Add SSZ Execution Blocks (EIP-7807) for Geth project proposal - #228
Conversation
taxmeifyoucan
left a comment
taxmeifyoucan
left a comment
There was a problem hiding this comment.
Thanks, the proposal reads very strong! No major issues, seems like well researched and realistic planned scope.
One issue is that lot of it reads like ai, it's very dense and would benefit from more of your own wording. But that's more of a format issue
| | --- | --- | --- | --- | | ||
| | through 6 | M0, ramp and design | Geth architecture ramp-up; the engine design document (API surface, per-type method conventions, the hashing backend interface, error handling for strict decoding) written and reviewed with the mentor; the conformance harness scaffold (pinned vector download, canary test, pass/skip reporting). Partly complete: the EIP set, the client survey, and the codebase mapping above are done. | Design document agreed; harness runs end to end with all cases skipped. | | ||
| | 6-10 | M1, static SSZ core | The static codec and classic merkleization per the engine section; both hashing backends with their equivalence test; the decoder fuzzer runs from here on. | The full classic ssz_generic suite passes, valid and invalid cases, on both backends. | | ||
| | 10-13 | M2, progressive types | Progressive merkleization, ProgressiveList/ProgressiveByteList, ProgressiveContainer; generalized-index resolution and single-field proof generation; the merkleizer fuzzer. | Progressive vectors green; differential comparison against remerkleable and dynamic-ssz agrees on randomized structures. | |
There was a problem hiding this comment.
python script that computes the same root from the same block data could be a good validation tool, comparing geth results to python reference as testing deliverable for this phase
There was a problem hiding this comment.
yeah, I've got that in the plan, will add it here as a clear deliverable...
|
|
||
| ### Current state of Geth | ||
|
|
||
| From going through the codebase: |
There was a problem hiding this comment.
Curious to see similarities with erigon/other ssz implementations fellows are working on
There was a problem hiding this comment.
Yeah, curious about that too, the progressive types, 7916 and 7495, are where the code should line up most across clients. And for 7807, Jack on Erigon and I end up computing the same block root, so we can cross-check each other's work..
531ec5a to
9dcc9c7
Compare
| | [Tosin](https://www.github.com/tosynthegeek) | [Update 2 & 3](https://hackmd.io/@tosynthgeek/rylPjYv7Mg) | [Update 4](https://hackmd.io/@tosynthgeek/BJjRLViEGg) | | [Full Fulu Compatibility and Cross-Client Interoperability for Ream](projects/ream-fulu-interoperability.md) | | ||
| | [Vansh](https://github.com/vanshsahay) | [Update 3](https://hackmd.io/@vanshsahay/Bk1rs4GXGx) | | [Update 4 & 5](https://hackmd.io/@vanshsahay/BJuLPYhEzx) | | | ||
| | [Vivek Singh Solanki](https://github.com/vivek-0509) | [Update 3](https://hackmd.io/@h2RUkQVmStGWLVKk0Ne1uQ/SJ-HrihXze) | [Update 4](https://hackmd.io/@h2RUkQVmStGWLVKk0Ne1uQ/HkA73ZV4ze) | | | | ||
| | [Vivek Singh Solanki](https://github.com/vivek-0509) | [Update 3](https://hackmd.io/@h2RUkQVmStGWLVKk0Ne1uQ/SJ-HrihXze) | [Update 4](https://hackmd.io/@h2RUkQVmStGWLVKk0Ne1uQ/HkA73ZV4ze) | | [SSZ Execution Blocks (EIP-7807) for Geth](https://hackmd.io/@h2RUkQVmStGWLVKk0Ne1uQ/ByggZxtEfl) | |
There was a problem hiding this comment.
this link needs to point to the project within the repo
9dcc9c7 to
b56a317
Compare
b56a317 to
2fe577c
Compare
2 participants
No description provided.