Genesis header mismatch

[davxy]

I see that we are not aligned about what are the genesis header values.

IIRC the GP doesn't define an explicit value right?

As I already discussed with @sourabhniyogi we're using the values defined here for genesis.

However, maybe makes sense to change the extrinsic hash as your is technically more correct ( it is the hash of the empty 'Extrinsic' struct correct?)

Also, in our definition:

• the author id is set to 0xffff as genesis block has not been authored by author 0.
• an epoch mark is includer, not sure if this makes sense

[jaymansfield]

Maybe we need a genesis header defined in the GP to avoid ambiguities?

Yes I think at this point its been open to interpretation but having an explicit defined genesis in the GP would make the most sense to avoid any ambiguity, or at least have a few rules surrounding it so all teams have the same expectations.

However, maybe makes sense to change the extrinsic hash as your is technically more correct ( it is the hash of the empty 'Extrinsic' struct correct?)

Yes our extrinsic hash is calculated from the extrinsic's struct.

I'll align to using the same format as you shared with @sourabhniyogi (with the exception of having a valid extrinsic hash) and will republish later today.

[jaymansfield]

I pushed an update and this is my new genesis block https://github.com/javajamio/javajam-trace/blob/main/blocks/0.json

[davxy]

@jaymansfield Looks like our hash for the empty extrinsic doesn't match.

• Your is: 0x189d15af832dfe4f67744008b62c334b569fcbb4c261e0f065655697306ca252
• Mine: 0xdc080ad182cb9ff052a1ca8ecbc51164264efc7dd6debaaa45764950f843acb8

Also I think the order of the validators in your genesis header epoch mark is not correct. It should be the same as the next epoch validators (i.e. the validators in safrole state).

These are the validators in the genesis header epoch mark:

7f6190116d118d643a98878e294ccf62b509e214299931aad8ff9764181a4e33
aa2b95f7572875b0d0f186552ae745ba8222fc0b5bd456554bfe51c68938f8bc,
5e465beb01dbafe160ce8216047f2155dd0569f058afd52dcea601025a8d161d
48e5fcdce10e0b64ec4eebd0d9211c7bac2f27ce54bca6f7776ff6fee86ab3e3
3d5e5a51aab2b048f8686ecd79712a80e3265a114cc73f14bdb2a59233fb66d0
f16e5352840afb47e206b5c89f560f2611835855cf2e6ebad1acc9520a72591d

These are the validators in safrole state: (i.e. next epoch vals)

f16e5352840afb47e206b5c89f560f2611835855cf2e6ebad1acc9520a72591d
48e5fcdce10e0b64ec4eebd0d9211c7bac2f27ce54bca6f7776ff6fee86ab3e3
7f6190116d118d643a98878e294ccf62b509e214299931aad8ff9764181a4e33
aa2b95f7572875b0d0f186552ae745ba8222fc0b5bd456554bfe51c68938f8bc
3d5e5a51aab2b048f8686ecd79712a80e3265a114cc73f14bdb2a59233fb66d0
5e465beb01dbafe160ce8216047f2155dd0569f058afd52dcea601025a8d161d

[davxy]

I also get a mismatch for what is the post-genesis state root.

In the first block I try to import (i.e. 593699) I see that the state root for parent is 0x5b44e4eb5a3b25da6630ee4e5009303a4a4d880e766b9dcca91e0ea3f4e16057 but my post-genesis state root (after importing your key-value items) is 0x5faeff18729a5008cf99b076f809206f339ce7789eb254dcb55f50405b7a3b32.

[jaymansfield]

@davxy thank you for reporting these.

I also get a mismatch for what is the post-genesis state root.

I missed this change in the GP (w3f/jamtestvectors#14).

We get the same value now after I made the update.

Looks like our hash for the empty extrinsic doesn't match.

I've updated my genesis extrinsic hash to be 0x0000000000000000000000000000000000000000000000000000000000000000, but we should still see why an empty extrinsic is not matching.

The one I derive seems to match jam_dunas: https://github.com/jam-duna/jamtestnet/blob/main/data/safrole/blocks/1_011.json

Here is how I'm doing it:

ET (ET ) = 0x00 (compact encoding of 0 for length)
H(ET (ET )) = 0x03170a2e7597b7b7e3d84c05391d139a62b157e78786d8c082f29dcf4c111314

EP (EP ) = 0x00 (compact encoding of 0 for length)
H(EP (EP )) = 0x03170a2e7597b7b7e3d84c05391d139a62b157e78786d8c082f29dcf4c111314

g = 0x00 (compact encoding of 0 for length)
H(g) = 0x03170a2e7597b7b7e3d84c05391d139a62b157e78786d8c082f29dcf4c111314

EA(EA) = 0x00 (compact encoding of 0 for length)
H(EA(EA)) = 0x03170a2e7597b7b7e3d84c05391d139a62b157e78786d8c082f29dcf4c111314

ED (ED ) = 0x000000 (the three lengths of verdicts, culprits, faults which are all zero)
H(ED (ED )) = 0xab29e6dc16755d0071eba349ebda225d15e4f910cb474549c47e95cb85ecc4d6

H(ET (ET )) + H(EP (EP )) + H(g) + H(EA(EA)) + H(ED (ED ))) =
0x03170a2e7597b7b7e3d84c05391d139a62b157e78786d8c082f29dcf4c11131403170a2e7597b7b7e3d84c05391d139a62b157e78786d8c082f29dcf4c11131403170a2e7597b7b7e3d84c05391d139a62b157e78786d8c082f29dcf4c11131403170a2e7597b7b7e3d84c05391d139a62b157e78786d8c082f29dcf4c111314ab29e6dc16755d0071eba349ebda225d15e4f910cb474549c47e95cb85ecc4d6

Previous value hashed:
0x189d15af832dfe4f67744008b62c334b569fcbb4c261e0f065655697306ca252

Also I think the order of the validators in your genesis header epoch mark is not correct. It should be the same as the next epoch validators (i.e. the validators in safrole state).

Fixed.

I've pushed new vectors with the updated state roots and genesis block matching JamBrains/jam-docs#13

[davxy]

@jaymansfield we're able to import your vectors.
Do you have any vector with WR accumulation?

[jaymansfield]

@davxy Thanks for confirming.

None that perform accumulation yet but I'll will be generating some soon and will ping you once they are ready.