.5 to include 0s in between, which escaped both of us it appears, I think you can confirm this

(We are doing a thorough review of 0.5.x + 0.6.x now that fluffy has pretty good annotations, but 0.4.x changes we might have to grind through them together... hope its not too annoying)